93001
Hydrogen          Final
Fluoride Study     ReP°rt
Report to Congress
Section 112(n)(6)
Clean Air Act As Amended

-------

-------
                              TABLE OF CONTENTS
                                                                                 Page

      Acronyms	  ix

      Executive Summary	  xiii


1.     INTRODUCTION			   1

      1.1    Background	   1
      1.2    Purpose	   2
      1.3    Approach	   3
      1.4    Organization of the Report	   4


2.     PROPERTIES AND HAZARDS OF HYDROGEN FLUORIDE	   7

      2.1    Description of Physical and Chemical Properties	   7
      2.2    Health Hazards	   8
      2.3    Environmental  Hazards	   12
      2.4    Release Characteristics	   12


3.     CHARACTERIZATION OF HYDROGEN FLUORIDE INDUSTRY	   19

      3.1    Production of Hydrogen Fluoride  	   19
      3.2    Uses of Hydrogen Fluoride	   21
      3.3    Market Outlook 	   25


4.     REGULATIONS AND INITIATIVES	    31

      4.1    U.S. Federal Regulation of Hydrogen Fluoride 	   31
      4.2    U.S. State and Local Regulations	   36
      4.3    International Efforts	   42


5.     HYDROGEN FLUORIDE INDUSTRY PROCESS DESCRIPTIONS 	   53

      5.1    HF Manufacture	   53
      5.2    Transportation and Storage	   56
      5.3    Fluorocarbon Production 	   57
      5.4    Alkylate Production for Gasoline	   61
      5.5    Uranium Processing	   67
      5.6    Aluminum Fluoride and Aluminum Manufacturing	   67
      5.7    Electronics Manufacturing 	   68
      5.8    Chemical Derivatives Manufacturing	   68
      5.9    Processes Using Aqueous Hydrogen Fluoride 	   69
                                                                                Page I

-------
                        TABLE OF CONTENTS (CONTINUED)
                                                                                  Page

       5.10   Dissolving Ores for Production of Tantalum and Columbium Metals  	  70
       5.11   Linear Alkylbenzene Production	  71
       5.12   Pharmaceutical Production	  71


 6.     HAZARDS OF HYDROGEN FLUORIDE PROCESSES AND INDUSTRY PRACTICES TO
       PREVENT RELEASES  	  77

       6.1     General Hazards 	  77
       6.2     General Industry Practices	  78
       6.3     Specific Industry Hazards and Practices	  85
       6.4     Research Efforts to Modify or Substitute Hydrogen Fluoride in Alkylation	  95


 7.     INDUSTRY PRACTICES TO DETECT AND MITIGATE HYDROGEN FLUORIDE RELEASES  101

       7.1     General Industry Practices to Detect Hydrogen Fluoride Releases	  101
       7.2     Hydrogen Fluoride Detectors Used by Specific Industries	  103
       7.3     General Industry Practices to Mitigate Hydrogen Fluoride Releases	  104
       7.4     Hydrogen Fluoride Mitigation Systems Used by Specific Industries	 .  107


 8.     CHARACTERIZATION OF HYDROGEN FLUORIDE ACCIDENTS  	  113

       8.1     Examples of Major Accidents	  113
       8.2     Analysis of HF Accident Databases	 .  114
       8.3     Overview of HF Accident Data	  120


 9.     MODELING HYDROGEN FLUORIDE RELEASES	  129

       9.1     Consequence Analysis	  129
       9.2     Models for HF Releases	  130
       9.3     Modeling Used in HF Study	  135
       9.4     Worst-Case Accident Scenarios	  139
       9.5     Descriptions of Scenarios Used in HF Study	  143
       9.6     Limitations of the Modeling Results	  148
       9.7     Results and Analysis 	  151
       9.8     Sensitivity Analysis	  155
       9.9     Summary	  160
Page //

-------
                       TABLE OF CONTENTS (CONTINUED)
10.
11.
COMMUNITY AND FACILITY EMERGENCY PREPAREDNESS AND PLANNING
                                                                                Page

                                                                                  167
10.1    Emergency Preparedness and Planning	  167
10.2    Industry Programs and Cooperation with the Community 	  168
10.3    Ways to Improve Facility Emergency Preparedness and Planning	  169
10.4    Community Efforts to Promote Emergency Preparedness and Planning	  170


FINDINGS AND RECOMMENDATIONS			  173

11.1    Findings	  173
11.2    Recommendations	  180
APPENDICES

       APPENDIX

       APPENDIX
            -  Summary Notes from Hydrogen Fluoride Roundtable

            -  Summary of Comments from Technical Reviewers on Draft Hydrogen Fluoride
              Study
       APPENDIX III -  Summary of Oral and Written Comments from Public Meeting on Hydrogen
                    Fluoride Study

       APPENDIX IV - Exposure Levels for Hydrogen Fluoride

       APPENDIX V -  Overview of Probit Equations

       APPENDIX VI - Example of Medical Treatment Guidelines from a Hydrogen Fluoride Producer

       APPENDIX VII - Facilities Reporting to TRI for Hydrogen Fluoride, 1990

       APPENDIX VIII- U.S. Producers of Fluorocarbons and of Other Chemicals Manufactured with
                    Hydrogen Fluoride or Chlorofluorocarbons

       APPENDIX IX - U.S. and Canadian Petroleum Refineries with Hydrogen Fluoride Alkylation
                    Units

       APPENDIX X -  Rule 1410:  SCAQMD Regulation on Hydrogen Fluoride Storage and Use

       APPENDIX XI - Containers for Transportation of Hydrogen Fluoride

       APPENDIX XII - Data Base Sources for Accident Information

       APPENDIX XIII - Descriptions of Hydrogen Fluoride Accidents
                                                                               Page Hi

-------
                       TABLE OF CONTENTS (CONTINUED)
       APPENDIX XIV - Deaths, Injuries, or Evacuations Caused by Hydrpgen Fluoride Accidents
       APPENDIX XV - Population Characterization
       APPENDIX XVI - Description of HQSYSTEM and SLAB
       APPENDIX XVII - Inputs for HGSYSTEM and SLAB Models
 BIBLIOGRAPHY
Page Iv

-------
                        TABLE OF CONTENTS (CONTINUED)
                                                                                    Page
LIST OF EXHIBITS
       EXHIBIT 2-1 -  Physical/Chemical Properties of Hydrogen Fluoride  	   7
       EXHIBIT 2-2 -  Exposure Guidelines For Several Toxic Substances Compared to Hydrogen
                     Fluoride	•	• • •   10
       EXHIBIT 3-1 -  U.S. Production Capacity for 1992	   20
       EXHIBIT 3-2 -  Non-U.S. North American Hydrogen Fluoride Producers and Shippers  ..   20
       EXHIBIT 3-3 -  Examples of Hydrogen Fluoride Uses	   22
       EXHIBIT 3-4 -  Types of Facilities Reporting to TRI for Hydrogen Fluoride	   23
       EXHIBIT 3-5 -  End Uses of Hydrogen Fluoride	   24
       EXHIBIT 3-6 -  Major Users and Producers of Anhydrous Hydrogen Fluoride	   26
       EXHIBIT 4-1 -  Regulatory Designations  of Hydrogen Fluoride	   32
       EXHIBIT 4-2 -  DOT Initial Isolation and Protective Action for Anhydrous Hydrogen Fluoride 35
       EXHIBIT 4-3  -  DOT Response Guide for Anhydrous Hydrogen Fluoride	   37
       EXHIBIT 4-4  -  DOT Response Guide for Aqueous Hydrogen Fluoride	   38
       EXHIBIT 5-1  -  Hydrogen Fluoride Manufacturing Process	   54
       EXHIBIT 5-2  -  Nitrogen Unloading of Hydrogen Fluoride From Tank Truck	   58
        EXHIBIT 5-3  -  Pump Unloading of Hydrogen Fluoride From Tank Car	  59
        EXHIBIT 5-4  -  CFC Manufacturing Process  	  60
        EXHIBIT 5-5  -  Phillips Alkylation Process	  63
        EXHIBIT 5-6  -  Phillips Hydrogen Fluoride System Reactor 	  64
        EXHIBIT 5-7  -  UOP Alkylation Process	  66
        EXHIBIT 6-1  -  Hazard Evaluation Procedures	  81
        EXHIBIT 8-1  -  Hydrogen Fluoride Events in the ARIP Database  	  115
        EXHIBIT 8-2 - Release Point From ARIP Data	  117
                                                                                    Page v

-------
                          TABLE OF CONTENTS (CONTINUED)
                                                                                      Page

  UST OF EXHIBITS (CONTINUED)

         EXHIBIT 8-3 -  Release Cause From ARIP Data	  1 -17

         EXHIBIT 8-4 -  How Was the Release Discovered?  	  118

         EXHIBIT 8-5 -  Source of Transportation Leaks	  1 -19

         EXHIBIT 8-6 -  Failure Types From HMIS Data  	  12o

         EXHIBIT 8-7 -  Industry Segment and Number of HF Releases 	  122

         EXHIBIT 8-8 -  Releases of Aqueous and Anhydrous HF Resulting in Injuries or Deaths   124

         EXHIBIT 8-9 -  Number of Reported Injuries or Deaths Associated with all HF Releases .  124

         EXHIBIT 8-10 - Analysis of Anhydrous Hydrogen Fluoride Incidents	  126

        EXHIBIT 9-1  -  Models Available for Simulation of Accidental Releases of
                      Hydrogen Fluoride	  133

        EXHIBIT 9-2  -  Release Mechanism Flowchart	  137

        EXHIBIT 9-3  -  Catastrophic Vessel Failure Scenarios	  142

        EXHIBIT 9-4  -  Range of Other Scenarios	  142

        EXHIBIT 9-5  -  Modeling Results of Catastrophic Vessel Failures	  152

        EXHIBIT 9-6  -  Modeling Results of a Range of Other Scenarios	  153

        EXHIBIT 9-7  -  Sensitivity Analysis - Wind Speed and Stability	  157

        EXHIBIT 9-8  -  Sensitivity Analysis - Release Rate	  157

        EXHIBIT 9-9  - Sensitivity Analysis - Relative Humidity	   158

        EXHIBIT 9-10 - Sensitivity Analysis - Surface Roughness	   158

        EXHIBIT 11-1 -  Reviewers Who Commented on Draft Hydrogen Fluoride Report	  11-10

        EXHIBIT 111-1  - Reviewers Who Submitted Written Comments on the Preliminary Findings 111-11

        EXHIBIT IV-1 - Comparison of Regulatory and  Guideline Exposure Levels for
                     Hydrogen Fluoride and Other Toxic Substances 	  IV-2

        EXHIBIT V-1 -  Probit Table	                    v_2
Page vi

-------
                         TABLE OF CONTENTS (CONTINUED)
LIST OF EXHIBITS (CONTINUED)

       EXHIBIT V-2 -  Coefficients for Four Probit Equations	  V-4

       EXHIBIT V-3 -  Results Based on Several Probit Equations for Five-Minute Exposures .  .  V-4

       EXHIBIT V-4 -  Results Based on Several Probit Equations for 30-Minute Exposures ...  V-5

       EXHIBIT V-5 -  Results Based on Several Probit Equations for 60-Minute Exposures ...  V-5

       EXHIBIT VIII-1 - U.S. Producers of Fluorocarbons	  Vlll-2

       EXHIBIT Vlll-2 - Other Chemicals Manufactured Using Hydrogen Fluoride or Based on
                     Another Hydrogen Fluoride Product	  VIII-3

       EXHIBIT IX-1 -  Petroleum Refineries With Hydrogen Fluoride Alkylation Units	  IX-2

       EXHIBIT XI-1 - Department of Transportation Regulations for the
                     Transportation of Hydrogen Fluoride  	  XI-2

       EXHIBIT XI-2 - Typical Class 105A300W Tank Car	  XI-4

       EXHIBIT XI-3 - Typical Railway Tank Car Specifications - Class 105A300W	  XI-5

       EXHIBIT XI-4 - Railway Tank Car Specifications	  XI-6

       EXHIBIT XI-5 - Motor Vehicle Tank Specifications	  XI-7

       EXHIBIT XI-6 - Cylinder Specifications	  XI-7

       EXHIBIT XIII-1 - General Descriptions of Hydrogen Fluoride Accidents
                     From the ARIP Database	  XIII-2

       EXHIBIT Xlll-2 - General Descriptions of Hydrogen Fluoride Accidents
                     From the AHE Database	  XIII-6

       EXHIBIT XIII-3 - General Descriptions of Hydrogen Fluoride Accidents
                      From the ERNS Database	  XIII-8

       EXHIBIT XIII-4 -  General Descriptions of Hydrogen Fluoride Accidents
                      From the HMIS Database	-	  XIII-13

       EXHIBIT XIII-5 -  General Descriptions of Hydrogen Fluoride Accidents
                      From the Other Sources	  XIII-15

       EXHIBIT XIV-1 - Deaths, Injuries, or Evacuations Caused by Hydrogen
                      Fluoride Accidents Listed in the ARIP Database	  XIV-1
                                                                                     Page vii

-------
                         TABLE OF CONTENTS (CONTINUED)
 UST OF EXHIBITS (CONTINUED)
        EXHIBIT XIV-2 - Deaths, Injuries, or Evacuations Caused by Hydrogen
                      Fluoride Accidents Listed in the AHE Database	  XIV-2
        EXHIBIT XIV-3 - Deaths, Injuries, or Evacuations Caused by Hydrogen
                      Fluoride Accidents Listed in the ERNS Database	  XIV-3
        EXHIBIT XIV-4 - Deaths, Injuries, or Evacuations Caused by Hydrogen
                      Fluoride Accidents Listed in the HMIS Database	  XIV-4
        EXHIBIT XIV-5 - Deaths, Injuries, or Evacuations Caused by Hydrogen
                      Fluoride Accidents Listed in Newspapers and Accident Reports	  XIV-5
        EXHIBIT XV-1 - Population Characterization 	  XV-2
        EXHIBIT XVI-1 - Descriptions of Models in HQSYSTEM			  XVI-2
        EXHIBIT XVI-1 - Descriptions of Source Types in SLAB	,	  XVI-4
        EXHIBIT XVII-1a,b - Inputs for HQSYSTEM Models: HFSPILL and HFPLUME, HEQADAS-S and
                      HEQADAS-T	  XVII-2
        EXHIBIT XVII-lc - Inputs for HQSYSTEM Models:  HFSPILL and HFPLUME or EVAP-HF, and
                     HEGADAS-S and HEGADAS-T	  XVII-6
        EXHIBIT XVII-2 - Inputs for SLAB Models	  XVII-8
Page vlli

-------
                                    ACRONYMS
AAR         Association of American Railroads
ACMH        Advisory Committee on Major Hazards
AEG         Atomic Energy Commission
AGIR         Aide a' la Gestion industrielle des Risques
AHE         Acute Hazardous Events (database)
AHF         Anhydrous Hydrogen Fluoride
AlChE        American Institute of Chemical Engineers
AIHA         American Industrial Hygiene Association
ANSI         American National Standards Institute
APELL       Awareness and Preparedness for Emergencies at the Local Level
API          American Petroleum Institute
ARCO        Atlantic Richfield Company
ARIP         Accidental Release Information Program
ASME        American Society of Mechanical Engineers
ASNT        American Society for Non-destructive Testing
ASO         Acid-Soluble Oil
ASTM        American Society for Testing and Materials

BOE         Buffered Oxide Etch
BP          British Petroleum

CAA         Clean Air Act
CAAA        Clean Air Act Amendments of 1990
CAMEO      Computer-Aided Management of Emergency Operations
CAS         Chemical Abstract Service
GCPA        Canadian Chemical Producers Association
CCPS         Center for Chemical Process Safety
CEFIC        Conseil Europeen des Federations de ('Industrie Chimique (European Chemical
              Industry Council)
CEPA         Canadian Environmental Protection Act of 1988
CERCLA      Comprehensive Environmental Response, Compensation and
              Liability Act
CFC          Chlorofluorocarbon
CFR          Code of Federal Regulations
CHEMTREC   Chemical Transportation Emergency Center
CIMAH       Control of Industrial Major Accident Hazards
CMA         Chemical Manufacturers Association
CMNIG       Chemical Manufacturing National Industry Group
CPQRA       Chemical Process Quantitative Risk Analysis
CTEF         Comite Technique Europeen de Fluor (European Technical Committee on Fluorine)

 DDBS        (Sodium) Dodecyl-Benzene Sulfonate        '
 DEGADIS     Dense Gas Djspersion Model
 Dl           Deionized water
 DOE         Department Of Energy
 DOT         Department Of Transportation
                                                                                  Page ix

-------
                               ACRONYMS (CONTINUED)
 EEC          European Economic Community
 EEGL         Emergency Exposure Guidance Level
 EHS          Extremely Hazardous Substance
 EPA          Environmental Protection Agency
 EPCRA       Emergency Planning and Community Right-to-Know Act
 ERNS         Emergency Release Notification System
 ERPG         Emergency Response Planning Guideline

 FEMA         Federal Emergency Management Agency
 FMECA       Failure Modes, Effects, and Criticality Analysis
 FTA          Fault Tree Analysis

 HazOp        Hazard and Operability (studies)
 HCFC         Hydrochlorofluorocarbon
 HF            Hydrogen fluoride
 HFC          Hydrofluorocarbon
 HMIS          Hazardous Materials Information System
 HMR          Hazardous Materials Regulations
 HMTA         Hazardous Materials Transportation Act
 HMTUSA      Hazardous Materials Transportation Uniform Safety Act
 HSDB         Hazardous Substance Data Bank
 HSE          Health and Safety Executive

 ICHMAP       Industry Cooperative HF Mitigation/Assessment Program
 IDLH          Immediately Dangerous to Life and Health

 LAB           Linear alkylbenzene
 LABS          Linear alkylbenzyl sulfonate
 LEPC          Local Emergency Planning Committee

 MIACC         Major Industrial Accidents Council  of Canada
 MIBK          Methyl-isobutyl ketone
 MSDS         Material Safety Data Sheet
 MTBE         Methyl tert-butyl ether

 NAS          National Academy of Science
 NIOSH        National Institute for Occupational Safety and Health
 NLM „         National Library of Medicine
 NOAA         National Oceanic and Atmospheric Administration
 NOV          Notice of Violation
 NPDES        National Pollutant Discharge Elimination System
 NRC          National Response Center
 NT1S          National Technical Information Service

 OECD         Organisation for Economic Co-operation and Development
 OSC          On-Scene Coordinator
 OSHA         Occupational Safety  and Health Administration
Pagex

-------
                              ACRONYMS (CONTINUED)
PEL
PHA
ppm
Probit
PSM
Permissible Exposure Limits
Preliminary Hazard Analysis
Parts per million
Probability unjt
Process Safety Management
QRA          Quantitative Risk Assessment

RCRA         Resource Conservation and Recovery Act
RMPP         Risk Management and Prevention Program
RP           Recommended Practice
RTU          Remote Terminal Unit

SAB          Science Advisory Board
SARA         Superfund Amendments and Reauthorization Act
SCAQMD      South Coast Air Quality Management District
SCBA         Self-Contained Breathing Apparatus
SERC         State Emergency Response Commission
SMCL         Secondary Maximum Contaminant Level
SPEGL        Short-Term Public Exposure Guidance Level
STEL         Short-Term Exposure Limit

TACB         Texas Air Control Board
TAME         Tertiary amyl methyl ether
TCAA         Texas Clean Air Act
TNO          Nederlandsche Organisatie voor Toegepast - Natuurwetenschappelijkonderzock
              (Netherlands Organization for Applied Scientific Research)
TPQ          Threshold Planning Quantity
,TRI           Toxic Release Inventory
TWA          Time-Weighted Average

UNEP         United Nations Environmental Programme
UOP          Universal Oil Products
UPS          Uninterrupted Power Supply

VOC          Volatile Organic Compound
VLSI          Very Large-Scale Integrated (circuits)
                                                                                Page xi

-------
                                  This page left intentionally blank
Page x//

-------
                                EXECUTIVE SUMMARY
Background

       Undef section 112(n)(6) of the Clean Air Act of 1990, as amended, Congress required EPA to
carry out a study of hydrofluoric acid (also called hydrogen fluoride (HF)), to identify potential hazards
to public health and the environment considering a range of events including worst-case accidental
releases, and to make recommendations for reducing the hazards, if appropriate. This report,
developed in response to the Congressional mandate, identifies and evaluates the hazards to the
public posed by the production and use of HF.  It is not intended to quantify risk to the public from
HF.  Analysis of public exposure to routine emissions was not included in this study because the
statutory language focuses on worst-case releases from accidents. EPA is submitting this report to
Congress in fulfillment of Section  112(n)(6) of the Clean Air Act as amended.

Summary Findings and Recommendations

       HF  is used industrially in large quantities throughout the United States (over 200,000 tons per
year) and in a great number of applications across a broad  range of industries (over 500 facilities).  It
serves as a major feedstock and source of the fluorine molecule for the production of fluorinated
compounds.

       An  accidental release of HF from one of these industrial facilities could have severe
consequences.  HF is toxic to humans, flora, and fauna in certain doses and can be lethal as
demonstrated by documented workplace accidents.  HF can travel significant distances downwind as
a dense vapor and aerosol under certain accidental  release conditions.  Because HF can exist as an
aerosol, the cloud can contain a substantially greater quantity of the chemical than otherwise would
be the case. Thus, the potentially high concentration of HF  in these dense vapor and aerosol clouds
could pose a significant threat to the public, especially in those instances where HF is handled at
facilities located in densely populated areas. Prompt and specialized medical attention is necessary to
treat HF exposure properly.

       However, the risk to the public of exposure to HF is  a function of both the potential
consequences and the likelihood  of occurrence of an accidental release; and the likelihood of an
accidental release of HF can  be kept low if facility owners/operators exercise the general duty and
responsibility to design, operate, and maintain safe facilities. In particular,  owners/operators can
achieve an  adequate margin  of protection both for their workers and the surrounding community by
assiduously applying existing industry standards and practices, existing regulations, and future
guidance and regulations applicable to various classes of hazardous substances in various settings.
The properties that make HF  a potentially serious hazard are found individually or in combination in
many other industrial chemicals; thus, HF does not require unique precautions.  Instead, within each
of the several different circumstances in which HF is handled, an appropriate combination  of general
and special precautions should result in:  (1) the safe management of HF and other hazardous
substances with an emphasis on accident prevention; (2) the preparedness to properly and quickly
respond to  chemical emergencies and to provide specialized medical treatment if necessary; and (3)
community  understanding of the risks involved.

       The EPA does not recommend  legislative action from the Congress at this time to reduce the
hazards associated with HF.  The Agency believes that the legislative authorities already in place
provide a solid framework for the prevention of accidental chemical releases and preparedness in the
event that they occur. The Agency recommends that facilities handling HF coordinate closely with
their Local Emergency Planning Committees (LEPCs).  LEPCs and facilities that handle HF should
                                                                                     Page xiii

-------
conduct drills and exercises to test mitigation, response, and medical treatment for a simulated HF
accident.  Furthermore, the Agency recommends that facilities actively conduct outreach efforts to
ensure that the community is aware of the hazards of HF, that protective measures are in place to
protect public health in the event of an accident, and that proper actions will be taken during an
emergency. Facilities should be able to rapidly detect,  mitigate, and respond to accidental releases in
order to minimize the consequences (e.g., through detection, monitoring, mitigation, and alert or alarm
systems).  Finally, the EPA will continue to support research and development efforts for process
safety Improvements and implementation, modeling and assessment improvements, and accidental
release monitoring and detection improvements.

Summary of Report

       HF is a very corrosive and toxic inorganic acid.  It can either be a gas or liquid in anhydrous
form (without water; 100 percent HF) or in aqueous solution (with water). Exposure to HF can cause
injury through inhalation, direct contact, or ingestion. HF is particularly caustic to tissue and exposure
may require special treatment.  HF is one of the more corrosive and toxic industrial chemicals, but it is
not unique among hazardous chemicals; other inorganic acids are similarly corrosive (e.g.,
hydrochloric acid), and some other relatively common chemicals are similarly toxic or more toxic than
HF (e.g., chlorine).  HF boils at 67°F, a temperature that is frequently exceeded under ambient
conditions. Consequently, if HF liquid is released, it may vaporize under ambient conditions.

       HF exhibits release characteristics in some circumstances that may make it particularly
hazardous to the public.  HF molecules may associate with one another (i.e., form larger molecules
like H4F4,  H6F6, H8Fg) via hydrogen bonding; such molecules may form a cloud that is heavier than
air. A vapor cloud of single, unassociated HF molecules will be lighter than air. A cloud that is lighter
than air is likely to disperse more readily than one that is heavier than air. In  addition, if HF is
released under pressure above its boiling point, droplets of HF may be carried into the air as aerosol
along with HF vapor. Anhydrous HF released under pressure above its boiling point may form a cloud
of vapor and aerosol that is heavier than air and that may travel for long distances close to the
ground, posing a threat to people in its path. Although an HF vapor cloud may form under some
conditions from a release of an aqueous solution of HF, depending on concentration and release
temperature, anhydrous HF is much more likely to form a vapor cloud and, therefore, is potentially
more hazardous to the public.

       HF has been a focus of interest to industry for several years.  Industry groups have carried out
research and tests to characterize the behavior of HF upon release, improve dispersion modeling
techniques, and to test systems for mitigation of HF releases.  A large accidental release of HF at a
petroleum refinery drew additional attention to the hazards of HF releases. The South Coast Air
Quality Management District (SCAQMD) studied the hazards of HF use and production in the Los
Angeles Basin and adopted regulations phasing out the use of anhydrous HF within the Basin.  These
regulations were litigated, during which time their implementation was suspended by the court.
However, a recent court decision permitted implementation of the rule after additional rulemaking
procedures are conducted.

       HF is produced at three sites in the United States:  Allied-Signal, in Geismar, Louisiana; Du
Pont Chemicals in La Porte, Texas; and Elf Atochem in Calvert City, Kentucky. Production capacity
was approximately 206,000 tons in 1992.  Both anhydrous  and aqueous HF have a wide variety of
uses. The largest use is the manufacture of fluorine-containing chemicals, particularly
chlorofluorocarbons (CFCs). Fluorocarbon manufacture consumes 63 percent of the total HF used.
HF also may be used as an alkylation catalyst for the production of gasoline blending components;
this use consumes 7 percent of the total.  Other uses include aluminum production (3 percent, with
additional HF produced and used captively) and nuclear applications (5 percent). A number of other
uses, including stainless steel pickling, manufacture of various chemical derivatives and products,
Page xiv

-------
electronics, specialty metal production, and glass etching and polishing, consume the remaining 22
percent of HF produced.

       HF is regulated under a number of.U.S. statutes.  It is listed as a hazardous substance under
the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA), as a
hazardous waste under the Resource Conservation and Recovery Act (RCRA), as an  extremely
hazardous substance under section 302 of the Emergency Planning and Community  Right-to-Know
Act (EPCRA), as a toxic substance under EPCRA section 313, as a hazardous material in
transportation under Department of Transportation regulations, as an air contaminant under the
Occupational Safety and Health Act (OSHA) Air Contaminants Standard, and as a highly hazardous
chemical under the OSHA Process Safety Management Standard. HF has been proposed  as a
regulated substance for accidental release prevention under section 112(r) of the Clean Air Act.  HF is
subject to risk management programs in several states, including California, Delaware, Nevada and
New Jersey. All of these regulations include  HF as one of a number of regulated substances.  The
South Coast Air Quality Management District  in the Los Angeles Basin is the only government agency
that has adopted specific regulations for HF;  these regulations would phase out use of anhydrous  HF,
would require interim control measures, and would impose reporting and inventory requirements.

       Industry has taken  steps specifically to address and minimize the hazards of HF use and
production. The American  Petroleum Institute (API), the major trade association of the petroleum
industry, has developed recommended practices for operating and maintaining HF alkylation units  at
refineries; the procedures and practices described are intended to minimize the potential for an HF
release, mitigate the effects of a release if it occurs, and provide for oversight and audit of the  entire
process. The National Petroleum Refiners Association endorses the API recommended practice. The
Chemical Manufacturers Association (CMA) sponsors an HF Mutual Aid Group comprised of specially
trained teams that respond to emergencies involving HF.  Companies that produce and use HF have
also formed an  HF Panel under CMA auspices. The purpose of the panel is to make safety, health,
and environmental information available to the entire industry. The panel appoints various Task
Groups to address aspects of HF safety, and develops and maintains guidelines for the safe handling
of HF.

       In the process of conducting the study and gathering information about HF, EPA visited a
number of facilities that produce or use HF and observed the procedures  used to promote process
safety. These procedures include designing  equipment for HF use to minimize hazards; regularly
testing, inspecting, and maintaining equipment; and training workers.  Some facilities have  installed
HF detection systems; however, reliable and  accurate HF detectors have been difficult to develop,
particularly for perimeter monitoring.  A number of facilities also have mitigation systems to reduce  the
quantity or concentration of HF if a release occurs. Systems include water spray systems to knock
down HF vapors in case of a release, scrubber systems to absorb HF vented from process streams,
and emergency de-inventory systems to rapidly move HF from failed equipment to safe equipment.
Facilities also use remotely-operated emergency isolation valves to prevent and mitigate releases.
Because EPA observed practices only at selected  sites, it is not clear to what extent practices to
promote HF safety are used at HF facilities in all industry segments.

       Special  equipment  is used in transportation to prevent releases in case of a transportation
accident.  U.S. HF producers transport anhydrous HF in rail cars that exceed DOT safety requirements
and have headshields and  shelf couplers to protect the tanks in the event of a derailment.  Safety
relief valves on tank cars and trucks are used to release HF gas in the event of overpressurization.
These valves are protected by extra heavy rollover type domes. Valves for loading and unloading are
also contained within the rollover protection dome on the top of the tanks. HF producers provide
rigorous training programs  for drivers of HF vehicles.  They  also may conduct route risk analysis. One
HF producer has installed a satellite tracking  system to track HF trucks. Loading and unloading of HF
from transport containers is often cited as a point where a release could occur, particularly as  a result
                                                                                     Page xv

-------
of failure of a transfer hose. To prevent releases, specially designed transfer hoses are used, and
precautions are taken to prevent corrosion of piping, valves, and vessels.

       A large release of HF from a refinery in 1987 led to formation of a vapor cloud that migrated
through a residential area, causing a number of injuries, a large-scale evacuation, and damage to
vegetation,  in general, however, there have been relatively few reports of accidents involving HF, and
only a small fraction of these caused impact to the public. There have been no off-site deaths
reported from HF releases although some worker deaths have occurred. EPA's analysis of accident
data is consistent with the expectation that releases of anhydrous HF or concentrated aqueous HF
solution (70 percent HF) pose more hazards both on-s'rte and off-site than less concentrated aqueous
HF.

       For its analysis of the hazards to the public from HF, EPA carried out consequence analysis,
using computer modeling techniques, for a range of worst-case accident scenarios. Modeling
Indicated that releases of large quantities of HF over a short period of time (e.g., resulting from
catastrophic vessel failure) could pose a hazard to people far beyond facility boundaries, particularly
under low wind speeds and stable atmospheric conditions.  This type of accident is highly unlikely,
but, based  on modeling results, has the potential to cause great harm.  Smaller releases may or may
not pose a  hazard beyond a facility fenceline depending on the circumstances of the release.
Mitigation systems (e.g., water spray, emergency de-inventory, automatic shutoff valves) were also
modeled and shown to reduce affected distances downwind.  EPA did not consider the probability
involved with these worst-case accident scenarios.

    While visiting HF facilities to observe  management practices, EPA also gathered information on
the interaction between communities and facilities for emergency preparedness and planning.  In the
event of a release of HF, coordination between the community and the facility would help community
officials react quickly and take proper actions to protect the public.  EPCRA (SARA Title III), mandated
the formation  of Local  Emergency Planning Committees (LEPCs) to develop emergency response
plans for chemical accidents.  Some HF facilities are members or supporters of LEPCs. HF facilities in
some industries have established mutual  aid agreements that may also involve community officials.
Some HF facilities cooperate with local government agencies in  activities such as conducting
emergency drills. The Chemical Manufacturers Association (CMA) has developed a community-
oriented program called the Community Awareness Emergency  Response (CAER) program which
recommends ways for chemical facilities to develop working relationships with communities to address
emergency situations involving many chemicals including HF. EPA's observations indicated that in
some areas near HF facilities, the public has not shown much concern or interest in the hazards of HF
and other chemicals, or in emergency preparedness and planning for chemical accidents. Also, some
facilities acknowledge that facility outreach can be greatly improved.
Page xvl

-------
                                  1.  INTRODUCTION
       The purpose of this report to Congress is to study the hazards associated with the production
and uses of hydrofluoric acid and to make recommendations about reducing these hazards based on
the findings. In this report anhydrous hydrogen fluoride will be designated as HF; aqueous solutions
of hydrogen fluoride will be called aqueous HF.  This distinction is important especially when
considering such things as severity of exposure, corrosivity, and likelihood of fuming and forming
dense vapor clouds upon release.  Although worker exposure and environmental impacts from
routine emissions are important issues, this report focuses primarily on the potential hazards to the
public from accidental releases of HF during production,  use, transport, transfer, and storage.
Analysis of public exposure to routine emissions was not included in this study because the statutory
language focuses on worst-case releases from accidents.

1.1    Background

       EPA was directed by Congress to carry out a study of hydrofluoric acid, a toxic, corrosive
material, which when released under certain conditions, can form a dense vapor cloud, travel
downwind, and pose a serious threat to the public. This report, developed in response to the
Congressional mandate, identifies and evaluates hazards to the public posed by the production and
use of HF.  EPA is submitting this report to Congress in fulfillment of Section 112(n)(6) of the Clean Air
Act as amended:

       "Hydrofluoric Add - Not later than 2 years after the date of enactment of the Clean Air Act
       Amendments of 1990, the Administrator shall, for those regions of the country which do
       not have comprehensive health and safety regulations with respect to hydrofluoric acid,
       complete a study of the potential hazards of hydrofluoric acid and the uses of hydrofluoric
       acid in industrial and commercial  applications to  public health  and the environment
       considering a range of events including worst-case accidental releases and shall make
       recommendations to the Congress for the reduction of such hazards, if appropriate."

       HF is manufactured and used in the U,S. primarily for the production of fluorocarbons  (63%);
for solutions used for glass etching, cleaning, stainless steel pickling, and chemical derivatives (9%);
as a catalyst for the production of gasoline (7%); for nuclear applications (5%); and for aluminum
production (3%). For the majority of the uses, there is no currently known viable alternative production
method or substitute chemical.

       HF is known to be a hazard because of its toxicity and corrosivity.  Exposure to HF can cause
injury through inhalation, direct contact, or  ingestion.  HF Is particularly caustic to tissue. HF exposure
may require special treatment.  HF can also form dense vapor clouds upon release and travel
downwind. However, such properties are not limited to HF. The formation of toxic, dense vapor
clouds can also potentially occur if chlorine (CI2), ammonia (NHg), and other toxic gases like
phosgene are accidentally released. In 1990, 34 and 22 billion pounds of NH3 and CI2 were
produced,1 respectively, while only 0.4 billion pounds of HF were produced.2

       During the  summer of 1986, Amoco, Allied-Signal, Du Pont, and Lawrence Livermore National
Laboratory voluntarily conducted a series of six experiments involving atmospheric releases of HF in
an attempt to characterize its behavior.  These studies, known as the Goldfish studies, were
conducted at the Department of Energy Liquefied Gaseous Fuels Spill Test Facility in Nevada  and
showed that the HF did not remain a liquid following the release. Instead, under the conditions
simulating a petroleum refinery HF alkylation unit release (i.e., HF above its boiling point and liquefied
under pressure), a cold, dense cloud containing aerosol was generated which traveled a substantial
distance downwind from the release point at ground level.  This result led industries involved in the
                                                                                       Page 1

-------
 use or production of HF to improve dispersion modeling techniques and release mitigation systems.
 The Goldfish tests facilitated the formation of an extended consortium of twenty energy and chemical
 companies which subsequently began a program to improve dispersion models and collect data on
 mitigation of HF releases. This group performed the 1988 Hawk series tests at the DOE Nevada site
 for this purpose.3

        Additional attention was focused on the use of HF as an alkylation catalyst at petroleum
 refineries, because of a large accidental release that occurred on October 30, 1987, at the Marathon
 Petroleum Company refinery in Texas City, Texas. The release occurred when a 50-foot, multi-ton
 heater convection unit was being moved for maintenance and repair and was accidentally dropped
 onto an HF acid vessel. The unit severed a 4-inch acid loading line and a 2-inch  pressure relief line,
 causing the release of between 30,000 and 53,000 pounds of HF over a 44 hour period.4'5 As a
 result of the high release rate immediately following the accident, the vapors initially migrated to an
 adjacent residential area.  Eighty-five square blocks and approximately 4,000 residents were
 evacuated; 1,037 residents were treated at three neighboring hospitals with skin, eyes, nose, throat,
 and lung irritation.   Vegetation was also damaged in the path of the vapor cloud, but no fatalities
 occurred.

        A 100-pound release of HF at a refinery in Torrance, California on November 24, 1987, further
 focused public concern in California on the hazards posed  by HF, especially at petroleum refineries.
 Studies by the South Coast Air Quality Management District (SCAQMD) in the Los Angeles basin led
 to adoption of specific regulations which phase out the use of anhydrous HF (unless its properties are
 modified), require interim control measures, and impose reporting and storage/use inventory
 requirements.  As the only regulation in the U.S. directed specifically at anhydrous HF, Rule 1410 is
 partly intended to eliminate the possibility of harm to the public in the Los Angeles basin due to an
 unmitigated accidental release of HF.  As a result of legal action, these regulations were temporarily
 suspended.

        EPA's evaluation of the processes and practices associated with the production and  uses of
 HF (see section 6.2) indicates that the techniques, processes, and equipment used in the various HF
 Industry segments are no different than those  commonly used in the chemical manufacturing and
 petroleum refining industries in the U.S.  EPA visited a number of facilities during the course of this
 study. The facilities visited were exemplary in  their approaches to handling HF; however, there have
 been serious problems at some facilities involving not only HF but other hazardous materials.  Such
 problems indicate the need for process safety management for HF and other hazardous chemicals, as
 well as the need to  communicate crucial information to stakeholders and the public on how to prevent,
 mitigate and respond to HF releases.

 1.2     Purpose

        EPA performed this study:

        »•      to gather information from producers, users, and  other stakeholders in
               the  HF issues, and compile that information into a document for public
               dissemination;
                                      #
        »•      to foster communication between the various stakeholders who have
               an interest in HF issues;

        >•       to gather information on technically sound methods with which to
               solve potential safety problems associated with the industrial
               production and uses of anhydrous HF; and          *
Page 2

-------
       >      to identify the issues and problems which remain to be solved.

       This study attempts to:

       >•      characterize how and where HF is produced and used in the U.S.;

       >      identify and characterize the hazards specific to those uses and
              processes;

       >      identify how industry segments try to manage and mitigate those
              hazards;

       >      assess the potential hazards posed to the public and the environment
              from HF releases; and

       »•      identify approaches to minimize hazards and maximize safety
              associated with HF and to identify issues which need additional
              evaluation.

These issues must be explored and dealt with effectively to protect the health and safety of the public,
and the environment.

1.3    Approach

       EPA believed that this study should reflect input from those individuals and organizations with
a "stake" or interest in its outcome.  Such stakeholders include environmental groups, labor, industry,
trade associations, professional societies, and state and federal government agencies.  Consequently,
EPA held a "Roundtable" meeting on October 17,  1991 in Fairfax, Virginia, with individuals representing
these interests. The goals of the Roundtable were to solicit input on the major issues surrounding HF
use, to develop ways to address critical HF issues, and to establish a group of technical reviewers for
the study.  A summary of the meeting notes from the Roundtable is provided in Appendix I.

       After the Roundtable, EPA met individually with some stakeholders to discuss specific issues
including quantitative risk assessment, realistic HF release scenarios, release prevention techniques,
release mitigation techniques, and any research efforts underway or contemplated concerning the
reduction of hazards associated with the use of HF. Stakeholders also provided EPA with documents
such as hazard and risk assessments, HF safe handling procedures, relevant articles about HF, and
release and dispersion modeling studies.  In addition, EPA conducted its own extensive literature
search and contacted numerous other potential stakeholders and international agencies and industrial
groups.  EPA also used several accidental release databases maintained by EPA, OSHA, DOT, and
other organizations to gather historical documentation on accidental HF releases, their causes and
consequences.

        EPA's preliminary analysis indicated that the greatest hazards of HF are associated  with the
manufacture and use of anhydrous HF as opposed to aqueous HF.  This finding was reinforced by
the Roundtable meeting and meetings with other stakeholders.  For this reason, EPA decided to focus
its effort on the assessment of major hazards associated with accidental anhydrous HF releases
during manufacture and use.

        EPA conducted site visits to various facilities that produce or use HF across the U.S.  These
visits provided a firsthand opportunity to obtain in-depth information about the industrial processes
involving HF, the facilities' process safety management programs, training programs, community
outreach programs, emergency preparedness and planning programs, hazard evaluation and risk
                                                                                       Paged

-------
  assessment methods, release prevention systems, and mitigation systems.  Tours of the process
  areas enabled EPA to observe process safety and HF handling techniques. Sites visited included
  facilities in the HF production industry, the petroleum refining industry, the chlorofluorocarbon
  manufacturing industry, the semiconductor industry, and the aluminum production industry.

         In the preparation of the study, the Agency consulted with contacts in Canada, the United
  Kingdom, France, the Netherlands, Sweden, and other countries to ensure that the most updated
  Information concerning international efforts to assess and manage HF was included in the report.
  EPA also had representatives from industry,  academia, and local governments review an early draft of
  the HF report (May 8,1992) to verify the technical accuracy and completeness of the information
  contained in the report.  A list of the reviewers and a summary of their comments are provided in
  Appendix II.

        The Hydrogen Fluoride Review Subcommittee of the Environmental Engineering Committee of
  the EPA Science Advisory Board (SAB) also  reviewed an early draft of the technical aspects of the HF
  report (May 8,1992). The SAB's primary suggestions included a more rigorous definition of the
  concepts of hazards, consequences, and worst-case scenarios, the development of a credible worst-
  case accidental release scenario, and the further consideration of exposure time in the dose response
  analysis. The SAB also made recommendations on the use of dispersion models as  they apply to
 various accident scenarios.7  As a result of SAB recommendations, the report was revised to clarify
 the definitions of certain concepts, to expand modeling input descriptions, to base consequence
 analysis on dose rather than on concentration, and to address the issues and limitations involved in
 developing worst-case scenarios.  A list of members of the SAB Hydrogen Fluoride Review
 Subcommittee is provided along with other technical reviewers in Exhibit 11-1 of Appendix II.

        EPA also held a public meeting on July 12, 1993 to present and discuss the preliminary
 findings of the HF report.  The meeting provided a forum for oral and written comments to be
 presented by  individual attendees. A summary of these comments is provided in Appendix III.

 1.4    Organization of the Report

        This report integrates information gathered about HF into a presentation that provides:

        »•      an overview of what HF is chemically and physically and how it reacts upon release;

        >      a picture of how, where, and  in what form and quantity HF is produced and used in
               the U.S.;

        »•      an overview of regulatory controls and industrial standards and guidelines already or
               soon to be in place to manage HF safely for protection of worker and  public health
               and safety and the environment;

        >      a characterization of specific HF industries and the processes involved in producinq
               and using HF;

       *       an evaluation of general process hazards as well as a focus on any special or unique
               hazards associated with the HF processes under consideration;

       >       a discussion of chemical process safety management and the specific HF industry
               practices in place to prevent or minimize the impact of accidental releases;

       »•       an overview of release detection and mitigation systems in place and under
               consideration in the event of an accidental  HF release;
Page 4

-------
an analysis of historical HF accidental releases to identify the causes and
consequences of releases, to determine how HF chemical and process hazards
contribute to accidental releases, and to determine how to prevent such releases;

a discussion of computer models to analyze the consequences of HF releases and a
consequence analysis performed by EPA on worst-case  accidental releases;

a discussion of emergency preparedness and planning considering both the industry
and community perspective;

identification of issues and questions that remain to be resolved; and

EPA's findings and recommendations.
                                                                        Page 5

-------
                                          ENDNOTES
  4.




  5.


  6.

  7.
 •Production by the U.S. Chemical Industry, Little Change in Chemical Output Last Year,"
 Chemical and Engineering News, June 24, 1991, pp 30-81. (372.7)

 SRI International, 1991 Directory of Chemical Producers United States of America, Menlo Park
 CA, 1991, p 704. (450)

 Seringer, Carolyn S., Du Pont Chemicals, comments from technical review of Hydrogen
 Fluoride Study, Report to Congress, Draft May 8, 1992, June 5, 1992.  (436.4)

 Memorandum, Subject:  OSC Report to the National Response Team Major Air Release of
 Hydrofluoric Acid Marathon Petroleum Company Texas City, Galveston County, Texas - October
 30 to November 1, 1987, From: Robert M.  Ryan, On-Scene Coordinator, U.S. Environmental
 Protection Agency Region IV, To: National Response Team, March 4, 1988. (370)

 Mason, R.J., Marathon Oil  Company, comments on the draft Chemical Emergency
Preparedness and Prevention Advisory on Hydrogen Fluoride, January 15, 1993. (82d)

"HP's Future is Up in the Air," Chemical Engineering, May 1990, p 39. (150)

U.S. Environmental Protection Agency, An SAB Report: Review of Hydrogen Fluoride Study-
Report to Congress, Science Advisory Board, Washington D.C., December 1992. (489.89b)
Page 6

-------
         2.  PROPERTIES AND HAZARDS OF HYDROGEN FLUORIDE
       This chapter discusses the physical and chemical properties and hazards of both anhydrous
and aqueous hydrogen fluoride, the potential health hazards posed by various routes of exposure, the
behavior of HF upon release, and the potential environmental hazards that could result from a release.

2.1     Description of Physical and Chemical  Properties

       Hydrogen fluoride (HF) is a strong inorganic acid. It can be either a colorless, corrosive liquid
or a colorless gas.1  Anhydrous HF is miscible in water.  HF in anhydrous form or in concentrated
solution fumes strongly when in contact with moisture in the atmosphere, forming a white mist. '
HF is commercially available in anhydrous form (without water) and aqueous form (in water solution).
Aqueous HF is often called hydrofluoric acid. Anhydrous HF is normally produced with a purity of 99
to 99.9 percent; aqueous HF primarily is produced commercially as a 70 percent solution, although
electronic and reagent grades of 5 to 52 percent are produced as well.4 Both forms have a sharp,
pungent odor;5 the odor threshold is 0.04 parts per million (ppm).6 A brief summary of selected
chemical and physical properties of anhydrous  and aqueous HF is presented in Exhibit 2-1.

                                        EXHIBIT 2-1
                      Physical/Chemical Properties of Hydrogen Fluoride
Property Anhydrous HF Aqueous HF (70%)
Boiling point
Melting point
Density at 25°C
Solubility in water
Vapor pressure
19.54°C
-83.55°C
0.9576 g/cm3
complete
922 mm Hg at 25°C
66.4°C
--„- 	
-69°C
1 .22 g/cm3
complete
150mmHgat25°C
      Source:
Kirk-Othmer Encyclopedia of Chemical Technology, 3rd ed., Vol. 10, 1980.
        Like other acids, aqueous HF is corrosive to a number of common industrial materials
 including iron, stainless steel, brass, glass, asbestos, concrete, and natural rubber. Aqueous HF
 corrodes steel and other metals at a high rate;7 in these cases and others,  specific materials of
 construction must be used. Carbon steels are commonly used for concentrations of aqueous HF 70
 percent and higher, while chlorobutyl rubber-lined equipment can be used for aqueous solutions up to
 70 percent.8  Equipment commonly used in an HF atmosphere such as hoses, gaskets, tanks, valves,
 pipes, and pumps must be resistant to corrosion caused by HF.

        HF is highly reactive, and in many cases, the reaction products are hazardous and may create
 dangerous situations.  In a manner similar to other concentrated inorganic acids, HF reacts with
 sulfides and cyanides generating the toxic gases hydrogen sulfide and hydrogen cyanide,
 respectively.  Reaction of HF with glass, concrete, and other silicon-bearing materials yields silicon
                                                                                      Page 7

-------
 tetrafluoride gas, and reaction with a number of common metals, such as steel, yields flammable
 hydrogen gas (HF itself is nonflammable).  Contact of HF with alkalies and some oxides can cause
 strong exothermic reactions. Adding relatively small amounts of water to either anhydrous HF or
 concentrated solutions of aqueous HF results in a violent reaction that produces heat and can cause
 spattering of the material;  however, in large excess, water can be an effective mitigation agent
 because of heat absorbing  properties, provided it is added promptly after the spill or release.9

 2.2    Health Hazards

        Exposure to HF can cause injury through  inhalation, direct contact, or ingestion.10
 Repeated ingestion of HF can cause increased bone and joint density typical of fluorosis or chronic
 fluoride poisoning.11  Acute exposure to HF will result in irritation, burns, ulcerous lesions, and
 localized destruction  of the tissues (necrosis) of the eyes, skin, and mucous membranes.12
 Concentrations of HF above 25 ppm in air can cause eye irritation; at 20 to 30 ppm, a reflex breathing
 difficulty is additional indication of the chemical's presence.13  HF is not considered to be
 carcinogenic.

        While acute exposure to high concentrations of HF can cause severe health effects and even
 death, one study indicates that individuals surviving such exposures do not suffer long-term effects.
 The health of seven workers who survived exposure to high concentrations of HF (approximately
 10,000 ppm) for several minutes in an industrial accident in Mexico was examined periodically for up
 to 11 years following  the accident to evaluate any long-term effects. Long-term effects were defined
 as illnesses or lesions that do not show any immediate symptoms or signs, but instead appear after a
 period of time, ranging from months to years, after the exposure.  The study looked for such effects as
 cancer, mutations, fluorosis, and neurological disorders.  Although the effects of the exposure were
 life-threatening, no long-term delayed effects were observed.  Particular attention was paid to possible
 long-term effects on the lungs from inhalation of high  concentrations of HF; however, tests showed no
 changes in lung function of the workers studied, other than changes that would occur during normal
 aging.15

        Another study, by the Galveston County Health District and the University of Texas Medical
 Branch, found indications of lingering disease symptoms, especially breathing problems, two years
 after exposure to HF during  the  Marathon Oil Refinery incident in Texas City on October 30, 1987.
 The Galveston  study was based on a sample of 2,000 people, including all highly exposed individuals
 and some of the individuals with intermediate, negligible and unknown exposure levels. Subjects were
 interviewed once after the release as part of the exposure study and again for the symptom  and
 disease prevalence study. The prevalence of severe symptoms two years after the release was
 significantly lower than it had been in the month after the release, but 24 percent of the highly
 exposed group still reported difficulty in breathing and sleep interruption due to headaches.  Some
 still reported  eye and  skin irritations. As a group, the highly exposed individuals reported more bone
 symptoms, which are  a  known systemic effect of fluoride exposure, than their less exposed
 counterparts.16

        Questions remain in  the medical and industrial communities regarding the validity of the
 Galveston study.  Lack of knowledge about the patients' prior medical histories to provide a medical
 baseline, disagreement regarding suitable definitions (e.g., severe exposure), and discrepancy over
 the causes of the eye and skin irritation are problems yet to be resolved.  The actual exposure levels
 and duration  are also  not known.  These were not  documented at the time of the incident,  and it is
 difficult to obtain reliable information from personal surveys taken two years after an incident.17
Page 8

-------
       2.2.1  Inhalation

       Inhalation of HF is particularly hazardous because HF readily dissolves in the mucous
membranes of the upper respiratory tract, nose, and throat.18 Dose, a function of the length of time
of exposure and the concentration to which one is exposed, is important in determining type and
amount of tissue damage incurred. Mild exposure to HF vapor can cause respiratory system irritation.
Respiratory exposure to high concentrations of HF characteristically results in disintegration of the
tissues of the upper respiratory system (ulcerative tracheobronchitis) and accumulation of blood in the
lungs (hemorrhagic pulmonary edema).19 Symptoms may include coughing, choking, chills, chest
tightness, fever, and bluish discoloration of the skin  due to lack of oxygen in the blood (cyanosis).
Severe exposure also can result in other systemic effects such as depletion of calcium levels
(hypocalcemia), if not treated promptly.20

       Various inhalation exposure guidelines, based primarily on health effects, have been
developed for chemicals like HF that are considered to be health hazards.  Some of these guidelines
are discussed in Appendix IV. These guidelines are intended to provide an exposure threshold;
however, actual health effects will vary from individual to individual based on various factors such as
age, health condition, etc. Exhibit 2-2 presents two  guideline levels for HF, the Immediately
Dangerous to Life or Health (IDLH) level developed by the National Institute for Occupational Safety
and Health (NIOSH) and the Emergency Response  Planning Guideline-3 (ERPG-3) developed by the
American Industrial Hygiene Association (AIHA).  Exhibit 2-2 also shows the IDLH  and ERPG-3 for
several common toxic substances for comparison. HF is clearly a chemical of concern, with an IDLH
of 30 parts per million (ppm) for 30 minutes and an  ERPG-3 of 50 ppm for 60 minutes.  However,
there are more toxic chemicals like phosgene  and less toxic chemicals like ammonia.

       For planning purposes, exposure guidelines, such as the IDLH and ERPG-3, are sometimes
used in conjunction with air dispersion modeling techniques to assess the potential consequences of
a release of a toxic vapor. Dispersion models are used to develop estimates of the concentration of
the vapor as a function of time, location, and distance from the point of release.  The exposure
guideline levels can be used as threshold concentrations or to determine dose levels to  estimate
areas in  which people exposed to the toxic vapor might be expected to be at risk. The IDLH is
defined for an exposure of 30 minutes, while the ERPG-3 is defined for an exposure time of 60
minutes.  In cases where the duration of exposure might be expected to be significantly shorter or
longer, these concentration levels might not be appropriate.

       Another approach to estimating potential effects areas uses "probit" (probability unit) analysis,
based on experimental animal lethality data, to estimate the percentages of humans affected as a
function  of concentration and time. The probit method is a useful tool; however, it is subject to the
same uncertainties as other analytical methods that use tcxicity data.  The quantity and  quality of
experimental data available vary by chemical,  making it difficult to compare toxicity. The experimental
animal data upon which equation coefficients  are based may vary between animal species, and the
correlation between animal and human responses may vary greatly from substance to substance;
therefore, there is likely to be uncertainty in applying probit equations. If animal data are not available
over a wide range of exposure periods, the probit equation might be particularly uncertain when
applied to exposures of much shorter or longer duration than the reported experimental exposure
times. Several different probit equations have been developed for HF.21'22'23'24  Depending on the
equation chosen, different results can be obtained for a given concentration and exposure duration.
As noted above, the IDLH and ERPG-3 are guideline concentration levels and are not intended to
represent potentially lethal concentrations. The results of probit analysis based on several different
equations are consistent with the definitions of these guideline levels,  indicating that concentrations
equal to the IDLH and ERPG-3 levels would not be  sufficient to cause a one percent fatality  rate in a
population exposed for one hour. Probit equations for HF are discussed in more detail  in Appendix V.
                                                                                        Page 9

-------
                                              EXHIBIT 2-2
                          Exposure Guidelines For Several Toxic Substances
                                            Compared to HF
           Lower concentration indicates higher concern, greater toxicity; higher concentration indicates
                                        lower concern, lower toxicity.
             Chemtcal Name
   IDLH
 (30 minutes)
ERPG-3
(1 hour)
             Phosgene
             Sulfuric Acid
2 ppm
(8 mg/nr)

20ppm
(80 mg/nrr )
1 ppm
(4 mg/m3)

7 ppm
(30 mg/m3*)
Hydrogen Fluoride
30 ppm
(25 mg/m3)
50 ppm
(41 mg/m3)
             Chlorine
             Hydrogen Chloride
             Sulfur Dioxide
             Ammonia
30 ppm
(87 mg/m3)

100 ppm
(149 mg/m3)

100 ppm
(262 mg/m3)

500 ppm
(348 mg/m3)
20 ppm
(58 mg/m3)

100 ppm
(149 mg/m3)

15 ppm
(39 mg/m3)

1000 ppm
(695 mg/m3)
     >       IDLH. The Immediately Dangerous to Life or Health (IDLH) level, developed by the
             National Institute for Occupational Safety and Health (NIOSH), represents the maximum
             concentration from which one could escape within 30 minutes without any escape-
             impairing symptoms or any irreversible health effects. (See the NIOSH Pocket Guide to
             Chemical Hazards)

     »•       ERPG.  Emergency Response Planning Guidelines (ERPGs) have been developed for a
             limited number of chemicals by the American Industrial  Hygiene Association (AIHA). The
             ERPGs are based primarily on acute toxicity data and possible long-term effects from
             short-term exposure.

             The ERPG-3 is defined as the maximum concentration in air below which nearly all people
             could be exposed for one hour without life-threatening health effects.

       Normally listed in mg/m3 rather than ppm. The likelihood of sulfuric acid vapor exposure is low due to very low vapor I
     pressure. Exposure levels are expressed in mg/m3 to account for the more likely acid mist (particulate) exposure route.

             Sources: National Institute for Occupational Safety and Health
                    American Industrial Hygiene Association
Page 10

-------
       2.2.2 Contact with Skin and Eyes

       Liquid HF can severely burn skin and eyes. Anhydrous HF gas or the fumes from
concentrated aqueous HF can also burn these tissues.   Like many other acids, liquid HF should
initially be diluted and rinsed from the skin surface with large quantities of water. However, additional
treatment is needed for direct contact by large amounts of concentrated HF because it may penetrate
the skin and tissue until it is neutralized  by reaction with calcium available in blood and body tissue or
a medically-introduced subcutaneous source of calcium.26  Industry experience regarding minor
exposure to HF vapor has been that a "mild" sunburn effect may develop on exposed skin.
          27
       Skin contact with anhydrous HF or solutions above 50 percent produce immediate pain and
tissue damage; exposure to solutions of 20 to 50 percent HF results in pain and reddening of the skin
that may be delayed one to eight hours, while reactions to more dilute solutions may be delayed up to
24 hours.28'29  The fluoride ion can penetrate skin and attack underlying tissues and bone. The
pain is said to be excruciating and unusually persistent. Healing often is delayed, and tissue
destruction (necrotic changes) may continue to occur beneath a layer of tough coagulated tissue to
produce deep penetrating ulcers.36 Hypocalcemia and other systemic effects can result from large
burns (over 25 square inches), and these effects may be fatal if proper medical treatment is not
obtained.31

       Both liquid and gaseous HF can cause severe irritation and deep-seated burns on contact
with eyes or lids.32 Corneal burns and conjunctivitis are common symptoms of exposure.  If not
treated immediately, permanent damage or blindness may result from direct contact.33'34
Solutions as dilute as 2 percent or lower may cause skin burns or eye irritation.
35
       2.2.3 Ingestion

       If ingested, HF can cause immediate and severe mouth, throat, and stomach burns.36 Even
small amounts and dilute solutions can lead to fatal hypocalcemia unless medical treatment is
initiated.37

       2.2.4 Recommended Medical Treatments

       Burns resulting from HF contact with skin, eyes, or mucous membranes require immediate and
specialized first aid and medical treatment from trained personnel.  This treatment differs from the
treatment of burns from other acids.  If untreated or improperly treated, permanent damage, disability,
or death may result. Treatment may involve introducing an agent to react with the fluoride ion and
prevent further or continuing tissue destruction. For skin contact, Du Pont, a major manufacturer of
HF, recommends five minutes of flushing followed by calcium gluconate treatment applied as a gel or
injection of a 5 percent solution.38 Calcium gluconate complexes with the fluoride ion to form an
insoluble product. Another form of treatment, as recommended by Allied-Signal, another major
manufacturer of HF, is prolonged  soaking in quaternary ammonium compound solution.39
Treatment by quaternary ammonium compounds has been recommended for topical dermal
treatment, but treatment by topical calcium gluconate gel is most commonly used.  Calcium gluconate
may be injected or given intravenously to treat more extensive dermal exposures.40 If HF is
ingested, the stomach may need to be lavaged with lime water.  Severe exposure to  HF by any route
can lower serum calcium levels (hypocalcemia) and can be treated intravenously with calcium
gluconate.41'42  Recommended first aid treatment of exposure to HF by inhalation is similar to the
treatment recommended for many other toxic gases and vapors. First aid recommendations include
immediately moving the victim to fresh air and getting medical attention; keeping the  victim warm,
quiet, and lying down; starting artificial respiration if breathing has stopped; and having oxygen
administered by a trained attendant.43 Promptness in administering treatment for exposure to HF  is
crucial. The medical treatment recommended by Allied-Signal is presented in Appendix VI as an
                                                                                     Page 11

-------
 example of the approach one company has taken to address the special concerns associated with the
 treatment of HF exposures.  Elf Atochem, also a manufacturer of HF, has contracted with a nationally
 prominent Poison Control Center to provide occupational health consultation following HF exposure
 Incidents.  This allows attending physicians to have instant access to specialized treatment
 protocols.44

 2.3     Environmental Hazards

        HF may be toxic to aquatic and terrestrial life, with the effect depending on the exposure
 concentration.  If HF was released to the environment in sufficient concentrations, the fluoride ions in
 the water could be toxic to surrounding plants and animals, while airborne HF in a vapor cloud could
 bum both plant and  animal tissue. Whether released to water, air, or land, HF does not  biodegrade.
 Calcium present in large enough quantities in soil or water will form an insoluble solid with the fluoride
 Ion,  removing it as an immediate environmental hazard.  Dilution or natural buffering capacities of soils
 or water will reduce the increased acidity created by the release of HF.

        HF is highly soluble in water. Fluoride ions, readily available in aqueous HF, were found to be
 lethal to fresh water fish at 60 milligrams per liter (mg/L). Fluoride ions are harmful to many other
 species of fish at concentrations of 40 mg/L and below.  Other more sensitive aquatic life are affected
 at levels as low as 10 mg/L. An aquatic toxicity rating for HF has not been assigned. According to a
 Canadian study, concentrations of fluoride equal to or exceeding  1.5 mg/L constitute a hazard in the
 marine environment,  while levels less than 0.5 mg/L present minimal risk of deleterious effects.45

        It should also be noted  that fluoride is added to drinking water to help prevent tooth
 decay.48  The Centers for Disease Control (CDC) recommend that communities fluoridate their
 drinking water systems at the optimum fluoride level, i.e., the level that results in the least staining
 and/or pitting of developing teeth (dental fluorosis) and the maximum reduction in dental decay. The
 CDC have defined the optimum fluoride level in drinking water as 0.7 mg/L to 1.2 mg/L.  EPA's
 standard Is different from the CDC recommendation because EPA standards are based on health
 effects, and the Agency considers dental fluorosis to be a cosmetic effect, not an adverse health
 effect.   Excessive amounts of fluoride can also lead to crippling skeletal fluorosis, however, which is a
 health effect.47  To reduce the risk of skeletal fluorosis, EPA has established a maximum
 contaminant level (MCL) for fluoride in drinking water at 4.0 mg/L. EPA set a Secondary  Maximum
 Contaminant Level (SMCL) at 2.0 mg/L to protect against dental fluorosis. The MCL is an enforceable
 standard that requires a system to install one  of the identified best technologies generally available,
 while the SMCL is a nonenforceable goal that requires a system that exceeds the level to give public
 notification.48

        Gaseous HF  can directly attack plant foliage, especially if  present in high concentrations.  In
 tow concentrations, HF is absorbed by the leaves. The most apparent effect of fluoride on vegetation
 Is necrosis or tip burn, but exposure to fluoride in sufficient quantities also may result in growth
 abnormalities or a decrease in reproductivity in both plants and animals.  Livestock that drink fluoride-
 contaminated water or eat contaminated foliage may have dental lesions, bone overgrowth, lameness,
 loss of appetite, a decrease in milk production, and reduced reproductivity.49

 2.4     Release Characteristics

        The behavior of HF in the event of a release depends on a variety of factors, including the
 conditions of the release and the atmospheric conditions. If HF is superheated and released under
 pressure, ft will form a cloud of HF vapor and aerosol, which reacts  readily with water vapor in the air.
 If HF vapor Is released directly or vaporizes from a liquid pool,  a visible cloud is often formed because
 of the reaction with moisture in the air.  Anhydrous HF can be  released as a vapor or as  a
 combination of vapor and  liquid droplets.  HF spilled as a liquid will evaporate at a rate that depends
Page 72

-------
on release temperature and atmospheric conditions. HF will vaporize quickly from a pool for the first
few minutes as heat is conductively transferred to the pool surface. The pool temperature will drop as
HF vaporizes causing a corresponding decrease in evaporation rate.50  Aqueous HF can also be
released from various operations and depending upon  release temperature and  concentration, can
vaporize and form dense white fumes.

        Anhydrous HF boils at approximately 20°C and is very soluble in water.  Because of its
volatility and low boiling point, it can be readily vaporized to form a vapor cloud.  Both anhydrous HF
and aqueous HF with a concentration greater than about 40 percent will react with the moisture in air
to produce white fumes.51 This reaction with moisture produces heat (exothermic reaction).  The
fumes have a pungent odor and are extremely  irritating if inhaled or contacted.52 Depending on the
size of the vapor cloud and prevailing meteorological conditions, an HF release could pose a severe
hazard to facility personnel and the nearby public. The visibility of the white fumes can permit a
process operator to detect small leaks and spills quickly and take action to prevent them from
worsening;53 unfortunately, HF fumes cannot easily be differentiated visually from common steam
leaks.

        Molecules of HF liquid and gas form hydrogen bonds (i.e., the hydrogen atom in one HF
molecule forms a bond with the fluorine atom in another HF molecule) to produce variable length
chains or polymers up to (HF)8 at ambient temperatures. At higher temperatures, however, single HF
molecules may exist.  HF liquid consists primarily of HF hexamer (HF)6.  The properties of HF vary
from what might be expected because of the hydrogen bonding; the density of HF vapor is greater
than would be expected,  and HF is likely to form vapor clouds that  are heavier than air (i.e., it behaves
as a dense gas) and travel at ground level following a release,  experiencing both gravity spreading
and turbulent flow. As the dense gas cloud mixes with air, the HF dissociates to the HF monomer, a
process that absorbs heat (endothermic reaction) and cools the cloud.  Evaporation of droplets of HF
aerosol in the cloud (see  below) also contributes to cooling effects. The cooling of the cloud
increases its density.  As  moist air is mixed into the cloud, HF reacts with the moisture to form
aqueous HF,  releasing heat, warming the cloud, increasing its buoyancy, and decreasing its  density.
Thus, these processes can lead to a cloud that can be either neutrally or positively buoyant (i.e., the
same density as air or lighter than air) depending on atmospheric conditions such as temperature and
humidity of the air and the rate of mixing between air and the HF cloud  itself.  Eventually, the cloud
becomes buoyant, dispersing vertically as well as horizontally in the atmosphere.54'55

        An HF release may, under certain conditions (i.e., superheated and released under pressure),
lead to aerosol formation  which  is a suspension of fine liquid particles in a vapor cloud.  Based on
spill tests, a release of gas liquefied under pressure could form a cloud  containing both HF vapor and
HF aerosol.56  Because liquid particles are airborne, aerosol formation adds greatly to the quantity of
HF contained in the cloud and thus adds to the hazards posed to workers and to the public. Aerosol
formation is not unique to HF. It depends both on the chemical and on the conditions of the release.
For example,  any gas liquefied under pressure, which flashes to a gas upon release, may carry liquid
with it as a fine spray or aerosol. The aerosolization properties of HF, H2SO4, and ammonia  (NH3)
have been investigated in periodic spill test studies.57'58 Similar to HF, releases of liquefied NH3
under pressure have resulted in clouds containing as much as 80 percent aerosol droplets of NH3.59
Spill tests for aerosol formation of sulfuric acid and sulfuric acid/isobutane mixtures have been
conducted, with test results indicating that a release of sulfuric acid under typical petroleum refinery
alkylation conditions would not form an aerosol.60

        During the summer of 1986, Amoco Oil  Company and Lawrence Livermore National Laboratory
conducted a series of six experiments involving atmospheric releases of HF. The studies, known as
the Goldfish test series, were conducted at the  Department of Energy Liquefied Gaseous Fuels Spill
Test Facility in Frenchman's Flats,  Nevada. In these tests, HF was released at a  temperature of 40°C
and a pressure of 110 to  120 pounds per square inch (psi)61 (conditions approximating petroleum
                                                                                      Page 13

-------
refinery HF alkylation unit operating parameters).  Upon release, the HF formed a cloud of vapor
(approximately 20 percent of the HF) and HF/water vapor aerosol (approximately 80 percent of the
HF) which traveled downwind as a dense gas.62 The tests were conducted under desert conditions;
therefore, the heat effects caused by reaction of HF with moisture in the air were probably smaller
than they would be in locations with higher humidity.  In an area of higher humidity, the cloud may be
heated more because of the reaction of HF with moisture in the air; additional water,  however,
reduces the volatility of the HF/water droplets which tends to keep the cloud dense longer.63
Page 14

-------
                                        ENDNOTES
1.      Growl, Daniel A., Wayne State University, comments from technical review of Hydrogen
       Fluoride Study, Report to Congress, Draft May 8, 1992, May 27, 1992. (133.3)

2.      Mark, Herman F., Donald F. Othmer, Charles G. Overberger, and Glenn T. Seaborg, eds., Kirk-
       Othmer Encyclopedia of Chemical Technology, Third Edition, Volume 10, John Wiley and
       Sons, New York, 1980. (285)

3.      Du Pont Chemicals, Hydrofluoric Acid, Anhydrous - Technical: Properties, Uses, Storage and
       Handling, Wilmington, DE. (137.5)

4.      Du Pont Chemicals, Material Safety Data Sheet (MSDS), Hydrofluoric Acid - Anhydrous,
       Wilmington, DE, June 1991. (138)

5.      Allied Signal, information from fact sheets.

6.      U.S.  Environmental Protection Agency, Reference Guide to Odor Thresholds for Hazardous Air
       Pollutants Listed in the Clean Air Act Amendments of 1990, Air Risk Information Support
       Center, Washington, DC, Document Number: EPA/600/R-92/047, March 1992. (139.54)

7.      Du Pont Chemicals, Hydrofluoric Acid, Anhydrous.

8.      Seringer, Carolyn S., Du Pont Chemicals, comments from technical review of Hydrogen
       Fluoride Study, Report to Congress, Draft May 8, 1992, June 5, 1992. (436.4)

9.      Hague, William J., Allied-Signal, comments from technical review of Hydrogen Fluoride Study,
       Report to Congress, Draft May 8, 1992,  June 10, 1992. (153)

10.    Mark, Herman F.

11.    Gosselin, R.E., R.P. Smith, H.C. Hodge, and J.E. Braddock, Clinical Toxicology of Commercial
       Products, Williams & Wilkins, Baltimore, 1987, pp 190-193. (145)

12.    U.S.  Environmental Protection Agency, EPA Chemical Profiles for Extremely Hazardous
       Substances, Emergency First Aid Treatment Guide for Hydrogen Fluoride, 1987. (139.5)

13.    Phillips Petroleum Company, comments from technical review of Hydrogen Fluoride Study,
       Report to Congress,  Draft May 8, 1992.  (370.92)

14.    Allied Signal, information from fact sheets.

15.    Trevino, Miguel A., Hydrofluoric Acid Exposures Long-Term Effects, October 1991 Draft. (485)

16.    Galveston County Health District and University of Texas Medical Branch, Hydrofluoric Acid
       Spill Report of the Exposure and Symptom and Disease Prevalence Studies,  Agency for Toxic
       Substances and Disease Registry, U.S. Department of Health and Human Services, Texas
       City, TX, Decembers, 1991. (140.5)

17.    Hague, William J., Allied-Signal.
                                                                                    Page 15

-------
 18.     Mudan, Krishna S., Acute Inhalation Toxicity of Hydrogen Fluoride, AlChE Summer Annual
        Meeting, Philadelphia, August 24-26,1989. (360)

 19.     Gosselin, R.E.

 20.     Allied Signal, information from fact sheets.

 21.     Mudan, Krishna S.

 22.     Perry, W.W., and W.P. Articola, Study to Modify the Vulnerability Model of the Risk Management
        System, Prepared for U.S. Department of Transportation, U.S. Coast Guard, Washington, D.C.,
        1980, No. CG-D-22-80 (489.91)

 23.     de Weger, D.E., C.M. Pieterson, and P.G.J. Reuzel, "Consequences of Exposure to Toxic
        Gases Following Industrial Disasters," Journal of Loss Prevention in the Process Industries, Vol.
        4, 1991 (July), p 272. (502)

 24.     ten Berge, W.F., A. Zwart, and L.M. Appelman, "Concentration-Time Mortality Response
        Relationship of Irritant and Systematically Acting Vapours and Gases," Journal of Hazardous
        Materials, Vol. 13,1986, p 301. (54.5)

 25.     Allied Signal, information from fact sheets.

 26.     Phillips Petroleum Company.

 27.     Phillips Petroleum Company.

 28.     National Library of Medicine (NLM), Hazardous Substances Databank (HSDB), on-line
        database, July 1992. (361.3)

 29.     Hague, William J., Allied-Signal.

 30.     Gosselin, R.E.

 31.     Allied Signal, information from fact sheets.

 32.     Joseph, Eileen Zola, ed., Chemical Safety Data Guide, The Bureau of National Affairs,
        Washington; DC, 1985, pp 475-476.

 33.     Joseph, Eileen Zola.

 34.     U.S. Environmental Protection Agency, EPA Chemical Profiles for Extremely Hazardous
        Substances.

 35.     Allied Signal, information from fact sheets.

 36.     Joseph, Eileen Zola.

 37.     Allied Signal, information from fact sheets.

 38.     Seringer, Carolyn S., Du Pont Chemicals.

 39.     Hague, William J., Allied-Signal.
Page 16

-------
40.    Rodgers, George C., University of Louisville, comments from technical review of Hydrogen
       Fluoride Study, Report to Congress, Draft May 8, 1992, June 5,1992.  (424.9)

41.    Allied Signal, information from fact sheets.

42.    Rodgers, George C., University of Louisville.

43.    Allied Signal, information from fact sheets.

44.    Laumer, John and Harold Lamb, Elf Atochem North America Inc., comments from technical
       review of Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992, June 4, 1992.
       (290.8)

45.    Environment Canada Environmental Protection Sen/ice, Hydrogen Fluoride and Hydrofluoric
       Acid, Environmental and. Technical Information for Problem Spills, Technical Services Branch,
       Ottawa, Ontario, July 1984. (220)

46.    U.S. Environmental Protection Agency, Fluoride Synopsis, Office of Ground Water and Drinking
       Water, Drinking Water Standards Division, Washington, D.C., July 22, 1992. (139.4)

47.    U.S. Environmental Protection Agency, Fluoride Synopsis.

48.    U.S. Environmental Protection Agency, "National Primary  and Secondary Drinking Water
       Regulations; Fluoride," 40 CFR Parts 141-143, {Volume 51, Number 63, "Rules and
       Regulations:  Part II}, April 2, 1986. (139.57)

49.    Environment Canada Environmental Protection Sen/ice.

50.    Hague, William J., Allied-Signal.

51.    Los Angeles County, Department of Health Services, Toxics Epidemiology Program, Health
       Effects Due to Hydrogen Fluoride Inhalation: A Literature Review,  1989. Prepared for the
       Hydrogen Fluoride Task Force of the South Coast Air Quality Management District. (325)

52.    Allied Signal,  information from fact sheets.

53.    Du Pont Chemicals, Hydrofluoric Acid, Anhydrous.

54.    Seringer, Carolyn S., Du Pont Chemicals.

55.    Phillips Petroleum Company.

56.    HF Mitigation Water Spray Project, created for a meeting  of the American Petroleum Institute
       and given to the U.S. Environmental Protection Agency, December 10,1991. (424.34ABC)

57.    Model Study of Sulfuric Acid Aerosol Formation for Phillips Petroleum Company. Prepared by
       Quest Consultants Inc., Norman, Oklahoma.  (340)

58.    HF Mitigation Water Spray Project.

59.    Kaiser, Geoffrey D., "A Review of Models for Predicting the Dispersion of Ammonia in the
       Atmosphere," Plant/Operations Progress, AlChE, Volume  8, Number 1, January 1989, p 58-64.
       (284.3)
                                                                                     Page 17

-------
 60.     Quest Consultants, Inc., Sulfuric Acid Aerosoling, Norman, Oklahoma, October 12, 1992.
        Document Number: 92-10-6065. (395)

 61.     Blewitt, D.N., J.F. Yohn, R.P. Koopman, and T.C. Brown, "Conduct of Anhydrous Hydrofluoric
        Acid Spill Experiments," International Conference on Vapor Cloud Modeling. (60)

 62.     Blewitt, D.N.

 63.     Koopman, Ronald  P., Lawrence Livermore National Laboratory, comments from technical
        review of Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992, June 3, 1992. (289)
Page 18

-------
       3.  CHARACTERIZATION OF HYDROGEN FLUORIDE INDUSTRY
       This chapter provides a brief overview and characterization of the industries that produce and
use HF, including production and use data and a brief discussion of the market outlook for HF.  The
quantities of HF produced or used by various industry segments should provide some perspective for
considering the discussion of hazards in subsequent chapters.
3.1
Production of HF
       U.S. total annual capacity for HF production was estimated to be approximately 206,000 tons
as of April 15, 1992, while non-U.S. North American capacity was approximately 103,000 tons as of
mid-1991.1 The HF industry has been operating at about 90 percent of capacity.  Additionally, three
North American aluminum producers make 60,000 tons of HF gas annually for their own on-site use.
Exhibit 3-1 presents U.S.  producers of HF for the commercial market. Exhibit 3-2 lists other North
American producers.  HF produced by the aluminum producers is not included in these exhibits
because the HF is not stored or available for the commercial market. It is produced as a result of the
aluminum manufacturing  process and then used immediately on-site.

       North American production is divided between Allied-Signal Inc., E.I. du Pont de Nemours &
Co., Elf Atochem North America, Inc., and a series of Mexican producers.3 Allied-Signal Inc. is the
largest HF producer in North America with capacity of 105,000 tons at its Geismar, Louisiana plant.
Du Pont is the second largest producer with a 75,000 ton plant in La Porte, Texas. Elf Atochem North
America,  Inc.  has a 26,000 ton plant in Calvert City, Kentucky.4  The Calvert City plant, which formerly
belonged to Pennwalt Corporation, was merged into Elf Atochem in  1990.5 The vast majority of HF
produced at Calvert City  is used captively to produce a variety of fluorochemicals and fluoropolymers.
A small amount of hydrogen fluoride enters the  merchant market.6 Alcoa of the U.S., Alcan of
Canada (Canadian producer has since ceased  production), and Industrias Quimicas de Mexico
produce and  use 60,000  tons of HF gas captively as an intermediate in producing aluminum fluoride
for aluminum production.7

       Western Europe and Japan also produce significant quantities of HF. Japan alone, for
example, has seven companies that have combined production capacities of 97,900 tons as of 1988,
while annual production capacity for Western Europe as a whole was 386,100 tons as of
January 1, 1990.8 France is the largest producer in Western Europe with 105,000 tons, followed by
Germany with 93,500 tons, and the United  Kingdom with 79,200 tons.  Italy, Spain, Greece and the
Netherlands also contributed to the Western European total stated above.  Other countries, such as
Finland and Sweden, produced less than 11,000 tons annually in the late 1980's. Trade data indicate
that the following countries exported less than 9,900 tons annually from 1987-1990:  Switzerland, the
Netherlands,  United Kingdom,  Portugal, Belgium, Luxembourg, South Korea, Irish Republic, Denmark,
Austria, Malaysia Federation, Taiwan, and Norway.9

       Imports of HF to  the United States  rose from 98,100 tons in  1980 to 130,000 tons in 1989,
according to  U.S. Department  of Commerce data.10'11   Imports for 1991 totaled 104,900 tons. Of
this total, the U.S. received 71,100 tons (68 percent) from Mexico, 18,700 tons (18 percent) from
Canada (Canadian producer has since ceased  production), 9,800 tons (9 percent) from Kenya, 4,200
(4 percent) from China and about 1 percent from four other Countries.12'13 (NOTE:  The quantities
listed for Kenya and China may reflect imports of fluorspar)""
                                              14
        In 1991, the U.S. exported approximately 9,000 tons of HF. Of this total, about 3,700 tons (42
 percent) were exported to Canada, 3,100 tons (34 percent) were tsxported to Mexico, 1,100 (13
 percent) to Venezuela, 226 (3 percent) to Taiwan, 200 (2 percent) to South Korea, and less than 105
                                                                                    Page 19

-------
  tons each (1  percent or less) to 24 other countries.15 Data regarding imports and exports of
  Individual HF producers were not found in the available literature.
                                             EXHIBIT 3-1
                                U.S. HF Production Capacity for 1992
Manufacturers
(Primary producers}
Allied-Signal Inc.
Engineered Materials Sector
Du Pont Company
Du Pont Chemicals
Site Location
Geismar, LA
La Porte, TX
Annual Capacity
(Thousands of tons)
105
75
      Elf Atochem North America, Inc.
        Fluorine Chemicals Division          Calvert City, KY

                                            TOTAL

      Source: SRI International estimates as of April 15,1992.
                                             EXHIBIT 3-2
                        Non-U.S. North American HF Producers and Shippers
      Company

      Fluorex

      Industrias Quimicas de Mexico

      Qulmica Fluor

      Quimibasicos
                       Site Location

                       Ciudad Juarez, Mexico

                       San Luis Potosi, Mexico

                       Matamoros, Mexico

                       Monterey, Mexico

                       TOTAL
Annual Capacity
(thousands of tons)
             Note:
Alcoa (U.S.), Alcan (Canada), and Industrias Quimicas de Mexico produce
60,000 tons of HF gas annually for captive use In the production of
aluminum fluoride for aluminum production.  In addition, according to the
Chemical Manufacturers Association HF Panel's comments on the Draft HF
Report, June 5,1992, another small Mexican company, Campanera Minera
LaValenclana (CMV) in Torreon, Mexico, appears to be producing HF, with
an annual production capacity of 6,000 tons.
     SourcBi  SRI International estimate as of mid-1991.
                                                                                  103
Page 20

-------
3.2    Uses of HF

       HF is the source of fluorine for most fluorine-containing chemicals.  It is either used directly in
the manufacture of such chemicals or in the production of intermediates for their manufacture. HF is
used to manufacture a wide variety of products, including refrigerants, gasoline, electronic
components, aluminum, and plastics.16 HF is used as a reactant or fluorinating source in the
manufacture of fabric and fiber treating agents, herbicides, pharmaceutical intermediates, inert
fluorinated liquids, and electronic grade etchants.17 Stannous fluoride, used in toothpaste, is
manufactured using HF. HF lasers have been tested for use in corneal transplants  and for use in
space.19 While the majority of HF used by industry is in the anhydrous or 100 percent form,
aqueous HF solutions with concentrations of 70 percent and lower are used in stainless steel ptekling,
metal coatings, chemical milling, glass etching, exotic metals extraction, and quartz purification.
See Exhibit 3-3 for some examples of uses of HF.

        Under section 313 of the Emergency Planning and Community Right-to-Know Act (Title III of
the Superfund Amendments and Reauthorization Act of 1986), facilities manufacturing, processing, or
otherwise using HF must report to EPA's Toxic Release Inventory  (TRI) if the quantity of HF
manufactured, processed, or used annually exceeds an established threshold. Facilities must report
the quantities of both routine and accidental releases of listed TRI chemicals as well as the amount
contained in wastes transferred off-site. The TRI reports exclude all non-manufacturing facilities and
those manufacturers with fewer than 10 employees.  In 1990, when the threshold for manufacturing or
processing was 25,000 pounds and the threshold for otherwise using was 10,000 pounds, a total of
531 facilities reported to the TRI for HF. These facilities represented a variety of industries, and
included Government facilities (e.g., Department of Energy facilities).  Facilities in the chemical,
primary metals, fabricated metals, and  electronic equipment industries each made up about 18 to 20
percent of the total.  Petroleum refiners accounted for 11 percent.  Other facilities reporting included a
number in the transportation equipment industry and stone, glass, clay, and concrete industries.
Exhibit 3-4 presents  the types of facilities reporting to TRI.  Many of these facilities, particularly those in
the metals and electronics industries, probably use aqueous rather than anhydrous HF.  Many also
report relatively small maximum on-site quantities.  About 80 percent of the facilities reported
maximum on-site quantities of less than 100,000 pounds.  Facilities reporting 100,000 pounds or more
were primarily chemical companies and refiners; some primary rnetal  companies and others also
reported quantities of 100,000 pounds or more. Exhibit 3-4 shows the distribution of facilities reporting
maximum on-site quantities of 100,000 pounds or more.

        According to the 1990 TRI data, the total of annual quantities of HF emitted, both accidentally
 and routinely, to the environment or waste transferred from reporting facilities was about 12,660,000
 pounds (6,330 tons). Appendix VII lists all facilities reporting to the TRI for HF in 1990, with maximum
 on-site quantity ranges and quantities  released. The released quantities reported include fugitive or
 point emissions to air, discharges to receiving streams or water bodies, underground injection on-site,
 releases to land on-site, discharges to POTW, and other transfers in waste to off-site locations.

        In 1991, HF  end uses were as follows (see Exhibit 3-5):

                K      63 percent (152,000 tons) as fluorocarbons,
                >.      7 percent (16,000 tons) as alkylation catalyst  for gasoline,
                ».      5 percent (13,000 tons) for nuclear applications (uranium),
                ».      3 percent (8,000 tons) purchased on the merchant market consumed in the
                       aluminum industry to produce aluminum fluoride,
                >.      22 percent (52,000 tons) in stainless steel pickling, various chemical
                       derivatives and products,  electronics, specialty metal production, and other
                       uses.21
                                                                                        Page 21

-------
                                                                   EXHIBIT 3-3
                                                   Examples of Hydrogen Fluoride Uses
ANHYDROUS HF
   •FLUORINE-
                  OXYGEN DIFLUORIDE -*• ROCKETPHOPELLANTS
                  RHENIUM HEXAFLUORIDE —•• RHENIUM METAL
                  TUNGSTEN HEXAFLUORIDE -»• TUNGSTEN METAL
                  SULFUR HEXAFLUORIDE -*• DUELECTRtO FOR ELECTRICAL AND ELECTRONIC EQUIPMENT
                  BROMINE FLUORIDE -*• CHEMICAL CUTT1NQ IN OIL WELL INDUSTRY
   • ALKYLATION CATALYST -
• ALUMINUM FLUORIDE -f ALUMINUM METAL
• URANIUM HEXAFLUORIDE -» SEPARATION AND ENRICHMENT OF URANIUM 235
                       lALKYLATES —*• GASOLINE
                      ' | LINEAR ALKYLBENZENES -» DETERGENTS

• CATALYST ANDSOLVEOTFOR PHARMACEUTICAL MAWFACTURE-*- ACETAMINOPHEN
• FLUOROSULFURICACID -* CATALYST FOR ALKYLATION AND POLYMERIZATION
                      CATALYST FOR ORGANIC REACTIONS
                    •  NUCLEAR TECHNOLOGY
                     [PREPARATION OF BORANES

                        [NEUTRALIZATION OF ALKALIS FOR LAUNDRY/TEXTILES
                    -*• METAL FLUORIDES -» BERYLLIUM FLUORIDE -*• BERYLLIUM
                        I ACID DIPS FOR STEEL
• STANNOUS FLUORIDE -«• TOOTHPASTE ADDITIVE
• HYDROCHLOROFLUOROCARBONS (HCFCs)	
  • AMMONIUM BIFLUORIDE -
-"• CHLOROFLUOROCARBONS (CFCs)-
                                               *
                                   • HYDROFLUOROCAHBONS (HFC«)
                                   •HALONS 	
                                                                                         • FLUOROBORIC ACID-
                                                     HEAVY METAL
                                                     FLUOROBORATE SALTS -
                                                     ELECTROPOUSHING
                                                     ALUMINUM
                                                     METAL CLEANING
                                                                                                                                  CATALYSTS
                                                                                                                                •ELECTROPLATING
                                                                                                                                 [METALLURGY
                                                                                         •ALKALI METAL AND
                                                                                          AMMONIUM FLUOROBORATE
                                                           HIGH TEMPERATURE
                                                             FLUXING IN METAL
                                                          |  PROCESSING
                                                   [CERAMICS
                                • LITHIUM FLUORIDE—•• FLUX COMPOSITIONS FOR
                                                   | METAL JOINING
                                • MAGNES.UM FLUORIDE _^N^™ MAGNESIUM METALLU
                                 POTASSIUM BiaUORIDE-» FLUORINE
                                                   [FLUORIDATION OF DRINKING WATER
                                 SODIUM FLUORIDE—» FLUX FOR STEEL
                                                   [ALUMINUM RESMELTING
                                                          [TANTALUM
                                                          ICOLUMBIUM
                                                                REFRIGERANTS
                                                                SOLVENTS
                                                                BLOWING AGENTS FOR FOAMS
                                                                ANESTHETICS
                                                                FIRE EXTINGUISHING AGENTS
                                                                                          • DIGESTION OF METAL ORES
                                                                                          • STAINLESS STEEL PICKLING
                                                                                          • GLASS POLISHING
                                                                                          > CLEANING SOLUTIONS
                                                     . AIRCRAFT PAINT STRIPPING
                                                      AND CLEANING
                                • MANUFACTURE OF SEMICONDUCTOR CHIPS
   CHLORODIFLUOROMETHANE •
                            • POLYTETRAFLUOROETHYLENE -
  • FLUOROAROMATIC COMPOUNDS _».
                                  HERBICIDES
                                  DRUGS
                                  GERMICIDES
                                  DYES
                                  POLYMERS
ELECTRICAL APPLICATIONS
SEALS, PISTON RINGS. BEARINGS, TAPES
PACKING
HOSE LINERS
MEMBRANES
  • VINYL FLUORIDE
                                                  [SURFACING FOR ALUMINUM AND STEEL SIDING
                   • POLYMNYL FLUORIDE -* FILMS -»•  LAMINATION TO PAPER. PLASTICS. RUBBER. FELT
                                                  [COATINGS
                                                  ^

-------
                                      EXHIBIT 3-4
                  Types of Facilities Reporting to TRI for Hydrogen Fluoride
               Electronic Equipment 19.2%
                     (102)
Primary Metal 18.6%
Industries (99)
Chemical Industry  20%
       (106)
       Fabricated Metal 18.1%
       Products (96)
                                                             me, Clay, Glass 4.9%
                                                          and Concrete (26)
         Transportation 5.5%
         Equipment (29)
      ther2.3%(12)
                                      Refiners 11. 5%
                                           (61)
                             All Facilities
                             (Total:  531)
       Refiners 41.4%
            (41)
                                            Chemical Industry 39.4%
                                                      (39)
                                                       Other 8.1%  (8)
                                               Primary Metal  11.1%
                                               Industries (11)
             Facilities Reporting Maximum Quantities
                On Site of 100,000 Pounds or More
                             (Total:  99)
    Source: 1990 TRI Data.
                                                                                Page 23

-------
                                              EXHIBIT 3-5
                                    End Uses of Hydrogen Fluoride
               Nudw Applications 6%

               13,000 tons A
                                                            * Other 22%

                                                              52,000 tons  • A
Aluminum Industry 3%
8,000 tons
                    • Aqueous Forms (basis: 100% weight HF)

                    A Anhydrous Forms
                       End uses include stainless steel pickling, chemical milling, glass etching
                       ore extraction, quartz purification, metal coatings, fabric and fiber treating
                       agents, herbicides, pharmaceutical intermediates, and inert fluorinated liquids.
   Source:  Will, Ray, R. Wlllhalm , and S. Mori, Chemical Economics Handbook Product Review: Fluorine Compounds,
           Preliminary Draft, SRI International, Menlo Park, CA, May 20,1992, p 53. (102)
in addition to the 8,000 tons of merchant HF consumed by the aluminum industry, between 53,000
and 67,000 tons of HF are produced and captively converted to various fluorides by aluminum
producers.

        The largest market for anhydrous HF (historically about two thirds)23 is in the production of
fluorocarbons and related substances, especially chlorofluorocarbons (CFCs).24  HF is a source of
fluorine in the manufacture of fluorocarbons and CFCs which are used as refrigerants,  solvents,
sources of raw material for production of fluoro-plastics, anesthetics, and fire extinguishing agents.25
The production of CFCs is being cut back, however, because of the damage these molecules cause
Page 24

-------
to the ozone layer. Many fluorocarbon replacements for CFCs themselves contribute to the depletion
of ozone, although at a slower rate than CFCs. Therefore, these replacements will also be phased out
during the next 30 to 40 years.""
26
       Appendix VIII provides a list of CFC and fluorocarbon manufacturers and also contains a few
examples of other chemicals that are produced from HF (including both anhydrous and aqueous HF).

       HF is also used as an alkylation catalyst in the petroleum refining industry to produce gasoline
blending components.  In 1991, 16,000 tons of HF were consumed in alkylation catalysis in the
U.S.27 The refineries in the U.S. and Canada which use HF for alkylation are identified in Appendix
IX.  Appendix IX also includes the type of HF alkylation process (Phillips or UOP) used at each
refinery; these processes are described in Chapter 5. Alkylate production capacity for U.S. refineries
fusing HF is 520,600 barrels per stream-day.28 Currently, HF alkylation processes produce about 48
percent of the alkylate produced in the United States: the remaining 52 percent is produced using
sulfuric acid.29  Clean Air Act Amendments (CAAA) regulations will mandate gasoline reformulation
to reduce motor vehicle emissions during the 1990s.  Alkylate production, a critical component of
gasoline reformulation and of cleaner burning fuels, may be important in meeting CAAA regulations.
Exhibit 3-6 presents an overview of the geographical distribution of major producers and users of HF
in the U.S. Chapter 5 provides information on the processes used by HF producers and the major
users of HF.

3.3    Market Outlook

       The future of the HF market depends primarily on the CFC and fluorocarbon markets. The
Montreal Protocol, signed in 1987 by the United States and 22 other countries, is a treaty that froze
production and consumption levels of CFCs at 1986 levels, beginning in 1989.  One of the
requirements of the 1990 London Amendments to the treaty is that CFCs are to be completely phased
out by January 1, 2000.  In the U.S., the 1990 CAAA fulfilled and in some cases surpassed the
requirements of the London Amendments.  The U.S. phaseout schedule has been further accelerated
by President Bush's announcement that CFCs would be banned by 1995.  The CAAA also require
recapture of CFCs and HCFCs when refrigeration equipment is serviced or scrapped, warning labels
on products containing CFCs or halons, and restricting sales of recharge canisters of refrigerant to
certified, trained mechanics.

       Hydrochlorofluorocarbons (HCFCs) are identified in the Montreal Protocol in a non-binding
agreement as the major interim substitutes for CFCs because they add less chlorine to the
stratosphere, and therefore cause less destruction of the ozone layer, than fully halogenated
CFCs. °  Although HCFCs use an average of about three times as much HF in their production as
CFCs, the demand for HF to produce HCFCs is expected to decrease.  This decrease is due to
several factors including the fact that the switch will be tempered by the relatively high cost of the
alternatives; consumers are turning to non-fluorocarbon replacements; and the increase in recycling
and conservation of refrigerants will decrease demand for these alternatives. Further, HCFCs are
identified  in the 1990 CAAA as Class II ozone-depleting substances, and will be restricted after 2015
and banned after 2030.  Based on the targeting of HCFCs  for phaseout, as well as the other factors
mentioned above, the total industrial demand for HF, and therefore its production, is expected to fall
by 1996.31
                                                         I
       Although the demand for HF increased by approximately 2 percent per year between 1978-
1987,32 it is expected to decrease by 4  percent between 1991 and 1996 due mainly to the
accelerated CFC phaseout schedule announced by  President Bush.33 In 1987 and 1988, North
American HF demand was estimated at 307,000 tons and 318,000 tons, respectively.34 U.S.
consumption was 241,000 tons in 1991, and is projected to fall to 231,000 tons in 1996, excluding
captive use by the aluminum industry.36
                                                                                     Page 25

-------
                                                 EXHIBIT 3*6
                      Major Users and Producers of Anhydrous Hydrogen Fluoride
                                                                    •  HF Production Facility
                                                                    A  Fluorocarbon Production Facility
                                                                    •  Petroleum Refineries with HF Alkylation Units
Source: 1991 Directory of Chemical Producers, SRI Intemationaland American Petroleum Institute and Morris, Jeff, Fina Oil and Chemical Company, comments
      from technical review of Hydrogen Fluoride Study Report to Congress, Draft May 8,1992, June 1,1991. (344)

-------
       Few HF producers are likely to enter the market in the 1990s. Industrial Oxygen Co. Ltd. has
indicated plans to become a major manufacturer of 5,000 tons per year of refrigerant gases and 2,200
tons per year of anhydrous hydrofluoric acid;36 however, there has been no indication  in the U.S.
marketplace of this activity moving forward. Industries Quimicas de Mexico SA de CV,  a subsidiary of
Paris-based Rhone-Poulenc SA, announced in 1990 that it would construct an HF and  xanthate plant
with an expected HF capacity of approximately 60,000 tons per year in Coahuila, Mexico, with startup
planned for the end of 1992.   No subsequent activity has occurred, however, which would indicate
this plant will be built.38
                                                                                     Page 27

-------
                                         ENDNOTES

 1.     Will, Ray, R. Willhalm, and S. Mori, SRI International, Chemical Economics Handbook Product
       Review: Fluorine Compounds, Preliminary Draft, May 20,1992, p 53. (102)
 2.     "Chemical Profile, Hydrofluoric Acid," Chemical Marketing Reporter, July 25, 1988, p 50. (110)
 3.     "Chemical Profile, Hydrofluoric Acid."
 4.     SRI International, 1991 Directory of Chemical Producers United States of America, Menlo Park,
       CA, 1991, p 704. (450)
 5.     Will, Ray.
 6.     Laumer, John and Harold Lamb, Elf Atochem North America Inc., comments from technical
       review of Hydrogen Fluoride Study, Report to Congress, Draft May 8,1992, June 4, 1992.
       (290.8).
 7.     "Chemical Profile, Hydrofluoric Acid."
 8.     Will, Ray.
 9.     Chem-lntell Services, Chem-lntell Database,  1991. (115)
 10.    U.S. Department of Commerce, U.S. Imports IM 145, Bureau of the Census, 1991. (489.86A)
 11.    Will, Ray.
 12.    U.S. Department of Commerce, U.S. Imports IM 145.
 13.    Will, Ray.
 14.    Hague, William J., Allied-Signal, comments from technical review of Hydrogen Fluoride Study,
       Report to Congress, Draft May 8,1992, June 10,1992. (153)
 15.    U.S. Department of Commerce, U.S. Exports EM 454, Bureau of the Census, 1991. (489.86B)
 16.    Chemical Manufacturers Association Hydrogen Fluoride Panel, The Hydrogen
       Fluoride/Hydrofluoric Acid Industry, Washington,  DC, May 7,1991.  Prepared for the U.S.
       Environmental Protection Agency. (270)
 17.    Chemical Manufacturers Association Hydrogen Fluoride Panel.
 18.    "Hydrogen Fluoride Laser Safe for Transplants, But Wound Is Larger," Ophthalmology Times,
       September 15,1990, p 7. (225)
 19.    "First Alpha Laser Tested," Defense Electronics, June, 1989, p 16. (139.7)
 20.    Chemical Manufacturers Association Hydrogen Fluoride Panel.
 21.    Will, Ray.
Page 28

-------
22.    Will, Ray.

23.    Will, Ray.

24.    Chemical Manufacturers Association Hydrogen Fluoride Panel.

25.    Chemical Manufacturers Association Hydrogen Fluoride Panel.

26.    U.S. Environmental Protection Agency, "Protection of Stratospheric Ozone, Proposed Rule,"
       40 CFR Part 82, September 4, 1991. (415)

27.    Will, Ray.

28.    "Annual Refining Survey," Oil & Gas Journal, March 18, 1991, p 85-105. (30)

29.    Morris, Jeff, American Petroleum Institute, Fina Oil and Chemical Company, comments from
       technical review of Hydrogen Fluoride Study, Report to Congress, Draft May 8,1992,
       Junel, 1991. (344)

30.    U.S. Environmental Protection Agency, "Protection of Stratospheric Ozone, Proposed Rule,"
       40 CFR Part 82, September 4,1991.

31.    Will, Ray.

32.    "Chemical Profile, Hydrofluoric Acid."

33.    SRI International, News Brief, "Demand for Hydrofluoric Acid Continues to Fall, Says SRI
       International Study," Menlo Park, CA, May,  1992. (451)

34.    "Chemical Profile, Hydrofluoric Acid," Chemical Marketing Reporter, July 25, 1988, p 50. (110)

35.    Will, Ray.

36.    Chem Weekly, September 13,1988, p 44. (117)

37.    "Rhone-Poulenc to Supply U.S. Firms From Mexican Hydrofluoric Plant," Journal of Commerce,
       May 16, 1990, p 1A. (423A)

38.    Leiva, F.S., Chemical Manufacturers Association, comments from technical review of Hydrogen
       Fluoride Study, Report to Congress,  Draft May 8, 1992, June 17, 1992. (293.12)
                                                                                    Page 29

-------
                                 This page intentionally left blank.
Page 30

-------
                       4.  REGULATIONS AND INITIATIVES
       This chapter presents an overview of the federal, state, and local regulatory framework with
which those industries producing, using, or transporting HF must comply; and a description of
international efforts and initiatives dealing with the management of HF, either specifically or as one
among several hazardous chemicals.  Although regulations, guidelines, and standards will not
eliminate the possibility of an HF accident, the purpose of this section is to determine how extensive
and specific regulatory programs, industry standards and guidelines, and international initiatives are in
terms of addressing the safety issues surrounding the handling of  HF.

4.1    U.S. Federal Regulation of Hydrogen Fluoride

       HF has been regulated by:

       >      the Environmental Protection Agency (EPA) under several regulations authorized by

                     the Comprehensive Environmental Response,
                     Compensation, and Liability Act (CERCLA),
                     Title III of the Superfund Amendments and
                     Reauthorization Act (SARA)  of 1986, also known as the
                     Emergency Planning and Community Right To Know
                     Act (EPCRA),  and
                     the Resource  Conservation  and Recovery Act (RCRA);

       >•      the Department  of Transportation (DOT) under the Hazardous Materials Transportation
              Act (HMTA) and the Hazardous Materials Transportation Uniform Safety Act
              (HMTUSA); and

       »-      the Occupational Safety and Health Administration (OSHA) under the Occupational
              Safely and Health Act and the Clean Air Act Amendments of 1990.

Note:  The Clean Air Act Amendments of 1990 (CAAA) require EPA to list HF among at least
       100 substances targeted for accidental release prevention regulations under Clean Air
       Act §112(r). In addition, Clean Air Act §112(b) lists HF as a Hazardous Air Pollutant.

A summary of federal regulations that specifically cover HF and hazard designations of HF is
presented in Exhibit 4-1.
                                           9
       4.1.1  EPA Regulations

       CERCLA. Under CERCLA, releases of listed hazardous substances in quantities equal to or
greater than their reportable quantity  (RQ) are subject to reporting to the National Response Center.
HF is listed under CERCLA with an RQ of 100 pounds.

       SARA/EPCRA. Under SARA/EPCRA section 302, EPA designated a list of chemicals as
Extremely Hazardous Substances (EHSs) on the basis of acute toxicity and assigned Threshold
Planning Quantities (TPQs) to these substances, based on toxicity and volatility.  If a facility has an
EHS in quantities above the TPQ, it must report to the State Emergency Response Commission
(SERC); the SERC notifies the Local Emergency Planning Committee (LEPC), and the facility will be
included in the local emergency plan. The facility must provide the LEPC with the name of a facility
                                                                                   Page 31

-------
                                         EXHIBIT 4-1
                         Regulatory Designations of Hydrogen Fluoride
        Regulating Agency
Regulation
Designation
EPA
EPA
EPA
EPA
EPA
DOT
OSHA
OSHA
CERCLA (RQ 100 Ibs)
SARA/EPCRA Section 302
(TPQ 100 Ibs)
SARA/EPCRA Section 313
RCRA
CAAA
HMTA
Occupational Safety and Health
Act, Air Contaminants Standard
(PEL 3 ppm, STEL 6 ppm)
Process Safety Management
Standard
(Threshold Quantity 1 ,000 Ibs)
Hazardous Substance |
Extremely Hazardous I
Substance |
Toxic Chemical |
Hazardous Waste (if discarded) |
Required to be included on List 1
of Regulated Substances for I
Accidental Release Prevention |
Corrosive Material |
Air Contaminant 1
Highly Hazardous Chemical •
representative and information requested by the LEPC that is necessary for planning.  HF is included
on the list of EHSs with a TPQ of 100 pounds; therefore, facilities with more than 100 pounds of HF
are included in local emergency plans and may be required to participate in local planning
efforts. HF is one of 360 acutely toxic chemicals that EPA has included on the list of EHSs. These
substances are listed because they have the potential to cause death in unprotected populations after
relatively short exposure periods at low doses. This list includes 24 substances that are gases under
ambient conditions, including chlorine, ammonia, and hydrogen chloride, as well as HF; it also
includes  liquids, such as sutfuric acid, and a number of solids.
                                               *
       Section 311 of SARA/EPCRA requires a facility to make a one-time submission to the LEPC,
SERC, and local fire department of either copies of, or a list of, material safety data sheets (MSDS) for
hazardous chemicals on the site in quantities above 10,000 pounds, or, in the case  of EHSs, above
the TPQ  (100 pounds in the case of HF).

       Section 312 of SARA/EPCRA requires that an annual inventory form be submitted each year to
the LEPC, SERC, and local fire departments. Information contained in the form includes the amount
of hazardous material on site and its location.  The annual inventory form is a simple reporting
requirement that tells the local government what is on site.
Page 32

-------
       Under SARA/EPCRA section 313, facilities are required to report annual emissions of over 300
listed Chemicals, including HF.  EPA maintains a database of the reported emissions, the Toxic
Release Inventory (TRI) database. See Appendix VII for TRI data from 1990.

       RCRA. HF may be listed as a hazardous waste (U134) if it is discarded as a commercial
chemical product, manufacturing chemical intermediate, or off-specification commercial chemical or
manufacturing intermediate. This hazardous waste designation includes container residues and spill
residues. HF may also be a hazardous waste if it exhibits the corrosivity characteristic (D002)  or the
reactivity characteristic (D003).

       Clean Air Act Amendments.  Under section 112(r) of the Clean Air Act (CAA) as amended,
Congress specifically mandated that HF be included on a list of at least 100 regulated substances for
accidental release prevention. On January 19, 1993 (58 FFI 5102), EPA published a proposed rule
listing 100 toxic substances, including HF,  as well as a list of flammable substances and high
explosives as a category.  Threshold quantities were proposed for all listed substances; the proposed
threshold for HF is 500 pounds. The list of substances is intended to focus on those that, when
released, can cause death, injury, or serious adverse effects to human health or the environment.
Facilities that use these chemicals in quantities exceeding their thresholds need to comply with new
CAA regulations on release prevention, detection, and emergency response. One accident prevention
provision of the CAA as amended mandates the development of regulations requiring facilities to
prepare and implement risk management plans. These regulations are to include a requirement for a
facility to conduct a hazard assessment; to develop a program, including maintenance and training,
for preventing accidental releases; and to develop a program for emergency response. EPA is
currently developing risk management plan regulations.

       Section 112(r)  of the CAA as amended also includes a general duty for owners and operators
of facilities producing,  handling, or storing any quantities of extremely hazardous substances, whether
or not they are specifically listed, to perform activities  to prevent and mitigate accidental releases.
Activities such  as hazards identification using appropriate hazard assessment techniques; designing,
maintaining, and operating a safe facility; and minimizing the consequences of accidental releases if
they occur are also included.

       As discussed below, OSHA has developed a Process Safety Management Standard for the
protection of workers from catastrophic chemical accidents.  The process safety management
regulations being developed  by EPA are intended to protect the public and the environment from
such accidents.  EPA recognizes that process safety management programs to protect workers and to
protect the public and the environment should be  essentially the same. Therefore, EPA is working
closely with OSHA to avoid duplicative and confusing rulemakings.

       Section 112(b) of the CAA lists HF as a Hazardous Air Pollutant (HAP). EPA is developing
standards to control emissions of HAPs from stationary sources in particular industries. These
standards may address equipment leaks and fugitive emissions.

       4.1.2 OSHA Regulations

        Hazard Communication Standard. Under the Hazard Communication Standard of the
Occupational Safety and Health Act, employers must provide information to employees about
hazardous chemicals to which they may become exposed. Employers are required to disclose
information about HF because  it is highly corrosive and acutely toxic.  Chemical manufacturers,
importers, and distributors must provide material safety data sheets (MSDS) to customers and have
them available on site to workers. A written hazard communication program is required.
                                                                                     Page 33

-------
        Protective Equipment.  Under the Occupational Safety and Health Act, employers must
 provide personal protective equipment wherever necessary, depending on the nature of the hazards
 to which workers might be exposed.  Since HF is extremely corrosive, certain protective gear is
 necessary to guard against exposure. Equipment must also be inspected, properly used, and
 maintained.

        Air Contaminants. Under the Occupational Safety and Health Act, employee exposure to
 about 600 air contaminants in the workplace is limited to specified concentrations. The eight-hour
 time-weighted average (TWA) permissible exposure limit (PEL) is a level that the employee's average
 airborne exposure cannot exceed in any eight-hour work shift.  For HF, this level is 3 ppm. The short
 term exposure limit (STEL) is a 15-minute TWA exposure that cannot be exceeded at any time during
 the workday. In its final rule of January 19, 1989, OSHA supplemented the PEL with a STEL for HF of
 6 ppm; however, the exposure limits set in the 1989 rulemaking were overturned in court.  The 3 ppm
 PEL was established in an earlier rulemaking and is still in effect.

        Clean Air Act Amendments.  Under the Clean Air Act Amendments section 304, OSHA was
 required to promulgate a chemical process safety standard to prevent accidental releases of
 chemicals which could pose a threat to employees.  OSHA published a final rule on February  24,
 1992, that requires development of a process safety management system for any process involving a
 highly hazardous chemical at or above its threshold quantity. The rule includes a list of highly
 hazardous chemicals and threshold quantities. Anhydrous HF is listed with a threshold quantity of
 1,000 pounds; aqueous HF is not covered (29 CFR Part 1910).

        The OSHA Process Safety Management Standard is intended to protect employees by
 preventing or minimizing the consequences of catastrophic releases of toxic, reactive, flammable,  or
 explosive chemicals.  The OSHA standard is designed to foster the implementation of comprehensive,
 integrated management systems at facilities handling highly hazardous chemicals.  This approach
 holds great promise for prevention of catastrophic chemical accidents.  Process safety management
 programs not only prevent deaths and injuries, but  also have the added benefit of increased
 productivity resulting in cost savings for employers.

        Requirements of the OSHA standard apply to processes involving highly hazardous chemicals
 in quantities at or above their threshold quantities.  These requirements include development of a
 compilation of written process safety information, including information about chemical hazards,
 technology of the process and equipment used, to  identify and understand the  hazards posed by
 processes involving highly hazardous chemicals.  Process hazards analysis, carried out by a team
 with expertise in engineering and process operations, is required.  Process hazards analysis, a central
 element of good process safety management, involves a systematic review of what could go wrong
 and what safeguards are in place or needed to prevent the accidental release of hazardous
 chemicals, including HF.  The OSHA standard requires partial completion of the initial process hazards
 analysis by May 26, 1994 and completion by May 26, 1997; the  analysis must be updated every five
 years. The standard also includes requirements for development of written operating procedures for
 processes involving highly hazardous chemicals,  employee training, performance-based
 responsibilities for contractor safety, pre-startup safety reviews for new and significantly-modified
 facilities, maintenance of mechanical integrity of critical equipment, and establishment of procedures
 for management of changes to process chemicals, technology, equipment, and  procedures. In
 addition, the standard mandates a permit system for hot work such as welding,  investigation of
 incidents involving an accidental release or "near miss" (a minor release that could have been worse
 or a major event that was luckily avoided), emergency action plans, and compliance and safety audits
 to ensure programs are in-place and operating properly.  Employee participation is required in
 process safety management programs developed under the standard. More details on the elements
 of process safety management and their role in prevention of accidents involving HF may be found in
 Chapter 6.
Page 34

-------
       4.1.3 DOT Regulations

       The Department of Transportation (DOT) regulates the transport of hazardous materials,
including HF. DOT classifies both anhydrous and aqueous HF as "corrosive materials" (hazard class
8) in accordance with 49 CFR, Section 172.1011  Anhydrous HF and aqueous HF in concentrations
greater than 60 percent are also designated as being in packing group I, the group that is associated
with the highest degree of hazard. Aqueous HF solutions of concentrations of 60 percent or less are
also considered "corrosive material" but fall in packing group II. Anhydrous and aqueous HF also
meet DOT criteria for the "poison" hazard class. Corrosive and poison labels are required.  According
to DOT, HF also meets the DOT definition of a poison gas (class 2,  division 3), because of its vapor
pressure at ambient temperature and its toxicity, and is a "poisonous by inhalation" material. It is
classified as corrosive (class 8) in order to conform to international transportation regulations.

         The DOT regulations include requirements for shipping containers, placarding of vehicles
and bulk packaging, and shipping procedures. For example, both anhydrous and  aqueous HF must
be  labeled "Corrosive" with a label that is white in the top half and black in the lower half.  Bulk
shipping containers must carry a "Corrosive 8" placard.

        DOT periodically issues guidance for emergency response associated with shipping
containers of any kind. The 1990 Emergency Response Guidebook provides information to first
responders at an accident for initial isolation and  protective action distances and information on
potential hazards and emergency actions.  Exhibit 4-2 shows the initial isolation distances for
anhydrous HF.  Guides for anhydrous HF  (Guide 15) and aqueous HF  (Guide 59)  are  presented in
Exhibits 4-3 and 4-4, respectively.3
                                         EXHIBIT 4-2
           DOT Initial Isolation and Protective Action foir Anhydrous Hydrogen Fluoride







Substance


Anhydrous
HF



ID No.


1052


Small Spills
(Leak or spill from a small
package or small leak from a
large package)
First
ISOLATE
in all
directions
(feet)

300

Then
PROTECT
those persons
in the
DOWNWIND
direction (miles)
1

|
Large Spills
(Leak or spill from a large
package or spill from many
small
First
ISOLATE
in all
directions
(feet)

900

packages)
Then,
PROTECT those
persons in the
DOWNWIND
direction (miles)

3

                                                                                       Page 35

-------
         DOT is also responsible for implementation of the Hazardous Materials Transportation Uniform
  Safety Act (HMTUSA).  Provisions related to regulations that might affect the shipment of HF include:

         >      regulations on highway routing that account for public safety;

         >•      modifications to shipping papers that would include disclosure to emergency response
                personnel; and

         >•      training for hazmat employers on loading, handling, storing, and transport of
                hazardous materials and emergency preparedness for accidents.

  4.2     U.S. State and Local Regulations

         In addition to Federal regulations, many states and local governments have regulations to deal
  with the hazards posed by the handling and use of HF. The state governments in California,
  Delaware, New Jersey and Nevada are active in regulating HF through their chemical accident
  prevention programs.  The South Coast Air Quality Management District (SCAQMD) in California is the
  only government agency to adopt specific HF regulations. On the local level, the City of Torrance,
  California has addressed the safe use of HF at a specific refinery within its jurisdiction.

         4.2.1  State Risk Management Programs: Delaware, New Jersey,  California and Nevada

         Delaware, New Jersey, California, and Nevada have implemented regulatory programs to
  promote risk management  planning for the prevention of chemical accidents  at facilities that handle,
  use, or produce certain hazardous chemicals.  If the quantity of the chemical on-site exceeds a
 specified threshold quantity, the facility must notify the regulating agency, and prepare and submit a
 risk management plan.  Although the specific requirements differ from state to state, a risk
 management plan generally consists of a description of the existing management program to  prevent
 accidents (e.g., preventive maintenance, training), the results of a formal hazard and/or risk
 assessment conducted by the facility, a summary of possible equipment or procedural actions to
 reduce risk, and a plan with a schedule to implement those actions. New Jersey and California
 require a facility to submit a complete risk management plan to the state or local government for
 approval. In Delaware, facilities are required to develop a risk management plan and make it available
 to inspectors during site inspections. Nevada requires that the owner or operator of a facility provide,
 at least every three years, a report on safety to the state government.

        HF  is a listed chemical in the risk management programs of all four states.  The threshold
 quantities for HF in California, New Jersey, Delaware, and Nevada are 100, 500, 900, and  1,000
 pounds, respectively. California has adopted the EPA TPQ of 100 pounds; facilities are covered if the
 maximum quantity on-site at any time exceeds 100 pounds.  New Jersey facilities are covered if
 threshold quantity of 500 pounds is exceeded by the maximum quantity on-site at any time. Delaware
 allows adjustment of the threshold quantity if the distance to the facility fenceline is greater than 100
 meters.  Delaware requires facilities to carry out calculations to determine the maximum quantity that
 could be released at one time in an accident.  If this quantity exceeds the threshold, the facility is
 covered.  Nevada facilities are subject if they, at any time, store or handle 1,000 pounds or more of
 HF.
Page 36

-------
ERG90
                               Exhibit 4-3
DOT Response Guide for Anhydrous Hydrogen  Fluoride

       ANHYDROUS HYDROGEEN  FLUORIDE
                                                                                                                GUIDE 15
                                                       POTENTIAL HAZARDS
                        HEALTH HAZARDS
                                  Poisonous; may be fatal if inhaled or absorbed through skin.
                                  Contact may cause burns to skin and eyes.
                                  Contact with liquid may cause frostbite.
                                  Clothing frozen to the skin should be thawed before being removed.
                                  Runoff from fire control or dilution water may cause pollution.
                        FIRE OR EXPLOSION
                        FIRE
                                  Some of these materials may burn, but none of them ignites readily.
                                  Cylinder may explode In heat of fire.

                                                       EMERGENCY ACTION

                                  Keep unnecessary people away; Isolate hazard area and deny entry.
                                  Stay upwind, out of low areas, and ventilate closed spaces before entering.
                                  Positive pressure self-contained breathing soparatus (SOBA) and chemical
                                   protective clothing which is specifically recommended by the shipper or
                                   manufacturer may be worn. It may provide little or no thermal protection.
                                  Structural firefighters' protective clothing Is not effective for these materials.
                                  Isolate the leak or spill area Immediately for at least 150 feet In all directions.
                                  See the Table of Initial Isolation and Protective Action Distances. If you find
                                   the ID Number and the name of the material there, begin protective action.
                                  CALL CHEMTREC AT 1-800-424-9300 AS SOON AS POSSIBLE,
                                   especially If there Is no local hazardous materials team available.
                                   Small Fires: Dry chemical or CO2.
                                   Large Fires: Water spray, fog or regular foam.
                                   Do not get water inside container.
                                   Move container from fire area if you can do it without risk.
                                   Apply cooling water to sides of containers that are exposed to flames until
                                    well after fire is out Stay away from ends of tanks.
                                   Isolate area until gas has dispersed.
                         SPILL OR LEAK
                                   Stop leak if you can do so without risk.
                                   Fully-encapsulating, vapor-protective clothing should be worn for spills and
                                    leaks with no fire.
                                   Use water spray to reduce vapor; do not put water directly on leak or spill
                                    area.
                                   Small Spills: Flush area with flooding amounts of water.
                                   Large Spills: Dike far ahead of liquid spill for later disposal.
                                   Do not get water inside container.
                                   Isolate area until gas Is dispersed.
                         FIRST AID
                                   Move victim to fresh air and call emergency medical care; if not breathing,
                                    give artificial respiration; if breathing is difficult, give oxygen.
                                   In case of contact with material, immediately flush skin or eyes with running
                                    water for at least 15 minutes.
                                   Remove and Isolate contaminated clothing and shoes at the site.
                                   Keep victim quiet and maintain normal body temperature.
                                   Effects may be delayed; keep victim under observation.
                                                                                                                            Page 37

-------
   ERG90
                              Exhibit 4-4
DOT Response  Guide for Aqueous Hydrogen Fluoride

        AQUEOUS  HYDROGEN FLUORIDE
                                                                                                                    GUIDE  59
                                                          POTENTIAL HAZARDS
                           HEALTH HAZARDS
                                     Poisonous If Inhaled or swallowed.
                                     Skin contact poisonous.
                                     Contact may cause burns to skin and eyes.
                                     Fire may produce Irritating or poisonous gases.
                                     Runoff from fire control or dilution water may cause pollution.
                          FIRE OR EXPLOSION
                          FIRE
                                    Some of these materials may bum, but none of them ignites readily.
                                    Some of these materials may Ignite combustibles (wood, paper, oil, etc.).

                                                         EMERGENCY ACTION

                                    Keep unnecessary people away; Isolate hazard area and deny entry.
                                    Stay upwind, out of low areas, and ventilate closed spaces before entering.
                                    Positive pressure self-contained breathing apparatus (SCBA) and chemical
                                     protective clothing which Is specifically recommended by the shipper or
                                     manufacturer may be worn. It may provide little or no thermal protection.
                                    Structural firefighters' protective clothing is not effective for these materials.
                                    CALL CHEMTREC AT 1-800-424-9300 AS SOON AS POSSIBLE,
                                     especially If there Is no local hazardous materials team available.
                                    Some of these materials may react violently with water.
                                    Smill Fires:  Dry chemical, CO2, water spray or regular foam.
                                    Urge Fires:  Water spray, fog or regular foam.
                                    Move container from fire area if you can do It without risk.
                                    Apply cooling water to sides of containers that are exposed to flames until
                                     well after fire is out Stay away from ends of tanks.
                         SPILL OR LEAK
                                    Do not touch or walk through spilled material; stop leak if you can do it
                                     without risk.
                                    Fully-encapsulating, vapor-protective clothing should be worn for spills and
                                     leaks with no fire.
                                    Use water spray to reduce vapors.
                                    Small Splits:  Take up with sand or other noncombustible absorbent
                                     material and place Into containers for later disposal.
                                    Urge Spills:  Dike liquid spill for later disposal.
                                    Isolate area until gas Is dispersed.
                         FIRST AID
                                   Move victim to fresh air and call emergency medical care; If not breathing,
                                    give artificial respiration; If breathing is difficult, give oxygen.
                                   In case of contact with material, immediately flush skin or eyes with running
                                    water for at least 15 minutes.
                                   Remove and Isolate contaminated clothing and shoes at the site.
                                   Keep victim quiet and maintain normal body temperature.
                                   Effects may be delayed; keep victim under observation.
Page 38

-------
       Once a facility meets or exceeds the specified chemical threshold in New Jersey, Delaware,
or Nevada as described above, the facility must develop a plan. Once the statutory quantity has been
met in California, it is up to the local implementing agency (e.g., fire department) to determine whether
a plan is required.  The requirements under the programs include:

              >•       registration with the state;
              +       up-to-date process safety information;
              »•       a hazard evaluation;
              >       standard operating procedures;
              *•       training in  standard operating procedures;
              >       preventive maintenance programs;
              >       safety audits;
              >       pre-start-up reviews before initiating new processes;
              >•       management of change procedures;
              *•       accident investigation; and
              *•       emergency response procedures.

       The state programs were implemented to require facilities to consider and  implement ways to
reduce risk associated with handling hazardous chemicals. These regulations anticipate the
development of federal risk management plan regulations required under the Clean Air Act
Amendments.4'5'6'7

       4.2.2  Texas Air Control Board

        Under the authority of the Texas Clean Air Act (TCAA) of 1965, the Texas Air Control Board
(TACB)  established a permitting program to control emissions from new or modified industrial facilities
which emit air pollutants. In September 1971, TACB Regulation VI was adopted which specifies that
all new or modified facilities must  obtain a permit prior to the start of construction.  Certain
requirements must be met before obtaining a permit as specified in Regulation VI.

        Over the years, Texas has seen significant air quality improvement through the use of best
available control technology (BACT) negotiated during the permit process; primarily through close
scrutiny  of traditional sources (e.g., process vents, bulk loading, and fugitive emissions). Beginning in
1979, some additional sources not previously reviewed in detail also became a focus of the TACB
permit review. These "non-traditional sources" are either continuous emissions of air contaminants
from systems not directly related to the process (e.g., cooling towers and wastewater treatment
systems) or accidental releases caused by upsets, maintenance, or disasters.

        HF is one of a list of chemicals which TACB staff presently considers to have disaster potential
and for which a "disaster review" may be required.  Disaster reviews are required for proposed
projects that either handle these chemicals in sufficient volumes to produce a life-threatening off-plant
impact in the event of a disaster,  or use processes that could increase the probability of off-site life-
threatening releases.  A disaster review may combine a health effects review of the predicted impacts
caused by a catastrophic release and an engineering analysis of the disaster prevention and control
systems employed at a facility.

        In a disaster review, process areas with the greatest hazard potential are identified and several
worst-case scenarios are defined. Catastrophic release rates are calculated, dispersion models are
 used to estimate maximum concentration, and predicted impacts of these releases are considered
 based on magnitude, duration, and movement of the plume. If the predicted impacts are considered
too great, the applicant must propose all  reasonable, possible design and operational changes to
 reduce the  probability, magnitude and duration of a catastrophic release.  If these changes satisfy the
                                                                                       Page 39

-------
 TACB-specified disaster review requirements, the permit engineer will incorporate special provisions
 into the permit which will ensure enforceability and maintenance of the applicant's disaster response
 and prevention program.8  The modeling and effects evaluation steps  may be unnecessary if the
 applicant concedes that a release would result in a life-threatening situation.  Whether modeling is
 required or not, the applicant works with the permit reviewer to incorporate changes into the design of
 the facility to reduce the probability, magnitude, and duration of such a potential release.

        The TACB staff has particular concerns about the use of anhydrous HF in urban areas.
 Because of recent HF related inquiries and the influx of many new permit reviewers, the TACB staff
 recently instructed permit reviewers that HF,  like many other toxic chemicals, is to be given an
 intensive review for disaster potential. This may include reviews of: safety features incorporated into
 the design of the facility to minimize off-site effects; emergency plans designed to mitigate off-site
 effects of any significant release; and the potential for future migration  of population into an area near
 the facility. Consideration is also given to the past performance of the facility in handling hazardous
 materials.  If TACB staff concludes that a facility, even with excellent engineering design, is simply too
 dangerous to be located at a proposed new site, denial of the permit may be recommended.  Overall,
 the disaster review  is conducted to assure that risk to the public is minimized. Review of HF facilities
 by TACB staff has included review of several petroleum refinery HF alkylation units and an HF
 production unit.9'10

        4.2.3  HF Regulations from the South Coast Air Quality Management District

        Although the CAA indicates that EPA should focus its efforts on "those regions of the country
 that do not have comprehensive health and safety regulations with respect to HF,"  EPA believes it is
 important to discuss the analysis performed by the South Coast Air Quality Management District
 (SCAQMD) as one method of evaluating HF hazards.  As a result of their analysis, SCAQMD proposed
 Rule 1410 to control HF in Southern California.  The purpose of this section is to describe some of the
 issues associated with HF by examining SCAQMD's rule. SCAQMD shared considerable information
 with EPA about SCAQMD's approach for evaluating HF. SCAQMD staff noted, however, that their
 decision with respect to HF is site-specific and does not necessarily reflect conditions associated with
 HF use elsewhere.  It is important to note that Rule  1410 was litigated,  during which time its
 implementation was suspended by the court. However, on July 30, 1993, a court decision permitted
 implementation of the rule if the SCAQMD corrects certain procedural errors that were made in the
 original promulgation of the rule.

        In the Los Angeles Basin, the SCAQMD is responsible for developing and enforcing air quality
 control rules. In response to concern about the safety of the use of HF in the Los Angeles Basin, and
 the HF accidents at the Marathon Oil refinery in Texas City, Texas, and at the Mobil refinery in
 Torrance, California, the SCAQMD adopted Rule 1410 on April 5, 1991.  Rule 1410 specifically
 regulates the storage, use, and transportation of hydrogen fluoride. It is the only regulation in the U.S.
 directed specifically at anhydrous HF.  Rule 1410 is partly intended to eliminate the possibility of harm
 to the public due to an accidental release of HF. It applies to fluorocarbon production facilities and
 petroleum refineries that use HF. The rule has three general requirements: 1) Phase Out, 2) Interim
 Control Measures,  and 3) Reporting and Storage/Usage Inventory Requirements. Additional
 details on each section of Rule 1410 can be found in Appendix X. The rule calls for the phase out of
 the use of anhydrous HF at one fluorocarbon facility on or before January 1, 1999, and four petroleum
 refineries on or before January 1, 1998. Rule 1410 also requires that these facilities implement interim
 control measures to prevent the release of HF until the phase out. Facilities that store or use either
 aqueous or anhydrous HF must also comply with the notification and storage usage report provisions
 of the rule.

        In developing Rule 1410, the SCAQMD conducted facility-specific evaluations of the hazards of
 HF at the five facilities that would be affected. These evaluations included computer modeling and
Page 40

-------
impact analysis.  In addition, the SCAQMD produced several technical and policy support documents.
In one document the SCAQMD discussed its concerns about HF safety at the five facilities in the
South Coast Air Basin (four refineries:  Golden West, Mobil, Powerine, and Ultramar, and one
fluorocarbon production facility:  Allied-Signal).11 The SCAQMD stated that although precautions
could be taken on these sites, available mitigating measures could not eliminate the consequences of
a major accident.  The SCAQMD cited both the Bhopal accident in India and the accident at
Marathon.  In examining current practices in transportation, storage and use, the SCAQMD found that
current safety practices were not sufficient to assure public safety and consequently recommended a
phase out of anhydrous HF at the five facilities. In addition, the SCAQMD conducted an
environmental assessment for the proposed Rule 1410 that describes the rule and its potential
impacts as well as alternative actions to those  required by the rule.12  In addition, the SCAQMD
published guidelines to the proposed Rule 1410, which outlined the general requirements of the rule
and compliance requirements for affected facilities.13

       Thepftase out section of Rule 1410 mandates that on or after January 1, 1998, refineries and
fluorocarbon production facilities cannot use anhydrous HF unless it is contained in a mixture, which
in a serious, near worst-case accidental release, will not result in atmospheric concentrations equal to
or greater than 20 parts per million  (ppm) for five minutes and 120 ppm for one minute at or beyond
the facility boundary.  The interim control measures  require that until the phase-out takes effect,
facilities must install safety equipment and implement the procedures required by the SCAQMD to
reduce the impact of an HF release.  In addition, Rule 1410 requires that after July 1, 1991, an owner
or operator must report to the SCAQMD any HF release that results in exposed persons requiring
medical treatment at an off-site facility, evacuation of any portion of the facility premises, or HF aerosol
transport beyond the facility property boundaries.

        Prior to suspension of Rule 1410, the five industrial facilities complied with the interim control
provisions. For example, four of the five affected facilities now have HF detectors connected via
remote terminal units (RTUs) which can transmit detection readings directly to SCAQMD. Also in
accordance with the interim provisions, the Allied-Signal facility in El Segundo had been working on
seismic upgrading to lessen the chance of release in case of earthquakes.

        A phase out of HF would result in either the closure of the HF facilities in the area, or in  a
retrofit of these facilities to accommodate an HF substitute unless HF can be modified to prevent
aerosol formation.  Refineries also produce alkylate, using sulfuric acid as a catalyst. However, there
may be substantial costs associated with switching from HF to sulfuric acid. Approximately 48%
percent of the total alkylate production capacity in the U.S. uses sulfuric acid as the alkylation
catalyst.14 At present there is no substitute for HF in fluorocarbon production.

        Alkylate is becoming increasingly important  for the production of clean burning high octane
motor fuels. However, the decision whether to use sulfuric acid or HF as an alkylation catalyst is not
straightforward, nor is the decision to switch from one particular catalyst to the other, as discussed
below.  The SCAQMD in its support documentation  for Rule 1410 asserted that sulfuric acid is a viable
alternative for Southern California because of particular circumstances surrounding sulfuric acid
supply, sulfuric acid regeneration capacity and safety issues associated with HF and sulfuric.  Some
stakeholders contend that sulfuric acid is far safer for the public located near refineries and that EPA ,
should require refineries using HF to switch to sulfuric. However, other stakeholders maintain that
there are other significant risk issues associated with sulfuric acid alkylation that must be considered.
For example, up to  100 times more sulfuric acid than HF must be transported for use in the alkylation
process.  The transport and transfer of large amounts of fresh and spent sulfuric acids must be
considered in terms of increased risk of worker exposure due to increased handling requirements and
risk to the public from transportation accidents.15 Further, the use of sulfuric acid generates waste
streams that require treatment and disposal and there is concern that one dangerous chemical  is
being substituted for another.  Upon release under  alkylation unit conditions, sulfuric acid will not form
                                                                                      Page 41

-------
 a dense cloud of vapor and aerosol, as HF may.16 However, releases from sulfuric acid production
 and regeneration plants could result in mists of sulfuric acid, sulfur dioxide (a toxic gas) and sulfur
 trioxide (a toxic, reactive liquid).17

        The intent of this study, as mandated by Congress, is not to determine the overall advantages
 or disadvantages of sulfuric versus HF alkylation, but primarily to identify the hazards of HF in its
 industrial production and use. Therefore, no additional comparison of the two alkylation processes will
 be made in this report. While both technologies are viable, the choice between the two involves
 consideration of complex site-specific factors, economics and safety.  The decision to switch from HF
 alkylation to sulfuric alkylation would also be very complex and would involve many of the same site-
 specific considerations, which are outside the scope of this study. This is an area, however, where a
 comprehensive analysis comparing all relevant issues is needed.

        4.2.4  City of Torrance, California

        The City of Torrance, California has attempted to ensure the safety of HF operations at the
 Mobil Oil refinery within its jurisdiction because of safety concerns.  These concerns were raised
 because of the history of accidents at the refinery as well as by the results of a 1986 industry-
 sponsored HF study regarding behavior of HF releases.

        In April 1987, the City Council of Torrance brought suit against Mobil in the California State
 Superior Court in Los Angeles, alleging that the Mobil refinery was a public nuisance and seeking
 both  an injunction barring Mobil from polluting the air with HF and other toxic chemicals, and an order
 requiring the company to operate the refinery safely.18  While the case was pending, another
 accident occurred at the refinery in which 100 pounds of HF were released in an explosion and fire.
 No one was injured as a result of exposure to HF.  Mobil chose to settle the lawsuit out of court.
 Mobii agreed to work on reformulating the HF catalyst to prevent it from forming a vapor cloud  in the
 event of a release; the reformulated catalyst is to be ready by the end of 1994. If the testing is not
 successful,  Mobil has agreed to discontinue operating the HF alkylation unit by the end of 1997 and
 consider other alternatives, including H2SO4 alkylation.19 The settlement also stipulates that an
 outside consulting firm be hired to study various plant operations and safety issues over a period of
 seven years. This outside consultant, whose actions are subject to court approval, will recommend
 ways to improve the plant's safety record.20
4.3
International Efforts
        The international community shares a concern about HF, however, there are no foreign
national or international regulations that address HF specifically. Instead, most countries regulate HF
as one of many hazardous materials of concern. A few countries regulate large quantities of HF as
part of their accident prevention regulations. A small number of other countries have conducted
studies of HF and explored regulatory options to reduce risk. The outgrowth of these international
efforts concerning HF includes the development of: publications such as technical guidance manuals
to promote safe handling of HF; models to determine the behavior, consequences, and risks of
accidental HF releases; and various non-regulatory programs and initiatives by both industry and
government agencies to prevent HF releases.

        4.3.1 Multinational Efforts

        The Seveso Directive.  Following two major hazardous materials accidents, the European
Community issued a Council Directive on the major accident hazards of certain industrial activities,
commonly called the Seveso Directive (82/501/EEC).  Effective January 8, 1984, the Directive is
concerned with the  prevention of major accidents and the limitation of their adverse consequences to
man and the environment.  HF is one of the listed chemicals subject to the Directive.
Page 42

-------
       Under the Directive, manufacturers must report if their activities involve the use of HF or other
listed chemicals in amounts greater than specified quantities. The current threshold quantity for HF is
50 metric tons (110,000 pounds).  In addition to the reporting requirement, the Seveso Directive also
requires that member countries develop emergency response plans and perform safety studies for
facilities handling hazardous chemicals. Although recent amendments to the Directive have moved
away from regulation of individually listed substances and toward regulation of categories of
chemicals, HF remains one of the few individually listed substances.

       The World Bank.  The World Bank and the International Finance Corporation  require
evaluation of the adequacy and effectiveness of measures to control major hazard accidents that
might affect people and the environment outside plant boundaries for projects that request funding.
For this purpose the World Bank developed "The World Bank Guidelines for Identifying, Analyzing, and
Controlling Major Hazard Installations in Developing Countries." These guidelines are based
substantially on the European Communities' directive on major accident hazards of certain industrial
activities, and regulations promulgated under the United Kingdom Health  and Safety at Work Act.  The
guidelines apply generally to  industrial processes, storage, and transportation of hazardous materials.
Hazardous chemicals are grouped into four categories: Very Toxic Substances, Other Toxic
Substances, Highly Reactive Substances, and Explosive or Flammable Substances. HF meets the
criteria as a Very Toxic Substance.

       To support the implementation of those guidelines, the World Bank has also developed the
document, "Manual of Industrial Hazard Assessment Techniques," which provides both hazard and risk
assessment methodologies that can be applied to existing operations as well as to rehabilitation and
expansion projects.  For those failures that would cause major severe damage or loss of life, on or off
the plant site, the first objective is to reduce the magnitude of the potential damage through
modification of the plant. If damage reduction is not possible, a risk analysis may  be  required to
determine how to reduce the probability of the hazardous event.222'23

       Industry Standard from the European Chemical Industry Council. The European Chemical
Industry Council (the Conseil Europeen des Federations de ('Industrie Chimique (CEFIC)),
headquartered in Brussels, Belgium, represents the chemical industry in 15 European countries.  A
sector group of the council called the European Technical Committee on  Fluorine  (dTEFj addresses
issues related to HF. The European HF producers, acting within the CTEF have issued a
recommended code to formalize a general standard for HF safety. The recommendations proposed in
the code are based  on the experience of HF producers and a compilation of various measures and
practices used by member companies in the CTEF. Over the last 10 years, CTEF  has produced
several documents as part of the code that covers specific equipment or  practices within the following
general topics: storage, transport, and safety equipment for HF; handling procedures for HF;
emergency response and planning for HF releases; physical properties of HF, and medical advice for
HF exposure. These recommendation documents form a comprehensive industry  standard for
producers and uses of HF, but are not intended as a substitute for the various national or international
regulations which already cover HF. The recommendations serve as a guide that  can be adapted and
utilized in consultation with an HF producer. __ Future recommendation  documents are  planned for
pumps and personal protective equipment.
24
       Awareness and Preparedness for Emergencies at the Local Level (APELL). Awareness
and Preparedness for Emergencies at the Local Level (APELL) is an initiative sponsored by the
Industry and Environment Office (IEO) of the United Nations Environment Programme (UNEP) in
cooperation with the Chemical Manufacturers Association (CMA) and CEFIC.  APELL is an
international program designed to create and increase community awareness of hazards within the
community from any industrial or commercial activities with the potential for fire, explosion, or release
of hazardous materials. Based on this awareness, APELL will further assist the community in
developing a cooperative plan to respond to any emergencies that these hazards might present.  In
                                                                                    Page 43

-------
 most countries participating in APELL, HF would be considered to pose a sufficient hazard to merit
 focus in the program.

       APELL works by providing information to concerned members of the community on industrial
 hazards and the measures taken to reduce these risks; by reviewing, updating, or establishing
 emergency response in the local area; and by increasing local industry involvement in community
 awareness and emergency response planning.  In addition, APELL serves to integrate industry
 response plans with  local emergency response  plans for the community to handle all types of
 emergencies, and to involve members of the community in the development, testing and
 Implementation of the overall emergency response  plan. Although APELL was primarily designed for
 governments of developing nations, UNEP provides APELL and APELL training to any nations, local
 governments, or facilities that wish to implement the program.  UNEP has provided training  in areas
 such as Eastern Europe, Tunisia, Egypt, and India.25

       Guiding Principles for Chemical Accident Prevention, Preparedness and Response.  The
 Guiding Principles for Chemical Accident Prevention, Preparedness and Response is a document
 produced by the Organisation for Economic Co-operation and Development (OECD).  The OECD is an
 intergovernmental organization in which 24 industrialized countries from North America, Western
 Europe, and the Pacific work together to respond to international problems.

       The Guiding  Principles provide general guidance for the safety, planning, construction,
 management, and operation of hazardous installations in order to prevent accidents involving
 hazardous substances.  Recognizing that despite prevention activities accidents may occur, the
 Guiding Principles recommend ways to mitigate adverse effects through effective land-use planning
 and emergency preparedness and response.  The Principles apply to all hazardous installations
 including fixed facilities that produce, process, use, handle, store, or dispose of hazardous substances
 and those that could have a major accident involving hazardous substances. Although there is no
 specific list of chemicals, facilities that handle HF would likely fall under the Principles. The  Principles
 also provide advice related to the role and responsibilities of public authorities, industry, employees
 and their representatives, as well as other interested parties such as non-governmental organizations
 and members of the  public potentially affected by an accident.26

       Convention on the Transboundary Effects of Industrial Accidents. The United Nations
 Economic Commission  for Europe developed a  convention that requires signatory countries, including
 the United States, to protect human beings and the environment against industrial accidents by:
 preventing such accidents as far as possible; reducing their frequency and severity; and mitigating
 their effects. The convention requires that signatory countries develop prevention, preparedness, and
 response plans.  A key element of the convention is a requirement for exchange of information.  Under
 the plan, if a facility contains more than a reportable quantity of a chemical it must notify the bordering
 country of the existence of that chemical in the facility. The bordering country can then develop a plan
 to deal with an accidental release. The convention covers many chemicals, including HF.27

       4.3.2 Great Britain

       Control of Industrial Major Accident Hazards (CIMAH) Regulation. Following an explosion
 at an industrial facility at Flixborough in 1974, the United Kingdom's Health and Safety Commission
 appointed a committee of experts, the Advisory  Committee on Major Hazards (ACMH), to consider the
 problems of industrial activities which pose  major hazards beyond the immediate vicinity of the work
 place. An underlying principle of the ACMH is that the primary responsibility for  controlling  and
 minimizing risks should lie with those who create the risks.  As an outgrowth of these efforts, Great
 Britain adopted the Seveso Directive. To implement the Directive, Great Britain passed the Control of
 Industrial Major Accident Hazards (CIMAH) Regulations 1984, which were designed to prevent major
Page 44

-------
industrial accidents. The CIMAH Regulations cover most of the chemical and petrochemical industry
that use substances which have dangerous flammable, explosive, or toxic properties, including HF.

       In addition to the CIMAH regulations, the United Kingdom's Health and Safety Executive has
been developing and using risk-based criteria for determination of planning decisions regarding major
hazard facilities, including HF.~
28
       Storage of a substance is treated differently from process use.  "A Guide to the Control of
Industrial Major Accident Hazards Regulations 1984" published by the Office of the Health and Safety
Executive (HSE) serves as a guide to the regulated community and to local authorities and emergency
services.29

       The HSE also conducts a program which offers assistance to local planning authorities in
developing planning strategies to minimize the risk to the public from hazardous facilities.  Planning
authorities are requested to consult the HSE before granting permission for development of a facility
that could pose a major hazard (such as an HF facility.)  Planning authorities are also advised to seek
guidance before permitting development which caters to numbers of people (e.g. housing) near
existing or proposed major hazards.

       A recent example of this planning assistance can be seen in the Public Inquiry into the
proposed Northwick Village Development that  was to be located near an existing HF facility.30  The
HSE provided consultation and issued advice  on how risk might be minimized through placement of
buildings  and restriction of certain types of development (e.g. schools,  major shopping developments,
housing for vulnerable groups of people.)

       Technical Guide to the Use and Handling of Hydrogen Fluoride. "A Guide to Safe Practice
in the Use and Handling of Hydrogen Fluoride" was published in 1988 by the Chemical Industries
Association, a group of representatives from chemical firms in Great Britain. The guide provides
instructions on a safe approach to building and operating a facility that  produces or uses HF.  The
guide covers  several topics in process safety including materials of construction, plant design, plant
operation, treatment of spillage, emergency procedures, and first aid."'
                                    31
       In 1986 Great Britain established provisions to control the emission into the atmosphere of
noxious or offensive substances from HF facilities in which: (1) HF is evolved either in the
manufacture of liquid HF or its components;  (2) mineral phosphates are treated with acid other than
in fertilizer manufacture; (3) mineral phosphates are defluorinated; or (4) anhydrous HF is stored and
handled in fixed tanks with an aggregate capacity exceeding one ton.
                                    32
       Great Britain has also initiated a three-month study by the Chemical Manufacturing National
Industry Group (CMNIG) to produce NIG guidance on health and safety in the large scale use and
manufacture of HF.  This effort will concentrate on the manufacture and use of HF in the chemical
industry including alkylation plants as well as other large scale users. The results of the study will
identify industry practices and identify optimum operational standards for safe use in both existing and
new plants, determine problems and the potential hazards associated with HF use.33

       4.3.3 Sweden

       The Swedish Act on Chemical Products regulates the control of HF and other chemicals in
Sweden. The Act is administered by the National Chemical Inspectorate which ensures that the
obligations created  by the legislation are fulfilled.  Under Swedish  law, manufacturers and importers of
chemical products are responsible for providing hazard information and safety advice for those  using
the chemicals.  In addition, there is a general obligation  to take precautions to prevent or minimize
                                                                                      Page 45

-------
 harm to human beings or the environment.  This obligation also includes a duty to make efforts to
 replace hazardous chemicals with less hazardous substitutes.

        Under this law, HF is classified as an "Extremely dangerous chemical product" due both to its
 high toxicity and strongly corrosive effects.  Even products containing low concentrations of HF
 receive this classification.  As an "Extremely dangerous chemical product," HF is subject to severe
 restrictions on importation, transfer, and non-professional handling. Also, there are general restrictions
 on the use of HF and a permitting system which allows authorities to  track the movement of HF.

        Several thousand tons of HF are used in Sweden every year,  primarily as a pickling agent in
 the metalworking industry, but also as an etching chemical in the electronics industry, and in some
 chemical products for cleaning and de-calcifying. Recently, it was discovered that in some cases
 there was misleading hazard classification information on HF-containing products (mainly those
 products used  in cleaning and de-calcifying).  As a result, proper restrictions did not apply and
 authorities were unable to track these  HF-containing products. The Inspectorate urged these
 companies to substitute HF-containing products with less harmful ones.  Within one year most of the
 products were withdrawn or had been replaced by appropriate substitutes.  Alternatives are being
 developed and tested for the remaining HF-containing products.34

        4.3.4 The Netherlands

        The Netherlands' Effort to Develop Risk-Based Regulations. The Netherlands has
 developed risk-based standards to regulate industry. Government policy is aimed at developing and
 evaluating  risk assessments to achieve effective  control of environmental and safety risks. The
 Netherlands is applying this risk-based regulatory approach to many new and existing facilities
 including several facilities that use HF. In the facilities themselves, the principal focus has been
 increased safety through improved design and management to reduce equipment failures and human
 error, respectively, and by a reduced inventory to reduce the field of hazard.35

        The Effort to Revise the Model of the Netherlands Organization for Applied Natural
 Science Research (TNO).  The Netherlands has recently revised, updated and improved the
 consequence modeling techniques described in  the 1979 document,  "Methods for Estimating the
 Physical Effects of the Escape of Dangerous Materials (liquids and gases)" commonly known as the
 "Yellow Book." The Yellow Book provides an approach to developing  a quantified consequence
 analysis to evaluate the safety of industrial installations using hazardous materials. The Yellow Book
 was developed by TNO at the request of the Directorate-General  of Labour of the Dutch Ministry of
 Social Affairs.

       The revision of the Yellow Book is part of the ongoing effort in the Netherlands to address
 hazardous chemicals such as HF. Two specific changes to the Yellow Book that would enhance
 consequence analyses for accidental releases of HF include: 1) improving the source terms for
 fuming liquids, and 2) updating the calculation of vapor cloud dispersion for dense gas releases.
 Another document published in 1990, "Methods for Calculation of Damage Resulting from the Physical
 Effects of Accidental Release of Dangerous Chemicals" commonly known as the "Green Book,"
 provides a model for predicting possible damage and is also applicable to HF.36

       4.3.5 France

       In France, the National Institute of Environment and Risks analyzes risks and  provides
 assistance  in the elaboration of standards and technical  regulations.   Industrial groups in France have
 the responsibility of submitting a safety case study in support of an application for authorization of the
 siting and construction of an industrial plant. One particular approach or technique submitted by a
 petroleum refinery was developed for use in studies on refineries. This technique consists of
Page 46

-------
developing a risk matrix and has been applied in studies carried out by various companies operating
facilities in France including ARCO.37

       In an effort related to HF safety, the Ministry of Environment has commissioned SNPE, a
consulting firm, to conduct an industry-wide study on HF storage tanks with capacities greater than 50
tons.  Currently there are no standards in France relating specifically to HF storage.  The SNPE study
will evaluate risks to the environment associated with HF storage areas, the potential safety impacts of
storage areas, and will make recommendations to improve storage safety. The study involves several
phases.  In the first phase, which has already been completed, SNPE visited ten storage sites with 50
tons or more of HF. Of the ten sites visited, eight are HF users and two are HF producers. These
storage sites are already subject to the Seveso Directive.  In the second phase, which was to have
been completed by March 1992, SNPE was to conduct a qualitative study of general hazards
associated with HF, consisting of an analysis of the risks attendant to the hazards, and a summary of
accident scenarios and consequences.  In the third phase, which was to have been completed  by
April 1992,  SNPE was to make recommendations on concepts and principles for safe operation and
safety equipment.  An interim report was scheduled for completion by the end of March 1992.38

       The ELF petroleum company  recently evaluated the possible use of HF alkylation in Lyon,
France, in a refinery, but decided that an H2SO4 alkylation unit would pose less overall risk to the
public in that location. The French government also had serious reservations regarding an HF
alkylation unit at that location. ELF has a second project involving either an HF or H2SO4 alkylation
unit under consideration for Marseilles in an industrial zone.  If the facility uses H2SO4, an acid
regeneration plant would be needed and the closest acid regeneration plant is in Belgium. The risk of
transporting large quantities of H2SO4 is currently being discussed.39

        In a study of HF alkylation, ELF is studying how to protect HF-containing vessels from fire by:

        »-      examining the quality of materials for exterior coverings and their resistance to
               fire;

        >      determining the best methods and standards for control (in
               conjunction with Ministry of Industry); and

        »•      conducting in situ trials with fire on tanks uncovered as opposed to
               those covered.

These results may be presented at a loss prevention symposium in Italy. ELF has also participated in
HF spill tests in Nevada.40

        Elf Aquitaine (ELF) has developed a proprietary mathematical model which contains enhanced
HF dispersion modeling capabilities.  The Aide aMa Gestion Industrielle des Risques (AGIR) model is
used at about a dozen industrial sites in Europe."'
41
        4.3.6  Canada

        Life-Cycle Management of Toxic Chemicals - Hydrofluoric Acid (HF) As a Case History. An
 approach used by Canada to manage toxic substances is the development of a comprehensive
 project to manage chemicals throughout their life-cycle. The Major Industrial Accidents Council of
 Canada (MIACC), a committee of government and industry representatives, and the Canadian
 Chemical Producers Association (CCPA), are developing a comprehensive project for life-cycle
 management of hazardous chemicals. The project initially has concentrated on safety and
 management of material handling.  It attempts to coordinate previous efforts to address individual
 components of the life-cycle (e.g., processing, storage, disposal) and combines these efforts into a
                                                                                      Page 47

-------
 comprehensive management plan. This life-cycle approach is the foundation of the Canadian
 Environmental Protection Act (CEPA) of 1988 and the CCPA's Responsible Care Codes of Practice.

        To provide a forum for sharing information about hazardous chemicals management, a
 workshop was held on January 24, 1991 in Toronto.  Hydrogen fluoride was selected to be the focus
 of the workshop and to be the first case study in the life-cycle management project. During the
 workshop, the following seven life-cycle stages of chemicals were evaluated:  research and
 development, introduction to the marketplace, manufacture, transportation, distribution, use, and
 disposal. The workshop examined the profile of the Canadian HF industry, manufacturing
 management systems and practices for HF, accident history, the health properties of HF,  a production
 overview and a review of transportation practices and primary uses of HF.  At the conclusion of the
 workshop, the participants identified specific HF safety issues and made several recommendations.
 The next step for the HF life-cycle management project will be the drafting of a generic management
 framework for all hazardous materials followed by additional workshops on specific chemicals such as
 chlorine in order to validate the framework.42

        Canadian Technical Manual for HF Problem Spills.  In 1984, a manual entitled, "Hydrogen
 Fluoride and Hydrofluoric Acid - Environment  and Technical Information for Problem Spills" was
 published. It is one of a series developed by Environment Canada's Environmental Protection Service
 to provide comprehensive information on chemicals that are frequently spilled  in Canada.  These
 manuals have been developed for many chemicals and are intended to be used by spill specialists for
 designing countermeasures for chemical spills and to assess their effects on the environment. The
 manual on HF contains technical information on commerce and production of  HF, material handling
 and compatibility, contaminant transport, environmental data, human health hazards, chemical
 compatibility, countermeasures, previous spill experience and analytical methods for detecting and
 measuring levels of HF.43

        Draft Canada-U.S. Joint Inland Pollution  Contingency Plan.  Canada and the United States
 are currently developing a response plan for releases of hazardous materials which pose  a threat.  HF
 may be one of the chemicals covered by the response plan.  Under the plan there would  be Joint
 Response Teams on both  the regional and Federal levels.  These response teams would facilitate
 response efforts in the event of a release.

        4.3.7  Mexico

        Joint United States of America-United Mexican States Contingency Plan for Accidental
 Releases of Hazardous Substances Along the Border.  The United States-Mexico Joint Contingency
 Plan for accidental spills and releases of hazardous substances, pollutants, or  contaminants along the
 border provides a significant framework for cooperation between the United States and Mexico in the
 case of a polluting incident. The plan applies to all pollution incidents along the inland border within
 an area 100 kilometers on  either side of the Inland International Boundary.  The objective of the plan  is
 to aid in the development of preparedness, reporting, and monitoring measures. In the event of a
 release of a hazardous substance, such as HF, the chairs of the Joint Response Team would be
 notified so that proper response action can be taken promptly.

        Mexican Buffer Zone Surrounding Hazardous Chemical Plant Sitings.  In response to
 concern about residential developments near  hazardous chemical plants, Mexico prohibits the
 construction of new residential buildings within a 1.2 kilometer zone of facilities which use  hazardous
 chemicals, including HF.
Page 48

-------
       4.3.8 Other International Efforts

       There are a number of other international efforts aimed at regulation of hazardous substances
such as HF. Norway has done work on risk-based standards, including promulgating new regulations
for the petroleum industry based on risk assessment with acceptance criteria to be proposed by
facility owners.  Belgium is currently in the process of developing risk-based standards similar to those
applied in the Netherlands.  In Denmark, qualitative and quantitative methods can be used but without
prescriptive acceptability criteria.  Greece is working on legislation. In the consideration of granting of
permits for the installation of new plants, Germany  requires that incremental as well as the total
emission values be calculated if the air emissions of HF and gaseous inorganic fluorides exceed
1 kg/hr.  Italy has implemented the  Seveso Directive but has not used quantitative risk criteria.
Portugal  is in the early stages of development of a risk-based  regulatory methodology. Spain, as a
member  of the European Community, is required to develop an approach under the Seveso Directive,
but has not yet done so.44  Western Australia, Victoria,  New South Wales, and Hong Kong have risk-
based criteria for all hazardous material handling facilities in place.

       4.3.9 International Information Exchange

       The International Conference on Vapor Cloud Modeling. The International Conference on
Vapor Cloud Modeling serves as a  worldwide information exchange for sharing data on spill testing.
is sponsored by the Center for Chemical Process Safety of the American Institute of Chemical
Engineers, The Health and Safety Executive of the United Kingdom, and the United States
Environmental Protection Agency.  This conference provides a forum for presentation of the latest
research on vapor cloud modeling  and mitigating the consequences of accidental releases.  Hydrogen
fluoride has been a focus of several papers  at this conference.  Ftesearch topics have included the
modeling of HF releases, the  modeling of water spray effectiveness on releases, and the chemistry of
mixing anhydrous HF with moist air.
It
                                                                                       Page 49

-------
                                          ENDNOTES
  1.
 3.



 4.



 5.


 6.



 7.



 8.


 9.



 10.


 11.



 12.


 13.

 14.
  •Performance-Oriented Packaging Standards; Changes to Classification, Hazard
  Communication, Packaging and Handling Requirements Based on U.N. Standards and
  Agency Initiative, Final Rule," 49 CFR Pan 107, et. a/., Department of Transportation Research
  and Special Programs Administration. (139.9)

  O'Steen, James K., Office of Hazardous Materials Technology, U.S. Department of
  Transportation, comments from technical review of Hydrogen Fluoride Study Report to
  Congress, Draft May 8, 1992, June 12, 1992.  (366)

  U.S. Department of Transportation, 7990 Emergency Response Guidebook, Research and
  Special Programs Administration, U.S. Department of Transportation, Document Number DOT
  P 5800.5. (489.87)

  Delaware Department of Natural Resources and Environmental Control, Regulation for the
 Management of Extremely Hazardous Substances, Division of Air and Waste Management
 September 25,1989. (550)

 New Jersey Department of Environmental Protection, Toxic Catastrophe Prevention Act,
 Division of Environmental Quality, June 20, 1988. (560)

 Lercari, Frederick, A., State of California Guidance for the Preparation of a Risk Management
 and Prevention Program, California Office of Emergency Services, Hazardous Material Division
 November 1989. (540)

 Nevada State Division of Environmental Protection, Adopted Regulation of the Division of
 Environmental Protection of the Department of Conservation and Natural Resources Mav 21
 1992, LCB File No. R078-92. (362.8)                                                '

 Letter, HF Production and Use in Texas, From:  Sam Crowther, Texas Air Control Board, To-
 Fred Millar, Friends of the Earth, August 17, 1992. (334b)

 Panketh, Joseph, K. Home, K.  Neumann, and S. Keil, Case Studies of Texas Air Control Board:
 Permit Reviews in which Potential for Catastrophic Releases Was Considered, Texas Air Control
 Board, Austin, TX, October 1986. (471)

 Du Pont Chemicals,  Du Font's La Porte, Texas Plant's Disaster Scenarios, Air Permit Submitted
 to TACB-1991, Wilmington, DE,  1991. (139)

 Hofman, Hazel, Shahid Khan, Chris Marlia, and  David Yeh, Supporting Document for Proposed
 Rule 1410: Hydrogen Fluoride Storage and Use, South Coast Management District, CA, March
 19, 1991. (170)

 Kurata, Susan, and Steve Smith, Final Environmental Assessment for Proposed Rule 1410:
 Hydrogen Fluoride Storage and Use, South Coast Management District, April 2, 1991. (290)

 Hofman, Hazel.

 Morns, Jeff, American Petroleum Institute, Fina Oil and Chemical Company, comments from
technical review of Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992 June 1
 1992. (344)
Page 50

-------
15     Technica Inc., Quantitative Risk Assessments of Generic Hydrofluoric Acid and Sulfuric Acid
       Alkylation Units for Phillips Petroleum Company, Management Summary, May 1990. (420)

16.    Quest Consultants, Inc., Sulfuric Acid Aerosoling, Norman, Oklahoma, October 12, 1992.
       Document Number: 92-10-6065. (395)

17     Puschinsky, Bob, Alkylation Technologies and Sulfuric Acid Production Consulting, comments
       from technical review of Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992, June
       28, 1992, p9-10. (377)

18.    Rabin, Jeffrey L., Torrance Asks Court to Declare Refinery a 'Public Nuisance'," Los Angeles
       Times, April 8, 1989, p 1. (421.4)

 1.9    Cole, T.D., Mobil Oil Corporation, comments from technical review of Hydrogen Fluoride Study,
       Report to Congress, Draft May 8, 1992, June 4, 1992. (126)

 20    Superior Court of the State of California For the County of Los Angeles, Consent Decree,
       •People of the State of California v.  Mobil Oil Corporation," October 19, 1990. (370.4)

 21    Commission  of the European Communities, Council Directive 82/591/EEC of 24 June 1982 on
       the Major Accident Hazards of Certain Industrial Activities, Substances Officially Classified to
       Date in the European Communities  Which Enter Into the Scope of Part II Annex II and
       Amendments, June 1990 . (132)

 22    The World Bank, Manual of Industrial Hazard Assessment Techniques, Office of Environmental
        and Scientific Affairs, October 1985. Prepared by Technica LTD, London. (328.5)

 23     The World Bank, World Bank and IFC Guidelines for Identifying, Analyzing and Controlling
        Major Hazard Installations in Developing Countries, Draft, Office of Environmental and Scientific
        Affairs, Projects Policy Department, Washington, DC, February 1985. (530b)

 24     European Technical Industry Council (CEFIC), code of industry recommendations for the
        European hydrogen fluoride industry, European Technical Committee on Fluorine (Comrte
        Technique Europeen du Fluor), Brussels, Belgium. (130.17)

 25     United Nations Environmental Programme, APELL:  Awareness and Preparedness for
        Emergencies at Local Level, A Process for Responding to Technological Accidents, Industry
        and Environment Office, Paris, France,  1988, Document Number: E.88.III.D.3. (467.8A)

 26.    Organisation for Economic Co-operation and Development, Guiding Principles for Chemical
        Accident Prevention, Preparedness and Response,  Paris, France, 1992. (369.1)

 27.    Convention on the Transboundary Effects of Industrial Accidents, United Nations. (135.9)

 28    Mudan, Krishna S., Technica Inc.,  comments from technical review of Hydrogen Fluoride Study,
        Report to Congress, Draft May 8, 1992, June 1, 1992, p 3. (360.2)

 29.    A Guide to the Control of Industrial Major Accident Hazards Regulations 1984, Health and
        Safety Executive, Her Majesty's Stationery Office, London, England. (146)

 30    Public Inquiry Into the Proposed Northwick Village Development, Applications by Aldersgate
         Developments Ltd to Castle Point  District Council, Proof of Evidence Submitted by the Health
         and Safety  Executive, October 1990. (374)
                                                                                       Page 51

-------
   31.




   32.


   33.




   34.





   35.


   36.




  37.




  38.




  39.




  40.


  41.





  42.





 43.





 44.
  Chemical Industries Association, Guide to Safe Practice in the Use and Handling of Hydrogen
  Fluoride, London, England, 1988. (147)                                        n/u/uyw,


  Chemical Industries Association.


  Workplan for the Hydrogen Fluoride Project, Great Britain, received from Craig Matthiessen
  U.S. Environmental Protection Agency. (515)                                  • «««*« i,


  Letter, From: Lolo Heijkenskjold, National Chemicals Inspectorate,  Solna, Sweden, To- Craiq
  Matthiessen, U.S. Environmental Protection Agency, Washington, DC, November 11, 1991.



  Mudan, Krishna S., Technica Inc.


 Proposa/ of the Project -Yellow Book Plus," TNO Prins Maurits Laboratory, The Netherlands.




 Nove/6b Meet'ng N°teS Fr°m the HVdr°9en Fluoride Roundtable, ICF Inc., Fairfax, VA,



 Personal Communication, Conversation with Bruno Lequime, SNPE France
 February 11, 1992. (468.7)


         Communication. Conversation with Georges Marlier, ELF,  France, February 6, 1992.



 Personal Communication, Conversation with Georges Marlier.
             and Harold Lamb Elf Atochem North America Inc., comments from technical

 /oon      y   9en Fluoride Study> ReP°H to Congress, Draft May 8, 1992, June 4 1992.
 (290.8)



f Jw SnJ°hn' ^d W^"e BiSSStt) Life-Cy°le Management of Toxic Chemicals- Hydrofluoric
Add (HF) as a Case History, Environment Canada, Prevention Division Environmental
Emergencies Branch. (440)
                    E"™°"men!a,' Protection Service- Hydrogen Fluoride and Hydrofluoric

                               lnformation for Problem Spi"s' Technical Services Branch-

Ultramar Inc., Quantitative Risk Assessment of the HFAIkylation Unit, Wilmington, CA, January
1992. Prepared by Science Applications International Corporation, McLean VA. (489 85)
 45.     Mudan, Krishna S., Technica Inc.
Page 52

-------
    5.  PROCESS DESCRIPTIONS OF HYDROGEN FLUORIDE INDUSTRY
       This chapter discusses the operations and processes currently used to produce, transport,
and store HF; the production of other chemicals from HF; and other uses of HF.
5.1
HF Manufacture
       HF use is reported as early as the seventeenth century when fluorspar and acid were used to
etch glass1  In 1856, the first synthesis of anhydrous HF was reported.  However, the chemical was
commercially manufactured only as the aqueous solution in concentrations of 38 percent or lower until
1931, when Sterling Products Company shipped the first bulk quantities of anhydrous HF.

       The two major U.S. producers of HF use slightly different manufacturing processes. One
producer uses a rotary kiln and the other uses a pre-reactor vessel with a reactor tube.  Although the
specifics of the processes are considered proprietary, enough information is known to provide a
general description of both. The chemistry, processes, and hazards of HF production are described
below.

        5.1.1  Chemistry of HF Manufacture

        HF is synthesized from calcium fluoride (CaF2) and sulfuric acid (HZSOJ according to the
following chemical equation:
                             H2SOA{liquid)
                                             heat
                                             2HF(gas)  + CaSO4(solid)
        Calcium fluoride is derived from the mineral fluorspar. Acid grade fluorspar is 97 percent
 calcium fluoride or greater.  Sulfuric acid in concentrations between 93 and 99 percent is typically
 used in the manufacture of HF.  The more dilute concentrations of sulfuric acid, however, are more
 corrosive to the reactors and may lead to the formation of HF solution rather than HF gas.  Typically, a
 slight excess of sulfuric acid over fluorspar is used.  The final product may be distilled to a purity of
 99.98 percent HF.2

        5.1.2   Manufacturing Process

        A generalized flow diagram of both the kiln and the pre-reactor/reactor tube manufacturing
 processes is shown in Exhibit 5-1.  Fluorspar, received by an HF manufacturing facility in bulk
 quantities  is stored in warehouses or silos, or it can be stored outside. The first step in the HF
 manufacturing process is fluorspar drying; the fluorspar is generally heated in a kiln to remove excess
 water which can cause corrosion during the manufacturing process.  The heating decreases the water
 content of the fluorspar to as low as 0.03 percent.3  Heated fluorspar passes through a second kiln
 for cooling and is then ground to a fine dust or powder.

         The second reactant, sulfuric acid, is frequently manufactured at the facility where  HF is
 produced  Oleum, a mixture of sulfuric acid and sulfur trioxide (SOg), can be mixed with water to form
 sulfuric acid at the appropriate concentration. At this point, the two HF manufacturing processes differ
 somewhat.
                                                                                      Page 53

-------
                   [Wast* Stream
    H2SO4
   (fluorspar)
                 CaS04
                 (gypsum)
                      to Pond
                                                              EXHIBIT 5-1
                                             Hydrogen Fluoride Manufacturing Processes
                                                           Waste Stream
   * Process may use either kiln reactor or pre-reactor/reactor tube
Source:


-------
       In the kiln manufacturing process, HF is produced by the reaction of fluorspar with sulfuric
acid in a horizontal, rotating, cylindrical reactor kiln which is externally heated by natural gas or fuel
oil.  A typical HF facility uses an air tight rotary kiln.  The fluorspar is conveyed to the reactor through
weigh bins.  The, concentrated sulfuric acid is sprayed onto the fluorspar. A screw conveyor facilitates
the mixing of the reactants as they pass slowly through the kiln. The fluorspar-sulfuric acid mixture
has a residence time in the kiln of between 30 minutes and one hour at a temperature between 200
and 250°C.  Temperature in the kiln is maintained as low as possible to minimize the corrosive effects
of the reactants and products. The gaseous HF is removed at the feed end of the kiln.  The entire
process is kept under vacuum to help pull off the product gases and to promote HF production.
Additionally, the vacuum maintained in the process and the specially designed kiln seals help to
prevent fugitive emissions of HF.4  A major solid byproduct of the reaction, calcium sulfate, is
removed from the opposite end of the kiln.5

       The other process for HF manufacture is very similar to the above except the heat is provided
internally in the reactor by the addition of SO3 and steam. This also adds part of the sulfuric acid by
the reaction:
                                SO3+H2O(st0am) -
Hence, no fossil fuel is necessary to provide the heat of reaction.   The temperature at the entrance
to the  reactor tube is 380°C and the residence time is several hours. Similar to the kiln process, the
entire pre-reactor/reactor process is kept under vacuum to help draw away the product gases and to
promote HF production.  Additionally, the vacuum helps to prevent fugitive emissions of HF.

       In both of the processes, the gaseous products are HF, water, sulfur dioxide (SO2), silicon
tetrafluoride (SiF^, and sulfuric acid. The solid byproduct, calcium sulfate (or gypsum), is removed
through an airlock into a water sluiceway, where it is slurried with water and neutralized.  The gypsum
is then passed to a drying bed  or pond.

       The gaseous products from the kiln process or the pre-reactor/reactor process are first
cleansed of entrained solids and acid mist and then fed to a series of precondensers and condensers.
The precondensers remove any high boiling impurities such as sulfuric acid and water from the HF
product stream. Liquid drips (e.g., unreacted sulfuric acid) from the precondensers are eventually
recycled to the kiln or pre-reactor.  Other condensates are piped to a series of scrubbers and
discarded through normal effluent waste streams. The precondenser also serves to cool the gas
stream before it is condensed.

        Gases remaining after the product stream has passed through the precondensers flow to a
condenser where the HF is liquefied by cooling.  The majority of the remaining HF is recovered by
absorption in H2SO4 which is in turn recycled to the process.7  Other gases, consisting mostly of
fluosilicates, do not condense and  are vented to  acid scrubbers. The resulting fluosilicic acid can be
recovered and sold commercially.  The crude HF is then piped to an intermediate storage facility.

        To make high grade HF (99.98 percent pure), the crude liquid HF is reboiled and distilled.
The distillation  process is used to remove any unwanted impurities by heating the HF to its exact
boiling point and collecting its condensate. The distillation process is carried out under  pressure so
that the HF will boil at higher temperature so water can be used to provide cooling for the
condensation.  Lower pressure would require refrigerated cooling.  Impurities from the distillation
process including SO2, SiF4, water, and sulfuric acid are sent to acid scrubbers and to water
scrubbers and are then removed through normal waste streams.  Potential emissions of  HF from the
                                                                                       Page 55

-------
 HF condensers or distillation columns are directed to the acid scrubber. A second counter-current
 scrubber is used to scrub any HF remaining.8

        After distillation, anhydrous HF liquid may be sold or used on-site as a feedstock for some
 other process such as production of chlorofluorocarbons (CFCs) or aluminum fluoride, or to produce
 various concentrations of aqueous HF.

        In addition to producers of anhydrous HF, there are several other types of facilities that
 process, repackage, and/or redistribute HF for certain use markets.  For example, some chemical
 companies purify and dilute anhydrous HF for use in the electronics industry. Some companies
 combine HF with other acids for applications such as cleaning. A substantial portion of  aqueous HF
 is marketed through chemical distributors as drummed or packaged product. Also, significant
 quantities of drummed 70  percent HF are imported from  Mexican  and off-shore sources.9

        To prepare aqueous HF, anhydrous HF passes through an aqueous mixer.  Water is added as
 needed to prepare the desired concentrations of HF which are then stored on site in specially
 designed alloy tanks.10  Electronic grade HF (usually 49 percent aqueous HF) is produced by
 purifying anhydrous HF in  a gaseous separation process. In this process, purified HF vapor is
 absorbed in a column supplied with deionized water. The aqueous HF product from the column is
 cooled and stored in tanks. Inert gases vented from the  column are cleaned in a scrubber before
 being discharged to the atmosphere.11

        To mix HF with additives to make a new wholesale product,  liquid HF is received in drums and
 Is pumped, along with other acids, water, or other chemicals to a batch mixing tank.  Acid  mix is then
 manually packaged in polyethylene bottles or drums. The distributor may store the packaged HF
 product in a warehouse before shipment to a customer.

 5.2    Transportation and Storage

        5.2.1   General

        The major HF producers, Du Pont and Allied-Signal, transport HF to users across North
 America mainly by rail car  or by tank truck. In 1987 approximately 274,000 short tons of HF were
 shipped in bulk in the U.S.12 Bulk shipments (rail or truck) of HF are limited mainly to anhydrous HF
 and 70 percent aqueous HF.  Aqueous HF in concentrations below approximately 60 percent are too
 corrosive to be compatible with steel.13

        Anhydrous HF is commonly transported in specially designed steel rail tank cars and tank
 trucks.  Rail tank cars in common use for anhydrous HF have capacities of approximately 20 to 91
 tons.  Tank trucks used for anhydrous HF typically have capacities of up to 20 tons. Anhydrous HF in
 smaller quantities may be shipped in cylinders of various  sizes. Anhydrous HF is shipped under its
 own vapor pressure as a liquid because of its relatively low boiling point and high vapor  pressure.
 Aqueous HF (concentration 70 percent) is also shipped in rail tank cars with capacities of 32 to 80
 tons and tank trucks with capacities up to 20 tons.  Aqueous HF may be shipped in polyethylene-lined
 drums and polyethylene carboys.  Aqueous HF is not generally transported under pressure.14
 Appendix XI presents DOT container specifications for HF.

        5.2.2  Loading and Unloading Procedures

        Facilities can use two methods to unload HF from a tank truck or rail car to a storage tank.
 Compressed gas (e.g., nitrogen) or a pump can be used  to move the HF from one container to
 another. Trained, professional drivers dedicated to HF transport from both Allied-Signal and Du Pont
 conduct unloading with varying degrees of involvement from employees at receiving facilities. The HF
Page 56

-------
producers also have dedicated fleets of tank cars and tank trucks that are equipped with vapor valves,
liquid valves, and relief valves (see Appendix XI). The tank car or tank truck is connected to the
storage tank by flexible hoses provided by the HF transporter or the receiving facility.

       In the compressed gas method, pressurized nitrogen is normally used.  For unloading
anhydrous HF, the compressed gas used by the facility must have a very low moisture content to
prevent the formation of hydrofluoric acid which is highly corrosive.  Maximum pressures for nitrogen
are typically between 80 and 100 psi.   Prior to nitrogen-pressurized HF unloading, all connections are
checked  fpr leaks when the valves are opened. In the unloading of a tank car, the liquid HF valve to
the storage tank is opened and the nitrogen flow to the tank car is started. To transfer the HF, the
tank car  isW a pressure approximately 25 psi higher than the storage tank.  The higher pressure is
maintained by venting HF gas from the storage tank to an acid absorption system. Exhibit 5-2 shows
a typical  unloading operation using compressed gas.  When unloading is complete, the line to the
storage tank is cleared with nitrogen.15

       Exhibit 5-3 shows a typical  pump unloading operation. During pump unloading, the piping
and flexible hose may be under vacuum. However, some facilities provide a small amount of dry
pressurized nitrogen gas to prime the pump and to prevent a vacuum from developing in the tank car.
For pump unloading, HF vapor from the top of the storage tank may be vented back into the tank car
so that a pressure equilibrium is maintained.  Pumps may be centrifugal, rotary, positive displacement,
or sealless types.

       5.2.3   Facility Storage

       Anhydrous HF is usually stored under ambient conditions in pressure vessels like cylinders or
tanks because of its high vapor pressure. Anhydrous HF may also be stored under refrigeration,
which reduces its vapor pressure.  Bulk storage vessels for HF are usually single walled horizontal
storage tanks ranging in capacity from 6,000 to 8,000 gallons to as large as 250,000 gallons or
spheres  with capacities up to 500,000 gallons.16'17

5.3    Fluorocarbon Production

       The largest use of HF is in the manufacture of chlorofluorocarbons, hydrofluorocarbons, and
hydrochlorofluorocarbons (i.e., CFCS, HFCs, and HCFCs).  Virtually all modern air conditioning and
refrigeration equipment is designed exclusively for fluorocarbon refrigerants.  These chemicals are also
 used as solvents,  sources of raw material for production of fluoroplastics, anesthetics and fire
extinguishing agents.18

         CFCs and HCFCs are produced by reacting anhydrous HF with chlorinated hydrocarbon in
the presence of a catalyst. A specific chlorinated hydrocarbon is used to produce a specific CFC or
 HCFC (e.g., carbon tetrachloride is used to produce CFC-11 and -12 and chloroform is used to
 produce HCFC-22).  A typical schematic of a CFC or HCFC manufacturing process uses a liquid
 phase reaction shown in Exhibit 5-4.   HF and the specific chlorinated hydrocarbon are pumped into a
 heated vessel in the presence of a catalyst.  The catalyst is typically antimony pentachloride.
 Although the fluorination step is slightly exothermic, the overall reaction is endothermic requiring
 additional heat to reach high conversions. Typical operating conditions in the catalytic reactor are
 temperatures between 21 and 38°C and pressures between 115 and 265 psi. The reaction goes
 nearly to completion and  leaves little unreacted HF. The crude CFC vapors from the reactor are fed
 directly  to the enriching column to purify the gas products by returning the underfluorinated material
 back to  the reactor.
                                                                                       Page 57

-------
                                               EXHIBIT 5-2
                               Nitrogen Unloading of HF from Tank Truck
     Source:  Du Pont Chemicals, Hydrofluoric Add, Anhydrous - Technical: Properties, Uses, Storage and Handlina
             Wilmington, DE. (137.5)
Page 58

-------
                                          EXHIBIT 5-3
                             Pump Unloading of HF from Tank Car
                                Pressure
                               Relief Valve
 Level
 Guage
 (Weigh
 Cell)
Source:  Du Pont Chemicals, Hydrofluoric Acid, Anhydrous - Technical: Properties, Uses, Storage and'Handling,
        Wilmington, DE. (137.5)
                                                                                              Page 59

-------
                                                 EXHIBIT 5-4
                                        CFC Manufacturing Process
          Anhydrou* HF •
          CCMorCHCO—^ HC Storage
          SbCOorSbas-
     Sources: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Prevention Reference
             Manual: Chemical Specific Volume 8: Control of Accidental Releases of Hydrogen Fluoride, Research
             Triangle Park, NC, August, 1987, Document Number EPA-600/8-87-034h. Prepared by the Air and Energy
             Engineering Research Laboratory, (400) modified as per Seringer, Carolyn S., Du Pont Chemicals,
             comments from technical review of Hydrogen Fluoride Study, Report to Congress, Draft May 8,1992, June
Page 60

-------
       After purification, the product stream containing hydrogen chloride, HF, and'CFC is sent to an
acid recovery column, typically operating at 115 to 265 pounds per square inch.  Hydrogen chloride is
concentrated at the top of the column and is recovered as a by-product.  The bottom contains the
CFC products and residual HF.  The HF is removed in an HF settler and it is recycled to the reactor
system. The mutual solubility of HF and the CFCs is temperature dependent, often requiring
temperature as low as -30°C for  separation. Trace acidic impurities are removed from the CFC
products by scrubbing with water and dilute caustic solution. After scrubbing, the product stream is
dried and fractionated into various CFC products.

       HCFC and HFC alternatives are produced through various routes which are different from and
more complex than CFC production. Routes to HFCs  proceed through CFC or HCFC intermediates.
They require approximately three times the HF per pound produced versus the CFCs they are
replacing. There are several potential HCFC and HFC products being developed. Because HCFCs
contain some chlorine (they have 95 to 98 percent less ozone depletion potential than the CFCs being
phased out), they are recognized as being transitional substitutes, and may be phased  out early next
century.  HFCs however, are seen as long term substitutes that can be used in equipment that was
developed to use CFCs. There is a strong likelihood that the HFCs may be usedwell into the next
century.  For this reason, HF may be needed to manufacture these compounds.""
19
5.4    Alkylate Production for Gasoline

       Perhaps the most highly publicized and controversial use of HF is as a petroleum refining
catalyst in the production of gasoline blending components.  Alkylate is an extremely valuable
gasoline component due to its high octane and low vapor pressure.  It constitutes approximately 11
percent of the gasoline pool.20 Octane is a measure of the anti-knock characteristics of a fuel when
burned in an internal combustion engine.  Higher octane results in less engine wear and more efficient
engine performance as a result of higher compression ratios.  The branched structure of alkylate is
responsible for its high-octane rating. The  low vapor pressure of alkylate is a valuable property to the
refiner because it allows the blending of higher vapor pressure butanes (and pentanes during certain
times of the year) to adjust gasoline vapor pressure to specifications.

       In addition, alkylate is the major component of gasoline produced for use in piston engine
aircraft. Alkylate is the only available blendstock that allows the refiner to meet the high octane and
paraffin requirements of aviation gasoline. Alkylate has also provided a means for producers to
reduce the lead content of aviation gasoline, and any future move to lead-free aviation gasoline will
increase its importance.21

       In the alkylation  process, three- and four-carbon light olefins (i.e., propylenes and butylenes)
are typically charged from the Fluid Catalytic Cracker (FCC) and a coker, if present in the refinery, and
are then reacted with isobutane in the presence  of an acid catalyst to produce branched, seven- to
eight-carbon  paraffins, collectively known as alkylate.22 The olefin feedstock may also contain
amylenes (five-carbon olefins). In this case, nine-carbon alkylate components will also be produced.
Alkylate components have boiling points in the gasoline boiling range.  Commercially, HF and sulfuric
acid are currently the only alkylation catalysts available. The alkylation catalyst serves to speed
reaction times and facilitate less severe reaction  conditions. A small amount of catalyst, however, is
consumed  as the result  of undesirable side reactions.

       In the past, organic lead compounds such as tetraethyl lead were added to gasoline to
improve octane. As lead was phased out due to environmental concerns regarding its toxicity,
gasoline blending components from the catalytic reforming (reformate) and alkylation processes
became more important for octane contribution.  Reformate contains aromatics (benzene, toluene, and
xylene), whereas alkylate contains no aromatics. The Clean Air Act Amendments of 1990 have
established limits on the aromatic content of gasolines due to toxicity concerns, as well as limits on
                                                                                      Page 61

-------
volatile organic compounds (VOCs), such as light olefins and butanes, due to their tendency to
contribute to ozone formation. Light olefins and isobutane are converted to alkylate in the alkylation
process.  Thus, the alkylation process reduces light components from the gasoline pool.23  As a
result of gasoline reformulation requirements, refiners are increasingly reliant on alkylate as a gasoline
blending component

       Alkylation units are the major primary means of conversion of light olefins produced in  FCC
units to liquid gasoline blending components. Oxygenates such as methyl tert-butyl ether (MTBE) and
tertiary amyl methyl ether (TAME) may be produced from the olefins isobutylene and isoamylene,
respectively, and are also used as gasoline blending components. Refinery production of these
olefins, however, may exceed the feedstock demands of the oxygenate units. Alkylate, on the  other
hand, is produced from a variety of olefin feedstocks, including isobutylene, isoamylene, and
propylenes and other butylenes and amylenes. Thus,  alkylation units can utilize a range of light
olefins produced at the  refinery.

       5.4.1   HF Alkylation  Processes

       The alkylation process incorporates three general steps: reaction, separation,  and treatment.
After the isobutane and olefins are reacted in a reactor in the presence of the acid catalyst, the
alkylate product, byproduct propane and butanes, unreacted isobutane, and acid are then  separated
through settling and distillation. Acid and unreacted isobutane are recycled to the reactor with fresh
acid, isobutane, and olefin feed makeup as needed.24  Alkylate, byproduct propane, and byproduct
butane are treated to remove fluorine compounds and/or residual HF before  being sent to storage.

       HF alkylation units account for about 52 percent of the alkylate produced in the United
States.25  HF serves as the reaction catalyst; nearly all of the HF is recovered from the process, with
only a small amount of HF consumed because of side reactions between the acid and impurities in
the feedstocks.26 Approximately 0.1 to 0.2 pounds of HF are consumed per  barrel of alkylate
produced.27

       The two major licensors of HF alkylation technology are Phillips Petroleum and UOP. The
Phillips design uses gravity to  circulate  HF between the settler and the reactor whereas the UOP
design uses a pump for circulation.  Both technologies require relatively small amounts of anhydrous
HF (i.e., roughly one truckload per month for a 10,000 barrel per day facility).  The key safety
advantage of the UOP acid circulation system is that smaller HF inventories are needed than the
Phillips system. The key safety advantage of the  Phillips process is that no pump is needed to
circulate the liquid HF between the reactor, the settler, and the heat exchangers, thus eliminating a
common leak source.28-29

       Phillips Process

       The Phillips process is used in over 85 refineries throughout the world. Exhibit 5-5  provides a
simplified  process flowchart. Exhibit 5-6 shows a closeup sketch of the Phillips reactor, settler, HF
storage, and cooler.  At the start of the  process, the isobutane and olefin feedstocks are mixed and
dried.  This mixture is then  introduced to the reactor at high velocities through nozzles to provide
contact with the HF catalyst. An  emulsion or suspension of HF and hydrocarbons is formed that rises
in the reactor toward the settler.  The reactor is essentially a vertical pipe from the acid cooler to the
settler. The alkylation reactions are exothermic, therefore, the reactor effluent is warmer than the
influent and flows to the settler due to the density difference  between cold feed and warmer effluent.
Most reactions occur as the emulsion rises through the reactor. The reaction mixture  of products,
byproducts, HF and unreacted isobutane exits the reactor and enters the settler where the  acid
separates from the hydrocarbon phase (i.e., alkylate, isobutane, propane).  HF settles by gravity to the
bottom of the settler, where it is removed and cooled in one or more heat exchangers.   Liquid
Page 62

-------
                                 EXHIBIT 5-5
                           Phillips Alkylation Process
 Isobutane

 aieflnfeed
                                      JJr-M
                                              Defluorkiatorand
                                               KOHTreeter
Source: Albright, Lyle F., "H2SO4, HF Processes Compared, and New Technologies. Revealed," Oil and Gas
      Journal, November 26, 1990, p 70-77. (10.1)
                                                                         Page 63

-------
                                             EXHIBIT 5-6
                                     Phillips HF System Reactor
                                                                        Reactor Riser
                                                                    Reactor Standpipe
                                                                Hydrocarbon
                                                                Feed In
    Source:  Albright, Lyie F., "H2SO4, HF Processes Compared, and New Technologies Revealed," Oil and Gas
            Journal, November 26,1990, p 70-77. (10.1)
Page 64

-------
HF flows to the reactor from the settler by gravity.  Fresh HF is stored in the settler bottom for use as
makeup for acid lost as a result of side reactions.  A slip stream of circulating HF is charged to the
acid regenerator to remove contaminants to maintain acid purity.

       The hydrocarbon phase from the settler contains mainly isobutane and alkylate with lesser
amounts of propane, n-butane, dissolved HF, and  isoalkyl fluorides. In the main fractionator, this
mixture is separated into four streams: (1) the overhead stream is a mixture  of HF and propane, (2) a
liquid side stream is mainly isobutane that is recycled to the reactor, (3) another liquid side stream is
n-butane rich and (4) the bottom stream  is alkylate product. An HF stripper is used to separate
propane from HF. The n-butane and propane must be treated with potassium hydroxide (KOH) to
remove any residual HF and the propane is further treated with alumina to remove isoalkyl
fluorides.31

       Polymers may form as the result  of side reactions in the alkylation process. Because they are
soluble in HF, polymers tend to accumulate in the  recycle acid stream.  To maintain acid purity, a
critical process parameter, a slip-stream  of acid is  drawn off the bottom of the heat exchangers and is
treated with a hot isobutane vapor stream to strip HF away from the polymer and water.  The
combined HF-isobutane mixture is recycled to the  settler.  Polymers are removed in the form of acid
soluble oil (ASO)."
32
       UOP Process

       The UOP process is used in 62 refineries throughout the world, 36 of which are in the United
States.33'34  Exhibit 5-7 gives a flow-diagram of a two reactor unit.  In this process, dried isobutane
and olefins are premixed before being fed to a cylindrical reactor.  HF is stored in a vessel in the
alkylation unit. Fresh makeup HF is added as needed to the recycled acid and is pumped into the
bottom of the reactor. The hydrocarbon mixture is introduced through several nozzles positioned at
various points in the reactor to achieve good dispersion and mixing.

       Since the alkylation reactions are exothermic, a water cooled heat exchanger in the reactor  is
used to maintain the reaction mixture at the desired temperature. The effluent mixture of alkylate
product, hydrocarbon byproducts, unreacted chemicals, and HF leaves the reactor and is charged to
a settler where the HF and  hydrocarbon phases are separated.  Facilities with relatively small capacity
use a single reactor; large facilities use two  reactors  in series.35  While the UOP main fractionation
column (or isostripper) is operated somewhat  differently from that of the Phillips process,
hydrocarbons and  regenerated HF are separated and treated in much the same manner.

       In a two reactor  system, the olefin feed is split to the two reactors. Olefins charged to the first
reactor are essentially completely reacted.  A stream of HF, alkylate, and isobutane proceeds from the
first reactor to the first settler.  The hydrocarbon stream from the first settler, consisting of unreacted
isobutane, propane, and alkylate product, is combined with additional olefins and fed to the second
reactor.  The HF stream from the first settler is also fed into the second reactor.  A second settler is
provided to separate the liquid phases in the stream.
                                 36
       HF inventories in UOP units range between 5 and 10 pounds of HF per barrel of alkylate
         i 37
                                                                38
produced.   These inventories are lower than those for Phillips units.    UOP units require lower
inventories because the settlers are typically small, the heat exchangers are installed in the reactor
(unlike the external heat exchanger in the Phillips reactor), and the pumped circulation of the emulsion
across the heat transfer surfaces in the reactors increases reaction and heat transfer efficiency.39
                                                                                       Page 65

-------
                                          EXHIBIT 5-7
                                     UOP Alkylation Process
                       UOP C3-C4 HF Alkylation Process*
                             Polymer and CBM
                             to Neutralization
                                                    D =  Depropanizer
                                                   HS-  HF Stripper
                                                   KT=  KOHTreater
                                                   AT m  Alumina Treater
                                AR - Acid Regenerator
                                  I - teostripper
     Split feed, series recycle
    Source: Albright, Lyle F., "H2SO4, HF Processes Compared, and New Technologies Revealed," Oil and Gas
           Journal, November 26, 1990, p 70-77. (10.1)
Page 66

-------
5.5    Uranium Processing

       Current technology for the manufacture of uranium reactor fuel used in commercial electric
power generation and DOE weapons development requires that the uranium be converted to gaseous
uranium hexafluoride (UF6) for enrichment.  In this process, HF is used as a source of fluorine.
Following enrichment, reactor fuel elements are manufactured by converting UF6 to uranium dioxide
(UO2).40  Only two facilities produce UF6 in the United States.

       To convert uranium ore concentrate (yellowcake) to gaseous UF6, two basic reaction routes
are used.  The first,  employed at the Allied Chemical Facility in Metropolis, Illinois, relies on a multistep
process. The first step is reduction of the ore concentrate to impure UO2. This is followed by the
hydrofluorination reaction between anhydrous HF and uranium dioxide to produce impure uranium
tetrafluoride (UF^.  The solid UF4 is then fluorinated with fluorine gas (electrolytically generated from
HF), yielding gaseous UF6.  Metal impurities forming volatile fluorides are produced and sent to the
UF  off-gas stream,  where they are next removed by fractional distillation.  The purified UF6 product is
then cooled, drained into a cylinder and allowed to solidify before transport to an isotope separation
plant for enrichment.41

       The process employed at the Sequoyah Fuels Facility in Gore, OK relies on nitric acid
dissolution of the ore concentrate followed by solvent extraction purification of the uranium-bearing
solution. Purified uranyl nitrate solution is evaporated and thermally denitrated to UO3, which in turn is
reduced to UO2. This UO2 product is then hydrofluorinated by anhydrous HF to UF4. The solid UF4 is
fluorinated to gaseous UF6, with fluorine gas. The UF6 is then cooled, condensed and solidified in
cylinders before transportation to an isotope separation plant.

        Following either process, the UF6 undergoes isotopic separation yielding UF6 enriched  in the
fissile isotope, uranium-235 (U235). The reject UF6> depleted in U235, is condensed and stored, while
the enriched UF6 is chemically converted to UO2 for use as reactor fuel.  This chemical conversion
involves a number of reactions that vary, depending on the type of process involved. Some
processes include reaction with ammonia; in these processes the UF6 is converted to ammonium
diuranate, which is  dried and thermally decomposed to UO2.  Processes are available for regenerating
HF from UF6, depending on the need for depleted uranium materials.

5.6    Aluminum Fluoride and  Aluminum Manufacturing

        Aluminum fluoride (AlFg)  is used to promote the fusing of aluminum, to prevent the formation
of oxides, and to suppress sodium ion formation in the electrolytic manufacture of  aluminum metal.
 Depending on the manufacturing process chosen, aluminum fluoride can be made using either HF or
fluosilicic acid (a byproduct from phosphoric acid production).43 Some large scale fully integrated
 aluminum producers can produce their own aluminum fluoride.

        In a typical process to make AIF3, gaseous HF emanating from the kiln is contacted directly
 with hydrated aluminum in a fluidized  bed reactor. To make aluminum, aluminum fluoride and
 aluminum oxide are added to an electrolytic cell. An electric current in the cell causes a reaction
 which releases the oxygen in the aluminum oxide to produce CO2 and aluminum metal. HF is also
 evolved in this process and is ducted to an air control system. The HF reacts with fresh aluminum
 oxide and is returned to the electrolytic cell.44 Thus, most of the HF used by aluminum companies
 for AIF3 production is generated and used captively in the gaseous state and is not isolated as a
 liquid product.  Only small quantities of aluminum fluoride are supplied to this industry by HF
 merchant producers.45
                                                                                       Page 67

-------
  5.7    Electronics Manufacturing

         Ultra-high purity aqueous HF is used in the manufacture of semiconductor chips.46 Its most
  concentrated form is typically 49 percent. While it is purchased in this concentration, it is typically
  used in much more dilute solutions,  1 to 10 percent. It is frequently mixed with ammonium fluoride
  (NH4F) to buffer its effect in what is known as Buffered Oxide Etch (BOE) to etch silicon dioxide and
  silicon nitride. It may be mixed with  an oxidizing agent such as nitric or chromic acid to etch
  polycrystalline or single crystal silicon.  Silicon is usually etched to delineate defects in the crystal, to
  define patterns in  integrated circuits, and to remove areas damaged by machining in the production of
  silicon ingots and wafers.

         HF mixed with a hydrocarbon-based surfactant effectively eats away the surface of a
  semiconductor wafer and easily covers its surface, greatly facilitating manufacture of very large-scale
  integrated circuits (VLSIs).^ Ultra-pure HF may also be used in the manufacture of 16-megabit
  dynamic random access memory, application-specific integrated circuits, and logic devices.49

         In a typical HF etching process, HF is pumped from 55 gallon drums or one gallon containers
  under nitrogen pressure, passed through a surge tank, and finally sent through a manifold which
  directs its flow to etching basins.  The wafers which are to be etched are set into trays fastened to
  automatic dipping arms.  A technician loads the wafers into the trays. After dipping, the wafers are
  rinsed in low conductivity deionized (Dl) water. The spent Dl water is pumped to an acid treatment
  plant for neutralization.00 A gaseous process using anhydrous  HF is also used in the
  semiconductor industry to etch integrated circuits.

  5.8     Chemical Derivatives Manufacturing

         HF is used directly or indirectly as a source of fluorine in the manufacture of many organic
 and inorganic compounds having highly specialized and valuable properties.  These compounds
 Include fabric and fiber treating agents, herbicides, pharmaceutical intermediates, and inert fluorinated
 liquids.  Other products include boron trifluoride (BFg), sulfur hexafluoride (SFg), and fluoride salts.51
 Processes used in the manufacture of some examples of these  derivatives are briefly discussed below.

        5.8.1   Inorganic Derivatives

        High purity anhydrous HF is the principal raw material for the production of fluorine.  Fluorine
 is generated in electrolytic cells, along with hydrogen gas. Electrolyte for the cells is prepared by
 mixing of potassium bifluoride (KF.HF), which is produced from  HF, with anhydrous HF to form
 (KF.2HF).  HF is stored in bulk and charged to a holding tank; it is continuously fed to the electrolytic
 cells from  the holding tank to maintain an  HF concentration of 40 to 42 percent.  HF is not used
 directly as the electrolyte because of its low conductivity.  Commercial fluorine-generating cells usually
 operate  at temperatures of 60 to 110°C and are cooled with water at 75°C.  The electrolyte level must
 be maintained at a set level below the cell head to maintain a seal between the fluorine and hydrogen
 compartments. Entrained electrolyte is removed from the product gas streams from the cells with
 demlsters and filters.  Most of the HF  is  removed from the gas streams by cooling to -110°C leaving a
 concentration of about 3 mole percent HF.  The condensed HF is recycled. For some uses no further
 purification of the fluorine is needed.  Depending on the intended use of the fluorine the HF
 concentration in the fluorine may be reduced to less than 0.2 mole percent by using sodium fluoride
 towers or further cooling to freeze out the HF.52

       Most inorganic fluorides are prepared by reaction of HF with oxides, carbonates, hydroxides
 chlorides, or metals. Ammonium bifluoride, which may be used for rapid frosting of glass and in
 metallurgical uses, is produced by a gas phase reaction of one mole of anhydrous ammonia with two
Page 68

-------
moles of anhydrous HF. Sulfur hexafluoride, which has a number of electrical uses, is produced by
reaction of sulfur vapor with fluorine.53

       Fluorosulfuric acid (HSO3F) (also called fluorosulfonic acid) is produced by mixing solutions of
HF and SO, in HSO,F or introducing HF and SO3 separately into a stream of HSO3F. HSO3F is used
as a catalyst for a number of processes and in the production of other fluorine derivatives.  Boric acid
and HSO3F may be use to produce boron trifluoride.  (Boron trifluoride may also be produced using
fluorspar,  borax, and sulfuric acid.)54

       Fluoboric acid, used as an intermediate for fluoborate salts, which have a number of
metallurgical uses, is produced by reaction of 70 percent aqueous HF with boric acid.  The reaction is
exothermic and is controlled by cooling.  The commercial produces usually a 48 to 50 percent
solution containing excess boric acid to eliminate any HF fumes.

        5.8.2   Organic Derivatives

       The most commonly used fluorinating agents for production of organic fluorine derivatives are
fluorides of alkali metals, which are generally produced from HF.  For example, sodium
monofluoroacetate, a rodenticide, may be produced by reaction of potassium fluoride with
chloroacetic acid ester. In one commercial process, ethyl chloroacetate,  purified by distillation to
remove traces of acid and water,  is mixed with potassium fluoride that has been oven dried and finely
powdered. The reaction takes place in an autoclave with stirring for 11 hours.

        HF may also be used directly as a fluorinating agent in the production of organic fluorine
compounds. Benzotrifluoride, used as an intermediate in the production of herbicides, drugs,
germicides, and dyes, is produced by reaction of benzotrichloride with anhydrous HF under high
pressure Typically, HF is reacted with benzotrichloride in a ratio of 4 moles HF to 1 mole      ^
benzotrichloride at temperatures of 80 to 110°C and pressures of 220 to  225 psi for 2 to 3 hours.

        Fluoroaromatics can also be produced by diazotization of substituted anilines with sodium
 nitrite in anhydrous HF, followed by in situ decomposition of the aryldiazonium fluoride.  The resulting
 aromatic fluorocarbon is further processed for pharmaceutical, agrochemical, and engineering resin
 applications.58

        A number of fluorine-containing polymers are produced from organic derivatives of HF.
 Tetrafluoroethylene, the monomer for polytetrafluoroethylene (Teflon™) is generally produced by
 pyrolysis of chlorodifluoromethane, an HF derivative (see Section 5.3, Chlorofluorocarbon Production).
 Vinylidine fluoride, used to produce polyvinylidine fluoride, may also be made from CFCs by several
 routes.59

 5.9    Processes Using Aqueous HF

        While the majority of HF consumed by industry is in the anhydrous or 100 percent form,
 aqueous HF solutions with concentrations of 70 percent and lower are used in stainless steel pickling,
 chemical milling, glass etching, exotic metals extraction and quartz purification.   Aqueous HF is
 used in combination with other acids for cleaning. As noted above, aqueous HF  is also used in
 electronics manufacture and in the production of some inorganic fluorine compounds.

         Stainless steel pickling requires mixtures of dilute HF and nitric (HNOg) acids to remove oxide
 scale formed on stainless steel during the annealing process.  Pickling gives stainless steel  its
 characteristic shiny appearance.  The concentrations of pickle acids used vary but  are typically 2 to
 3 5 weight percent HF and 6 to 10 weight percent HNO3,  Over time, the metal concentration in the
                                                                                        Page 69

-------
  pickling bath increases, leading to a decrease in bath activity. The spent acid must be replaced
  periodically, either by new acid or recycled acid.

         Spent pickling acid can be neutralized with lime. The resultant sludge is landfilled, and soluble
  nitrate salts are discharged with wastewater.  This has been the conventional method of dealing with
  HF/HNOg waste acid  At least until the early  1980's, no hydrofluoric acid recovery processes were in
  operation in the U.S. '

         In the late 1980's, AQUATECH developed an acid recovery system that can be used to recycle
  waste HF pickling acid.  Waste acid may be recovered by neutralizing it with potassium hydroxide
  filtering the resulting KF/KNO3 salt solution to remove metal hydroxides for recycle, and converting the
  clean potassium salts into a mixed acid (HF/HNOg) and potassium hydroxide base  using
  electrodialysis.  This system has been successfully used at Washington Steel Corporation to recycle
  valuable components of spent pickling acid, and reduce the  amount of waste HF acid.62

         Mixtures of aqueous HF and HNO3 are used in the aerospace industry for paint stripping and
  cleaning aircraft surfaces. HF-based cleaning solutions are approved by the Federal Aviation
  Administration (FAA) and are required for some applications.

         Aqueous HF is used in the manufacture of glass articles.  A mixture of HF and sulfuric acid
  may be used for acid-polishing in the mechanical finishing of glass.  Dilute aqueous HF may be'used
  for acid etching or frosting to produce articles with good light-diffusing properties 63
 5.10
Dissolving Ores for Production of Tantalum and Columblum (Niobium) Metals
         Aqueous HF is used to digest metallic ores to produce tantalum (Ta) and columbium (Cb)
 metals. These metals are used in many different applications including military electronics VCRs TVs
 and other electronic systems, aerospace superalloys (used in jet engine components), medical
 diagnostic equipment, pacemakers, and anti-armor ballistics.  One of the unique properties of these
 metals is their extreme resistance to corrosion.64

        The only process currently  used commercially to produce these metals involves HF  The use
 of HF is essential because the Ta and Cb contained in the ores are basically insoluble in most acids
 except HF. Also, HF maintains a high,degree of purity in the liquid phases of the process. This purity
 is essential to the quality of the finished and intermediate products.  Furthermore, slightly radioactive
 contaminants in the ore, such as uranium and thorium, react with HF to form insoluble fluoride
 compounds which are safer and more easily handled than liquid waste streams containinq these
 radioactive elements.

        The production process involves converting Ta/Cb-bearing ores and slags (purchased raw
 materials) to pure chemicals, metals, and alloys through a series of chemical and metallurgical
 operat tons. These include grinding, chemical digestion and dissolution, filtration, solvent extraction
 crystallization, drying, calcination, pressing, sintering, chemical reduction, melting, forging, swaging'
 rolling, and drawing.                                                                       a'

        The Ta/Cb-bearing ores are first ground and then fed to digestion tanks containing
 hydrofluoric acid. The acid dissolves the Ta and Cb from the ore to produce fluorotantalic acid
 ,£f7> fS fi"or°°olumbic acid (H2CbF7)- Ore impurities (e.g., Al, Ca, Mg,  U, Th, etc.) also react to
 form insoluble fluoride compounds.  After a sufficient dissolution period, the slurry is filtered, removing
 the insoluble compounds and the leftover solution of Ta and Cb acids is pumped to the metal
 separation area.  The Ta and Cb are continuously extracted from the feed solution by a solvent
 extraction process utilizing contact with methylisobutylketone (MIBK), hydrofluoric acid, sulfuric acid
 and  water. This process separates the Ta/Cb solution into two separate intermediate product
Page 70

-------
streams, one containing H2CbF7, and one containing H2TaF7. There is also a liquid waste stream
comprised of an aqueous solution of sulfuric and hydrofluoric acids.67  This acid stream is either
pumped to intermediate holding tanks^or to on-site waste treatment plant tanks.  Wastewater from the
HF process is neutralized using lime.°°
                  68
       After further processing, the Ta is reacted with potassium chloride (KCI) and HF followed by
cooling in a crystallizer to form potassium tantalum fluoride (KgTaFy). The KgTaFy is further reacted
with metallic sodium in a sealed reaction vessel utilizing electric furnaces followed by cooling to form a
mixture of Ta metal  and  salts, potassium fluoride (KF) and sodium fluoride (NaF). The fluoride is
extracted from this mixture by leaching with water followed by drying in a steam-heated dryer.  About
half of the Ta product is sold at this stage and the other half is pressed, purified, and converted to
form a variety of other end products.  The remainder of the Cb processing does not involve HF.

5.11   Linear Alkylbenzene Production

       HF is used as a catalyst in the production of linear alkylbenzene (LAB) which is ultimately
used to make industrial  and household detergent.  Detergents fall into two categories:  hard or DDBS
(sodium dodecylbenzene sulfonate) detergents, that are slow to biodegrade: and soft or LABS (linear
sodium alkylbenzene sulfonate) detergents, which  rapidly biodegrade.  Since the 1960's, the use of
DDBS or hard detergents has been largely phased out. Linear alkylbenzene is used in the production
of soft detergent.
70
        5.11.1  Chemistry of Alkylation with HF Catalyst

        The starting materials in the manufacture of soft detergents are benzene and linear paraffins,
which are long single chain hydrocarbon molecules. The alkylation process results in the attachment
of the linear paraffin to the benzene molecule.  There are two commercial production processes
available: one involves the use of an aluminum chloride catalyst, and the other the use of HF as the
catalyst.71  In the United States, only two companies use HF as a catalyst for linear alkylbenzene
production.

        5.11.2 Manufacturing Process

        Typically, in the detergent industry, the UOP process is utilized in HF alkylation.  It is similar to
that used in the petroleum refining industry, and is described in detail in section 5.4.1 of this report.
In a UOP detergent alkylation process, where HF is the catalyst,  paraffins are converted to olefins
through catalytic dehydrogenation (i.e., the removal of hydrogen).  Anhydrous HF and benzene are
then added to the olefins. HF catalyzes the reaction between the olefins and benzene, producing
crude linear alkylbenzene. The HF used in the reaction is recovered and recycled.  The  final step in
linear alkylbenzene production is purification.  The product alkylbenzene sulfonate or soft detergent is
obtained by sulfonation of the LAB intermediate with oleum  or SO3, in a separate  process unit.

        Vista Chemical Company's Lake Charles LAB Plant in Westlake, Louisiana uses HF as a
catalyst in linear alkylbenzene production. Vista typically has on hand from 65,000 to 70,000 gallons
of HF to produce 200 million pounds annually of LAB.  Of these, 40,000 are used in the actual
chemical process, while the remaining 25,000 to 30,000 gallons are stored  at ambient temperature.
Vista is attempting to remove the excess and maintain only  a minimum inventory. HF is  transported to
their facility once or twice a year, by tank truck.72

 5.12   Pharmaceutical Production

        HF is used as a catalyst and solvent in the production of an  intermediate  in acetaminophen
 production. The steps in the production of the intermediate are:73
                                                                                       Page 71

-------
         (1)     Combination of two compounds to form an intermediate (the reaction is non-catalytic);

         (2)     The catalytic rearrangement of the intermediate formed in (1), using anhydrous HF as
                both catalyst for the reaction and solvent for the products formed in this step and
                maintaining low temperatures;

         (3)     Recovery of the HF catalyst/solvent and subsequent recycle to step (2); and

         (4)     Purification of the product formed in (2).

         HF is used in this process for its reaction selectivity (i.e., HF does not catalyze other,
 undesirable reactions that might produce unwanted byproducts) and because no other solvent is
 required when HF is used. Low temperatures can be maintained, which is conducive to production of
 fairly pure reaction products, and very little waste is produced.74

         The Hoechst Celanese Corporation's bulk acetaminophen unit  at Bishop, TX, receives
 shipments by truck of 5,000 gallons (about 40,000 pounds) approximately once per year for use as a
 catalyst and solvent as described above. The total HF storage capacity on site is 8,400 gallons
 (64,000 pounds).  The maximum quantity of HF in the process is currently 3,200 gallons (24,300
 pounds); however, Hoechst Celanese indicates that production rates will be  increased and the
 quantity of HF in the process will rise to 5,200 gallons (39,000 pounds).75
Page 72

-------
                                         ENDNOTES

1.      Mark, Herman F., Donald F. Othmer, Charles G. Overberger, and Glenn T. Seaborg, eds., Kirk-
       Othmer Encyclopedia of Chemical Technology, Third Edition, Volume 10, John Wiley and
       Sons, New York, 1980. (285)

2.      U.S. Environmental Protection Agency, Screening Study on Feasibility of Standards of
       Performance for Hydrofluoric Acid Manufacture, October 1978, Document Number EPA-450/3-
       78-109. (435)

3.      U.S. Environmental Protection Agency, Screening Study on Feasibility of Standards of
       Performance for Hydrofluoric Acid Manufacture.

4.      Allied-Signal Inc., Hydrogen Fluoride Production Process Description, The HF Products Group,
       Morristown, NJ. (235)

5.      Allied-Signal Inc.

6.      Seringer, Carolyn S., Du Pont Chemicals, comments from  technical review of Hydrogen
       Fluoride Study, Report to Congress, Draft May 8, 1992, June 5, 1992. (436.4)

7.      Hague, William J., Allied-Signal, comments from technical  review of Hydrogen Fluoride Study,
       Report to Congress, Draft May 8, 1992, June 10, 1992. (153)

8.      Allied-Signal Inc.

9.      Chemical Manufacturers Association  Hydrogen Fluoride Panel, The Hydrogen
       Fluoride/Hydrofluoric Acid Industry, Washington, D.C., May 7, 1991.  Prepared for the U.S.
       Environmental Protection Agency.  (270)

10.    U.S. Environmental Protection Agency, Section 305(b) Survey Special Responses, 1987-1988.
       (430)

11.    General Chemical Corporation, Bay Point Plant brochure,  April, 1992. (141.58)

12.    Starry, Claire, Kirtland E. McCaleb, and William A. Stock, Study of Truck Transportation of
       Hazardous Chemicals, SRI International. (460)

13.    Mark, Herman F.

14.    Environment Canada Environmental  Protection Service, Hydrogen Fluoride and Hydrofluoric
       Acid, Environmental and Technical Information for Problem Spills, Technical Services Branch,
       Ottawa, Ontario, July 1984. (220)

15.    Du Pont Chemicals, Hydrofluoric Acid, Anhydrous - Technical: Properties, Uses, Storage and
       Handling, Wilmington, DE. (137.5)

16.    Hague, William J., Allied-Signal.

17    Seringer, Carolyn S., Du Pont Chemicals.

18.    Chemical Manufacturers Association Hydrogen Fluoride Panel.
                                                                                      Page 73

-------
 19.


 20.



 21.



 22,



 23.




 24.


 25.


 26.


 27.


 28.



 29.



 30.


 31.


 32.



 33.



 34.


 35.


 36.


 37.


 38.


 39.


 40.
 Seringer, Carolyn S., Du Pont Chemicals.

 Chapin, L.E., G.C. Liolios, and T.M. Robertson, "Which Alkylation - HF or H2SO4?" Hydrocarbon
 Processing, September 1985, p 67.

 Phillips Petroleum Company, comments from technical review of Hydrogen Fluoride Study,
 Report to Congress, Draft May 8, 1992.  (370.92)

 Encyclopedia of Science and Technology, Third Edition, Volume 1, McGraw Hill Book
 Company, New York, 1987, p 355-356. (328.7)

 Morris, Jeff, American Petroleum Institute, Fina Oil and Chemical Company, comments from
 technical review of Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992, June 1
 1992. (344)

 Morris, Jeff, American Petroleum Institute, Fina Oil and Chemical Company.

 Morris, Jeff, American Petroleum Institute, Fina Oil and Chemical Company.

 Chemical Manufacturers Association Hydrogen Fluoride Panel.

 Morris, Jeff, American Petroleum Institute, Fina Oil and Chemical Company.

 Union Pacific Resources Company, Hydrofluoric Acid Process Orientation, Wilmington, CA,
 1987. (205)

 Albright, Lyle F., "H2SO4, HF Processes Compared, and New Technologies Revealed," Oil and
 Gas Journal, November 26,  1990, p 70-77. (10.1)

 Albright, Lyle F.

 Albright, Lyle F.

 Technica Inc., Quantitative Risk Assessments of Generic Hydrofluoric Acid and Sulfuric Acid
Alkylation Units for Phillips Petroleum Company, Management Summary, May 1990. (420)

 Mudan, Krishna S., Acute Inhalation Toxicity of Hydrogen Fluoride, Technica Inc., Columbus,
 OH, March 1989. (360)

 Morris, Jeff, American Petroleum Institute, Fina Oil and Chemical Company.

Albright,  Lyle F.

Albright,  Lyle F.

Morris, Jeff, American Petroleum Institute, Fina Oil and Chemical Company.

Albright,  Lyle F.

Albright,  Lyle F.

Chemical Manufacturers Association Hydrogen  Fluoride Panel.
Pag© 74

-------
41.     Emerson, Steven D., Kerr-McGee Corporation, comments from technical review of Hydrogen
       Fluoride Study, Report to Congress, Draft May 8, 1992, June 5, 1992.  (139.514)

42.     Emerson, Steven D., Kerr-McGee Corporation.

43.     Chemical Manufacturers Association Hydrogen Fluoride Panel, The Hydrogen
       Fluoride/Hydrofluoric Acid Industry, Washington, DC, May 7, 1991.  Prepared for the U.S.
       Environmental Protection Agency. (270)

44.     Carwile, Roy, Aluminum Association, comments from technical review of Hydrogen Fluoride
       Study, Report to Congress, Draft May 8, 1992, May 21, 1992.  (77.5)

45.     Chemical Manufacturers Association Hydrogen Fluoride Panel.

46.     Chemical Week, February 6,  1991, p 14. (118)

47.     McCue, Thomas C., Wacker Siltronic Corporation, comments from technical review of
       Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992, June 2, 1992. (328.645)

48.     Japan Chemical Week, "Hydrofluoric  Acid Usable for 16M-DF1AM Production," August 9, 1990.
       (215)

49.     Electronic News, "Tl to Make Ultra-Pure HF Acid," September 24, 1990. (475)

50.     Motorola, Inc., Discrete and Materials, Chemical Safety Audit Profile, Phoenix, AZ, April 9, 1991.
       (355)

51.     Chemical Manufacturers Association Hydrogen Fluoride Panel.

52.     Mark, Herman F.

53.     Mark, Herman F.

54.     Mark, Herman F.

55.     Mark, Herman F.

56.     Mark, Herman F.

57.     Mark, Herman F.

58.     McFadden, Thomas H., Mallinckrodt Specialty Chemicals Company, comments from technical
       review of Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992, June 1, 1992.
       (328.648)

59.     Mark, Herman F.

60.     Chemical Manufacturers Association Hydrogen Fluoride Panel.

61.     U.S. Environmental Protection Agency, Development Document for Effluent Limitations
       Guidelines and Standards for the Iron and Steel Manufacturing Point Source Category, Effluent
       Guidelines Division, Volume  V, p 251, Washington, D.C., May 1982. (489.93)
                                                                                    Page 75

-------
 62.     Thomburg, Gary, John C. McArdle, Joseph A. Piccari, and Carolyn H. Byszewski, "Recovery
        and Recycle of Valuable Constituents in Spent Pickling Acids," Environmental Progress,
        Volume 9, Number 4, November 1990, p N10. (470)

 63.     Mark, Herman F.

 64.     Letter, Information on the Unique Value of Hydrogen Fluoride (HF) to Industry and Society,
        From: Anthony T. Camprtelli, Cabot Corporation, Boyertown, PA, To: Craig Matthiessen,  U.S.
        Environmental Protection Agency,  January 24, 1992. (294.3a)

 65.     Letter, Information on the Unique Value of Hydrogen Fluoride (HF) to Industry and Society.

 66.     Letter, Information on the Unique Value of Hydrogen Fluoride (HF) to Industry and Society.

 67.     Letter, Information on the Unique Value of Hydrogen Fluoride (HF) to Industry and Society.

 68.     Letter, Information on the Unique Value of Hydrogen Fluoride (HF) to Industry and Society.

 69.     Letter, Information on the Unique Value of Hydrogen Fluoride (HF) to Industry and Society.

 70.     Meyers, Robert A., ed., Handbook  of Petroleum Refining Processes, McGraw Hill Book
        Company, New York, 1986, "UOP Linear Alkylbenzene (LAB) Manufacture," by P.R. Pujado, p
        1-37-1-42. (416)

 71.     Meyers, Robert A., ed.

 72.     Vista Chemical Company, HF Overview, Lake Charles LAB Plant, Westlake, LA. (497.8)

 73.     Hanlon, Richard G., Hoechst Celanese Corporation, comments from technical review of
        Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992, May 28, 1992. (192)

 74.     Hanlon, Richard G., Hoechst Celanese Corporation.

 75.     Hanlon, Richard G., Hoechst Celanese Corporation.
Page 76

-------
             6.  HAZARDS OF HYDROGEN FLUORIDE PROCESSES
               AND INDUSTRY PRACTICES TO PREVENT RELEASES
       This chapter discusses the hazards associated with the loss of containment of HF during
production, use, and storage at industrial facilities, and during transport. Also discussed are industry
accidental release prevention practices specifically for HF processes.  Before the hazards and
practices specific to HF are addressed, however, it is important to develop a general frame of
reference in terms of the general hazards associated with any chemical release in the chemical and
petroleum refining industries and the general practices for safety and chemical accident prevention.
This will facilitate a more thorough understanding of any unique hazards or practices associated with
HF processes.                               ,

6.1 General Hazards

       The primary concern in any industry that handles or produces hazardous chemicals is a loss
of containment. Releases occur primarily because of human error or equipment failure or some
combination of both.  For example, a release may occur if an operator turns a wrong valve or a pump
seal fails.  It is often difficult to attribute a release solely to human error or equipment failure;  many
times releases result from  a combination of both.

       Human error is a result of the human factors involved in researching, designing, constructing,
and operating a process.  Poor decisions, misjudgment, or lack of skills by operators, maintenance
workers, process designers,  or management can contribute to the potential for a release.  Typical
human errors include operator error, inadequate equipment design, and inadequate maintenance.
These errors in turn may be caused by management failure to provide adequate operator training,
clear standard operating procedures, or adequate resources for proper equipment.  In terms of human
factors contributing to human errors, there is no difference  between HF processes and other chemical
processes.

       Failure of equipment such as pipes,  vessels, pumps, hoses, seals, and valves can result in a
chemical release.  Further, process instrument failure can contribute to process upset conditions.
Most equipment failures are due to chemical hazards and process hazards.  A hazard is "a chemical
or physical condition that has the potential for  causing damage to people, property or the
environment."1 Often, process hazards are increased by the general configuration or operating
conditions (e.g., high temperature or pressure) of the process.  Accidents can be initiated by process-
related events such as overpressurization, overfilling, and loss of utilities.  In the chemical processing
and refining industries, typical chemical hazards include toxicity, corrosivity, flammability, and
reactivity.

       The process and chemical hazards that may result  in equipment failure in HF processes are
not categorically different than the hazards common to other chemical processes. For example,
overfilling storage vessels is a process hazard  common to many industrial facilities.  Sites may take
steps to ensure that when the liquid level in  a storage vessel reaches a set level, no more material is
introduced into the vessel. For some hazardous materials,  including HF, the consequences of an
overfill incident may be severe, and additional precautions may be necessary to eliminate the
possibility of overfilling vessels.

       Industry practices are used to prevent  releases  by addressing the human factors related to
human error and the chemical and process hazards related to equipment failure. Industry addresses
the prevention of releases through the development  and implementation of industry-wide standards
and practices, and process safety management programs at individual sites.
                                                                                    Page 77

-------
6.2 General Industry Practices

       6.2.1 Process Safety Management

       Facilities that handle hazardous materials have a responsibility for understanding the hazards
present at their sites and for taking steps to ensure that chemical accidents due to these hazards are
prevented. Analysis by many organizations, including the American Institute of Chemical Engineers -
Center for Chemical Process Safety (AlChE-CCPS), EPA, and others, has indicated that major
chemical accidents could be prevented not by hardware and technology alone but by comprehensive
management systems designed to identify and control hazards.2'3 These  management systems are
known today as Process Safety Management  (PSM), consisting of "comprehensive sets of policies,
procedures, and practices designed to ensure that barriers to major incidents are in place, in use and
effective. The management systems serve to  integrate process safety  concepts into the ongoing
activities of everyone involved in the process - from the chemical process operator to the chief
executive officer."4
facility:
       PSM consists of several essential elements that work together to allow safe operation of a
              Management Commitment:  Management must adopt a philosophy that makes safety
              an integral part of operation from the top down; an attitude that all accidents can be
              prevented and that business must always be conducted properly.

              Process Hazards Analysis or hazard evaluation: The purpose of the process hazards
              analysis is to examine, systematically, the equipment, systems, and procedures for
              handling a hazardous substance; to identify the mishaps that could occur, analyze the
              likelihood that mishaps will occur, evaluate the consequences of these mishaps; and
              to analyze the likelihood that safety systems, mitigation systems, and emergency
              alarms will function properly to eliminate or reduce the consequences of the incident.
              A thorough process hazards analysis is the foundation for the remaining elements of
              the PSM system.

              Process Knowledge and Documentation: Facilities document the details of the
              technology and design of the process, its standard conditions and consequences of
              deviation from these standards, the known hazards of the  chemicals and processes
              involved and protective systems for protection of workers,  public and environment.

              Standard Operating Procedures (SOPs): These are procedures that describe the
              tasks to be performed by the operator or maintenance worker to ensure safety during
              operation and maintenance.

              Training:  A program to teach those responsible for designing, operating, and
              maintaining the unit or plant. Elements in a management training system include
              development of training programs, training of instructors, measuring performance and
              determining the effectiveness of training.  Training  is typically carried out by  plant
              managers and training staff.

              Maintenance (Process and Equipment Integrity): A formal program to ensure that
              equipment is constructed according to design, installed properly, and adequately
              maintained.
Page 78

-------
       *      Prestartup Review:  The purpose of this review is to ensure that all elements of
              process safety, including hardware, procedures, and control software, are in place
              prior to startup, and that all prior issues of concern have been resolved.

       »•      Management of Change:  Management must instruct personnel to recognize change
              and to evaluate change with regard to process safety.

       >•      Safety Audits:  The purpose of safety audits is to measure facility performance, to
              verify compliance with a sound process safety program, and to determine that risks
              are being appropriately managed.

       *      Accident Investigation: Accident investigation is a management process by which
              the underlying causes of incidents are identified, and steps are taken to prevent
              similar incidents.

       >•      Emergency Planning and Response:  Emergencies involving the processing of highly
              hazardous chemicals can have catastrophic results if not handled properly.
              Employees need to know and be trained in proper emergency procedures, evacuation
              requirements, and notification steps.

       Recently, the  Occupational Safety and Health Administration (OSHA) issued a final standard
which establishes procedures for process safety management to protect workers from chemical
accidents. The standard, Process Safety Management of Highly Hazardous Chemicals (29 CFR Part
1910, February 24, 1992) emphasizes the management of hazards through process safety
management to prevent or mitigate the consequences of chemical accidents involving highly
hazardous chemicals. Facilities that handle highly hazardous chemicals in certain quantities will be
required to follow the procedures in the rule to develop, document, and follow the elements of process
safety management mentioned above.  The standard covers processes that involve anhydrous HF in
quantities at or above 1,000 pounds. Other areas covered by PSM which are applicable to HF risk
assessment and management include contractors, emergency planning and response, work permits,
and human factors assessment.

        For the most  part, industries that produce or use HF follow the same process safety
management practices used at other chemical or petroleum industries.  Like other industries, HF
industries are trying to minimize inventories of hazardous chemicals, conduct rigorous accident
investigations,  and establish programs to test the quality of new equipment and materials before they
are installed. There are some unique  approaches to process safety management in some HF
processes.  For example, Du Pont has developed a safety guideline specifically for HF entitled,
"Anhydrous  Hydrogen Fluoride Safety Guardian Manual," and this manual is issued to all sites that
handle HF.  It  is a corporate performance standard to ensure safe manufacture, handling, storage,
and shipping of anhydrous  HF, including specifying  both minimum requirements for existing facilities
and state-of-the art design for new HF facilities.  The manual covers special properties of HF, first aid
and medical treatment, design information, mechanical integrity, operation and handling, process
safety management,  environmental  considerations, transportation, and customer safety.

        6.2.2   Hazard Evaluation

        Facilities that handle, produce, and use hazardous chemicals may perform a variety of
procedures for identifying and evaluating process hazards. Hazard evaluation is a particularly
important element of process safety management and is extensively used in the chemical  and
petroleum industries. This section discusses various ways to perform hazard evaluations.
                                                                                     Page 79

-------
         Hazard evaluation procedures may be used to:

                >•      identify existing hazards;

                »•      identify potential consequences of the hazards;

                >      estimate the likelihood that events might occur to
                       cause an accident resulting in the consequences
                       identified; and

                >      estimate the likelihood that the systems  in place at the
                       facility would eliminate or reduce the consequences of
                       an accident.6

         Facilities use hazard evaluation techniques to review equipment and procedures, identify
  potential accident scenarios, and identify actions that can be taken to reduce the likelihood and
  mitigate the consequences of accidents. Hazard evaluation has been extensively applied to identify
  potential "weak" spots in processes and reduce the potential for posing risk to public health and safetv
  and the environment.

         There are many different approaches to hazard evaluation, of varying degrees of complexity,
 that are commonly used in industry to identify and assess hazards. Exhibit 6-1 describes several
 procedures discussed in the AlChE document Guidelines for Hazard Evaluation Procedures.7 These
 methods are applicable to the evaluation of hazards wherever HF is managed. Hazard evaluation is a
 complex endeavor and care must be taken to ensure that appropriate experts are involved so that a
 meaningful result is obtained. Interpretation of results is also complex. For example, relying only on
 consequence reduction without considering impacts on the entire process could lead to an operating
 mode that actually increases total risk.

        The selection of a hazard evaluation technique depends on several factors, including the
 phase of process or plant development and the complexity of the process or plant, the purpose of the
 evaluation, the potential consequences of the hazard or hazards being evaluated (e.g., the potential
 for a large release of a highly toxic substance such as HF might warrant a detailed hazard evaluation)
 availability of data required for the hazard evaluation, and time and cost requirements.

        Computer modeling techniques may be used in conjunction with the hazard analysis
 techniques described.  Computer modeling is often used to predict the dispersion of dense gases
 such as HF in air over time following a release, and may provide an estimate of the potential
 concentration and downwind travel distance from the point of release.

        Chemical Process Quantitative Risk Analysis (CPQRA) is  a relatively new methodology that is
 used to supplement other hazard evaluation techniques by providing quantitative estimates of risks
 Risk is a measure that is a function of both the probability that a hazard will result in an event with the
 potential to cause damage to life, property,  and the environment  and the severity of the consequences
 of the specific event. Quantitative estimates of consequences may be obtained using computer
 models for toxic chemicals such as HF. Source and dispersion models can provide quantitative
 information on release rates and dispersion of vapor clouds to some concentration level.  Quantitative
 risk estimates are derived by combining the estimates of incident consequences and frequencies
 using various techniques.  Risk may be presented in terms of risk indices, single numbers or
 tabulations that provide measures of individual or societal risk; these indices may be used in either an
 absolute or a relative sense.
Page 80

-------
                                                               EXHIBIT 6-1
                                                      Hazard Evaluation Procedures
                                                                                  DESCRIPTION
   Process/System Checklists
Checklists are written lists of items or procedural steps to verify the status of a system.  They are
intended to identify common hazards and ensure compliance with standard procedures.	
   Safety Reviews
Safety reviews are inspections that can vary from routine visual examinations to extensive, formal
examinations of plant conditions or operating procedures.  They are intended to ensure that operating
and maintenance practices match the design intent and standards, and identify any new hazards.
   Relative Ranking -- Dow and Mond
   Hazard Indices
Relative ranking should normally be performed before design completion, or early in the development
of an existing hazard analysis program. The Dow and Mond Indices are examples of relative ranking.
   Preliminary Hazard Analysis (PHA)
The PHA is used in the conceptual design or R&D phase of process plant development to aid in
hazard reduction during final design.  It focuses in a generalized way on the hazardous materials and
major process areas of a plant.	_^_^_	——	
   "What If" Analysis
'What If is a brainstorming approach intended to consider unexpected events that would produce an
adverse consequence.  The analysis uses questions beginning 'What if' to identify possible accident
event sequences, hazards, and consequences, and results in possible options for risk reduction.
   Hazard and Operability (HazOp) Studies
 HazOp studies are intended to identify hazards and operability problems in a process plant by
 identifying deviations from the plant design.  An interdisciplinary team carries out the analysis and
 recommends changes or further studies.        	         .
   Failure Modes, Effects, and Criticality
   Analysis (FMECA)
 FMECA identifies the way equipment and systems fail (failure modes), based on an assessment of
 risks using a system of penalties and credits assigned to plant features.	
    Fault Tree Analysis (FTA)
 Fault tree analysis focuses on one particular accident event; a graphic model is used to identify
 combinations of equipment failures and human errors that can cause that accident.	
    Event Tree Analysis
 Event tree analysis is used to evaluate possible accident outcomes in terms of the sequence of
 events that follow an initiating event. The results can be used to specify safety features in plant
 design or to assess adequacy of existing safety features.	
    Cause-Consequence Analysis
 Cause-consequence analysis is a combination of fault tree and event tree analysis. It is used to
 identify potential accident consequences and their causes.	
    Human Reliability Analysis
 Human Reliability Analysis is used to evaluate the factors that influence the performance of plant
 personnel and identify potential human errors and their effects, as well as the causes of observed
 human errors.
Source:  American Institute of Chemical Engineers, Guidelines for Hazard Evaluation Procedures, 1992.

-------
         The strength of Quantitative Risk Assessment (QRA) is in determining incidents of high risk
  contribution at a specific facility.  Through QRA, potential incidents can be ranked according to risk
  contribution, and mitigation methods can be developed to lower the probability and/or consequence
  of the potential incident. Typically, the highest ranked risk incidents will require different mitigation
  approaches. These approaches may vary from simple operational changes to addition of redundant
  equipment.

         The results of quantitative risk analysis are strongly dependent on the data used.  There is a
  variety of models available for estimating consequences, at many levels of complexity,  and results may
  vary depending on the model used, the assumptions made, and the input data used.  An assessment
  of the consequences of a toxic gas release depends on the dispersion model,  the release conditions
  and environment, and the interpretation of toxicity data, which may be limited and subject to
  substantial error. In addition, historical data on incident frequency may be sparse or inappropriate
  (e.g., historical data on frequency of failure of equipment may not reflect changes in technoloqv that
  have occurred).

          The uncertainties in the models, the data, and the general analytical techniques should be
  considered. In addition, since there is no accepted standard QRA methodology or database, QRA
  should not be used to try to determine absolute facility risk. Absolute values can vary by several
  orders of magnitude as a result of differences in input data and assumptions. Use of QRA for
  comparison of overall risk among different facilities using different methods would require much care
  in interpreting the results.

        Additional information on  models specifically designed to address the behavior of HF upon
  release is provided in Chapter 9.

        6.2.3 Industry-Wide Standards

        Many industry-wide standards for design, testing, and maintenance of equipment have been
 published by various organizations, including the American National Standards  Institute (ANSI)  the
 American Society for Testing  and Materials (ASTM), the American Society of Mechanical Engineers
 (ASME), and other standards development organizations and industry associations.  Organizations
 and industry associations such as the American Petroleum Institute (API) and the Chemical
 Manufacturers Association (CMA)  have been  or will be developing standards and guidelines
 specifically for HF processes.

        American Petroleum  Institute

        API is the petroleum industry's major  trade association and has set a  number of industry
 standards and performance requirements. API has also produced documents concerning the use of
 HF. In January of 1990,  API  issued Recommended Practice (RP) 750, Management of Process
 Hazards.  This document outlines the key elements of a comprehensive program for managing all
 potentially hazardous processes.  Focusing on  HF alkylation units, in March of 1990 API issued The
 Use of Hydrofluoric Add in the Petroleum Refining Alkylation Process, a background paper that
 outlines four systems that minimize the risks associated with the HF alkylation process.^

        Because of recent increased concern about HF safety at petroleum refineries, API formed the
 HF Alkylation Committee to offer guidance to  those facilities with HF alkylation units.  The HF
 committee developed a recommended practice for safe operation of HF alkylation units which has
 been approved and given a designation of RP 751.  The API Recommended Practice 751  Safe
 Operation of Hydrofluoric Acid Alkylation Units, was published in June 1992. RP 751  outlines many of
 the procedures and practices  used effectively in the industry to minimize the process hazards of HF
Page 82

-------
alkylation  The engineering systems and procedures described, when properly implemented, minimize
the potential for an HF release, mitigate the effects of a release in the unlikely event that one occurs,
and provide for oversight and audit of the entire process. The RP contains sections on:

              >      hazards management;
              +      operating procedures and worker protection;
              »•      materials, maintenance, and inspection;
              +      transportation and inventory control; and
              ».      relief, utility, and mitigation systems.10

The National Petroleum Refiners Association, an association of domestic refiners, including large
integrated petrochemical and refining companies as well as small and independent refiners, supports
API's RP on hydrofluoric acid alkylation and has sent the RP to its member refineries.

        In addition to written guidance, API gives lectures on the safe handling of HF in refineries.11

        Chemical Manufacturers Association

        CMA supports research on specific chemicals germane to the industry and has taken steps to
deal with HF. In the early 1980's, HF producers and shippers developed mutual aid agreements to
ensure round-the-clock emergency response to HF transportation incidents.  Subsequently, the HF
Mutual Aid Group was formed under the sponsorship of CMA.  The HF Mutual Aid Group is comprised
of specially trained teams that respond to emergencies involving HF. The group is activated by a call
to the Chemical Transportation Emergency Center (CHEMTREC), the CMA hotline used by fire
departments and other emergency responders across the United States to deal with chemical
transport emergencies.12

        To enhance safety in the manufacture, transportation, and emergency response to HF, HF
 producing and using companies chartered the HF Panel under CMA. The Panel was formed to
 develop and maintain guidelines for the safe handling of HF.  HF producers provide general and
 specific guidance to their customers. Through the Panel, HF producers and users cooperate to make
 safety, health, and environmental information available to the entire industry. The intent of the HF
 Panel is to enable all the participants in the North American HF industry to share expertise in the safe
 handling and use of HF.  The Panel  appoints experts to various Task Groups to improve specific
 aspects of safe handling.  Current Task Groups address: Materials of Construction; Medical and
 Toxicology; Mutual Aid; Personal Protective Equipment; Storage Systems; Transportation; and
 Advocacy/Communications. In 1991, the panel expanded from its original membership of
 manufacturers and shippers to include HF users and suppliers of raw materials.

         6.2.4  Industry-wide Practices for HF Processes

         The following are general industrial practices commonly used to address equipment failure
 and human error that also may be addressed by  specific codes and standards. These practices
 typically address the special concerns of the HF industry (e.g., corrosion). The information was
 gathered from HF stakeholders and visits to facilities with established programs for safely handling HF.
 As a result, the practices described do not provide comprehensive  overviews of the areas described.
 Rather, the following practices are a sample of what can be done to reduce hazards.

         Equipment Failure

         For the most part, industries that produce or use HF follow the same practices to guard
 against general equipment failure as other industries managing hazardous  chemicals. The HF
 industry conducts regular testing, inspection and maintenance on equipment.
                                                                                      Page 83

-------
         Corrosion. Facilities that manage anhydrous HF (e.g., petroleum refineries with HF alkylation,
  CFG manufacturers, HF producers) have significant concerns regarding equipment integrity.
  Corrosion can lead to loss of containment and must be guarded against through proper selection of
  materials, process operation, and maintenance of equipment. Carbon steel is used for anhydrous HF
  and aqueous HF 70 percent or greater in concentration. These steels are specifically chosen to
  minimize blistering.  Proper welding materials and methods are required. Additional treatments may
  be required. Steels can develop a film of fluoride scale, which must be considered in the design and
  specification of equipment.  Proper material selection is also required for process equipment, with the
  need determined by the specific chemical and process conditions. Often corrosion resistant'alloys
  such as Monel, Hastelloy B, Alloy 20, and other materials are used.  For aqueous solutions up to 70
  percent HF, chlorobutyl rubber is often used.  Teflon™ has been shown to be an acceptable material
  within certain temperature limits for service in  process equipment such as transfer hoses
  Polypropylene should not be used in anhydrous HF service.13

         Because corrosion  may present a problem in  HF processes, inspection and maintenance is
  especially important  It is industry practice to monitor for corrosion by methods such as ultrasonic
  and acoustic emission testing. To identify cracks, fractures, or bad joints in metal equipment due to
  corrosion or mechanical stress, techniques such as eddy current testing, hydrotesting, radiography
  and leak testing are utilized. Corrosion probes and visual inspection are also used.1* Hydrogen
  which  may be generated by the action of HF on steel, can induce cracking in improper welds in '
  pressure vessels or areas of extreme stress or hardness. Hardness testing and corrective stress
  relieving procedures are widely used in the industry.15 Cracking in welds in carbon steel pressure
  vessels in HF service can be prevented by taking proper care to reduce weld and heat affected zone
  hardness.  Wet fluorescent  magnetic particle inspection may be used to find cracks and other
  fractures."

         Inspection and Maintenance. For particularly critical equipment such as pumps, seals, and
  hoses, HF facilities typically have equipment-specific inspection and maintenance programs.  A major
 focus is the integrity of flexible hoses that are used for HF loading/unloading. Typically, the hoses are
  hydrotested and also replaced frequently. Many HF facilities use guidelines from API, ASME, and
 ANSI to develop maintenance and inspection programs for equipment such as pressure vessels and
 heat exchangers.    Non-process equipment, such as  actuation points for alarms and interlocks is
 also routinely tested.1"  Although the maintenance and inspection program for equipment in HF
 service varies from industry  to industry (i.e., from CFC  manufacturer to alkylate producer), all of the HF
 facilities visited during this study seemed keenly aware of the need to have effective preventive
 maintenance programs for equipment in HF service.

        Because of the considerable hazards posed by HF to human health, many HF facilities test,
 maintain, and inspect personal protective equipment at frequencies that meet or exceed OSHA
 requirements and to ensure worker safety. Most facilities have extensive written procedures to inspect
 HF personal protective equipment and train personnel  to perform the inspection.  Typically  HF
 protective gear is inspected  and tested after each  use  to prevent worker exposure/®

        Equipment Design.  HF facilities can eliminate problem release points by sound equipment
 design.  For example, pump seal failure has been mentioned frequently by facilities as a possible
 release point.  Consequently, a few facilities are installing sealless pumps. This technology, however
 is not yet proven where high pressure pumps are required (e.g., refinery  alkylation). An alternative is'
 double  mechanical seals or dual or tandem seals which can help prevent releases.

        In addition to pump seals, concerns about possible defective pipe seams, welds and flange
 connections have led  HF facilities to strengthen designs by specially treating all welds, using seamless
 pipes wherever possible, or using special gaskets in flanges.  Welded piping is used wherever
 possible to minimize threaded connections and gasketed joints. Other practices are designed to
Page 84

-------
avoid liquid HF traps, to minimize dead legs, and to equip piping with valves to isolate leaking
equipment.

       Overpressurization due to thermal expansion of HF is a major concern with liquid-full piping
systems and vessels because it could lead to rupture.  Where necessary, relief valves and rupture
discs are used to prevent overpressurization.

       Heat exchangers in which HF is cooled with water present the potential for contact between
water and HF in the event of leakage. This can create corrosive acid. Proper maintenance and
m?teria"of construction will guard against such equipment failure.  Such heat ^^er^e
Silly designed so that the HF liquid is at a higher pressure than the water, and if a leak occurs,
Se HF wHHeak into the water solution rather than vice versa, diluting the acid,  reducing the heat
generated, and ensuring that water does not enter the HF  process stream.  Potential eakage ,n heat
exchangers can be monitored by equipping cooling water wrth fluor.de or pH detectors.

        Loss of Utilities.  Utilities may be lost as the result of equipment failure. The loss of power
water  steam, or air at a facility handling HF and the effect this loss could have on HF Processes and
containment systems is an important safety consideration. Different industnes may have drfferent
criticalI svstems depending on the processes used.  Facilities may have  concerns regarding how
emergent^TaterdSuge systems will be supplied and powered in the event of a storm or earthquake.
A common Sition toL problem of powering critica. equipment is to provide backup electnc
 generators or an uninterruptable power supply (UPS).

        Human Error

        Training programs for HF equipment maintenance personnel are largely developed by the
 individual facility  The programs may include lectures, field work, on-tne-job training, or other
 apTroacries.  The length of training will depend on the equipment and the  type of fac, ^ _Manr
 Sties use the training programs approved by ASME, American Soc,ety for Non-Destruct,ve Testmg
 (ASNT), and API.

        Facility operators are also rigorously trained to operate equipment  properly and to perform
 particular tasks according to standard operating procedures.  Many sites use a combinationid_class-
 room, on-the-job training and proficiency testing to ensure the competency of control room and field
 operators and mechanics.

 6.3   Specific Industry Hazards and Practices

        The following sections discuss the hazards associated with equipment failure and the factors
 associated with human error in  specific HF industry segments. Also, specrfic hazard evaluajons and
 specific HF industry practices used to assess and address these hazards  are discussed. The
 pSL and eTaluaSons are provided as examples of efforts by specific facilities  These examples
 may not be unique; however, the extent to which these practices and hazard evaluat-ons are used
 throughout the industry is not known because these  examples are based  on  a limited number of
 facility site visits and on select facility documentation.
                                                                                       Page 85

-------
          6.3.1  Hydrogen Fluoride Manufacturing

          Equipment Failure
          As discussed above, the integrity of equipment in HF service is a concern of the industry To
  assure equipment reliability, the Allied-Signal HF production facility in Geismar, Louisiana uses a
  preventive maintenance program that identifies particularly critical equipment to protect from failure
  Critical equipment includes pressure vessels, process relief valves, boiler relief valves, bolting for HF
  service, liquid HF piping and electrical substations. Also, Allied-Signal has a relief valve testing
  program that includes pretesting, disassembly, overhaul, bubble-tight pressure check, and set

  Knf ^^C-  ?   ^ an ai?iaiy t0°' f°r testing electrical systemns' infrared thermography is used to
  spot electrical problems characterized by a rise in temperature.20  Further, when HF service
  equipment fails, Allied-Signal conducts a failure analysis to determine the cause of the failure and to
  J™T,!nrt de!fcnDchf9es' » necessary. To identify defective or poor quality equipment or materials
  ±Ht fh   uc    PSnt US6S, a TeX3S Nuclear Alloy Anflyzer whicn verifies the composition of alloys
  used in their HF production facility in La Porte, Texas.21

  ho   H ' Mthe r°tary kiln Process used bv Allied-Signal, failure of the seals on the kiln is a process
  SH! ,'  t0rU!a £,   ! pr°?ess is kePt under vacuum; however, if the vacuum  is lost and  pressure
  bunds up in the kiln the kiln seals could fail. Additionally, the seals could fail if the kiln is plugged due
  n SEE re/H°c   ? ^0*°$?* 9yPSUm> Unl6SS Precautions are taken, a seal rupture  could result
  wnn.H nnff,,      7* ^ k"n ^ Even th°ugh' accordin9 to *«***&**, such an HF release
  a^lhi« th 6f  a£9Hr th8 PU^ beCaUSe the h0t gaS6S in the kiln would disPerse hj9n into the
  atmosphere, the facility has installed water spray scrubbers at the kiln seals.  In the internal
  heater/reactor production method used by Du Pont, the main process hazard is also failure of the
  reactor sea!.

         Both the rotary kiln and the internally heated reactor production  processes have similar
  hazaros n condensing, purifying/distilling, and storing HF.  The condensation step involves a flow of
  cooling fluid to cool the product gases. If HF is not cooled and condensed properly, the downstream
  S£ !£ T3™ focesses can be uPset- Coolin9 water systems can fail for several reasons including a
  toss of electricity, corrosion, and a loss of flow due to pump failure or other plugged equipment
  Subsequent to this failure, an upset condition could result causing undue stress (eg
  overpressurization) on the process equipment.  Likewise, in the purification/distillation' step
  overheating and overpressurization could  result because this step requires large inputs of thermal
 SPHrfS'S !!a0pera;ed unde,r vacuum'  Anv of these hazards c°uld result in a loss of containment of
 «r J5 ow£L , P /'    9     agaiPSt thSSe hazards' HF Producers have backup equipment available
 or the ability to stop reactant supply to the reactor.  Critical process vessels are also provided with
 safety relief valves to prevent vessel rupture.

 ra^ i, T?8re arS c°mPan.ies that Purchase HF and repackage it for sale in varying concentrations In
 repackaging operations, simple dilution or mixing operations may be subject to pipe failure hose
            f      ' f Hd hUma" err0r' After use' the PumP should be cleaned and any excess HF
             -f T        ^ Because °f lhe relB«vely sma" quantities of HF used during these
          ' mitigatlofn measures are usually limited to leak plugging. The major hazards involved in
                  f                                           .
th« «        °Pf     t are the rupture °r failure of tne HF c°ntainer during transport or storage in
              ' A Wrehouse IS
 t« «
 ho SS -H   f .'  A Wfrehouse IS tyP'^'y Provided with absorbent material, a dike to prevent spread of
 the liquid and leak plug equipment, and may have sprinklers to knock down vapors
Page 86

-------
       Human Error

       The two largest HF producers, Allied-Signal and Du Pont, have a large stake in ensuring that
industries using their products handle them safely.  Consequently, they both sponsor and conduct
safely training for their customers.  These HF producers also share their information resources on HF
and provide guidance materials. Du Pont has set up steering committees to address the human and
equipment factors surrounding  HF safety.  Du Pont has also developed a safety manual specifically for
anhydrous HF.

       6.3.2  Fluorocarbon Production

       The HF used in the production of CFCs is consumed early in the process.  Consequently, the
HF hazards of concern focus on the initial stages of the process, including the safe flow of HF into
large storage vessels, the pumping of HF from the storage vessels into the reactor, the purification of
the CFC products, and the control of the elevated temperature and pressure in the reactor. HF
hazards are of less concern during the recovery step when only a small amount of unreacted HF is
recycled. HCFC facilities are similar to CFC production units in that HF is consumed early in the
process.  Routes to HFCs typically begin with CFC  or HCFC raw materials.

       Equipment Failure

       CFC reactor equipment consists of vessels, pipes, pumps, valves, heat exchangers, and
instruments containing HF.  Failure of any of these  due to mechanical or chemical stress can lead to
an HF liquid or vapor release.  Many CFC producers prefer to use sealless pumps, thus eliminating
the potential leak through pump seals.  Other areas of concern are flange leaks and corrosion of
piping and vessels.  Fluorocarbon producers are aware of these potential hazards due to the
increased corrosivity of aqueous HF and take extreme care to avoid and correct any water in the feed
stock.

       CFC reactors are operated at elevated temperatures and pressures in order to maximize
conversion and reduce energy consumption.  The fluorinsition step is slightly exothermic; however, the
overall reaction is endothermic, requiring  heat to be added in order to obtain high yields. As in  other
HF-consuming processes,  the amount of HF in the reaction vessel is kept at a minimum.  The reaction
vessel contains mainly the chlorinated hydrocarbon and catalyst, with the majority of HF consumed as
it is added.22

       Operator error or equipment failure could cause overpressurization of the process.  As in other
HF-consuming processes,  this  is addressed primarily through standard operating procedures (SOPs),
training,  inspection, maintenance,  automatic shutdown systems, and relief valves.

       A loss of electricity in the CFC manufacturing process would cause the HF feed pump and the
process  cooling system to stop. As soon as the HF feeds are stopped, the fluorination reaction stops
reducing the potential for overpressurization of the reaction system. Loss of a cooling system is
addressed with SOPs and electrical back up of key emergency vent systems to abatement devices.

       Human Error

       Du Pont's rigorous HF  training program covers CFC production.  This training program is also
offered to other CFC producers.
                                                                                     Page 87

-------
         6.3.3  Alkylate Production

         Concern about hazards and potential HF release points are especially critical where large
 quantities of HF are stored or used. Particular areas of concern in an alkylation unit include the
 reactor, the settler, the acid circulation circuit, and the HF storage vessel.  The following discussion
 presents some examples of hazards.  It is not intended to be a description of all possibilities.

         Equipment Failure

         Some facilities have installed isolation valves in the reactor section for isolation of equipment in
 emergencies.  The isolated equipment is vulnerable to overpressurization in certain circumstances,
 and relief systems are necessary. Relief valves in the reactor section would release HF and
 hydrocarbons to an HF scrubbing and neutralization system and to flare in the event of
 overpressurization.  Pressure would otherwise be relieved at a low elevation in the isostripper. This
 allows for relief of mostly hydrocarbons and minimizes the potential for acid relief.23

        The HF in the reactor can become contaminated with water and impurities introduced to the
 process with the feedstocks.  If left unchecked, this may lead to corrosion of process equipment and
 the potential for release of HF. Consequently, feedstocks are typically dried before entering the
 reactor. Neutralization systems are particularly susceptible to corrosion. Radiographic examination
 can be used on piping to determine the effect of corrosion on wall thickness.  A Mobil  refinery checks
 for flaws in all  its welds by conducting radiographic examination and  dye penetrant testing.  In other
 parts of both alkylation processes where HF is found in small quantities, the corrosivity of HF can
 cause equipment failure and accidental releases. Thus, pH meters are installed in cooling water
 systems for early detection of HF concentration.24 Also, in the presence of low HF concentrations,
 valves can fail to seal  because of fluoride deposits.

        In the UOP process, an acid circulation pump seal failure could result in an HF release.
 Another potential cause of an HF release is the failure of attachments or connections (e.g., piping,
 nozzles, or instruments) to HF process or storage vessels.  In the past, several refineries have had
 releases because of broken sight glasses on HF process vessels. Consequently, many facilities are
 replacing sight glasses with magnetic and nuclear level indicators on vessels. The UOP  design calls
 for double-seals in the acid circulation  pump and remote shut-off valves to minimize  a potential HF
 release. Additionally, HF sensitive paint may be used to identify small releases or leaks at a flange
 joint that may not be visible otherwise.25

        The Phillips reactor employs an acid circulating design with no acid circulation  pump and no
 sight glasses but does contain a larger HF inventory compared  to some UOP designs.  However, the
 Phillips process runs at a lower pressure which can reduce the  rate and quantity of a release.

        The use and production of flammable hydrocarbons in the alkylation process adds the
 potential for fire and explosion hazards. Such events could impact vessels containing HF. To guard
 against fire  hazards, detection of flammable gases is a general priority at petroleum refineries.  In the
 event of a fire or explosion,  HF stored or located nearby could be released and heated by the fire.
 Such a release may not pose  a significant hazard to the public,  because the hot HF would be highly
 buoyant and disperse  easily.  However, other explosions could cause rupture of distant HF pipes or
 vessels, where the HF  might not be heated by the fire upon release.

        UOP and Phillips, the major licensors of HF alkylation processes, provide their licensees with
 lists of approved HF service equipment and schedules for maintenance.  Also, in a BP refinery, as an
 example of quality control, materials intended for use in alkylation units undergo special inspection
 upon receipt to assure proper material  and parts are used.
Page 88

-------
       API's recommended practice, Safe Operation of Hydrofluoric Acid Alkylation Units, provides
petroleum refineries with suggested maintenance methods, recommended inspection frequency for
various equipment in HF service, and the latest and most reliable techniques for equipment
replacement and repair. Individual refineries can use these resources to develop their own preventive
maintenance programs to meet individual needs.

       Contact between incompatible materials can cause an HF release.  In the HF alkylation
process, the uncontrolled contact of HF with caustic and alumina is an example of such a concern.
An example of the results of such an incident is the explosion and fire that occurred in  1987 at Mobil
Oil Corporation's refinery in Torrance, California. The incident apparently occurred because
procedures and instruments failed to control the HF level which resulted in a flow of HF to the
propane treater.  The resulting reaction between HF and potassium hydroxide created high pressure
in the treater, which caused the vessel to rupture and release flammable hydrocarbons.2
Subsequently, a fire started that was fueled by a mixture of propane and butane coming from nearby
pipelines that were also ruptured in the explosion of the propane treater.27

       Process upsets can present a hazard during alkylation.  For example, an upset in the feed
dryers of the alkylation unit can result in increased water levels in the HF circulation streams. This
could cause corrosion or, as in one case, increase the rate of acicl soluble oil (ASO) production. High
ASO levels could  result in HF regeneration upsets.  Such an upset could increase the amount of HF
sent to the process heater where ASO may be  burned, leading to a possible release of gaseous HF
through the process heater stack.28

       Loss of electric power can lead to process upsets and, therefore, is also a significant concern.
Many of the processes in an alkylation unit are run by electricity. For example, pumps used to
circulate cooling water could be disabled, allowing the process to heat up and possibly lift a relief
valve. A standard response to an electrical failure in an alkylation unit, however, would be to cut out
the olefin feed in order to stop the reaction.  The power failure that disabled the cooling pump would
likely take out the olefin feed pump as well, eliminating one of the reactants.  In addition, motor  driven
pumps are often backed up by pumps driven by steam turbines.29 Further, refineries are  usually
provided with backup uninterruptable power systems (UPS) to guard against a loss of electric power.

       Human Error

       To minimize human error contributing to an HF release, the API;s recommended practice,
Safe Operation of Hydrofluoric Acid Alkylation Units, suggests training programs for HF operators, non-
operating personnel, maintenance workers, emergency responders, medical response  personnel, and
workers using personal protective equipment (PPE). To ensure operating procedures are
standardized at HF alkylation units, the recommended practice specifies the need for a facility
operating manual that covers procedures on HF release detection  and response, first aid, acid
sampling, unit neutralization  and dryout, unloading HF, emergency procedures, and testing of critical
alarms, isolation,  and mitigation devices.  In addition, the recommended practice  provides procedures
and training guidance for inspecting, testing, cleaning, and maintaining PPE.

        For maintenance personnel, some general training and SOPs are provided by companies that
license the HF alkylation unit or design the facility.30 In the New Orleans area, several refineries
jointly sponsor a training academy called the Greater New Orleans Industrial Education Council. The
purpose of the council is to teach contractors about general maintenance and process safety
techniques to ensure high standards for contractor performance.  After such training, the contractor
still needs site-specific training at the individual refinery.31  Similar industry-sponsored training
programs exist in Texas City, Texas.
                                                                                       Page 89

-------
        The British Petroleum (BP) refinery in Louisiana has developed a particularly comprehensive
 program for shutdown maintenance performed by contractor personnel.  One part of the BP program
 provides maintenance procedures for contractors to safely handle HF and to assure proper material
 installation during a maintenance shutdown.32'33

        Hazard Evaluation

        Hazard evaluations have been conducted by several petroleum refineries on their HF alkylation
 units. The following are descriptions of hazard evaluations and risk assessments conducted by
 several petroleum refineries.

        Powerlne Oil Company. Powerine Oil Company, as part of a California-required Risk
 Management and Prevention Program  (RMPP) for its  use of HF in the alkylation unit of its Santa Fe
 Springs, California, refinery, carried out a hazard evaluation study to identify possible hazards that
 could be caused by operator error, equipment failure, or external events resulting in a release. The
 hazard evaluation consisted of a safety review, or HazOp (What If analysis,  Guide word analysis), and
 seismic assessment.

        Consequence analysis was carried out for several scenarios  identified by the HazOp analysis
 as potentially leading to HF releases.  These scenarios represented conditions that could occur during
 truck unloading, storage, and  processing of HF, and were considered to represent the highest
 potential for significant impact on public health and safety. The scenarios modeled were a truck
 unloading  accident, ruptures at the bottom and top of the acid storage drum, a rupture at the top of
 the isostripper, and failure  of the seal on the acid recirculation pump.

        Ultramar.  Ultramar, a petroleum refinery in the Long Beach,  California, area, carried out a
 HazOp and Fault Tree analysis at its HF alkylation unit to identify potential hazards, and to satisfy the
 RMPP requirements of the  state  of California.  Ultramar systematically analyzed all parts of the unit
 and the operating procedures to determine ways in which HF could be released.34 Some "most
 likely hazards" that were identified from the HazOp and Fault Tree analyses included pump seal
 rupture, releases during truck  unloading, and a release following a severe earthquake.  This analysis
 led to 111  recommendations for design or procedural changes to the unit.  Examples of some of the
 recommendations that were implemented include updating and reissuing HF truck  unloading
 procedures, conducting a pressure survey for all HF unit pump seals, and replacing certain valves.

        Ultramar also carried out a quantitative risk assessment of its HF alkylation  unit, using air
 dispersion  modeling and quantitative Fault Tree analysis, considering local conditions and population,
 as well as HF toxicity data.  The  results were used to  calculate the risk of a  number of different HF
 releases in terms of mean societal risk or fatalities per year.  The types of releases found to represent
 the highest risk were rupture of acid settlers, serious leakage from settlers, and serious leakage as a
 result of fire or explosion.

        Phillips  Petroleum  Company.  Phillips Petroleum Company conducted a quantitative risk
 assessment of both the HF alkylation and the sulfuric acid alkylation processes. The study evaluated
 the direct risks from the unit, associated risks from acid transportation, and the benefits of risk
 mitigation for the Phillips design  HF alkylation unit in densely and sparsely populated  areas.   The
 analysis indicated that the risks are sensitive to plant siting.  It also showed  that risks  from the
 alkylation unit could be reduced  through mitigation measures (e.g., emergency shutoff valves, acid
 dump system), design modifications (e.g., remove acid circulation pumps), and proper process
 management.35

        BP OH.  BP Oil has an HF Alkylation QRA Program to conduct quantitative risk assessments
 (QRA)  for their refineries worldwide that have HF alkylation units (three are located in the U.S.). BP's
Page 90

-------
QRA process involves extensive audits of the units and their operation, evaluation of the frequency
and consequences associated with all potential HF release sources, determination of risk and
appropriate acceptability criteria, cost-benefit analyses, identification and evaluation of risk mitigation
measures, and determination of sensitivities and uncertainties.  The BP program stresses that QRA is
a decision aiding tool and should not to be used as the sole basis for decisions that could have
significant cost and operational impacts.  BP has used the results of QRA not only in concept studies
and detailed design for new installations and major projects, but also to evaluate existing installations
and operations.

       6.3.4 Transportation and Storage

       HF can be released during transit and during unloading/loading operations. Most industry
sources consider the greatest potential risk of a release to occur during loading/unloading rather than
during transport or processing because of the use of temporary connections and multiple handling
operations. This is confirmed by various hazard evaluations from facilities with HF alkylation units. 6
Most HF releases during loading/unloading operations result from corrosion of the equipment or a
failure to follow standard operating procedures.
37
       Equipment Failure

       Both HF producers and users frequently mention their concern about the failure of transfer
hoses during loading/unloading operations. For this reason, extensive procedures and hazard
reduction techniques are followed when hose transfer takes place.  Pressure testing the hose with
compressed nitrogen gas is generally performed before placing the hose in HF service and the use of
quick-acting, remotely-actuated shut-off valves are employed to minimize HF transfer problems.38
For the compressed gas unloading method, there  is the hazard of overpressurization; for the pump
method, there is also the hazard of pump or pump seal failure.

       Facilities have had to choose loading/unloading equipment and transport packaging materials
carefully  because of the corrosive properties of HF in the presence of moisture.  Many metals will
corrode when in contact with aqueous solutions of HF or with anhydrous HF in the  presence of
moisture. Therefore, transportation piping, valves, vessels, and hoses may fail unless precautions are
taken to  prevent corrosion.39  In addition, the transfer hose used is a specially designed heavy duty
hose.  Facilities are also concerned with valves that may be susceptible to fluoride scaling and
subsequent inability to seat as a result of scale accumulation.

       HF can be released as a result of a transportation accident. When compared with other
hazardous substances produced and transported  in high volumes, such as sulfuric acid, chlorine, or
ammonia, the frequency of HF shipment is low, and therefore the likelihood of transportation accidents
involving release of HF is expected to be lower.  In fact, according to incident data from the
Department of Transportation's (DOT) Hazardous Materials Information Systems (HMIS) for
transportation accidents involving hazardous materials, only 0.27 percent of all incidents reported in
1987 involved HF. Nevertheless, facilities receiving HF shipments recognize that a release of HF from
a truck or tank car could pose significant off-site impacts (see Chapter 8 for discussion of
transportation accidents).

       To prevent a release of HF, the transport containers and equipment provided by Allied-Signal
and Du Pont are overdesigned for safety.  In fact, the HF producers that transport HF throughout the
country often  exceed what is required of them by law to ensure safety. For example, according to
DOT regulations rail cars in anhydrous HF service must be constructed of steel that is at least 0.4 inch
thick.  Du Pont uses rail tank cars that are constructed of one inch thick steel as recommended by the
American Society of Mechanical Engineers (ASME).  Du Pont tank trucks are constructed of 0.5 inch
thick steel. In addition, both Du Pont and Allied-Signal utilize headshields and shelf couplers on their
                                                                                       Page 91

-------
 HF rail cars (even though not required by DOT regulations) to protect the tanks in the event of a
 derailment.

         In addition to DOT requirements, the Association of American Railroads (AAR) has published a
 Manual of Standards and Recommended Practices, which details requirements for tank cars that will
 transport anhydrous HF. These include prohibition of bottom openings, specific heat treatments for
 tank material and welds, use of non-corrosive valves and fasteners, and specifications for safety relief
 valves.40 Safety relief valves on tank cars and trucks are used to release HF gas in the event of
 overpressurization.  These valves are protected by extra heavy rollover type domes which provide
 mechanical protection to these valves.  In conformance with AAR standards, bottom outlet valves are
 not used on trucks in anhydrous HF service.41 Another mitigation  procedure involves transfer
 hoses, which are pressure tested every six months, replaced yearly, and specially stored to prevent
 kinking. 2 In the event of a release during transportation, hazardous chemical  information and
 emergency response is available through the CHEMTREC  hotline sponsored by CMA.

        The raised valves used for loading and unloading HF are contained within a roll-over
 protection dome on top of transport vessels. To stop an in-progress HF release from defective valves,
 both HF producers have developed emergency capping kits. Du Pont and Allied-Signal have adapted
 the chlorine capping kit for use on HF assemblies. Regardless of the design, the emergency capping
 kit is placed over the leaking/defective valves on top of the rail car  or truck.  Du Pont also uses valves
 on some tank trucks that were designed such that if the valves on  the top of the vehicle are sheared
 off, a secondary internal valve will prevent an HF release.43 This is an European design that Du
 Pont would like to further evaluate for use in the U.S.

        HF storage vessels are susceptible to hydrogen blistering,  weld hardness, and stress
 corrosion cracking.  To guard against vessel failure, HF producers  use corrosion resistant equipment
 and conduct regular inspection and maintenance.  HF storage tanks are also manufactured to comply
 with current ASME code for Unfired Pressure Vessels which includes specifications for corrosion
 allowance and  minimum thickness.44 HF tanks are installed above ground and are usually
 supported by structural  steel or concrete saddles.  Because HF can pool and fume if released from
 the tank when the HF is stored at atmospheric temperature and pressure, facilities that have a bottom
 outlet have developed drainage patterns to divert the spillage to a containment area away from the
 tank.

        Overfilling of HF storage tanks also is a hazard that can result in an HF release.
 Overpressurization of an HF storage tank, which can cause HF to be released through a relief valve,
 might occur if the tank is overpressurized with nitrogen, for example, during the unloading of HF or if
 an HF tank is overheated. The Allied-Signal facility in El Segundo keeps its HF storage cool using an
 internal refrigerated coil  and insulation.  Also, the storage vessels are enclosed  in a building. At 3M,
 HF storage tanks are enclosed in a cooled building, and the HF is kept at -40 degrees C.   Cooling
 the HF below its boiling  point will result in pool formation rather than a vapor release.  Industry also
 uses redundant level indicators, safety interlocks, pressure  gauges, and alarms to address these
 hazards.  For each tank, a facility usually develops an individual plan for acid delivery, inventory,
 maintenance, cleaning, monitoring, and emergency response.

        Human Error

       Training standards for drivers and for personnel who unload and load HF are provided mainly
 by Allied-Signal and Du Pont rather than HF users. To assure highly trained and experienced
 personnel and well-maintained equipment, both HF producers have professional drivers and dedicated
 fleets of trucks and rail cars.  Both have rigorous training programs for drivers of HF vehicles which
 address the hazards of HF,  first aid,  unloading procedures,  and operation of the safety features of
 shipping containers.  Allied-Signal selects drivers based on road tests, drug tests, and other screening
Page 92

-------
methods.  Once selected, the drivers must attend classes on such topics as proper unloading/loading
techniques, personal protection equipment, DOT requirements, Community Right-To-Know laws, and
emergency procedures.47  Hands-on experience is gained donning PPE (e.g., Scott air pack),
handling emergency equipment (e.g., chlorine capping kit), completing safety checklists, and
observing  and participating in unloading/loading operations.  Certification and recertification are
required.

       To reduce the risk posed by transporting HF, HF shippers conduct route risk analysis.  Allied-
Signal has conducted extensive route risk analysis for all its HF deliveries which average distances of
1,000 miles.49'50'51  In addition, Allied-Signal has installed a satellite tracking system to track HF
truck transport vehicles.  If there is a problem en route, drivers can communicate immediately with
headquarters and emergency response personnel. This system allows for better control of shipments
and possibly faster response in an emergency.*0
52
       Hazard Evaluation

       Du Font's CFG facility in Antioch, California, prepared an Off-Site Consequence Analysis
Report which included a hazard analysis for HF based on HazOp studies as a supplement to a
required Risk Management and Prevention Plan submitted to Contra Costa County.  Potential HF
events were rated according to probability of release and severity of consequences, using a qualitative
rating and the most severe events were chosen for modeling analysis.  None involved the CFC
process.  Of the three events chosen, two were related to transportation and the third was related to
storage.  For HF, the event considered to have the largest potential consequence was a corrosion .
hole in a transfer line between  a tank car and a storage tank or the tank recirculation line. Such a
hole could be caused if excessive amounts of  moisture entered the system. "High" severity
consequences were defined as potential serious injuries or death to exposed individuals. The
probability of occurrence was considered low,  however, because of operating procedures and
because the transfer line is only used eight to  ten hours per week.  "Low" probability was defined as
unlikely occurrence during the expected lifetime of the facility assuming normal operation and
maintenance.  This event was modeled to estimate consequences both before and after completion of
a planned HF simplification and mitigation project at the  Du Pont facility.

       The second event modeled was a corrosion leak in a storage tank because it was considered
to have high potential consequences. The likelihood of this event was also considered low. Failure of
a tank car angle valve was the third event modeled. The likelihood of this event was considered high,
defined as likely to occur at least once during  the expected lifetime of the facility (in fact, such an
event had actually occurred at the facility, but the consequences were not severe).

       6.3.5. Other Uses

       Uranium Processing.  In uranium processing plants, hazards associated with HF arise from
the potential for a release of either HF or UF6.  If UF6 is released, it will immediately decompose into
the airborne toxic products, uranyl fluoride and HF. The Nuclear Regulatory Commission (NRC) has
been studying the side reactions and decomposition of UF6.  To ensure safety in operating HF
uranium processing plants,  NRC is requiring risk assessments as part of the permitting process.

       Aluminum Fluoride and Aluminum Manufacturing. The major process hazards associated
with aluminum fluoride manufacture center on  the reliability of the power distribution system.  The
major process hazards in the aluminum reduction electrolytic cell (pot) and in the air control system
are a loss of electricity and plugged equipment.  Monitoring devices related to these hazards include
a meter to detect vibration or a loss of electricity and detection equipment to measure ambient HF.53
                                                                                      Page 93

-------
        Production of Electronic Aqueous HF.  The General Chemical Bay Point Corporation, Bay
 Point Works, which uses anhydrous HF to produce electronic aqueous HF, carried out an off-site
 consequence analysis for two scenarios, both of which were considered to be of low probability and
 high consequence. One scenario was the failure of the railcar unloading line (see Section 6.3.4,
 Transportation and Storage, for a discussion of unloading and loading operations); the other was
 failure of the anhydrous HF storage tank pressure relief system.  Neither of these scenarios would be
 unique to the electronic HF production process.  The Bay Point Plant also reported a number of
 planned mitigation/risk reduction steps for its HF processes, including alarms if water flow to the
 absorption column is reduced or stopped, or if the primary or secondary scrubber liquid lines have
 low pressure; locks on pressure relief block valves to insure that these valves are not inadvertently
 closed; development of a preventive maintenance program for instruments and safety equipment,
 including annual testing and maintenance of pressure relief devices; development and updating of
 operating manuals, including loading and unloading procedures; and formalization and improvement
 of operator training for handling HF, including operating procedures and emergency response.54

        Electronics Manufacture.  Because electronics manufacturing uses aqueous HF below the
 fuming concentration and because the dipping process is highly controlled to ensure system purity,
 the industry is relatively free from process hazards.  To eliminate any possible corrosion problems,
 teflon tubes and polypropylene pump housings are used. HF releases would be very unlikely to have
 off-site impacts.

        Pickling, Etching, and Coating.  Because the steel  pickling process uses solutions of HF in
 concentrations of 70 percent and less, the hazards to the public are not as great as those associated
 wfth processes using anhydrous HF, such as alkylation. An accidental release of aqueous HF would
 stay in liquid form  rather than aerosolize and therefore could be contained and the release mitigated.
 However, at these concentrations, HF poses an increased corrosion problem and still maintains the
 potential to cause  damage to the environment.

        Dissolving Ores to Produce Tantalum and Columbium Metals. Aqueous HF is extremely
 corrosive to certain equipment, making equipment corrosion the major hazard in the process to
 produce tantalum and columbium processes.  In addition, as aqueous HF  is transferred between
 process areas, the potential of pump seal failure or pipe corrosion exists which could cause a release.

        Linear Alkylbenzene  (LAB) Production.  To prevent HF releases, Vista Chemical uses only
 HF-compatible materials in maintenance, design and construction activities, such as in vessels, pipes,
 fittings, gaskets,  nuts,  bolts, flanges, valves, vents and bleeders.  The company inspects HF
 equipment at least every two years.  The HF process area at Vista has a full-time maintenance person
 who checks for leaks and spills. In the past three years, they have taken a major step towards
 reducing HF hazards by eliminating several major process vessels from HF service. This change
 reduces HF process volume by almost 50 percent, reducing the sources of potential leaks and spills,
 such as pumps, flanges, valves, vents, bleeders and the vessels themselves.  In addition, the
 company conducts HF training annually, requires LAB  personnel to attend a yearly Du Pont seminar
 on HF safety, and carries out  personal protective equipment training and emergency drills.55

        Pharmaceutical Production. Hoechst Celanese, which uses HF in a bulk pharmaceutical
 process, reports  that all equipment that may vent HF is connected to a potassium hydroxide scrubber
 system that absorbs and neutralizes HF.  This scrubber vents to  a 200 foot flare.  The potassium
 hydroxide is routinely analyzed once per shift.  The HF equipment is also connected to an emergency
 scrubber system through rupture disks and pressure release valves.  If vessels or pipes have to be
 opened to the atmosphere, an evacuation system that vents through a sodium hydroxide scrubber
 system is used.  A water curtain/deluge system can be activated if a leak to the atmosphere develops.
 A design HazOp study was performed before the plant was built, and Hoechst Celanese performs a
 Process Safety Review every five years.  The plant has 17 HF monitors that alarm both locally and in
Page 94

-------
the control room; there is also a remotely operated camera that can be used to monitor the HF area
from the control room.56

6.4  Research Efforts to Modify or Substitute Hydrogen Fluoride in Alkylation

        Another method of preventing releases of a hazardous substance is to modify the properties
of the substance to reduce its hazard or to identify a less hazardous material that can be used as a
substitute. Research and development is being conducted on modification of the properties of
anhydrous HF,  use of alternative liquid catalysts, and use of alternative solid catalysts.  Research on
modification of the properties of anhydrous HF includes work being done by Mobil Oil Company, as
well as a joint effort by UOP and Texaco,57 to use additives to reduce aerosolization and encourage
rain out. In small scale tests, ninety percent reduction of aerosolization with no loss of alkylation
performance has been claimed.58  Future work will include large scale release tests, process
demonstration,  modeling for risk reduction, and commercial demonstration.  Another example of
research designed to reduce the hazards of HF is Phillips Petroleum's work on system modifications
to reduce volatility of HF and significantly lower inventories of HF.

        Research into liquid and solid catalysts that can serve as substitutes for HF in alkylation has
been underway for more than fifteen years.  Both boron  trifluoride and zeolites were studied in the late
1970's as alternatives to HF or sulfuric acid.  These solid catalysts, although they offer a safe
alternative to HF, were subject to rapid deactivation and  were determined to be too costly for large
scale use.59  Research has been performed to determine the alkylating ability of various catalyst
complexes such as ethyl fluoride-antimony pentafluoride complexes, hydrogen fluoride-tantalum
pentafluoride mixtures, tantalum pentafluoride-aluminum  pentafluoride, antimony pentafluoride-
graphite, and a fluorosulfurous  acid-antimony pentafluoride complex.60 Aluminum trichloride has
been studied as an alternative liquid catalyst, with limited success.*"
61
        As recently as February 1992, three companies, Catalytica; Conoco, a subsidiary of Du Pont;
 and the Finnish oil company, Neste Oy, announced a joint venture for the development of a
 commercial catalytic alkylation process based on a proprietary solid catalyst developed by Catalytica.
 If pilot tests are successful, a solid catalyst could become available in the production of alkylate.  The
 joint venture anticipates an operational pilot plant in Finland by Fall 1992. A solid catalyst would
 eliminate the potential for airborne HF release.   Another alternative, a super acid catalyst
 supported on a solid medium, is being developed by M.W.Kellogg and Haldor Topsoe A/S.
                       63
                                                                                        Page 95

-------
                                         ENDNOTES
 1.


 2.

 3.




 4.

 5.



 6.




 7.

 8.



 9.



 10.



 11.




 12.




 13.



 14.


 15.


 16.

 17.
 American Institute of Chemical Engineers, Technical Management of Chemical Process Safety,
 Center for Chemical Process Safety, New York, 1989. (10.467)

 American Institute of Chemical Engineers, Technical Management of Chemical Process Safety.

 U.S. Environmental Protection Agency, Review of Emergency Systems, Report to Congress,
 Section 305(b) Title III Superfund Amendments and Reauthorization Act of 1986, Office of Solid
 Waste and Emergency Response, Washington, D.C., May 1988. (489.92)

 American Institute of Chemical Engineers, Technical Management of Chemical Process Safety,

 Du Pont Chemicals, Draft AHF Safety Guardian Manual, Wilmington, DE,  December 18,1991.
 (137.8)

 American Institute of Chemical Engineers, Guidelines for Hazard Evaluation Procedures, Center
 for Chemical Process Safety, New York, NY, 1985. Prepared by Battelle Columbus Division.
 (10.465)

 American Institute of Chemical Engineers, Guidelines for Hazard Evaluation Procedures.

 American Institute of Chemical Engineers, Guidelines for Chemical Process Quantitative Risk
 Analysis, Center for Chemical Process Safety, New York, NY, 1989. (10.46)

 API HF Alkylation Committee, Hydrofluoric Acid Alkylation Unit Safety, API Recommended
 Practice  (Draft), September 11,1991. (190)

 American Petroleum Institute (API), Safe Operation of Hydrofluoric Acid Alkylation Units, API
 Recommended Practice 751, First Edition, June 1992. (10.6)

 Safe Handling of Hydrofluoric Acid in Refineries, Presented to Manufacturing Safety
 Subcommittee, American Petroleum Institute, Committee on Safety and Fire Protection
 Meeting, Ft. Lauderdaie, Florida, April 6,1983. (426)

 U.S. Environmental Protection Agency, The Hydrogen Fluoride/Hydrofluoric Acid Industry,
 Washington, DC, May 7,1991. Prepared by the Chemical Manufacturers Association Hydrogen
 Fluoride Panel. (270)

 Seringer, Carolyn S., Du Pont Chemicals,  comments from technical review of Hydrogen
 Fluoride Study Report to Congress, Draft May 8,1992, June 5, 1992. (436.4)

 Mobil Oil Corporation Torrance Refinery, Hydrofluoric Acid Alkylation Unit Risk Management
 and Prevention Program Report, Torrance, CA, March 16,1989. (185)

 Hague, William J., Allied-Signal, comments from technical review of Hydrogen Fluoride Study
 Report to Congress, Draft May 8, 1992, June 10, 1992. (153)

 Seringer, Carolyn S., Du Pont Chemicals.

API HF Alkylation Committee.
Page 96

-------
18.     Ultramar HF Alkylation Unit Risk Management and Prevention Program, Ultramar Refining, Inc,,
       Wilmington, CA, April 1990.  Prepared by Science Applications International, McLean, VA.
       (424.33)
19.     Mobil Oil Corporation Torrance Refinery.
20.     Allied-Signal Inc., Allied-Signal Inc., HF Production-Geismar Plant, Maintenance/Reliability
       Programs, HF Products Group, Geismar, LA, February 1992. (147.1 B)
21.     Facility visit by U.S. Environmental Protection Agency to Du Pont, La Porte, TX,
       December 12, 1991. (620)
22.     Seringer, Carolyn S., Du Pont Chemicals.
23.     Puschinsky, Bob, Alkylation Technologies and Sulfuric Acid Production Consulting comments
       from technical review of Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992,
       June 28, 1992. (377)
24.     Phillips Petroleum Company, comments from technical review of Hydrogen Fluoride Study,
       Report to Congress, Draft May 8, 1992. (370.92)
25.     Morris, Jeff, American Petroleum Institute, Fina Oil and Chemical Company, comments from
       technical review of Hydrogen Fluoride Study, Report to Congress, Draft May 8,1992,
     -  June 1, 1992. (344)
26.    Mobil Oil Corporation Torrance Refinery.
27.    Rabin, Jeffrey L, "Mobil Refinery Explosion Rekindles Safety Debate," Los Angeles Times,
       December 20, 1987, p B12. (421.25)
 28.    U.S. Environmental Protection Agency, U.S. EPA Release Prevention Questionnaires,
       1987-1988. (490)
 29.    Mudan, Krishna S., Technica Inc., comments from technical review of Hydrogen Fluoride Study,
       Report to Congress, Draft May 8, 1992, June 1, 1992. (360.2)
 30.    API HF Alkylation Committee.
 31.     Facility visit by U.S. Environmental Protection Agency to British Petroleum, New Orleans, LA,
       January 13, 1992. (630)
 32.     Facility visit by U.S. Environmental Protection Agency to British Petroleum,
 33.     BP Oil) BP Oil US Refining Contractor Safety Program.  (75)
 34.     Ultramar HF Alkylation Unit Risk Management and Prevention Program.
 35.     Technica Inc., Quantitative Risk Assessments of Generic Hydrofluoric Acid and Sulfuric Acid
        Alkylation Units for Phillips Petroleum Company, Management Summary, May 1990. (420)
 36.     API HF Alkylation Committee.
                                                                                      Page 97

-------
  37.



  38.

  39.



  40.



  41.


  42.



  43.



 44.


 45.


 46.



 47.



 48.


 49.




 50.




 51.




 52.


 53,


 54.



 55.
 U.S. Environmental Protection Agency, U.S. EPA Release Prevention Questionnaires 1987-
 1988. (490)

 Hague, William J., Allied-Signal.

 Du Pont Chemicals, Hydrofluoric Acid, Anhydrous - Technical: Properties, Uses Storage and
 Handling, Wilmington, DE. (137.5)

 Association of American Railroads, Manual of Standards and Recommended Practices:
 Specifications for Tank Cars, Section 2.1.7, "Hydrogen Fluoride, Anhydrous." (48)

 Seringer, Carolyn S.,  Du Pont Chemicals.

 Phillips Petroleum Company,  comments from technical review of Hydrogen Fluoride Study,
 Report to Congress, Draft May 8,1992. (370.92)

 Facility visit by U.S. Environmental Protection Agency to Du Pont, La Porte, TX,
 December 12,1991. (620)

 Allied-Signal, Inc., Anhydrous Hydrofluoric Acid, The HF Products Group, Morristown, NJ. (17)

 Du Pont Chemicals, Hydrofluoric Acid, Anhydrous.

 Chow, C.S., 3M, comments from technical review of Hydrogen Fluoride Study Report to
 Congress, Draft May 8, 1992,  June 3, 1992.  (122)

 Facility visit by U.S. Environmental Protection Agency to Allied-Signal, Geismar, LA
 October 11,1991. (650)

 API HF Alkylation Committee.

 Allied-Signal Inc., Risk Assessment of Hydrogen Fluoride Transportation Routes (Routes
 Originating atEISegundo, CA), Morristown, NJ, March 28, 1989. Prepared by ICF Technology
 Inc., Fairfax, VA. (424.1)

 Allied-Signal Inc., Risk Assessment of Hydrogen Fluoride Transportation Routes (Routes
 Originating at Amherstberg, ONT), Morristown, NJ, March 28,  1989.  Prepared by ICF
 Technology Inc., Fairfax, VA. (424.2)

 Allied-Signal Inc., Risk Assessment of Hydrogen Fluoride Transportation Routes (Routes
 Originating at Geismar, LA), Morristown, NJ, May 23,  1989.  Prepared by ICF Technology Inc
 Fairfax, VA. (424.3)                                                              ay    "

 Facility visit by U.S. Environmental Protection Agency to Allied-Signal.

 U.S. Environmental Protection  Agency, U.S. EPA Release Prevention Questionnaires.

 General Chemical Corporation, Bay Point Off-Site Consequence Analysis, Engineering and
 Research, Inc., pp 8-44. (141.5C)

Vista Chemical Company, HF Overview,  Lake Charles Lab Plant, Westlake, LA. (497.8)
Page 98

-------
56     Hanlon, Richard G., Hoechst Celanese Corporation, comments from technical review of
       Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992, May 28, 1992. (192)

57.     Meeting between U.S. Environmental Protection Agency and UOP/Texaco, June, 1992.

58.     Meeting between U.S. Environmental Protection Agency and Mobil, January 28, 1992.

59.     Albright, Lyle F., "H2SO4, HF Processes Compared, and New Technologies Revealed," Oil and
       Gas Journal, November 26, 1990, p 70-77. (10.1)

60.     Olah, George A., G.K. Surya Prakash, and Jean Sommer, Superacids, John Wiley & Sons Inc.,
       New York, 1985. (369)

61.     Albright, Lyle F.

62.    Chemical and Engineering News, "Gasoline Alkylation Process to be Commercialized,"
       February 10, 1992, p 24. (141)

63     Letter, Preliminary Findings-Hydrogen Fluoride Study, From:  William J. Hillier, The M.W.Kellogg
       Company, To:   Ed Freedman, U.S. Environmental Protection Agency, July 9, 1993. (298)
                                                                                     Page 99

-------
                                  This page intentionally left blank.
Page 700

-------
            7   INDUSTRY PRACTICES TO DETECT AND MITIGATE
                        HYDROGEN FLUORIDE RELEASES
       Chemical accident prevention programs are a critical element in a firm s overall strategy for
managing chemical releases.  Equally important are strategies to deal wrth releases '^nd when they
occur  This chapter focuses on selected industry practices to detect and mit,gate anhyd ou; HF
releases. It is important to note that a release of concentrated aqueous HF may also cause a vapor
cloud  Some of the following detection and mitigation options may be appropriate for erther type of
HF release.

7.1    General Industry Practices to Detect HF Releases

       Manv HF facilities recognize the importance of detecting HF leaks quickly. Detection can
indicate which system or equipment is malfunctioning, and thereby enable facility personnel to use
adeauate personal protective gear and alleviate the problem in a safe and t.mely manner.  Some
dSs canSmatically trigger shutoff swrtches and water negation systems. Others can provide
eanV warning to employees and community officials who can begin emergency response procedures
sound a^Is  and b'egin protective measures, like evacuation,  as needed.  There are several methods
and systems that can be used to  detect HF releases, including visual observation and detection
equipment and systems.

        7.1.1  Visual Observation

        HF forms a visible white cloud and, therefore, operating and maintenance personnel can spot
small leaks around flanges, valves, and places that might lead to more serious^'eaJs-HF vapor
clouds, however, are similar in appearance to steam clouds.  Thus, visual detection of HF ,s more
difficult in a plant area where there are steam clouds from leaks, condensate drams or condensate
traos  Conversely, an HF cloud may be mistaken for steam,  or the observer may not be able to
 differentiate it  from background.  Aqueous ammonia may be used as a detector around suspected
 leak areas to confirm or highlight the visible sign of a small HF release point.  Many facilities use  HF
 sensitive paints which change color when contacted by HF to identify leak sources.  One
 manufacturer  of HF sensitive paints, Valspar Corp., sells a paint which is sensitive to 5 percent
 Concentrations of HF or greater.  Some facilities have chosen not to use HF sensrtive paintsBecause
 HF is corrosive to paint in general; therefore, leakage becomes readily apparent wrth degradat.on of
 the paint surface.

        Many facilities use closed circuit television systems so that operators in the control room can
 identify and determine the exact  location of a release with remotely-operated visual monitoring. These
 systems are also valuable for directing and observing the effects of mitigation measures.  Other types
 of camera systems can identify HF leakage based on the motion of the release.  The drawback of a
 closed circuit television system is that without sharp contrast, HF vapor clouds can be mistaken for
 background  Also, the closed-circuit television screens in the control room may not be constantly
 monitored.  New technologies  rely on the use of more expensive infrared camera systems which
 produce a thermal image. These systems will  be discussed further below.

         7.1.2   Detector Equipment and Systems

         Reliable and accurate  HF detectors have been difficult to  develop. The corrosive nature of HF
 causes detectors to deteriorate fairly rapidly, and materials that resist corrosion, used routinely in
 detector systems, are costly. Several types of detector systems, including multi-detector systems,
 fixed detector systems, mobile detectors, open path systems, and thermal imaging systems, have
 been and are currently being tested to demonstrate reliability.  There are  limitations to  using detection
                                                                                    Page 101

-------
   methods, including the following: detector equipment and systems may not be reliable, purchasing
   and maintaining the high technology equipment may be expensive, the impact of a false positive mav
   be costly and dangerous (e.g., water spray mitigation systems automatically turned on), other gases
   may interfere with the detector, and an HF cloud could "evade" sensor locations if sensors are not
   properly placed.

          Multi-Detector Systems

         The types of HF specific detectors most commonly used in multi-detector systems are based
  on electrochemical cells. In these detectors, the chemical properties of HF change the parameters of
  the sensor, generating an electrical signal. Interference from other chemicals like hydrogen chloride
  and sulfur dioxide,  however, can cause false positive responses.  Electrochemical detectors produced
  by Sensidyne Inc. and Gas Tech have been installed at some HF facilities.  These systems are also
  sensitive to temperature and humidity, and must be calibrated every one to three months. A newer
  technique for HF detection is the ion mobility spectrometer. In this device, an air sample is drawn into
  the test chamber and ionized.  The HF ions are separated from other ions by an electric field  and the
  concentration is computed.  A typical measurement range for this device is 0 to 10 ppm HF  These
  systems have low maintenance requirements; however, high humidity may result in false low readings
   on mobility spectrometers supplied by either Sensidyne Inc. or Environmental Technologies Group   '
  Inc. have been installed at a few HF facilities.                                          «,uuH

         A r.f.Cent develoPment by Exxon Corp., marketed by Environmental Technologies Group Inc
  utiHzes a silicon oxide chip as an HF detector. As HF passes over a thin film of silicon oxide the  "
  surface is etched, increasing light reflection from the film surface. An optical device senses the
  reflected light and generates a current which  can be used to activate an alarm. HF concentration is
  not calculated, but response to concentrations above 500 ppm occurs in seconds.  The silicon sensor
  is disposable and can be used only once, but can be replaced for a nominal cost. It is designed to
  monitor HF leaks at flanges, pump seals, and similar sources.1  One benefit of this system is that it is
  Hi* SpSClflCi

        Fixed Detector/Multi-Sample Point Detectors

        Fixed-detector/multi-sample point systems use a single detector or instrument.  Samples from
 several process points are pumped to the instrument and introduced in rotation for analysis This
 te nl^S!81 ?E If "I-?-81 °.l!en usedwnen an expensive or complicated analytical method or instrument
 fe needed HF facilities that use this system may use a gas chromatograph with an electron capture
 detector. This instrument separates and measures specific gas concentrations.2 An ion mobility
 spectrometer may also be used as an HF detector with a multiple sample point system.3

        Mobile Detectors

        Indicator tubes, such  as Draeger tubes, can provide quick spot checks for HF concentrations
 In contrast to the gas chromatograph, indicator tubes may be less precise, but are simple to operate'
 anl°T £tm     6aSily f°r measurements at the process area or at the facility fenceline.  In this
 SI fjn  PumPed.tnrou_9h  a detector tube produces a stain; the length of the stain is proportional
 to the HF concentration.  The tubes can be easily moved to detector locations and are used to
 confirm a release and its concentration.

        Open Path Detectors

      f °Pf1 P3th .Ol7emote sensir|9 systems can detect a chemical plume when it crosses the open
 m«LaHVh    H°r, in'rared''9nt beam- The absorption or scattering of the light by the chemical is
 measured by a detector.  However, vapor clouds may also scatter light and cause a false positive
Page 702

-------
Research is still being conducted to develop reliable and accurate open path sensors such as laser
detectors which can detect HF across large areas, especially at the fenceline.  MDA Scientific is
developing a Fourier Transform Infrared Spectrometer which can be used as an open path detector.
This system is designed to detect 160 chemical compounds and provide concentration data.  It has
been successfully tested at an aluminum smelting facility and Mobil Oil is scheduled to test this sensor
at its plant  in Torrance, California.

       For facilities that use real time emergency response computer systems, the sensors can be
integrated to provide data to estimate impacts of an HF release.  Regardless of the type of detector
system, the detectors need to be inspected, tested, and maintained to ensure reliability of operation.
In the case of power failure, the sensor systems are provided with backup electrical power or power
supply which cannot be interrupted.4 Drawbacks of open path sensors are that they are limited to
monitoring  perimeters, and presently  cannot detect concentration levels.

       Thermal Imaging

       The thermal imaging monitoring technique detects the presence of a cloud cooler than
ambient conditions. .Based on tests conducted by Amoco, thermal imaging has been found to be very
sensitive and can provide monitoring of an entire process unit. The results are displayed on a video
monitor and software has been developed which will provide interpretation from the camera and can
provide alarms to unit operators.  While this monitoring technique is not HF specific, it was tested
during the  Goldfish experiments on atmospheric releases of HF and provided data on cloud
position.5  (The Goldfish experiments are discussed further in Section 7.3 of this chapter and in
Chapter 9.)

7.2    HF Detectors Used by Specific Industries

       The use of HF detectors  is not standard practice in the HF industry. This is an emerging
technology and many facilities are uncertain about the reliability and accuracy of detector systems
and equipment.  In fact, Du Pont is testing HF detectors at their HF production facility in La Porte,
Texas, before the company installs the same detectors at their fluorocarbon production facility. In
California,  however, the South Coast Air Quality Management District  (SCAQMD) has mandated the
use of HF detectors at the five facilities that use anhydrous HF in the District.  The detectors described
below are  provided as examples to illustrate efforts at specific facilities. Because these examples are
based on a few facility site visits  and on select facility documentation, it is not known the extent to
which these types of detectors are used throughout the industry. Facilities that do not have any HF
detection systems generally rely  on visual observation by operators.

       7.2.1   HF Manufacturing

       To detect leaks, the Du Pont HF production facility in La Porte, Texas, has installed
electrochemical detectors and TV monitors around the process area.6

       7.2.2   Refinery Alkylate Production

       Since ambient air levels of hydrocarbon would normally be associated with an acid leak in an
alkylation unit, hydrocarbon detectors have been installed at alkylation unit sites for early detection of
a release.  This currently available technology consists of both wide area and fixed point detectors
located throughout the alkylation unit site.

        Several refineries use various types of detectors to identify HF releases.  The American
Petroleum Institute's (API) RP 751 recommends using leak detection  systems deemed appropriate for
the unit, including closed-circuit television, point sensors, open-path sensors,  and other imaging
                                                                                      Page 103

-------
 systems.7 For example, a Mobil refinery is evaluating an ion mobility detector as a point sensor to
 detect HF at a concentration of 10 to 20 ppm.  the same facility is testing the use of silicon detectors
 to trigger deluge systems at HF concentrations of 200 to 400 ppm.8  Amoco has installed throughout
 their alkylation units several types of detectors including open path infrared absorption detectors and
 thermal imaging cameras.9

 7.3    General Industry Practices to Mitigate HF Releases

        Mitigation measures are designed to reduce the quantity or concentration of HF released after
 a loss of containment, before the HF migrates off-site, reducing the potential for chemical exposure to
 workers and the general public.  Effective mitigation measures are specific to the site, location
 conditions, process characteristics, and scale of operation. For example, a release of superheated
 liquid HF is likely to result in an aerosol vapor cloud and, therefore, diking would not be an effective
 mitigation technique.  Alternatively, a release of low pressure subcooled HF liquid may form a liquid
 pool.  Diking and vapor suppression may be useful; however, liquid HF vaporizes very rapidly.10

        Because HF can be released as a vapor or as a liquid, mitigation systems have  been
 developed to address both atmospheric releases and liquid spills.  The series of Goldfish tests in 1986
 (see Chapter 9 for additional discussion) showed that for accidental releases of HF at alkylation unit
 temperature and pressure, and at the  ambient conditions at the test facility, the HF flashed and
 generated a denser-than-air cloud. The cloud also  had a high aerosol and cold HF vapor content with
 no liquid drop-out observed.  In 1988,  another series of HF tests called the Hawk HF Test Series (also
 discussed in Chapter 9) was devised to measure the effectiveness of water sprays.  For atmospheric
 releases of HF gases  or vapors, the tests indicated  that water spray systems could be effective in
 reducing airborne HF. At water to HF  ratios of 40 to 1, water sprays have been documented as
 reducing the concentration of HF in the air by up to 90 percent.11'12  The following is a summary
 of the mitigation systems used in the HF industry.

        7.3.1   Water Spray Systems

        Following the  Goldfish Test series in 1986 which examined the source and dispersion
 characteristics of  HF released under alkylation unit conditions, the test participants formed the Industry
 Cooperative HF Mitigation/Assessment Program (ICHMAP).  This group of 20 companies in the
 chemical and petroleum industries then sponsored and conducted a water spray test  program in 1988
 called the Hawk Test series. Approximately 80 experiments were conducted in a flow  chamber with
 the release source a horizontal jet of HF pointed at the mitigation device.  These experiments
 considered variations  in wind speed, humidity, acid  type,  and the mitigation device.

       These tests demonstrated that high HF removal rates could be achieved under controlled
 ideal conditions; however,  the rate could also be reduced by non-optimal interaction between the
 cloud and the spray.  Further, a high ratio of water to HF would not guarantee effectiveness of a water
 mitigation system, and the issue of scale-up from the experimental design to a plant-scale design
 system needed further work. To address these issues a computer program to model water spray
 removal was developed to assess overall effectiveness of a water spray given facility-specific
 configuration and conditions. This model called HFSPRAY was verified against the 1988 Hawk Test
 experiments and found to agree well with both field  test data and wind tunnel data.14

       A water spray system can be used to mitigate a major release of HF. An alternative to water,
 although rarely used, is a mild alkaline solution such as ammonia water or a sodium carbonate
 solution. A water spray system may include several types of water spray subsystems. A fire water
 subsystem consists of a stationary dense water stream that is used primarily for fighting  fires and
 could also be used to knock down HF  vapors.  Other remotely operated or portable water monitors
 are primarily used to knock down HF vapors.  In contrast to the water monitors, a water deluge
Page 104

-------
subsystem provides a narrow water umbrella that is directed towards specific equipment such as
pumps and flanges to knock down an HF leak before it becomes a vapor cloud.15

       To be effective,  HF facilities use sensors or other methods to identify an HF leak, and then
quickly apply water in sufficient quantities. To prevent overflow of the facility effluent system, HF
facilities are beginning to design containment systems that collect the contaminated water and drain
quickly. A subsequent system is needed to neutralize the acid-contaminated water from expected pH
levels between 1 and 2  to pH levels between 7 and 11. For effective, safe remote operation, the HF
mitigation equipment may have a dedicated control panel in the control room.  The American
Petroleum Institute's RP 751 also recommends that any water mitigation system be fully testable, with
operating procedures specifying test procedures and frequencies. 6  The SCAQMD requires an
automated water spray system, or an SCAQMD-approved alternative at facilities that use anhydrous
HF in southern California.17

       Several facilities are concerned that the mitigation systems pose unworkable design
requirements, do not add significantly to the protection of the public, and that the systems have the
potential to cause more harm than good. Water spray mitigation systems do pose many design
challenges.  Facilities need to obtain sufficient water resources and to install large water supply tanks
near the existing HF units. Many facilities have problems obtaining the amount of water required.  The
water spray nozzles have to be designed to provide adequate coverage and droplet size to knock
down the leaking HF vapors effectively.  Also, the high water flows suggested by  the Hawk tests or
required by regulation could damage process equipment further, or even injure workers in the unit. If,
however, the water spray rates are reduced to prevent further  equipment damage or delayed to allow
workers to exit the area, the effectiveness of the mitigation technique could  be diminished.  If a leak
rate exceeds the design specifications, the effectiveness of a water mitigation system may be reduced
significantly.18 Finally, to accommodate water spray systems,  facilities must be able to direct and
collect the acid-contaminated water quickly.  The collected wastewater would need to be neutralized
and disposed of according to applicable regulatory requirements.  Because of the high costs and
design complexities associated with development and implementation of water spray systems,  many
facilities may not choose this mitigation approach.

        7.3.2   Scrubbers

        Scrubber systems can be used to absorb HF gas or vapors vented from  process streams.
Scrubbers are commonly used in HF facilities to absorb HF vapors released from vents, pressure relief
valves, transfer lines, rupture discs, and other devices.  Because HF reacts  readily with water, a water
spray effectively removes the HF gas or vapor from an effluent gas stream,  and the HF-contaminated
water can then be treated.  In some cases, e.g., alkylation units, an alkaline solution is used to obtain
additional HF absorption and to neutralize the HF.  Types of scrubbers that have been used include
spray towers, packed bed scrubbers, trayed towers and Venturis.

        The Chemical Manufacturers Association (CMA) HF committee recommends that all anhydrous
HF storage tanks be equipped with some type of vent gas scrubbing equipment.  Scrubbers are also
required by the SCAQMD at facilities using anhydrous HF in the District.19  During tank filling
operations, fumes will be displaced from the storage tank to the atmosphere unless they are absorbed
by water or alkaline scrubbers.20 Scrubbers designed for emergency releases may be on-line
continuously or activated by leak detectors, pressure or temperature sensors, or  remote manual
response.21 The API notes that in most units, scrubbers are designed to stay on stream at all times,
and therefore do not need to be activated by leak detectors.22
                                                                                    Page 105

-------
        7.3.3   Emergency De-inventory

        Emergency de-inventory or rapid acid transfer systems may mitigate an HF release by
 removing the volume of HF from the failed equipment to safe equipment.  For some facilities, including
 petroleum refineries, the rapid transfer would require removing HF in a stream with other chemicals.
 Factors to consider in a rapid acid transfer system include de-inventory time, the extent of acid
 movement, the force required to move the acid ( e.g., pump), the vessels to include in the de-
 inventory system, the receiving vessels, the venting of HF vapors generated during the acid
 movement, and pressure relief.23 Some facilities are developing  plans to empty a leaking storage
 tank rapidly.  Concern has been voiced that the amount of time required for rapid acid transfer of
 large HF vessels would be longer than that required for all the HF to leak out in a catastrophic
 release. 4 The addition of equipment such as large pumps, lines, and valves involved in rapid de-
 inventory systems may actually increase risk by  providing additional locations for loss of
 containment.    In addition to possible long de-inventory times, the transfer operation itself has
 Inherent release hazards.26

        7.3.4   Secondary Containment

        Secondary containment is an industry practice used to contain releases of many hazardous
 chemicals.  HF has a boiling  point of 19.7°C (67.4°F), which is at or below temperatures commonly
 found under ambient conditions.  Because HF will volatilize rapidly above this temperature, secondary
 containment of HF has limited applicability in climates or processes which may have higher
 temperatures.

        Even though a liquid  pool of HF will volatilize rapidly, containment of liquid releases of HF can
 minimize both ground contamination and vapor cloud size. Examples of containment systems include
 dikes, impounding basins, and enclosures around one or several tanks with a capacity to contain the
 largest tank. The most common containment type for HF use is a low wall dike having a minimum
 capacity of at least 110 percent of the capacity of the  largest storage tank on the site or in a tank
 farm.   Drainage systems underlying a storage vessel is another type of containment system which
 can provide a direct feed from the vessel to a neutralization basin.

        For many volatile hazardous chemicals, impermeable flotation devices and foam applied to a
 pool can further reduce vapor emissions; however, commercially viable flotation devices and foams
 are not available for HF.

        Enclosures are containment structures which surround storage and process equipment to
 capture HF if spilled or vented.  The spilled liquid contained in the enclosure is then treated and the
 gaseous HF is neutralized by scrubbers.  The use of specially designed enclosures for HF storage or
 process equipment does not appear to be widely practiced.  For  processes that use HF and
 flammable chemicals (e.g., HF alkylation), the potential for  explosion in the enclosure may increase the
 risk of a large  release of HF.28

        7.3.5   Remotely-Operated Emergency Isolation Valves

       The magnitude of an HF release can be reduced by using valves that can quickly isolate
 major HF inventories from the source of the leak  or spill.  An increasing number of facilities have
 installed remotely operable emergency block valves. Remote shutoff allows workers to shut down a
 system or piece of equipment from another location.  Remote shutoff eliminates the need for access to
 equipment.  Also, operators in the control room may be able to spot a release or upset condition
 before the field personnel and shut the unit down more quickly using the remote shutoff. Shut-off
 systems can also position valves to direct process HF to a storage drum, and isolate vessels within
 minutes.29  For example, before a field operator can identify a leak from a pump, personnel in the
Page 106

-------
control room may see the leak with TV cameras or detectors and be able to shut down the pump
automatically.  These remote-controlled valves are also used on either side of an unloading hose,
critical pipe, or HF vessel to isolate the equipment in the event a leak develops.30 Particularly in the
unloading/loading area, the two HF producers, fluorocarbon manufacturers, many refineries, and other
industry segments are installing remote shutoff valves on either side of the flexible transfer hoses to
limit the amount of HF released if a hose failure occurs.

       7.3.6   Automatic Valves

       Some systems that handle HF are being equipped with automatic shutoff switches to
deactivate malfunctioning equipment and/or valves to isolate a leak area more quickly. If there is a
process upset or a release occurs, a monitor or detector can be programmed to shut off the
equipment automatically. For example, if overfilling occurs, the level indicator may trigger the
automatic valves on a loading hose to close.  Automatic shutoff valves may be placed in the field, in
the control room, or other convenient locations. To  accompany the automatic shutoff controls, proper
procedures need to specify the conditions for the automatic shutoff of any equipment. Manual shutoff
systems usually back up automatic shutoff systems.

       7.3.7.  Relief Valves

       Relief valves located  on HF storage and process equipment are designed to release HF to
relieve excess pressure that could eventually lead to a vessel rupture. These relief valves are common
in most facilities and are mandated to be installed on vessels to various codes and standards. Most
facilities vent relief valves to scrubbers or other devices to  absorb any HF released.

       7.3.8  Capping Kits

       To stop HF  leaks in valves on tank cars or tank trucks, both HF producers have developed
emergency capping kits for use on most vehicles. The producers and shippers of HF have adapted
the chlorine capping kit for use on HF assemblies.  Regardless of the design, the emergency capping
kit is placed over the leaking/defective valves on the top of the tank car or truck. Du Pont also uses a
European design for valves on tank trucks to prevent an HF release if the valves are sheared off the
top.31

7.4    HF Mitigation Systems Used by Specific Industries

       Most of the  HF facilities that were visited or supplied written information use some form of
mitigation system.  The mitigation systems described below are provided as examples to  illustrate
efforts at specific facilities. Because these examples are based on a few facility site visits and on
select facility documentation, the extent to which these mitigation systems are used throughout the
industry is not known.

       7.4.1   HF Manufacture

       Du Pont requires that its facilities have the capability to empty both process and storage
equipment on short notice in the event of an emergency.  The La Porte facility is installing tank car
motion detectors; an HF transfer is shut down if_a tank car moves during transfer operations.  Du Pont
conducts inventory drills at least once per year.
32
        At Du Pont, diked areas are designed to divert a spill for treatment rather than to contain it.
Asphalt is used rather than concrete under storage vessels because HF reacts with concrete. The
company also requires that water dilution facilities be adequate to dilute the largest credible HF leak
to below 20 percent HF concentration, to minimize the amount of HF that is vaporized.33
                                                                                     Page 107

-------
        To further reduce the hazards associated with large quantities of HF, Allied-Signal has
 reduced their typical HF inventories from 2,000 tons to 1,200 tons.

        7.4.2   Fluorpcarbon Manufacturing

        Because of the large HF storage vessels typically in use at fluorocarbon manufacturers, de-
 inventory may not be a viable alternative.  For example, at a Du Pont fluorocarbon manufacturing
 facility in Corpus Christi, Texas, it would take over 13 hours to assemble the number of rail cars
 required to empty a large 250,000 gallon storage vessel.  Instead, the facility emphasizes
 comprehensive vessel inspection, preventive maintenance and repair of the vessel.  They maintain a
 quick weld program to affix metal plates to cracks or holes in the vessel,  but have never had to use it
 In an emergency situation.

        7.4.3   Refinery Alkylate Production

        One recommendation to API's Committee on HF Alkylation was to enhance mitigation systems
 to lower the risk to the community and the environment.  In their position paper "The Use of
 Hydrofluoric Acid in the Petroleum Refining Alkylation Process," API developed safety guidelines
 containing options to mitigate the impact of a release.  These included:

        *-       providing a monitoring system for early detection of an HF release;

        >       implementing a system for applying large quantities of water to acid
                releases;

        >•       removing or segregating acid inventory rapidly from the process to
                minimize the amount of acid released;

        >•       using a system of remotely operated isolation valves in acid-containing
               parts of the process to limit the potential HF release in case of a leak;
               and
        »•      following operating procedures and design of facilities to minimize the
               inventory of HF and to minimize the number of points in the process
               where HF might be released.

These guidelines also discuss the need to train and work with emergency response personnel and the
community to determine the proper response to an HF release.  All of these mitigation steps depend
on management's overall commitment to ensure that procedures, training, and audits are
implemented.

        Several refineries, including Ultramar, Mobil, Phillips, Chevron, and Amoco, have or are
planning to develop and install water spray mitigation systems.  For example, after extensive research,
Amoco is installing a massive water spray system surrounding their HF alkylation unit in Texas City,
Texas.  The system is able to deliver 33,000 gallons per minute from a water curtain and a total of up
to 40,000 gallons per minute from both the water curtain and fire monitors.34

        The Chevron refinery in Salt Lake City is installing a deluge system consisting of five elevated
water towers  that can deliver a total of approximately 12,900 gallons per minute. The monitors, which
are 50 feet high, can be adjusted to deliver a spray, or a direct stream for greater distance. Monitors
are controlled by joysticks in the control room.  A large lined sump will collect the water, where it will
be neutralized before being treated in the wastewater treatment system.
Page 108

-------
       For fully installed systems, data are hot available on water spray effectiveness on an actual HF
release; however, several refineries, including Ultramar, have tested the systems in simulation
exercises. In addition, HFSPRAY, a verified mathematical model, has been developed to quantify
effectiveness of water sprays at specific installations, given specific release scenarios and weather
conditions.36

       7.4.4  Linear Alkylbenzene Production

       Vista Chemical has installed monitors for leak detection, an emergency relief scrubber system,
a portable potassium hydroxide vacuum system, and emergency shutdown and isolation systems to
mitigate leaks and spills. Additionally, they have put in place a water spray system, an emergency
pump-out system, and a closed circuit TV-monitoring system for the HF unit.
                                                                                     Page 109

-------
                                         ENDNOTES

 1.      Seringer, Carolyn S., Du Pont Chemicals, comments from technical review of Hydrogen
        Fluoride Study, Report to Congress, Draft May 8, 1992, June 5, 1992. (436.4- Jim Small)

 2.      U.S. Environmental Protection Agency, Review of Emergency Systems, Report to Congress,
        Section 305(b) Title III Superfund Amendments and Reauthorization Act of 1986, Office of Solid
        Waste and Emergency Response, Washington, D.C., May 1988. (489.92)

 3.      Seringer, Carolyn S., Du Pont Chemicals.

 4.      Du Pont Chemicals, Draft AHF Safety Guardian Manual, Wilmington, DE, December 18, 1991.
        (137.8)

 5.      Morris, Jeff, American Petroleum Institute, Fina Oil and Chemical Company, comments from
        technical review of Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992,
        June 1, 1992.  (344)

 6.      Facility visit by U.S. Environmental Protection Agency to Du Pont, La Porte, TX,
        December  12, 1991. (620)

 7.      API Recommended Practice 751, First Edition, "Safe Operation of Hydrofluoric Acid Alkylation
        Units," American Petroleum Institute, Washington, D.C., June 1992. (10.6)

 8.      Presentation by Mobil Oil to U.S. Environmental Protection Agency, January 28, 1992. (610)

 9.      Presentation by Amoco to U.S. Environmental Protection Agency, February 6, 1992. (640)

 10.     API Recommended Practice 751.

 11.     HF Mitigation Water Spray Project, created for a meeting of the API and given to the U.S. EPA,
        December 10, 1991. (424.34ABC)

 12.     Presentation by Mobil Oil.

 13.     Blewitt, Doug,  N., Design of Water Spray Mitigation Systems for Amoco HF Alkylation Units. (56)

 14.     Blewitt, Doug,  N.

 15.     Mobil Oil Corporation Torrance Refinery, Hydrofluoric Acid Alkylation Unit Risk Management
        and Prevention Program Report, Torrance, CA, March 16, 1989. (185)

 16.    API Recommended Practice 751.

 17.     Hofman,  Hazel, Shahid Khan, Chris Marlia, and David Yeh, Guideline to Comply with Proposed
       Rule 1410:  Hydrogen Fluoride Storage and Use, South Coast Air Quality Management District,
        CA, June 14, 1991. (160)

 18.    Seringer, Carolyn S., Du Pont Chemicals.

 19.     Hofman,  Hazel.

 20.    Allied-Signal, Inc., Anhydrous Hydrofluoric Acid, The HF Products Group, Morristown, NJ. (17)
Page 110

-------
21.     Du Pont Chemicals, Draft AHF Safety Guardian Manual.

22.     Morris, Jeff, American Petroleum Institute, Fina Oil and Chemical Company, comments from
       technical review of Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992,
       June 1, 1992. (344)

23.     API Recommended Practice 751.

24.     API Recommended Practice 751.

25.     Seringer, Carolyn S., Du Pont Chemicals.

26.     Hofman,  Hazel.

27.     Allied-Signal, Inc., Anhydrous Hydrofluoric Acid.

28.     Whittle, David K., Donald K. Lorenzo, John Q. Kirkman, "Results of Alkylation Unit Hazard,
       Operability Study Are Analyzed and Summarized," Oil and Gas Journal, July 10, 1989, p 96.
       (365)

29.    Personal Communication, Conversation with Mike Brown and Bill Davis, Salt Lake City Chevron
       Refinery, Salt Lake City Utah, April 14, 1992. (76.7)

30.    API Recommended Practice 751.

31.    Facility visit by U.S. Environmental Protection Agency to Du Pont.

32.    Du Pont Chemicals, Draft AHF Safety Guardian Manual.

33.    Du Pont Chemicals, Draft AHF Safety Guardian Manual.

34.    Wade, Robert C., HF Mitigation Strategies atAmoco's Texas City Alkylation Unit, Amoco Oil
       Company, Chicago, IL (499.5)

35.     Personal Communication, Conversation with Mike Brown and Bill  Davis.

36.     Fthenakis, Vasilis M., Brookhaven National Laboratory, comments from technical review of
       Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992, June 8, 1992. (139.69)
                                                                                    Page 111

-------
                                  This page intentionally left blank.
Page 112

-------
      8.  CHARACTERIZATION OF HYDROGEN FLUORIDE ACCIDENTS
       Despite practices to prevent, detect, and mitigate releases of HF, as discussed in Chapters 6
and 7 there have been a number of accidental releases of HF. This chapter reviews such accidents.
A compilation and characterization of accidents associated with HF can serve several useful purposes:

       >.      an understanding of how and why an incident occurred can help
              determine how HF chemical and process hazards contribute to
              accidental releases;

       *      a review of the events that contribute to an incident may also  provide
              useful insight into how accidental releases can be prevented;  and

       ».      an evaluation of probable root causes may also help to determine if
              certain types of incidents are more likely in certain HF industry
              segments.

       This chapter first includes a discussion of a few of the more notable incidents involving HF
that have occurred in the U.S. Following this is an analysis of the accident data available for HF.  To
the extent possible, EPA has tried to distinguish between incidents  involving aqueous and anhydrous
HF.

8.1    Examples of Major Accidents

       8.1.1  Marathon Petroleum

       Attention has recently been focused on HF mainly because of a large accidental release of HF
that occurred on October 30, 1987, at the Marathon Petroleum Company refinery in Texas  City, Texas.
This incident occurred when a 50-foot,  multi-ton heater convection unit was accidentally dropped onto
the top of an HF storage vessel.  The unit was being moved for repair and maintenance during a
 general plant turnaround.  The dropped convection unit severed a 4-inch acid loading line and a 2-
 inch pressure relief line causing the HF to be released.1  An estimated 30,000 to 53,000 pounds of
 HF vapors were released.2'3

        Although most of the HF was released during the first few minutes in the form of an aerosol as
 the storage vessel depressurized, vapors continued to discharge at a much lower rate from the vessel
 for the next 44 hours and  migrated northwest through an adjacent residential area. Eighty-five square
 blocks and approximately 4,000 residents were evacuated; 1,037 residents reported to three
 neighboring hospitals. Injuries included skin burns and irritation to the eyes, nose, throat,  and lungs.
 Vegetation also was damaged in the path of the vapor cloud. No fatalities occurred. •

        The accident and the cause of the discharge were investigated by OSHA; Notices of Violation
 (NOVs) were subsequently sent to Marathon and two of its contractors. The specific problems cited
 included:  not instituting accepted engineering control measures to prevent the release (i.e., emptying
 HF vessel before hoisting a heavy load over it; not hoisting a heavy load over an HF tank); the crane
 was not properly blocked (wooden blocks supporting  crane outriggers were  crushed); crane
 inspection documentation was not prepared; and the crane safety devices were not inspected prior to
 use and a malfunction occurred.6
                                                                                     Page 113

-------
         8.1.2  Mobil Oil

         A 165-pound release of HF occurred in 1987 at Mobil Oil's refinery in Torrance, California. The
  release occurred following an undetected excess flow of HF to the alkylation unit's propane treater.
  The propane treater uses potassium hydroxide (KOH) to neutralize trace amounts of HF in liquid
  propane, an alkylation byproduct. An excess of HF was charged to the treater, and was not detected
  because a series of controllers and alarms was inoperable.  The probable cause of the accident,
  determined by examining the damaged equipment, was that the presence of excess HF resulted in an
  exothermic reaction and created abnormal pressure,  causing the KOH treater to fail. The upstream
  cooler may have failed as well for the same reason.  When the treater failed, it  released HF and
  propane, which exploded and started a large fire.6

         While the two events discussed above, along with spill tests by industry, are often cited as the
  stimuli for focussing attention on HF, consideration of other incidents involving  HF can provide a more
  balanced and complete picture.

         8.1.3  Sequoyah Fuels Corporation (Subsidiary of Kerr-McGee)

         On January 4, 1986, a cylinder containing uranium hexafluoride (UFg) ruptured at the
 Sequoyah Fuels Facility in Gore, Oklahoma.  Once released, the UF6 hydrolyzed to paniculate uranyl
 fluoride (UOgF^) and  HF. The reaction produced a white plume that dispersed  from the accident site
 with the prevailing winds.  Most of the solid reaction products were deposited on-site. In this incident,
 the UF6 cylinder was overfilled because of improper positioning on a scale designed to gauge the
 capacity of the container; one wheel of the dolly holding the cylinder was on the solid floor and  not on
 the scale.  As UF6 is normally a solid at room temperature and pressure, cylinders must be heated to
 liquefy the chemical before it can be removed.  Although official company policy prohibited the
 heating of overfilled tanks to remove excess chemical, the day shift supervisor instructed an operator
 to place the cylinder in a steam chest for six hours to liquefy the UF6. About two hours later, the
 cylinder ruptured releasing 29,500 pounds of UF6 and generating a large cloud of HF and uranyl
 fluoride. Subsequent investigation by the Nuclear Regulatory Commission Interagency Public Health
 Assessment Task Force revealed that the cylinder failed because of excess hydraulic pressure and
 approximately 3,350 pounds of HF was released as a result of the reaction of UF6 in the atmosphere.
 One worker died, about 35 people were injured, and more than  100 workers and residents were
 screened at local hospitals as a result of the accident.7'8 Favorable wind and weather conditions
 dispersed the cloud.  Many employees  at the site reported to evacuation points. The public was
 notified by radio and several residences in the area were evacuated.

        8.1.4  Great Lakes Chemical Corporation

        Another release of HF occurred on June 27, 1989, from the Great Lakes Chemical Corporation
 in El Dorado, Arkansas.  The company uses HF to produce brominated fire retardant chemicals.
 Several bolts on a diaphragm isolator between a pressure gauge and a valve on an  HF storage tank
 failed due to corrosion, and tank pressure forced the two halves of the isolator apart, releasing 1,320
 pounds of HF.  Although there were no injuries, evacuations, or other consequences associated with
 this release, it is illustrative of an equipment failure situation in which HF vapor was released.9

 8.2     Analysis of HF Accident Databases

        Several databases were examined to identify accidents involving HF:

        *•      EPA's Accidental Release Information Program (ARIP) database;
        >•      EPA's Acute Hazardous Events (AHE)  database;
Page 114

-------
       *•      Emergency Response Notification System (ERNS) database;
       ».      DOT'S Hazardous Materials Information System (HMIS);

       Newspaper accounts and facility accident reports were also collected.  A detailed description
of each database listed above is provided in Appendix XII. Specific accident data from ARIP, AHE,
ERNS  and HMIS are provided in Appendix XIII, Exhibits 1 to 4, respectively. Information taken from
newspaper accounts and facility accident reports can be found in Appendix XIII, Exhibit 5. Exhibits 1
to 5 in Appendix XIV contain information solely on the accidents that resulted in death, mjuiy, or
evacuation from each of these databases and other sources. The data sources for Appendices XIII
and XIV provide varying levels of information about each incident; therefore, not all entries are
complete.  Because of the variety of information provided in the databases, each database will be
analyzed and discussed separately below.

       8.2.1 ARIP Data

       The ARIP database provides accurate and facility-verified information on the causes and
prevention of accidental releases.  However, because the database is designed to incorporate only
the most severe accidents,  it does not represent a nationwide statistical sample of all  accidents that
have occurred. The following analysis should be viewed with this in mind. See Appendix XII for more
details on the ARIP database.

        Of the approximately 2,700 events reported in ARIP, 33 (about one percent) involved HF.
Exhibit 8-1 shows the total number of ARIP and HF events recorded annually for 1986 through 1991.
                                         EXHIBIT 8-1
                                HF Events in the ARIP Database
                                                                    HF Events as a
                                                                 Percentage of the Total
Total Number of
  ARIP Events
Number of HF
   Events
              1991
              i^m
         Total of All Years
         The ARIP data were analyzed to determine the number of release events for each HF industry
  segment. The Chemicals segment had the largest number of releases (10), followed by Petroleum
  Refining (7) Metal Production (6), Aircraft/Aerospace (4), Television/Semiconductors (3), Metal
  Fabrication (2), and Transportation (1).  The total quantity released in the Chemicals segment was
                                                                                      Page 115

-------
  about 9,500 pounds. The Petroleum Refining segment had releases totaling approximately 58,000
  pounds, all but 4,500 pounds of which were from the Marathon accident. The Metal Production
  segment released about 23,000 pounds of HF,  nearly all of which was in aqueous form.

         ARIP data show that anhydrous HF was released in 13 events; the remaining 20 events
  involved aqueous HF.  Generally, releases to air were reported for both releases of anhydrous HF and
  releases of aqueous HF in concentrations 70 percent and above. The releases of aqueous HF in
  concentrations less than 70 percent were reported primarily as releases to land.

         ARIP data can also be used to characterize the duration and rates of release events  The
  overall durations of HF events reported in ARIP  ranged from 5 minutes to 78 hours.  Release rates
  (averaged  over the duration of the release) ranged from less than 1  pound per minute up to 320
  pounds per minute.  Overall, most HF events reported in the ARIP database are of relatively short
  duration and small quantity.

        ARIP data provide information on the point in the process where a release occurred how the
  release was discovered, and the cause of the release. Exhibit 8-2 shows the distribution of locations
  in a facility where releases occurred. Leaks associated with piping (process and storage vessel
  piping, joints, and  instruments) account for 10 of the 33 releases, while vessel leaks account for much
  of the remainder.

        Exhibit 8-3 illustrates the primary and secondary causes of these releases. Equipment failure
  was given as the primary cause in 55 percent of the 33 incidents and as the secondary cause in 9
  percent.  Corrosion and inadequate maintenance and inspection are likely to be root causes of these
  incidents. Human  error was also listed frequently  (27 percent) as the primary cause of release
  Overall, there were few secondary causes  identified.

        Early discovery is an important part of mitigating the adverse consequences of an accident
 As shown in Exhibit 8-4, 55 percent of the  HF releases identified in ARIP were discovered through
 operator observation.  Gauges and other devices,  such as monitors, were responsible for detectinq 15
 percent of the releases.                                                                   a

        More than half of the HF releases reported in ARIP involved spills of aqueous HF. Releases of
 HF vapor occurred almost as often.  Releases of concentrated aqueous HF can produce vapor
 clouds, but not to the same extent as anhydrous HF. Aqueous HF releases caused more
 environmental damage; releases of pickle liquors, etchant solutions, and other HF  mixtures have
 caused fish kills and soil contaminations. According to ARIP data, releases of anhydrous HF are more
 likely to pose exposure hazards off-site than releases of aqueous HF. Releases of aqueous HF on
 the other hand, pose greater environmental damage hazards.
Page 116

-------
                               EXHIBIT 8-2
                       Release Point From ARIP Data
                  (Percentage of Total Events)

                                             Process Vessel 12

                                                          Waste System  12
     Storage Piping  6

     Instruments 6
         Valves 6

              Other 6
          Shipping Container 6
    Storage Vessel 24
     • ARIP Database, U.S. Environmental Protection Agency, 1986 to 1991. Data as of January 29,1992.
                                EXHIBIT 8-3
                       Release Cause From ARIP Data
                                                       Human Error  27 %
                  (Percentage of Total HF Events)
Equipment Failure
        55%
Other  9%
                                                         Process Upset 6%
                                                          Maintenance 3%
    Source:  ARIP Database, U.S. Environmental Protection Agency. 1986 to 1991. Data as of January 29,1992
                                                                           Page 117

-------
                                          EXHIBIT 8-4
                               How Was The Release Discovered?
                             Initial Release Discovery
                           (Percentage of Total ARIP Events)
                  Operator Observation
                         55%
                                                                 Unknown  18
                                                                     Other  12%
           Injury or Death 3%
                        Gauge or Device
                            15%
        Soorca: ARIP Database, U.S. Environmental Protection Agency. 1986 to 1991. Data as of January 29,1992.
Foreman Observation
          15%
        Injuries and evacuations, as reported in the ARIP database, are relatively infrequent in
 incidents involving HF. Excluding the Marathon event, ARIP data report only six employees injured no
 members of the public injured, and no evacuations.

 A mo   ThS ARIP database does not reP°rt ar(y fatalities from HF incidents; however, the version of
 ARIP used in this study does not include a recent incident at Southwestern Refining in Corpus Christi
 Texas, where two employees were killed in an HF incident that resulted from maintenance activities  '
 conducted in violation of standard procedures. Pump mechanics at the facility had removed all but
 three bolts that held the HF acid circulation  pump casing together,  in preparation for maintenance
 Upon  removal of the next bolt, the remaining two bolts failed and the pump casing housing blew off
 releasing HF.  The mechanics did not know that a discharge valve leaked, allowing system pressure to
 build in the pump casing.  One employee was killed from injuries caused by the impact of the pump
 casing and the other from exposure to HF.  Six other employees were injured in this incident. 1°

        Another observation that can  be made from the ARIP  database concerns the use of formal
 hazard evaluation methods. As noted in Chapter 6, a formal hazard evaluation (process hazard
 review) is critical to chemical accident prevention and the development of a good process safety
 management program as  required in the recently promulgated OSHA process safety manaqement
 standard. ARIP data show that 23 of the 33 facilities in this data set have not performed a formal
 hazard evaluation.  The remaining ten  have performed one or more formal hazard evaluation
 procedures.

       In 17 events in the ARIP database, HF is cited as the secondary chemical released  These
 events primarily involve HF solutions.  About seven of the events occurred at chemical facilities and
 another five occurred at aerospace facilities.  Releases of HF as a secondary chemical also occur
 primarily from piping and storage.
Page 118

-------
       8.2.2 AHE Data

       The AHE database contains information on approximately 6,000 events, including 27 incidents
involving HF. These data cover HF releases that occurred between 1980 and 1987. Refer to
Appendix XII for details on the AHE database.  Fourteen of the HF incidents occurred at fixed facilities
and the remainder occurred during transportation.  Of the 17 transportation incidents, five involved
releases from tank cars, three from tank trucks, six from drums, and three were unknown  (Exhibit 8-5).
Six injuries associated with the HF events were reported, three in fixed-facility incidents and three
during transportation incidents. Two fatalities were reported.  Evacuation was performed  in five
events, two of which were associated with a fixed facility and three with transportation  incidents.

                                         EXHIBIT 8-5
                                Source of Transportation Leaks
                                     (Number of Events)
                                                                Tank Truck 3
                                                                 Unknown  3
                    Source: AHE Database, U.S. Environmental Protection Agency, 1960-1987
                         Data as of Junes, 1991.
        8.2.3  ERNS Data

        The ERNS database contains records of incidents from 1986 to the present. Refer to
 Appendix XII for details on the ERNS database. During the period up to November 10, 1991, 97 HF
 events were reported, representing less than 0.1 percent of the approximately 150,000 incidents
 recorded in the database.  Most of the HF events (70 percent) occurred at fixed facilities.  The
 distribution of releases from shipping containers reported in ERNS includes nine from drums, eight
 from tank trucks, and one each from tank cars, jugs, and ship containers.  The quantity released
 reported in ERNS ranged from 1 to 82,000 pounds.  No fatalities were reported in the database;
 however, there were eight events involving injuries and  six events reporting evacuations. Fifty-one
 percent of the HF events were associated with aqueous HF  releases, and 49 percent with anhydrous
 HF releases.
            11
                                                                                      Page 119

-------
         8.2.4  HMIS Data

         During the period from January 1, 1980 to December 31, 1990, 19 HF incidents were reported
 to the U.S. Department of Transportation's (DOT) HMIS.  Refer to Appendix XII for details on the HMIS
 database.  As shown in Exhibit 8-6, eight incidents occurred due to dropped or failed packaging (i.e.,
 cylinders, plastic drums and bottles, metal drums), and eight incidents resulted from valve failure on a
 tank car or other package.  The remainder of the incidents occurred due to derailment, tank car piping
 failure, or a weld failure on a tank car.

                                          EXHIBIT 8-6
                                 Failure Types From HMIS Data
                                   (Number of Incidents)
            Packaging Failure 8
Derailment  1
    Piping Failure 1
                                                                      Weld Failure
                                                                       on Tankcar 1
                                                      Valve Failure  8
            • Includes dropping or failure of cytltndars, plastic drums or botttes.or metal drums
            ** Includes failure on tank cars or packaging
            Source: Hazardous Materials Information System (HMIS), Department of Transportation. January 1,1980. Data as of Nov. 11,1991.
 8.3     Overview of HF Accident Data

        For analyses of release cause, facility type, consequences, and transportation releases, the
 release information from all data sources were combined. The total number of incidents described in
 Appendix XIII is 155, corrected for duplicate reports. All releases are identified as either anhydrous or
 aqueous. Because the hazards to facility employees and the public are different for anhydrous and
 aqueous HF, releases of anhydrous and aqueous HF are analyzed separately when considering on
 and off-site consequences (e.g, deaths, injuries, evacuations). The release data are similarly
 separated in examining transportation releases.

        8.3.1  Release Cause

        The 155 HF releases reported in the data sources discussed above appear to be caused
 primarily by equipment failure and human error,  or a combination of both.  These causes are often
 Interrelated; for example, equipment failure (e.g., malfunctioning valve) may be due to or exacerbated
 by human error (e.g., inadequate maintenance).  Consequently, the primary or root cause of a release
Page 120

-------
is not always accurately identified by the facility in release reporting. Analysis of accident data,
however, does provide useful insight into some of the events leading to releases.

       HF releases were most often attributed to the failure of process and transportation equipment,
including valves, piping, pumps, unloading hoses, and storage vessels, with failed valves the most
frequent cause.  Also, failure of packaging was identified  in a large portion of the incidents. Corrosion
is probably the major hazard leading to equipment and packaging failures. For example, in a release
at Great Lakes Chemical in Arkansas, corrosion from HF caused failure of the pressure gauge on a
storage vessel which led to the release of 1,320 pounds of HF.12

       Human error, such as failure to follow standard operating procedures, was also frequently
cited as the cause for HF releases. Human errors can include such things as opening flanges on a
process line before ensuring that the line is cleared,  mistakenly opening a valve  on a line under
pressure,  dropping cylinders or bottles of HF, and leaving containers open during transport. Failure to
follow standard operating procedures or accepted engineering control measures was cited in two of
the largest accidents involving HF (i.e., Marathon Petroleum and Kerr-McGee).

       In another large release attributed to human  error, employees at Consolidated Rail
Corporation misjudged the severity of the situation and did not respond to a leaking tank car.  A weld
eventually failed on the pressurized vessel, resulting  in a vapor cloud containing 6,400 pounds of HF
that traveled 2.5 miles.   Another release attributable to human error occurred when a drum was
punctured by a fork lift at the Inland Container Corporation in Missouri, resulting in a release of 413
pounds of liquid HF.14

        EPA recognizes that errors in equipment design, standard operating procedures or by
management can lead to operator or maintenance errors that cause accidental releases.  Information
on such errors and their significance in HF incidents are not available.

        8.3.2   Types of Facilities

        Accidental releases of HF have occurred in every HF industry segment.  The hazards posed
by these  releases depend on the type of facility and the  amount and concentration of HF released.
The HF release data in Appendix XIII show the number of accidents associated with various industry
segments or processes, as indicated in Exhibit 8-7.  The designations of facility  type for this exhibit
are, in some instances, different than those made in the  specific databases,  especially for
transportation-related incidents.  Transportation incidents include those during loading and unloading
as well as those during transit.  Also, there may be some overlap between the chemical manufacturing
and fluorocarbon manufacturing categories.

        The 55 transportation incidents (approximately 35 percent of the total) include loading and
 unloading, equipment failure incidents during transportation, and package failures during transport to
 or from a facility.  Most facilities recognize that unloading/loading is a vulnerable operation because of
the number of "handlings," or transfers of material involved (i.e., from storage to transport vessel, and
from transport container to storage.).  Several refineries with HF alkylation units have examined at
 least one transportation-related incident (e.g., failure of unloading hose) in their hazard evaluations.

        The release hazards associated with tank car shipments of anhydrous HF are related primarily
 to the likelihood of derailment and subsequent loss  of tank car integrity.  A leak can occur during
 shipment due to normal operational mechanical stresses, which can lead to loosened fittings,  gasket
 leaks or,  leaking valves.15 The rate of derailments has been declining since 1978.  For example, the
 frequency of derailments  in  1978 was ,15 per million train-miles; this was reduced to eight by 1983 and
 tofivein1990.16
                                                                                      Page 121

-------
         Chemical production incidents include those which occurred during HF production and use in
 the manufacture of various chemicals for such applications as Pharmaceuticals, rocket fuel oxidants,
 plastics, dyes, and electroplating baths.  Industries that use dilute HF solutions (i.e., aircraft
 manufacturing, stainless steel manufacturing,  electronics manufacturing, and glass etching) report
 fewer HF releases.

                                           EXHIBIT 8-7
   	                 Industry Segment  and Number of HF Releases
           Industry Segment *
              Transportation **
             Petroleum Refining

         Chemical Manufacturing

        Electronics Manufacturing

     Stainless Steel Manufacturing

          Aircraft Manufacturing

            CFG Manufacturing

                Glass Etching
         Aluminum Manufacturing

                   Cleaning

                Undesignated
                                  10       20       30       40
                                                Number of Releases
50
         60
                  70
         * Designations may not correlate with those in the databases, and categories

         (such as Chemical Manufacturing and CFC Manufacturing) may have some interchange
         ** Includes loading, unloading, and in transit.

         Source: Compilation from ARIP, AHE, ERNS. HMIS, newspaper articles, and facility
                  accident reports - corrected for overlap.
        8.3.3 On- and Off-site Consequences

        The potential consequences of an accidental release at a facility that handles hazardous
 chemicals are largely determined by the conditions of the release, the quantity, the behavior of the
 hazardous chemical in the environment,  and the proximity and sensitivity of populations potentially
 exposed  The concentration of the hazardous material being handled is also critical in assessing the
 hazards posed to the public by the release.  For example, a small release of anhydrous HF may pose
 a greater  hazard to the public than a large release of aqueous HF at a concentration less than 40
 percent. There have been HF releases involving anhydrous HF and various concentrations of
 aqueous HF.  The potential hazards of HF exposure to the public depend largely on the concentration
 of HF used in various industry segments or processes.  Releases from refineries using HF alkylation,
 for example, would likely involve anhydrous HF under conditions which could lead  to formation of a
 dense vapor cloud that could migrate off-site; a release from etching operations would involve HF in
 concentrations less than 70 percent which would be less likely to migrate off-site.
Page 122

-------
       In order to determine the severity of release consequences, releases of anhydrous and
aqueous HF have been analyzed separately.  In the 155 releases examined from all data sources,
there were 82 incidents involving anhydrous HF and 73 involving aqueous HF. The analysis of HF
release migration and the evacuations, injuries, and deaths attributable to these HF releases indicate
that anhydrous HF and aqueous HF with concentration 70 percent or more pose the greatest potential
hazard to the public and facility employees.

       Migration Off-Site

       When released, anhydrous HF and aqueous HF in concentrations of 70 percent or higher may
migrate off-site and pose a threat to the public.  The ARIP database includes ten HF incidents in
which migration off-site of the released material was reported. These include seven releases of
anhydrous HF or 70 percent aqueous HF. In addition, three incidents were reported involving
aqueous HF in concentrations less than 70 percent; in one of these incidents, migration of HF off-site
in the air was reported, but  in the other two, HF was reported released to water.

       Evacuation

       Evacuations resulting from accidental HF releases were analyzed because they may indicate a
perceived hazard to the community. Of all reported HF releases,  12 releases, or seven percent,
resulted in evacuations.  Ten of  12 evacuation events involved anhydrous HF. This indicates that it
may be more appropriate to focus attention on releases involving anhydrous rather than aqueous HF.
Transportation-related incidents  accounted for most of these  evacuations, perhaps because no
containment or mitigation measures are likely to be in place,  as would be the case at a fixed facility.
For example, a Consolidated Rail Corporation tank car leaked 800 gallons of HF, causing a 2.5-mile
vapor cloud and required the evacuation of 1,500 people in a 1.1  square mile area.  Many of the
incidents requiring evacuation involved large HF releases of more than 5,000 pounds.  The largest
release (Marathon Petroleum) caused the largest evacuation  incident in which approximately 4,000
members of the general public were evacuated. The smallest release incident which had an
evacuation involved a spill of three pounds.

       Deaths and Injuries

       Deaths and  injuries  have resulted from 28 reported HF releases. Exhibit 8-8 shows the
number of releases that caused injury/death from releases of HF at concentrations below 70 percent
and those equal to or greater than 70 percent, including anhydrous HF. Two releases resulted in
fatalities. In one case, two employees died when  a pipeline ruptured releasing 150 pounds of
70 percent HF at McDonnell Douglas.  In the other case, an employee died and 100 other employees
were injured, as a result of a 29,500-pound vapor release of UF6 from a cylinder failure at Kerr-McGee,
that generated about 3,350  pounds of HF.

       Of the releases that caused at least one injury,  seven were releases of aqueous HF  less than
70 percent concentration and 20 were releases of anhydrous HF or HF 70 percent or higher.  The only
case where injuries to the public were reported was due to a release of anhydrous HF.

       Exhibit 8-9 shows the number of injuries or deaths associated with either aqueous and
anhydrous HF incidents.  A total of 167 on-site injuries and 1,108 off-site injuries were reported for the
incidents examined.  All of the off-site injuries were attributable to two incidents. The Marathon HF
release caused 1,037 off-site injuries.  The remaining 71 injuries resulted from the Consolidated Rail
tank car leak created by a failed weld. Another 22 injuries were not differentiated as off- or on-site.
Injuries occurred from accidents involving as little as one pound of HF  (Markair) and concentrations as
low as one percent  (Learjet). Most injuries to the public or facility employees, however, occurred
following large vapor releases.
                                                                                     Page 723

-------
                                           EXHIBIT 8-8
              Releases of Aqueous and Anhydrous HF Resulting In Injuries or Deaths
                                                                 Aqueous HF (7)
                   ARIP, ARE, ERNS, HMIS Data Bases and Other Sources (In Appendix XIII), and
                   Segregation of Data Bases for Anhydrous and Aqueous HF by W.J.Hague, Allied-Signal, 8/1/92
             Anhydrous HF (21)
         (Including >70% HF concentration)
                                           EXHIBIT 8-9
              Number of Reported Injuries or Deaths Associated with all HF Releases
                                       Off-site (71)
                      On-sfte (167)

   Injuries (unspecified)  (22)
                                                                   Off-site (Marathon) (1037)
                   ARIP, AHE, ERNS, HMIS Data Bases and Other Sources (In Appendix XIII)
                   27 incidents caused at least one injury.
Page 124

-------
       The average number of injuries per release also provides an indication of the relative hazard
posed by anhydrous versus aqueous HF releases.  With the Marathon incident included, there were
68 injuries per anhydrous HF release, and 2.3 per aqueous HF release.  Without the Marathon
incident, the average number of injuries per release for anhydrous HF is 13, which is still more than
5 times the average number of injuries  per aqueous HF release.  Exposures due to aqueous releases
are probably localized and limited to the facility employees working directly with the HF, whereas the
vapor released in an anhydrous HF release may also injure other workers. In fact, most of the
accidents involving aqueous  HF were releases of small quantities, which were contained within the
facility.

       Thirty-five percent of the HF incidents reported from refineries resulted in  injury. Twenty-three
percent of releases involving the transport of HF resulted in injury. In addition, 50 percent (i.e., 3 of 6)
of the HF incidents at aircraft manufacturing facilities resulted in  injury.  It is unclear why this industry
segment has such a high rate of employee injuries related to releases of HF.  Injury rates in other
industry segments were somewhat lower.

       8.3.4 Transportation Incident Analysis

       There were 55 transportation incidents out  of the 155 HF releases reported in the data
sources considered.  These included releases during HF transport and loading/unloading operations.
Transportation releases included 27 releases of aqueous HF and 28  releases of anhydrous HF.
Quantities of anhydrous and aqueous HF released during transportation incidents are generally small.
The amount released was reported for 40 of these  incidents; 53  percent of the releases were less than
100 pounds.

       The  average amount of anhydrous HF released during transportation incidents, approximately
425 pounds, is also smaller than the average anhydrous release at fixed facilities. One
uncharacteristically large incident involved a release of 6,400 pounds. When this release is excluded
from consideration, the average release quantity decreases to 273 pounds.  Releases from the
refining, chemical, and metal manufacturing industries averaged 5,200 pounds, 1,400 pounds, and
2,900 pounds, respectively.

       A recent study prepared for the Hydrogen  Fluoride Panel of the Chemical Manufacturers
Association  (CMA) that focused on incidents involving tank cars and tank trucks  revealed that almost
five billion pounds of anhydrous HF have been shipped by rail and tank truck over the past ten years,
with very few release incidents and no fatal or serious injuries as a result of anhydrous HF releases
during transportation.17 The study also reported that over the past 20 years, a total of 13 anhydrous
HF releases from tank cars and tank trucks have been reported to DOT, as shown in Exhibit 8-10.18
There were no injuries resulting from these releases.19 The study also concluded that the number
and frequency of tank-car and tank-truck incidents involving loss of anhydrous HF has decreased over
the last 10 years. It also found that release amounts were small, with 10 of 13 losses in the past five
years involving less than 200 pounds.  According to the study, transportation of anhydrous HF, using
existing road and rail equipment and current safety standards, has resulted  in minimal hazard to the
public and others.
                                                                                     Page 125

-------
                                      EXHIBIT 8-10
                            Analysis of Anhydrous HF Incidents
Mode of
Transportation
Rail
Truck
* As estimated by CMA
Number of
Incidents
Involving
HF
12
1

Number
of Trips
40,000*
20,000*

Pounds of
HF Shipped
(Thousands)
5,000,000*
650,000*

Incidents
per
10,000
Trips
3
0.5

Incidents
per 650 M
Ibs.
1.6
1.0

Source: Chemical Manufacturers Association, Transportation Safety of Anhydrous Hydrogen Fluoride, Hydrogen
Fluoride Panel, Washington, D.C., March 30, 1992. (103D)
Page 126

-------
                                        ENDNOTES
1.      Memorandum, Subject: OSC Report to the National Response Team Major Air Release of
       Hydrofluoric Acid Marathon Petroleum Company Texas City, Galveston County,  Texas - October
       30 to November 1, 1987, From: Robert M. Ryan, On-Scene Coordinator, U.S. Environmental
       Protection Agency Region IV, To: National Response Team, March 4, 1988. (370)

2.      Mason, R.J., Marathon Oil Company, comments on the draft Chemical Emergency
       Preparedness and Prevention Advisory on Hydrogen Fluoride, January 15, 1993. (82d)

3.      Memorandum, Subject: OSC Report to the National Response Team.

4.      Memorandum, Subject: OSC Report.

5.      Chemical Engineering, "HP's Future is Up in the Air," May 1990, p 39. (150)

6.      Memorandum, Subject: OSC Report.

7.      Mobil Torrance Refinery, Report on Cause and Origin of November 24, 1987 Fire and
       Explosion, Torrance, California, December 18, 1987. (181b)

8.      Hazardous Materials Intelligence Report, "Leak of Toxic Gas at OK Uranium Processing Plant,"
       January 10, 1986, p 1.  (520)

9.      U.S. Nuclear Regulatory Commission, Assessment of the Public Health impact from the
       Accidental Release of UF6 at the Sequoyah Fuels Corporation Facility at Gore, Oklahoma,
       Washington, D.C., March 1986,  Document Number NUREG-2289-VI. (491)

 10.    U.S. Environmental Protection Agency, U.S. EPA Release Prevention Questionnaires,
       1987-1988. (490)

 11.    U.S. Environmental Protection Agency, ARIP Report - Southwest Refining, Corpus Christi, TX,
       1991. (Not logged into the ARIP database).

 12.    Letter, Concern over Segregation of Data Bases for Aqueous and Anhydrous HF, From: William
       J. Hague, Allied-Signal, To: Craig Matthiessen, U.S. Environmental Protection Agency, August
       18, 1992. (293.3)

 13.    U.S. Environmental Protection Agency, U.S. EPA Release Prevention Questionnaires.

 14.    National Transportation Safety Board, Anhydrous Hydrogen Fluoride Release from NATX 9408,
       Train No. BNELSYat Conrail's Receiving Yard Elkhart, Indiana, February 4,  1985, Hazardous
       Materials Accident Report, Washington, D.C., November 27, 1985, Document Number:
       NTSB/HZM-85-03. (20)

 15.    AHE Database, U.S. Environmental Protection Agency, 1960-1987. (05)

 16.    Chemical Manufacturers Association, Transportation Safety of Anhydrous Hydrogen Fluoride,
       Hydrogen Fluoride Panel, Washington, D.C., March 30, 1992. (103D)

 17.    Chemical Manufacturers Association, Transportation Safety of Anhydrous Hydrogen Fluoride.
                                                                                   Page 127

-------
  18.     Strickland, Gordon D., Chemical Manufacturers Association, comments from technical review
         of Hydrogen Fluoride Study, Report to Congress,  Draft May 8, 1992, June 5, 1992. (466.9)

  19     Chemical Manufacturers Association, Transportation Safety of Anhydrous  Hydrogen Fluoride.

  20.     Strickland, Gordon D., Chemical Manufacturers Association.
Page 128

-------
               9.  MODELING HYDROGEN FLUORIDE RELEASES
       The statutory language requiring the HF study states that EPA should examine the potential
hazards of HF to the public, considering a range of events, including worst-case accidental releases.
In response to the Congressional mandate, EPA has carried out consequence analysis, using
computer modeling techniques, for a range of scenarios for accidental releases of HF. This chapter
briefly describes the elements of a consequence analysis, the purpose of such an analysis, and
presents background on the types and characteristics of models used to analyze the consequences
of HF releases.  This is followed by a discussion of the selection of models by EPA and the
development of worst-case accident scenarios used in the HF study. The chapter concludes with a
statement of the limitations of the modeling results, presentation and analysis of the results, and
evaluation of the sensitivity of the modeling results to certain input parameters.


9.1     Consequence Analysis

       Consequence analysis is used to estimate the potential health impact or damage to property
and environment from a release  of a hazardous substance. It usually  involves a determination of the
amount of substance released, the rate at which it enters the air, and dispersion of the airborne
material downwind under particular meteorologic conditions to a certain exposure level, damage level,
or dose (concentration multiplied by duration of exposure).

       Consequence analysis can be used:

       ».       to compare impacts of results from different release scenarios with
               each other to assess the significance of various release scenarios;

        »-       to illustrate how  model variations or input uncertainties influence the
               results; and

        ».       to demonstrate how site-specific parameters can significantly alter the
               results of such analyses.

        Consequence analysis is not designed to determine real-time impacts associated with an
event; the greatest value of consequence analysis is in determining the potential impacts of  a range of
conditions and scenarios for planning purposes and to examine ways to minimize those  impacts.
However, caution must be used when reviewing consequence analysis results.  Consequence analysis
ignores the probability of a release occurring.  If used as the sole basis for an operational decision,
consequence analysis may lead to actions which could actually increase the overall risk  at a facility.

        Consequence analysis for a specific facility would likely include the population potentially
affected by a release. The number and location of people around the facility is an important
consideration for planning.  Appendix XV presents the number of people estimated to live within
circles around selected HF facilities within one-mile and five-mile radii. In the event of a release of HF,
it is likely that only a fraction of the people within these circles would be exposed. The population
potentially exposed to a release would be determined by factors such as the weather conditions (i.e.,
ambient temperature, atmospheric stability, wind direction, and wind speed)  and the accident scenario
 (i.e., the rate of release, type of release, conditions of release, and the plume width).
                                                                                     Page 129

-------
         One common method to estimate the potential consequences of a release is to model the
  dispersion of the release.  Models are valuable tools for analyzing and estimating the behavior and
  effects of chemicals accidentally released into the atmosphere. Most dispersion models generate data
  on contaminant concentration at downwind distances. Modeling information has been used by
  industry and government to determine the possible consequences of a release and to plan for
  emergency response and release notification. Also, modeling has been used to assess the relative
  effectiveness of various accident prevention, hazard reduction, and mitigation techniques on the
  consequences  of the release.


  9.2     Models for HF Releases

         Many models are available for predicting the dispersion of chemical releases.  However, only a
  few are able to  account for many of the particular properties and release characteristics of HF.
  Current and ongoing research has focused on developing models for HF and conducting HF spill
  tests both to validate the models and to increase knowledge about HF release characteristics. This
  section  discusses HF field tests and describes the models that may be suitable for estimating HF
  dispersion.

         9.2.1  Field Tests

         During the summer of 1986, Amoco Oil Company and Lawrence Livermore National Laboratory
 conducted a series of six experiments involving atmospheric releases of HF.  The studies, known as
 the Goldfish test series, were conducted at the Department of Energy Liquefied Gaseous Fuels Spill
 Test Facility in Frenchman's Flats, Nevada.  The purpose of the experiments was to examine source
 characteristics, dispersion properties, and water spray mitigation techniques. The test results
 indicated that approximately 20 percent of the liquid released flashed to vapor and the remaining 80
 percent of HF was transported downwind, along with the HF vapor, as an HF/water vapor aerosol.
 The cold dense  gas was detected at substantial distances downwind with no drop-out of HF.  It was
 also determined that to estimate HF dispersion, thermodynamics and interactions with atmospheric
 water vapor must be considered. In addition, the tests showed that water sprays were effective in
 reducing HF in the vapor cloud.1'2'3-4

        To formalize the effort to continue HF field tests, the participants in the Goldfish tests formed
 the Industry Cooperative HF Mitigation/Assessment Program (ICHMAP), which then sponsored a
 series of tests of the effectiveness of water sprays on mitigating anhydrous HF releases. The tests,
 sponsored and funded by twenty companies from the chemical and petroleum industries, were known
 as the Hawk HF Test Series. These studies of water spray mitigation showed that water to HF ratios
 of 40:1 could result in 90 percent reduction in the amount of airborne HF.5

        9.2.2  Model Characteristics for Estimating HF  Dispersion

        Computer modeling systems are often used to model accidental releases of hazardous
 materials and estimate the effects of such releases.  In order to deal realistically with a release of HF,
 a computer model should be able to take into account the unusual properties of HF and other factors
 including the following:6

       *•     HF Thermodynamics.  When HF is released into the atmosphere from a
              pressurized vessel at a temperature above its boiling point, it first flashes to
              vapor, entraining liquid droplets of HF as aerosol. The HF cloud then is
              subject to various thermodynamic effects which sometimes produce opposing
              results. The HF polymerizes, resulting in slumping and gravity spreading of
              the cloud.  As air mixes with the cloud, the HF dissociates in  a process that
Page 130

-------
absorbs heat and cools the cloud.  The evaporating aerosol also cools the
cloud.  The cooling of the cloud increases its density. As moist air is mixed
into the cloud, HF reacts with the moisture to release heat, warming the cloud
and decreasing its density.  As the HF aerosol absorbs water from the air, its
volatility decreases and it consequently travels further as an aerosol. Thus,
the final density of the cloud is dependent on the aerosol fraction, the humidity
of the air and the rate of mixing of the air with the cloud of HF.  The model
must consider all three of these factors.

Dense Gas Dispersion. When a dense gas such as HF disperses, it
undergoes gravity spreading as a result of the difference in density between
the cloud and the ambient air.  At the same time, ambient air is entrained into
the cloud.  The model must consider the amount of gravity spreading and the
rate of entrainment of air; it must also consider the effect of surface roughness
on these processes (e.g., the effect of the turbulent conditions encountered as
a result of flow over surface structures at an industrial site and the terrain
surrounding the facility).

Release Duration and Averaging Time. Averaging time  is the time over which the
concentration of the contaminant is averaged.  Averaging time is usually chosen to
correspond to an exposure time for which an individual would suffer a certain health
impact.  For releases of short duration, under windy conditions, the cloud would likely
pass by quickly and not expose an individual for the full averaging time. To avoid
situations where there is no exposure to an HF cloud during  much of the averaging
time, the preferred averaging time (and therefore exposure time) should be less than
the release duration.

 Jet Releases. When a gas flows out of a hole  at a high velocity (e.g., when a tank or
 pipe containing compressed or liquefied gas is punctured or broken), a gas jet is
formed. Jet behavior affects the degree of cloud dispersion through entrainment of air
 and subsequent dilution.  Modeling of HF gas  jets, therefore, must account for this
 type of dispersion.

 Effects of Mitigation Devices. Models should have near-field modeling capabilities to
 assess of the effects of mitigation systems in place or planned. For example, to model
 water mitigation systems, the cloud properties at the downwind distance where the
 water mitigation system would  be located should be known.  The cloud properties at
 that distance are then used as the baseline for modeling the water mitigation system.

  Effects of Surface Roughness.  There is controversy among  modelers regarding the
  influence of surface roughness on the dispersion and gravity spreading of a dense
  gas cloud. Surface roughness is a measure of the irregularity of the terrain over which
  a cloud passes. Irregularities include mountains, trees, and facility structures (process
  vessels).  Flat, rural areas are characterized by a lack of irregularities.  The size of the
  surface feature relates to the size of a surface roughness factor (e.g., large
  irregularities are assigned large factors). The phenomenon  of neutrally buoyant
  clouds passing over surface irregularities increases turbulence and consequently
  increases dispersion. Alternatively, dense gas clouds  like HF hug the ground and may
  not pass over these surface features but channel around them, potentially leading to
  less dispersion. As a general  "rule of thumb" for dense gases, surface roughness
  factors that represent a large fraction of or are equal to the cloud height should not be
  used. Based on limited wind tunnel tests, some analysts assert that surface structures
  do create mixing and help disperse a cloud; others contend that the wind tunnel test
                                                                         Page 131

-------
                 results are not conclusive and that only very low surface roughness factors should be
                 used. Additional research is necessary to resolve this issue/'8'9'10

          9.2.3 Available Models

         There are several computer modeling systems in the public domain that can be used to model
   an accidental release of HF. Most are not specific only to HF releases. Exhibit 9-1 presents some
   dense gas models that are used to model HF releases. The ways these models incorporate the
   factors necessary for modeling HF releases (discussed above) are briefly noted in the exhibit.

   .K .u  ^ !,8pt f°' HGSYSTEM- the dense gas models discussed above do not take into consideration
   the thermodynamic properties of HF.  Because of the modeling difficulties associated with HF a
   subcommittee of the ICHMAP developed HGSYSTEM, designed to simulate the consequences of an
   accidental release of HF and assess the effectiveness of mitigation systems 11  HGSYSTEM is
   composed of a group of models which can be used in various combinations to describe the type of
   SflSl&t; .vaP°r Jet, evaporating pool) and the area of dispersion (i.e., near-source, far-field)
   HGSYSTEM includes two spill models, one (HFSPILL) used to compute the amount of HF released
  £"! PS=S|m!? d StCLra9e' and the  °ther (EVAP) used to estimate HF emissions from an evaporating
  pool,  HFSPILL can be used with the HFPLUME model to predict the jet behavior and near-source
  dilution from a pressurized release.

         HGSYSTEM also includes two models for prediction of downwind dispersion. HEGADAS-5 is
  a version o 'the HEGADAS model for simulating heavy gas dispersion from area sources modified to
  oeai with HF.  HEGADAS-5 includes heat and water vapor effects and allowance for HF
  thermodynamics, as well as revisions to the treatment of gravity spreading of dense clouds, treatment
  ?r.DMlS8J?19e^     °e rou9hness conditions, and improved treatment of along-wind dispersion
  PGPLUME  is a Gauss.an model for treatment of HF plumes that do not exhibit dense gas behavior'
  (e.g., if the plume is released vertically or the release is elevated). PGPLUME can be used for such a
  Lei6Da1?Ma=er     , dilUted sufficient|y so that the thermodynamic and jet entrainment calculations in
  HFPLUME are no longer required.  HGSYSTEM also contains models that can be applied to anv
  heavy gas. A separate model from HGSYSTEM, HFSPRAY, describes the mitigation of HF releases
  witn water sprays.

        9.2.4  Comparison of Modeling  Results with Spill and Wind Tunnel Tests
 nf mnHDn-r,              °f ^^^ involvin9 modeling of HF and comparison of the results
 of modeling with field and wind tunnel test results. Some of these studies are described in this
 section. The studies frequently have contrasting conclusions about verification of certain models
 This attests to the different assumptions and methods used in running the models and the
 uncertainties in the  algorithms used to describe HF dispersion.
 SLAR                 a"d H<;SYSTEM Comparison with Ro.Hfigh Toeto  Tne evaluation of the
 SLAB and pEGADIS dense gas dispersion models using data obtained from HF spill tests is
 presented in several reports.12-13'" In a 1987 report, several simplifications were made in order

 £ h!S          • t0^6 G°ldfiSh Spi" tSSt data>  Because neither SLAB nor DEGADIS was designed
 L ^    a pressunf ,d jet release 
-------
                                                                        Exhibit 9-1
                                           Examples of Models Available for Simulation of Accidental Releases of HF
                                                                          Unite Duration,
                                                                          Averaging Time
       HF
Thermodynamics
                                                Gravitv Spreading
                                                                                                                               Does not simulate
                                                                                                                               effects
                                                                               Vertical and
                                                                               horizontal jets
                                                    Appropriate treatment
Treatment may not be
applicable to large
surface roughness
Assumes ratio of vapor
flashed and liquid;
computes resulting
temperature
                                                                                                                               Does not simulate
                                                                                                                               effects
Must develop concentration  Treatment may not be
vs. density relationship      applicable to large
outside of model           surface roughness
                                                     Not appropriate for
                                                     long travel times
                                                                                                                               Does not simulate effects but
                                                                                                                               can start model at any
                                                                                                                               downwind distance with
                                                                                                                               estimated cloud properties
                                                                               Jet release at
                                                                               all angles
                                                     Appropriate treatment
Treatment applicable to
industrial settings
Effects computed at each
downwind distance
                                                                                                                               Does not simulate
                                                                                                                               effects
                                                                                Direction of jets
                                                                                unspecified
                                                     Assumes unimpeded
                                                     flow of material
Treatment does not
incorporate large
obstacles such as
buildings and plant
structure
Does not incorporate
thermodynamics
Additional Notes on Models:

         SLAB is a dense gas model developed by the Lawrence Livermore National Laboratory. It can be used to model continuous, finfe duration and instantaneous releases.













 model which has been replaced for the UK Health and Safety Executive.

         Sources:          Memo from D.N. Blewttt to R.C. Wade, April 1,1991.
                           User Guide for BP CIRRUS.                                                         .       „
                           Koopman, Ronald P., Lawrence Livermore National Laboratory, Comments on Draft HF Study (Doc 289).
                           Fryman, Chuck, BP Oil, Comments on Draft HF Study (Doc 139.63).

-------
         SLAB is a steady state model that was modified to apply to the short duration of the Goldfish
  spill tests. The modified SLAB model gave good agreement with experimental results for the first
  Goldfish test but underpredicted HF concentrations based on the second and third Goldfish tests
  Other modifications were made to the model that improved agreement with the experimental data'
  including addition of an aerosol evaporation model, modifications to the entrainment and velocity
  equations, and changes to the assumed vertical concentration profile. For example, the addition of a
  simple aerosol evaporation model to the SLAB model allowed the vapor and droplets to be treated as
  a single fluid with a modified molecular weight to account for the HF/water aerosol, thus increasing the
  density above that of the pure vapor. Evaporation  of the aerosol was assumed to occur linearly with
  downwind distance, resulting  in a steady decrease in molecular weight as the aerosol
  evaporated.16-17

         In a 1988 report, the steady state model DEGADIS predicted maximum HF concentrations
  within a factor of two compared to the Goldfish spill test results. In some cases DEGADIS
  underpredicted and in other cases overpredicted maximum concentrations, with the ratio of DEGADIS-
  predicted to observed maximum concentration ranging from 0.6 to 1.3.18 The 1987 report cited
  above, found the  model results difficult to interpret because the length of time for which the average
  concentration was calculated varied, depending on downwind distance.19  It was also difficult to
  generalize about the model's tendency to overpredict or underpredict at various distances because
  the test data were not consistent  in this regard.

  * ^ ^  A 1992 study compared the results of DEGADIS, SLAB, HGSYSTEM, and other models to field
  test data from five sites,  including the Goldfish tests.20  DEGADIS and HGSYSTEM were within a
  factor of two of measured concentrations from two Goldfish tests. This study further found that SLAB
  tends to underpredict the Goldfish test data within a factor of five.21

    22  HGSYSTEM results were compared with the results of the first three Goldfish spill tests of
  HF.    HFPLUME and HEGADAS were used for modeling.  The  modeling results for plume centerline
  concentrations compared favorably with observed results, considering standard deviations and the
  uncertainties in modeling a limited number of experimental releases.  In addition HGSYSTEM
  predictions of cloud width agreed well with observed data.

        HGSYSTEM Results Compared to Wind Tunnel Simulation Results  The dense gas
 dispersion modeling program contained in HGSYSTEM (discussed  above) was used to model an HF
 release from a petroleum refinery.  The results of modeling were compared to the results obtained
 from wind tunnel experiments on the dispersion of simulated HF releases in a scaled-down model of a
 petroleum refinery. These experiments used an inert gas mixture having the same density relative to
 air as calculated for a cloud of HF in air at its coldest temperature.  Concentrations of the simulant
 gas, which was released  from pipes near the center of the wind tunnel model,  were measured as a
 time series; the times for  each concentration measurement in the model were then scaled to simulate
 the equivalent times at a full scale facility.  Several release scenarios were used.  The same scenarios
 were used with HGSYSTEM, and roughness  parameters derived from the wind tunnel scale model
 refinery and terrain were also used in HGSYSTEM.  Concentrations predicted by the model were
 compared to the wind tunnel concentration results.  The model predictions of downwind and
 crosswmd concentrations were found to agree fairly well with the wind tunnel results.23

        Water Spray Mitigation Modeling. The model HFSPRAY simulates the momentum, mass and
 energy reactions between a water spray and a turbulent plume of HF in air. HFSPRAY can be a tool
 to quantify the effectiveness of water sprays  and may be used to evaluate specific installations
 release scenarios, and weather conditions.24 The model comprises two sets of equations one to
 describe the gas phase and the other to describe the liquid drop phase.  The model is capable of
 predicting flow velocities, temperature, water vapor and HF concentrations in two dimensions  for
Page 134

-------
spraying in any direction. The Hawk water spray mitigation tests, a series of field tests using wind
tunnels conducted in Nevada in 1988, have been simulated using the HFSPRAY model. Model
predictions agreed well with the wind tunnel test results.25 „ Recent studies show that the latest
version of HFSPRAY has greater ability to describe mass, momentum, and energy transfer between
water sprays and unconfined releases of HF, and that HFSPRAY correctly reproduces the flow fields
induced in wind tunnel modeling of actual industrial systems.26'27'28


9.3    Modeling Used in HF Study

       EPA selected and used  models to evaluate the consequences of HF releases.  The approach
and considerations for the HF modeling are discussed below.

       9.3.1  Basis for Selecting Models

       For this study, several modeling systems including the models listed in Exhibit 9-1  (BP
CIRRUS, DEGADIS, HGSYSTEM, and SLAB) were considered for estimating the dispersion of HF upon
release. In addition, EPA also considered the Areal Locations of Hazardous Atmospheres (ALOHA)
model, the air dispersion modeling component of the Computer Assisted Management of Emergency
Operations (CAMEO™) system.
       ALOHA was eliminated from consideration because it contains a simplified version of
DEGADIS and it does not sufficiently account for transient releases.  Also, ALOHA is designed
primarily as a screening model for planning emergency response.29  BP CIRRUS was eliminated
because HGSYSTEM contains the same dispersion model (HEGADAS), modified to specifically
address HF releases. DEGADIS was briefly evaluated along with HGSYSTEM and SLAB. For releases
longer than at least seven minutes, it was discovered that in the steady state mode, DEGADIS results
will closely resemble the results from HGSYSTEM and SLAB. However, for relatively short duration
releases of about one minute, DEGADIS in the steady state mode appeared to overpredict the HF
concentrations and in the transient state mode appeared to underpredict the HF concentrations,
compared with HGSYSTEM and SLAB.  Consequently, DEGADIS modeling was not pursued further.

       The HGSYSTEM dense gas dispersion model, HEGADAS, accounts for many of the critical
physical/chemical processes that are considered important for HF dispersion.  Some critical and
unique features of HF dispersion include aerosol formation, polymerization, and hydrolysis. SLAB
simulates atmospheric dispersion of denser-than-air releases including a ground-level evaporating
pool, an elevated horizontal jet, vertical  jet or stack release, or instantaneous or short-duration
evaporating pool release. Both HGSYSTEM and SLAB can handle complex dispersion concepts such
as aerosolization, transient releases, surface roughness, gravity spreading, and entrainment.
Therefore, HGSYSTEM and SLAB were  selected to estimate consequences of accidental releases of
HF for this study. These models are discussed in more detail in Appendix XVI. The purpose of using
two models was not to validate the models but to calculate potentially affected distances at certain
dose levels using different models and comparing the results. The range of results demonstrates that
even the complex dense gas algorithms in these models still have uncertainties in predicting HF
dispersion. The range of numbers de-emphasizes the certainty or importance of individual output
values while allowing comparison and evaluation of input release scenarios.

        9.3.2  Considerations for Modeling

        Modeling and source term  assumptions and calculation procedures depend on both pre-
release storage conditions and meteorologic conditions.  Based on these conditions, HF may generate
a liquid pool or a positively, neutrally, or negatively buoyant cloud upon release. The chart in Exhibit
                                                                                   Page 735

-------
 9-2 Indicates some possible release mechanisms and considerations that should be applied when
 determining a model to use to assess impacts.  This chart could be applied to any hazardous
 chemical release.30

        Significant research has been conducted on the behavior of substances when they are
 released to air, including HF (see Section 9.2.1 on field tests on HF). Prediction of the behavior of a
 material released under accident conditions is extremely complex and difficult because of the many
 variables that influence the results. For example, the amount released and the rate at which a
 chemical enters the  air are not precisely known  and are usually estimated based on the storage and
 meteorologic conditions at the time of the event. Winds can shift and meander,  causing unpredictable
 movement of a cloud, and obstacles along the path can channel a dense gas cloud or increase
 turbulence, affecting the degree of dispersion.

        Model verification  or validation is an issue in determining the credibility of these models.
 Although not performed in this study, model verification has been conducted using spill test data (i.e.,
 Goldfish test) and wind tunnel data. Neither HGSYSTEM nor SLAB has been extensively tested for
 verification under different field conditions of humidity, release slope, surface roughness, or channel
 flow. However, modeling, when applied correctly, can provide valuable  insight regarding potential
 exposure to a nearby community from various release scenarios.

        Models are complex tools that can be difficult to use correctly. Moreover, the model
 predictions of travel  distances of a release are only estimates, regardless oHhe model used.
 Therefore, no single  number or result from a specific model can be asserted to be the correct answer
 or representative of other release conditions.31

        9.3.3 Modeling Inputs

        HGSYSTEM  and SLAB require the user to enter a variety of data about storage, release, and
 ambient conditions.  Consistent with the Congressional mandate, model input  data were chosen to
 represent worst-case conditions and produce  a large HF release.  The modeling  inputs were compiled
 from diverse sources including local and state governments, facilities that use  HF, accident data,
 process design data, and  other modeling efforts.32'33'34'35'36'37'38  Some characteristics
 of HF that are important for modeling are discussed in Section 9.2.2.  Many of the inputs are based on
 the configurations  and conditions of HF process operation and storage vessels at actual facilities. It
 must be pointed out  that these data may not represent configurations and conditions for HF facilities
 as a whole. For example,  the operating conditions and flows of a Phillips licensed HF alkylation unit
 may not be the same as the operating conditions and flows of a UOP licensed HF alkylation unit.

        The general guidelines for choosing certain inputs are discussed below.  The specific data
 used in modeling depend on the scenario modeled. Also, the modeling inputs were selected to
 produce the worst-case release scenario.  The scenarios are discussed in  detail in Section 9.5.  For a
 complete listing of specific model inputs for HGSYSTEM and SLAB, see Appendix XVII.

        »•     Temperature. Storage temperatures reported at HF facilities were reviewed and from
              these, a temperature was chosen which would lead to a rapid release rate.  HF boils
              at 19°C (67°F), which is normal room temperature. HF stored above its boiling point
              would flash on release, while below its boiling point, it would likely form a pool.
              Consequently, typical storage and ambient temperatures above the boiling point of HF
              were  selected.
Page 136

-------
                                                Exhibit 9-2
                                       Release Mechanism Flowchart
                                             RELEASE MECHANISM
DOES VAPOR FORM
WHEN DEPRESSURIZED
FROM STORAGE
CONDITIONS TO
AMBIENT CONDITIONS?
                                                 2-PHASE
                                                                          GASES
                        NO
                                      i
                        LIQUID POOL EVAPORATION MODEL
                        - CONFINED / UNCONFINED
                        - BOILING / NON-BOILING
                        = WATER / EARTH /CONCRETE/ ETC
                        -MULTI COMPONENTS
    JET
ENTRAPMENT
   MODEL
IS THE RELEASE
INSTANTANEOUS?
ADIABATIC
EXPANSION
  MODEL
                                                        NO
v^NEUTHALLY BUOYANTT^^
1 YES
GAUSSIAN DISPERSION
MODEL


DENSE GAS
DISPERSION MODEL
                                                                                                          Page 137

-------
                Pressure. The pressure selected for modeling was based on the highest pressure for
                the particular process or storage operation described in the scenario.  The highest
                pressure would result in the largest release of HF to air.

                Hole Size.  One to two inch diameter holes were used in modeling based on common
                pipe diameters in use (release due to corrosion, embrittlement, fatigue, accidental
                shearing, etc.). For modeling catastrophic vessel failure, the holes were either based
                on the puncture hole created by a train rail in the case of a derailment or on the hole
                size necessary to empty a vessel in approximately 10 minutes.

                Release Duration.  Duration of the release was based on the time for the maximum
                quantity (i.e., the entire contents) in a vessel to be released from the hole given the
                capacity of the vessel as well as the location and diameter of the hole. The maximum
                quantities in the process  and storage vessels were determined from data provided by
                SCAQMD on actual facilities, information obtained during facility visits, and on studies
                conducted by the HF facilities themselves. When the release point is not directly
                associated with a large storage volume of HF,  (e.g., pump seal failure), a default value
                of 20 minutes was used as the maximum time  estimated to empty or shut down the
                system/equipment and stop the flow.

                Prevention and Mitigation Systems.  Water sprays and deluge systems were
                assumed to reduce an HF release by the amount observed in field tests.39
                Emergency de-inventory and automatic shutoff times were assumed to reduce the
                duration of the HF release. The time to stop the release through mitigation is based
                on facility estimates obtained during site visits.

                Release Height.  Release height is an estimate based on facility visits and the general
                dimensions of HF process and storage vessels.

                Relative Humidity. Relative humidity of 50 percent was chosen to incorporate the
                effects of HF reacting with water in the atmosphere in the modeling results.

               Surface Roughness.  Surface roughness is an estimate of the effect of surface terrain
               and the presence of buildings or other man-made structures on the movement and
               dispersion of a vapor cloud. A surface roughness of 0.03 meters, simulating a rural
               area, was chosen because a small surface roughness value is generally expected to
               yield larger HF doses downwind.  Section 9.2.2 discusses the controversy surrounding
               use of large surface roughness values that simulate industrial settings.

               Meteorological Conditions.  Except where noted, two sets of meteorological
               conditions were assumed for modeling.  The conservative case, which would result in
               less dispersion, includes low wind speed (3.4 miles per hour or  1.5 meters per
               second), stable atmospheric stability conditions (F stability), and rural surface
               roughness conditions.  An inversion, a specialized nighttime condition that prohibits
               the vertical dispersion of surface releases, was  not considered in this study. A more
               likely case involves wind speed of about 12 miles per hour (5.0 meters per second),
               neutral atmospheric conditions (D stability), and rural surface roughness conditions.
               Atmospheric stability is a measure of the turbulence and  mixing in the atmosphere
               near the ground. The most turbulent conditions (A stability) occur on sunny days with
               wind speeds less than 3 meters per second.  The most stable conditions (F stability)
               occur in calm, pre-dawn hours.
Page 138

-------
       A sensitivity analysis was performed on several input parameters (i.e., wind speed and
stability, emission rate, relative humidity, and surface roughness) to determine the relative influence
that changes in the input parameters have on modeling results.  The sensitivity analysis is described
in Section 9.8.

       9.3.4 Exposure Guidelines for Modeling

       The primary hazard to the public associated with HF is inhalation of vapor or fumes, which can
damage mucous membranes and lungs and possible cause death. Dispersion modeling can be used
to estimate the concentration of HF in a vapor cloud at certain distances.  However, to determine the
potential health effects associated with inhalation of a particular concentration of HF, the amount of
time an individual is actually exposed to the HF must also be known.  The exposure concentration for
a specific time period is called a dose. There are several accepted inhalation exposure concentration
guidelines, as described in Chapter 2, that can be used as a dose threshold for modeling inhalation
exposure. These include the Immediately Dangerous to Life and Health (IDLH) level, based on a 30
minute exposure, and the Emergency Response Planning Guideline level 3 (ERPG-3), based on a 60
minute exposure. .

       Accidental releases are often of short duration. It may not be appropriate to model a short
duration accidental release using exposure concentration guidelines established for long exposure
periods.  In short duration releases (e.g., one minute) under windy conditions, the HF cloud would
likely pass by quickly and not expose an individual for the 60 minutes required to  reach the ERPG-3.
Using ERPG-3 in this case would not be appropriate if the dose-response  relationship of the chemical
is non-linear (i.e., the effect of exposure to a 30 ppm concentration for 30 minutes is not equivalent to
the effect of exposure to 900 ppm for one minute).  However, the dose-response relationship for HF
has been described both as non-linear40'41 and as linear.42'43 Consequently, the  exposure
time associated with the dose should be as close to the duration of the release as possible.
Unfortunately, for short duration releases, few exposure guidelines with short  exposure periods appear
to be available in the literature or widely accepted.  For example, the United Kingdom's Health and
Safety Executive has proposed a "Dangerous Toxic Load" of 12,000 ppm HF for one minute; this level,
however, has not yet been published officially/'
44
        EPA's modeling efforts used the IDLH as the main dose threshold.  This exposure level would
represent conditions associated with a potentially life-threatening  event. Specifically, the modeling
was carried out based on the scenarios discussed in the next section, to determine the greatest
distance to the IDLH level, as an indication of the area in which people might be exposed to a life-
threatening  dose.  In select scenarios, distances were calculated  to the ERPG-3 level.  Additional
information on equations used to relate concentration, duration of exposure, and potential fatalities
may be found in Appendix V.

        The HGSYSTEM and SLAB models have the capability to determine average concentration
over a defined exposure time.  In estimating dose, both models assume a linear relationship between
effective dose and exposure time. This means that the IDLH is reached when all of the HF
concentrations averaged over a 30 minute period equal 30 ppm.


9.4     Worst-Case Accident Scenarios

        The statutory language for the HF study requires that EPA examine the potential hazards of
HF to the public considering a range of events including worst-case accidental releases. This section
describes one approach for developing scenarios of worst-case HF releases.
                                                                                     Page 139

-------
         9.4.1 Possible Definitions of Worst-Case Accident Scenarios

         There is no single definition for a worst-case accident scenario. Facilities, government
 agencies, and the public  may each have different interpretations. Sonie have indicated that worst-
 case is represented by the high consequence event of a "total loss of containment" of the largest
 vessel or container on-s'rte.  However, others argue that instantaneous vessel failure is highly unlikely
 and therefore other more probable events are worst-case events even  if the consequences are not as
 large as the catastrophic  vessel failure.  There is a debate over the importance of probability and
 consequence in determining the worst-case scenarios. Therefore, different groups (e.g., industry and
 the public) may consider  different accident scenarios to represent the worst case.

         For regulatory programs, government agencies sometimes need to define worst-case
 scenarios.  The Texas Air Control Board (TACB) considers the worst-case  release as a release that
 causes the maximum impact or consequence, including maximum area or distance to the chemical's
 IDLH.    Alternatively, the  South Coast Air Quality Management District in California has used serious
 near worst-case conditions which it considers to be  less likely to occur and of greater magnitude than
 those  release scenarios considered in California's Risk Management and Prevention Programs to
 determine risk of HF facilities in the Basin.46 EPA developed an approach based on "credible worst-
 case assumptions" in the Technical Guidance for Hazards Analysis (Green Book) to help local officials
 screen and identify zones around a facility for which the community is potentially vulnerable from a
 chemical release.  For gases, the total quantity of a vessel is assumed  to be released to air over a
 period of 10 minutes.  For liquids, the total quantity of a vessel is assumed to be spilled
 instantaneously, spreads out into a pool one centimeter deep, and evaporates.  These scenarios,
 however, are not sufficiently specific or complex for defining worst-case scenarios for HF.47 The
 Green Book calculation only provides a generic screening tool for prioritizing emergency planning and
 was not intended to account for the complex release scenarios particular to different industries or
 operations involving HF.

        9.4.2 Worst-Case Accident  Scenarios based on Congressional  Mandate

        Because the Congressional mandate emphasized hazards posed to the public, the worst-case
 accident scenario should reflect the most severe consequences from a release.  Releases in this
 scenario category would include catastrophic vessel failures under worst-case conditions of wind and
 atmospheric stability and topography. However, Congress also required consideration of a range of
 events, not just the events involving the largest quantity of HF. The range  of events can reflect the
 range  of operations and equipment that involve HF (e.g., unloading, pumping, reacting, purifying).
 Within each of these operations, the release scenario that results in the worst-case consequence can
 be developed. For example,  a worst-case scenario for an unloading/loading operation would be the
 complete failure of the hose.  Scenarios other than catastrophic vessel failure  may also represent
 potential incidents with a greater likelihood of occurrence.

        For this study, the worst-case scenarios were developed to reflect the conditions under which
 the public might be exposed to the greatest hazard from an HF release. These worst-case scenarios
 will be categorized as:

        *•     those scenarios causing worst-case consequences from catastrophic vessel failure
        *•     those scenarios causing worst-case consequences from a range of other events or
              situations in which HF is produced, used, stored, or transferred.

 For the purposes of this study, the worst-case consequences of a release will  be defined as the
furthest distance at which there could be potentially lethal effects.
Page 140

-------
       9.4.3 Basis for Selecting Scenarios

       Scenario development is process, situation, and site specific. The scenarios for this report
include a range of worst-case conditions that may not be representative of any specific site.
Consequently, these scenarios should be viewed as generic worst-case situations involving the
release of HF. The inputs and release conditions can always be refined; however, the scenarios used
to model consequences in this report provide a conservative range of conditions for approximating
doses and potentially affected distances.

       EPA first developed a list of potential release scenarios based on actual accident events,
industry input concerning process areas considered most vulnerable to loss of containment, current
industry practices for release prevention and mitigation, site visits and analyses of processes, and risk
assessment/hazard evaluation information collected during the course of this study.  The quantity of
HF lost from containment and the resulting release to air are dependent on the mass flow of HF, the
process and storage conditions, and whether anhydrous or  aqueous HF is involved.  For this analysis,
all of the releases modeled involved leaks of liquid HF because liquid releases result in a higher mass
flow rate than vapor releases.48 The maximum quantity of HF released  was based on the capacity
of HF vessels at various types of facilities. Because anhydrous HF has a greater potential for vapor
cloud formation than aqueous HF, all of the releases modeled involve anhydrous HF except for one
large release of 70 percent HF.

       Worst-case scenarios are also dependent on environmental conditions such as stability of the
atmosphere, wind speed, humidity, roughness of the terrain, and the air and ground temperature.
Worst-case environmental conditions were chosen based on estimates  that would cause the greatest
downwind exposure. The catastrophic vessel failures were only modeled with D stability because the
models were not able to handle huge release rates together with the most conservative meteorological
conditions associated with F atmospheric stability.

       To address the worst-case scenario in the category of catastrophic vessel failure, a release of
a large HF bulk storage vessel is modeled (Exhibit 9-3). This type of bulk storage vessel is typical of
that at an HF production facility. Within this category of catastrophic vessel failure, a derailment of an
HF rail car is also modeled.  To compare the consequences of an anhydrous versus aqueous HF
release, the derailment scenario is  modeled first with anhydrous HF and then with 70 percent aqueous
HF.  For all of these catastrophic failures, release mitigation was not considered in the modeling
because mitigation is generally considered not feasible or available for  these releases. The
probabilities of a major vessel rupture and other such catastrophic accidents are generally very low.
This type of worst-case event (e.g., storage vessel splitting open, spilling the entire contents) is not
unique to facilities handling HF, but could be evaluated for any facility that handles hazardous
materials. Severe damage due to catastrophic vessel failure could occur with any of a number of
highly toxic chemicals that are handled in bulk quantities, such as chlorine or ammonia.

       To address the range of worst-case scenarios, several HF facility locations and situations were
examined. The range of other worst-case scenarios that were modeled involve a diversity of HF
operations, a variety of release conditions (e.g., high/low pressure) and a range of release rates
(Exhibit 9-4).  In addition, various scenarios were selected to reflect potential releases at different
types of industries that use or handle  HF (e.g., HF manufacturers, refineries, fluorocarbon producers).
Several types of release events can be grouped and addressed in one scenario (e.g., pump seal
failure, flange and gasket leaks, threaded connection or weld leaks) because of similar failure
mechanisms and because the release might be of similar magnitude. The probability of the releases
are difficult to determine other than to say that certain types of releases (e.g., pump seal failure, hose
leak) are  reported with some frequency in accident data.
                                                                                      Page 141

-------
                                                   Exhibit 9-3
                                    Catastrophic Vessel Failure Scenarios
Scenario Scenario Description Atmospheric Initial Flow Total Quantity
Number Stability Rate, kg/s to Air, kg
1
2
3
Vessel Rupture-Bulk Storage
Derailment
Derailment 70 percent HF
Empty vessel
Empty rail car
Empty rail car
D
D
D
1100*
77
2.5**
1 ,800,000
65,000 I
3,000
      average pool evaporation rate from first 1,200 seconds (20 minutes).

     "average pool evaporation rate from first 1,000 seconds; rate is the effective evaporation of anhydrous HF determined
     from the ratio of the partial pressures of anhydrous and 70 percent aqueous HF at 25°C.
                                                   Exhibit 9-4
                                          Range of Other Scenarios
       Scenario
       Number
         Scenario
                Atmospheric   Initial Flow
Description        Stability      Rate, kg/s
Total Quantity
  to Air, kg
                   Hose Failure
                             Empty tank truck
                                                                                     43
                   Hose Failure (mitigated with
                     automatic shutoff valves)
                             One minute
                             release
                                   43
                   Hose Failure
                                                Empty tank truck
                                                                  43
                   Hose Failure (mitigated with
                    automatic shutoff valves)
                             One minute
                             release
                                   43
                  Settler Leak-Bottom
                             Empty settler
                                                                                     50
                  Settler Leak-Bottom
                             Empty settler
                                                                                     50
          10
Settler Leak-Bottom
(mitigated with water sprays)
                                               90% flow
                                               reduction
                                   4.9
          11
Vessel Leak
                                               Nearly empty
                                               vessel
                                                                  12
          12
Vessel Leak (mitigated with
  emergency de-inventory)
                                               3 minute release
                                                                                     12
          13
Settler Leak-Inlet Pipe
                                               20 minute release
                                                                                    9.5
          14
Settler Leak-Inlet Pipe
                                               20 minute release
                                                                                    9.5
          15
Pump Seal Failure
                                               20 minute release
          16
Pump Seal Failure
                                               20 minute release
Page 142

-------
       The consequences from these other release scenarios can be reduced by implementing
different mitigative measures. For example, some facilities attempt to decrease the rate of the release
with water sprays and others attempt to reduce the duration of a release with emergency de-inventory
or remotely activated valves. To evaluate the effectiveness of these methods for reducing the HF dose
downwind, several scenarios were modeled to incorporate these mitigative measures.

       Consequences estimated for these releases depend not only on the conditions but the
number of people that might be exposed.  For example, a worst-case release probably would have
little impact on public health in a sparsely populated area, while a release in a densely populated
area, under conditions not considered worst-case,  might have a greater potential for harming the
public. A site-specific characterization of the population around a Facility would be necessary for
evaluating the potential impact of a release on the  public.  Appendix XV presents a sample
characterization of the populations around  selected facilities that manufacture or use anhydrous HF.

       It  is important to emphasize that the scenarios described here are meant to  illustrate possible
accidental release events and the role that industry practices (or lack thereof) might play  in quickly
detecting, mitigating, and responding to such events.  For each scenario, the furthest distance is
determined at which an HF dose based on IDLH is reached.  The actual risk of fatality or  other serious
health effects to populations surrounding facilities that handle HF, or any other hazardous substance,
depends on a site-specific analysis of the likelihood that a release will occur, the magnitude and
severity of a release, the site-specific meteorologic  conditions, and the level of exposure the public
might receive. The modeling analysis based on the scenarios presented does not assess or estimate
public risk but rather provides an indication of the severity of effects and potential doses that could
result from an accidental HF release.

       The scenarios used for the modeling are described in detail below.


9.5    Descriptions of Scenarios Used in HF Study

       Except when indicated, all of the scenarios are modeled assuming the released HF becomes
airborne.  This means that all of the HF flashes to vapor or part of it flashes and the  rest aerosolizes
so that no liquid pools or rains out after the release. Two scenarios, one involving an anhydrous
release from a vessel rupture and another involving a release of 70 percent aqueous HF from a
derailment! are assumed to form evaporating pools.

       9.5.1  Catastrophic Vessel Failure

       The following descriptions cover the scenarios in the category of catastrophic vessel failure.
These types of scenarios are extremely unlikely.  The releases from catastrophic vessel failures are
expected to last for many hours and have large impacts downwind. It is unrealistic to assume that
during the several hours that the plume travels, the wind speed and direction would  remain constant.
In fact, wind speed and wind direction would not be expected to remain constant beyond the time it
takes the plume to pass about 10 kilometers. The  models were not able to change wind speed or
wind direction during the course of one release.  Consequently, it was assumed that the modeling
results would not be accurate past 10 kilometers.  If during D atmospheric stability, a catastrophic
vessel failure resulted in an HF cloud that passed or came close to 10 kilometers, then the same
scenario was not modeled with F atmospheric stability.  The most conservative meteorological
conditions associated with F stability was not modeled because the HF plume would be expected to
go much further than 10 kilometers.

              Accident Scenario 1 - Catastrophic Failure of Bulk HF Storage Vessel
                                                                                    Page 143

-------
       In this scenario, which may only be applicable to the manufacturers of HF that maintain bulk
storage of HF, a 500,000 gallon HF storage sphere ruptures.  In 10 minutes, the entire contents spills
onto the ground and begins to evaporate.  The release duration is chosen based on the EPA Green
Book's description of a release with "credible worst-case assumptions." In reality, a catastrophic failure
of a bulk HF storage vessel would likely result in a fraction of the HF flashing, forming a vapor-aerosol
cloud, and the remainder forming a pool from which HF would evaporate.  The hole size for the
catastrophic failure of the HF sphere is based on the hole size that would release all of the HF in 10
minutes.  HF evaporates from the pool at an average rate of about 18,000 gallons per minute (1,100
kilograms per second) over a 20 minute period.  In such a catastrophic release, nothing is currently
available to stop the release to the air or to contain the spreading pool.  Because of the large size of
the spill, the release is modeled as an evaporating pool at D stability.

               Accident Scenario 2 - Derailment of Rail Car Containing Anhydrous HF

       A  rail car containing anhydrous HF derails in a rural area.  The rail car is punctured; HF spills
from a six inch diameter hole. The hole size was based on the puncturing of a tank car by a rail.
Because of the isolated location of the derailment,  no mitigation could be applied. The rail car
empties in 14 minutes at an initial rate of 1,250 gallons per minute (77 kilograms per second).  This
scenario represents a very large release under pressure and temperature close to ambient (pressure
22 psl and temperature 27°C). The release was modeled assuming D atmospheric  stability class
conditions.

       Although the accident databases did not specifically indicate  derailment as  an accident cause,
one documented rail accident involved  a failed weld on a pressurized tank car.  About 2,900 kilograms
of HF were released. In this worst-case scenario, a rail car was emptied, spilling about 65,000
kilograms.

               Accident Scenario 3 - Derailment of Rail Car Containing 70 Percent HF

       A  rail car carrying 70 percent HF derails in a rural area. Like  scenario 2, the rail car is derailed
and punctured, and HF spills from a six inch diameter hole.  Because of the isolated location of the
derailment, no mitigation could be applied. The rail car empties at an initial rate of  1,850 gallons per
minute (115 kilograms per second). The release lasts for about 20 minutes. A liquid pool forms at the
site and HF begins to evaporate at a rate characteristic of a 70 percent HF solution.

       This scenario is modeled as evaporation from a liquid pool. This scenario represents a very
large release under pressure and temperature close to ambient (pressure 22 psi and temperature
27°C). The release was modeled assuming D atmospheric stability class conditions.


       9.5.2  Range of Other Scenarios

       The following descriptions cover a range of other scenarios based on a diversity of HF
operations, a variety of release conditions (e.g., high/low pressure) and a range of release rates.

               Accident Scenario 4 - Transfer Hose Failure  (F Stability)

       This scenario applies to nearly all HF industry segments where HF is transferred to or from
transport or storage vessels (e.g., HF manufacture, petroleum refineries,  fluorocarbon manufacturers).
Many industry representatives consider the loading/unloading operation to be particularly hazardous
because large quantities of HF are moved from one vessel to another using hoses and temporary
connections.  This scenario also may be representative of a piping failure.
Page 144

-------
       Both loading and unloading operations via pumping and under nitrogen pressure were
considered; however, because loading/unloading under nitrogen requires higher pressures than
pumping, the release flow rate from this unloading/loading method would be greater.  The selected
pressure for modeling a release during loading/unloading was based on facility information provided
to the South Coast Air Quality Management District.49'50

       This scenario depicts the failure of the flexible hose used to transfer liquid anhydrous HF to or
from a tank truck. The  HF transfer rate is assumed to be about 680 gallons per minute (43 kilograms
per second) under nitrogen pressure.51  The hose suddenly splits open and liquid HF is released at
the same rate as the transfer.  (A hose failure would more likely begin with a small leak and if ignored,
it might expand catastrophically.) The HF liquid vaporizes and aerosolizes, forming a cloud that
begins to drift downwind.52 A liquid pool may also form. It is assumed that the release is not
discovered in time to stop the flow of HF, and the tank truck empties in just over seven minutes.  The
transfer hose failure was modeled assuming that all of the released HF becomes airborne.  The
release was modeled assuming F atmospheric stability class conditions.

               Accident Scenario 5 - Mitigated Hose Failure (F Stability)

        Some facilities that use HF  are equipped with automatic shutoff valves and shutdown switches
to mitigate the release from a hose failure.  This scenario assumes that in less than a minute, HF
sensors, tank truck drivers or  standby operators who monitor the HF transfer will observe the release
and shutdown transfer operations by activating emergency remote shutoff valves and/or turning  off
pumps. An area or plant alarm will be sounded, initiating response actions.   Assuming that the
leak is quickly identified and valves on either end of the  hose are quickly closed, approximately 680
gallons of HF are released.

        The transfer hose failure was modeled assuming a one minute release where all of the
released HF becomes airborne. The release rate, 680 gallons per minute (43 kilograms per second),
is the same as that in scenario 4.  This means that before the leak is isolated with automatic shutoff
valves,  all of the HF flashes to vapor or part of it flashes and the remainder aerosolizes so that no
liquid is remaining after one minute.  For processes involving HF, this scenario represents a release
under moderate pressure (95 psi).  The release was modeled assuming F atmospheric stability class
conditions.

               Accident Scenario 6 - Transfer Hose Failure (D Stability)

        This  accident is the same as scenario 4 except the release was modeled assuming D
 atmospheric stability class conditions.

        Many HF hose failures appear in the accident history, including a 90-kilogram release when a
 transfer pipe between two tanks corroded, and a 58-kilogram discharge when a line  split during
 pumping.  Such releases did not involve the complete hose failure as  do the worst-case scenarios
 modeled here, in which 18,000 kilograms (scenarios 4 and 6) of HF are released.

               Accident Scenario 7 - Mitigated Hose Failure  (D Stability)

        This accident is the same as scenario 5 except  the release was modeled assuming D
 atmospheric stability class conditions.
                                                                                      Page 745

-------
                Accident Scenario 8 - Settler Leak-bottom (F Stability)

        This scenario could occur at a petroleum refinery with an HF alkylation unit.  During operations
 of the alkylation unit, the settler vessel separating HF from  hydrocarbons develops a two inch diameter
 hole below the HF liquid level.  The HF flows from a hole at the bottom of the settler at about 800
 gallons per minute (50 kilograms per second). The facility  is not equipped with mitigation equipment.
 The alkylation process is stopped, however, the HF continues to flash and aerosolize from the settler.
 The HF is drained from the settler in nine minutes.

        The settler leak-bottom was modeled assuming that all of the released HF becomes airborne.
 This means that all of the HF flashes to vapor or part of it flashes and the reminder aerosolizes so that
 no liquid is remaining after the  release.  For processes  involving HF, this scenario represents a release
 under moderate pressure (125  psi) and temperature (40°C).  The release was modeled assuming F
 atmospheric stability class conditions.

        A similar accident occurred at a refinery, when the  bleeder valve on an acid tank opened,
 releasing 725 kilograms of HF.  This is a much smaller amount than was released in the worst-case
 scenario  modeled here, where a settler completely  empties, spilling 26,000 kilograms of HF.

                Accident Scenario 9 - Settler Leak-bottom (D Stability)

        This accident is the same as scenario 8 except  the release was modeled assuming D
 atmospheric stability class conditions.

               Accident Scenario 10 - Mitigated Settler Leak-bottom (D Stability)

        A facility that experiences the release in scenario 9  applies water sprays almost immediately to
 the release. Based on the high flows of the water sprays, it is assumed that the air release was
 reduced by 90 percent. For simplicity, the 90 percent reduction of the air release was modeled as a
 90 percent reduction of the initial flow in scenario 9. The release was stopped in nine minutes and
 was modeled  assuming D atmospheric stability class conditions.

               Accident Scenario 11 - Vessel Leak  (D Stability)

        This scenario is applicable to any facility that stores HF.  HF begins to leak from a crack in a
 weld caused by corrosion or a defect in the bottom of a 5,000 gallon vessel.  The crack is about
 1/16th of  an inch wide by 5 feet long.54 Since the leak is located below the liquid level, liquid HF
 begins to spill from the tank.  Under worst-case conditions,  all of the HF becomes vapor.  The size of
 the vapor cloud is based entirely on the rate of HF release into the air. In this scenario, the amount  of
 HF in storage  determines the duration of the event.

        At the release site, HF detectors are either not working or not installed. There are no video
 monitors and operators are attending to problems elsewhere in the unit.  HF begins to leak unabated
 from the crack in the vessel. The  release is eventually discovered after 5 minutes because of the
 significant vapor cloud.  Area alarms are sounded, and response actions are initiated.  The release
 continues for a total of 20 minutes at about 195 gallons  per minute (12 kilograms per second) until the
 vessel is nearly empty.

       The effective diameter of the vessel crack area is based on information supplied by Du  Pont
 on a permit application to the Texas Air Control Board.55 The vessel leak was modeled assuming
 that all of the released HF becomes airborne. This means that all or some part of the HF flashes  to
 vapor and the remainder aerosolizes so that no liquid  remains. Realistically, depending on ambient
Page 146

-------
conditions and the conditions of the release, a liquid pool could form during this event that, over time,
would eventually evaporate and form a cloud which would travel downwind.  This scenario represents
a release under pressure and temperature close to ambient (pressure 22 psi and temperature 27°C).
The release was modeled assuming D atmospheric stability class conditions.

       Vessel leaks appear frequently in the accident history as a cause of HF releases; an example
is the 600 kilogram release of anhydrous HF due to a corroded pressure gauge. This release was
much smaller than this worst-case scenario, which is based on a 5,000-gallon vessel which spills
about 3,900 gallons (14,400 kilograms) of its contents.

               Accident Scenario 12 - Mitigated Vessel Leak (D Stability)

       Upon discovery of the release in scenario 11, the operators begin an emergency de-inventory
operation on the cracked vessel. Within three minutes, the total remaining content of the failed vessel
is dumped to a spare vessel, stopping the release. The leak is modeled similarly to scenario 11
except that the release duration is three minutes.

               Accident Scenario 13 - Settler Leak-Inlet Pipe (F Stability)

       This scenario could occur at a petroleum refinery with an HF alkylation unit. During operations
of the alkylation unit, the inlet pipe to the acid settler develops a one inch diameter hole.  The liquid in
the pipe is approximately 50 percent  HF and 50 percent hydrocarbons. The total HF/hydrocarbon
leak rate is initially 300 gallons per minute (19 kilograms per second).  In this scenario, the
HF/hydrocarbon mixture  does not ignite. Ignition of the mixture could  decrease off-site consequences
as a result of combustion and of increased HF dispersion from thermal effects.

        This scenario was modeled with an effective diameter to produce an HF initial flow rate of 150
gallons per minute (9.5 kilograms per second), roughly half the combined flow of HF and the
hydrocarbons. The release  duration  is approximately 20 minutes. Model inputs were based on
SCAQMD data which was provided by HF facilities. For processes involving HF, this scenario
 represents a release under high pressure (235 psi) and moderate temperature (40°C). The release
was modeled assuming F atmospheric stability class conditions.

               Accident Scenario 14 - Settler Leak-lnlei:  Pipe (D Stability)

        This accident is the  same as scenario 13 except the release was modeled assuming D
 atmospheric stability class conditions.

               Accident Scenario 15 - Pump Seal Release (F Stability)

        In HF processes, pumps may be used to transfer HF for loading and unloading trucks and
 tank cars, to circulate acid in alkylation units, and to feed HF raw material to other processes (e.g.,
 fluorocarbon manufacturing). Pump seal failure is a possible cause of HF releases.  In operations
 using aqueous HF, the corrosive nature of the acid may contribute to  the breakdown of the seal.

        The release rate of HF from a pump seal failure varies depending upon the initial conditions
  (i.e., phase, flow rate, pressure, temperature).  One estimate places the release rate at 10 to 100
  gallons per minute for a pump seal failure in an HF process.56 Depending on the initial conditions,
  HF can be released as a vapor, aerosol or liquid. A small  leak can be detected by vapors emanating
 from the pump seal.  One refinery using HF alkylation estimated that a seal failure on a pump which
  has a shaft diameter of 1.5  inches (clearance of  0.02 inches) would yield a release rate of about 20
  gallons per minute at a typical operating pressure of 125 psi.57
                                                                                      Page 147

-------
         In this worst-case scenario, a pump seal in an alkylation unit leaks at an average rate of
  29 gallons per minute (about 2 kilograms per second). The HF detector fails to respond or no HF
  detector is present and the unit operator is checking on a problem elsewhere in the unit. A mechanic
  at the change room notices fumes near the pump and calls the central control room. The control
  room operator radios the unit operator to check the pump. The operator confirms the leak and tells
  the control room operator to shut down the pump and sound the unit alarm.  The pump is shut off,
  but there are  no remotely-activated valves, consequently, HF continues to leak until valves can be
  closed manually. Response actions are  initiated.  Responders don personal protective equipment to
  enter the process area to close valves and stop the leak. After 20 minutes, the valves on the pump
  are finally closed and the leak is stopped. Approximately 580 gallons of HF are released.

         Given the pump pressure and the relatively small clearance of the leak, it is likely that the
  released HF would be sprayed into droplets and mixed with the air. Consequently, for this worst-case
  scenario, the pump seal failure was modeled assuming that all of the released HF becomes airborne.
  The release is assumed to last 20 minutes.  Depending on the  nature of the spill, if mitigation such as
  water sprays were quickly applied to the  spill, the downwind concentrations could be considerably
  reduced.  However, mitigation measures  were not considered in this scenario.  For processes
  involving HF, this scenario represents a release under moderate pressure (125 psi) and temperature
  (40°C). This release was modeled assuming F atmospheric stability class conditions.

                Accident Scenario 16 - Pump Seal Failure (D  Stability)

        This accident is the same as scenario 15 except the release was modeled assuming D
  atmospheric stability class conditions.

        A pump seal failure at a refinery HF alkylation unit occurred, in which 68 kilograms of HF were
  released.  Although worst-case scenarios 15 and 16 were based on a system with moderate pressure
  and temperature, the amount released, 2,200 kilograms, is significantly higher.


        The specific inputs for each of the above scenarios for HGSYSTEM and SLAB are provided in
 Appendix XVII.


 9.6    Limitations of the Modeling Results

        The modeling results for the above scenarios should be examined in light of several limitations
 or difficulties in developing worst-case scenarios and in modeling the scenarios.

        9.6.1  Limitations of the Worst-Case Scenarios

        The scenarios considered in this report are generic and not site specific. Process conditions
 (e.g., temperature, pressure), the effectiveness of any mitigation systems, and the HF inventories will
 vary from facility to facility. Therefore, it is unlikely that the specific combination of worst-case
 conditions used in these cases would be experienced in a release at a specific site. In defining and
 developing the worst-case scenarios, EPA acknowledges that the scenarios may be extremely
 conservative. Actual releases would most likely result in lower flow rates and shorter durations than
 assumed in the worst-case scenarios.

        In some cases, simplifying assumptions were made to accommodate the dispersion models
 used. For example, the flow rate was assumed to remain  constant at the initial flow rate until the
Page 148

-------
vessel was emptied even though a variable emission rate is consistent with fluid dynamics. The
impact of this assumption is evaluated in the sensitivity analyses later in this chapter.

       The worst-case scenarios did not include certain scenarios, such as releases from refrigerated
HF storage where the vaporization rate would be expected to be significantly less. In addition, certain
facility situations were not covered, such as modeling cases where a combination of mitigative
measures (e.g., automatic shutoff and water sprays) are used. The choice of a rural surface
roughness value also does not reflect the structures typically found at a facility.

       Finally, the likelihood of the releases was not quantified in any of the different scenarios
modeled.  The probability of an accidental release is critical  in evaluating risk to facility employees and
the public. Consequently, the results presented here are only useful in comparing the relative
potential severity in terms of concentration, distance, and dose for a variety of  release scenarios
including worst-case.

       9.6.2  Limitations of the Models

       It  is fortunate for this study that a comprehensive model, HGSYSTEM, has been specifically
tailored to the complexities of HF releases.  SLAB also has been shown to effectively  reproduce
experimental spill data of HF releases.58 However, these models cannot incorporate all of the
complexities of the release, dispersion, and chemical interactions associated with a spill event.  Each
model has its own limitations.  SLAB does not account for the association/dissociation of HF  in air or
the reaction of HF with water vapor.  HGSYSTEM makes the simplifying assumption that if any HF
flashes (flashing is temperature and pressure dependent), then all of the HF will aerosolize and
vaporize.  This is a worst-case assumption which may not reflect cooling factors and  other complex
atmospheric reactions occurring during a release.

        The model algorithms are based primarily on a theoretical understanding of the concepts of
 release and dispersion. However, some of the model parameters are derived from small-scale
 laboratory experiments. The models must account for a wide range of release conditions and the
 interaction of complex factors (e.g., effect of humidity and molecular disassociation on gas density).
The purpose of field tests is to evaluate and validate the models and to better understand the
 influence of complex chemical reactions and thermodynamic effects during a full-scale release.
 However, the validity of HGSYSTEM and SLAB models may be largely untested, because there are
 few HF field tests to date, and the tests that do exist do not reflect a wide range of release conditions.
 Some algorithms and assumptions contained in the models cannot be adequately evaluated with
 currently  available field test data. For example, in the Goldfish spill tests, HF was released only at a
 temperature of 40°C and pressures between 110 to 120 psi, which are conditions approximating
 petroleum refinery HF alkylation unit operating parameters.  The rates of release in the Goldfish tests
 were 10 to 30 kilograms/second for a total  release quantity  of approximately 3,600 kilograms. Both
 HGSYSTEM and SLAB have been validated with downwind HF concentrations from such a release.
 Both models agree with the field data; however, other release conditions that are typical of other
 possible  release scenarios (e.g., hose transfer failure, catastrophic vessel failure) have yet to be
 validated.

        The scenarios in this study were developed to include a range of release conditions that
 represented worst-case releases.  For most of these scenarios, the models have not been validated
 with field data specific for HF releases. Of the scenarios modeled in this study, only scenarios 8, 9,
 15, and 16 simulate the temperature and pressure conditions of the Goldfish tests.  Only scenarios 3,
 11, 12, 13, and 14 are within the range of release rates in the Goldfish tests.  In fact, scenario 1  -
 catastrophic failure of a bulk storage vessel - assumes release of a quantity at least 40 times greater
 than the largest release quantity in a field trial.   Also, none  of the field trials were conducted under the
                                                                                       Page 149

-------
  most conservative atmospheric conditions of stability F, which is specified in scenarios 4, 5, 8, 13, and
  15. The Goldfish desert conditions of extremely low surface roughness, high temperatures, and low
  relative humidity are not often reflected in sites where HF is produced, used, or stored. Also, because
  field data on concentration were collected no further than about 3 kilometers from the release, the
  results of modeling the scenarios in this study are outside the range of available empirical data.

         For large releases, both models estimate that several hours would elapse before the HF  cloud
  reaches the distance at which the IDLH is reached. It is unrealistic to assume that during the several
  hours that the plume travels, the wind speed and direction would remain constant. In fact, wind
  speed and wind direction would not be expected to remain constant beyond the time it takes the
  plume to pass about 10 kilometers. The models were not able to accommodate changes to wind
  speed or wind direction during the course of one release.  Consequently, it was assumed that the
  modeling  results would not be accurate past 10 kilometers.  In the scenarios where the release
  passed or came close to 10 kilometers during D stability (e.g., catastrophic vessel failure), the release
  scenario using F stability was not modeled because the HF plume would be expected to qo much
  further than 10 kilometers.

        In some  cases, however, simplifying assumptions used in the modeling may give rise to less
  conservative estimates.  In the mitigation scenarios with water sprays, the percent reduction in flow is
  assumed to start at the beginning of the release. However,  this does not account for the unmitigated
  flow during the time prior to the  initiation of the water sprays. Also, the models may not calculate
  accurate dose estimates in situations when the release duration is shorter than the  exposure
  averaging time.  In these situations, an individual will not be exposed to the cloud for the full
  averaging time and therefore the models will be left to average in some zero concentrations  into the
  dose calculations.  Consequently, exposure averaging times of 60 minutes for ERPG-3 will have to
  average in more  zero concentrations than  shorter averaging times of 30 minutes for IDLH. This may
  make distances to ERPG-3 somewhat less accurate than distances to IDLH.

        In  running the catastrophic vessel failure scenarios, the models are being pushed to the limits
 of their capability to estimate concentrations and downwind distances.  Several data field ranges  built
 into the HGSYSTEM model  (e.g., rate, hole size, vessel capacity) had to be expanded to account for
 the more extreme inputs from these catastrophic vessel failures.  Because of the long  dispersion times
 associated with releases from catastrophic vessel failures during which time wind speed and direction
 are uncertain beyond 10 kilometers, these  catastrophic releases were only modeled with D stability.

        9.6.3 Impact of the Limitations on the Results

        As mentioned above, many assumptions and estimates were made when determining the
 various inputs for modeling the accident scenarios.  In evaluating the results, therefore, it is important
 not to attach too  much significance to  precise numerical results.  Two different dispersion models
 were used for comparison of results.

        Because of the uncertainties in the models and in the unlikelihood that wind speed and
 direction will last several hours for significant distances, the modeling results are presented as
 distances up to a maximum of 10 kilometers. Greater distances are expressed simply  as greater than
 10 kilometers. Also, the distances are  approximated to the nearest one-half kilometer.  The distances
 for all of the scenarios are based on the IDLH. Only a few select scenarios are modeled using the
 ERPG-3 exposure level.  Caution  should be used in  interpreting the distances to the exposure levels.


 9.7    Results and Analysis
Page 750

-------
       This section compares the results obtained for different types of release scenarios, models,
and mitigation options for reducing release consequences,  The results are first discussed generally
and then evaluated as they pertain to catastrophic vessel failure and to a range of other scenarios.
The data are presented as distances to the IDLH (30 ppm for 30 minutes) or the ERPG-3 (50 ppm for
60 minutes). Estimates of the area covered by the resultant plume are also given for several
scenarios.  Exhibit 9-5 shows the modeling results for catastrophic vessel failures using the
HGSYSTEM and SLAB models.  Exhibit 9-6 shows the modeling results for the range of other
scenarios using the same two models.

       9.7.1  General Discussion of Results

       Distances to Exposure Level. The calculated distances to IDLH for  various scenarios suggest
that an HF  release can pose a hazard far beyond facility boundaries.  The modeling results indicate
that for many worst-case scenarios, HF has the potential to travel into populated areas. HF can
aerosolize,  form a dense gas cloud, and the plume can remain largely intact over substantial
distances.  This should be of concern to facilities that manage HF and to local communities that need
to develop  emergency plans for possible HF releases. Distances to ERPG-3, although somewhat
shorter than distances to IDLH, could be sufficient to pose a hazard to the public around many HF
facilities.  However, it should be noted that for emergency response planning, distances of concern
should be estimated using site-specific analyses.

       The SLAB model also provided information on maximum width of the plume which has
reached the furthest distance to a specific dose threshold.  For example, in the mitigated hose failure
at F stability level (Scenario  5),  the plume spread to a maximum width of about 700 meters. When D
stability level was assumed (Scenario 7), the maximum plume width was only 150 meters.
HGSYSTEM also can be used to determine plume widths at various concentrations of interest; this
capability was not examined for this study.  Factors that influence the width of an HF plume are
concentration at centerline, horizontal (or crosswind) wind speed, stability class, and averaging
time.
59
       Area of Plume.  The area covered by a cloud may be a more important indicator of the extent
of potential consequences of a release than the distance covered by the plume. Plume area which is
a function of both the plume distance and width, will better describe the populations threatened by the
release.  Data on maximum plume width and distance (length) to the IDLH from SLAB runs were
analyzed to estimate the area of the plume  created by a release.  An elliptical shape was assumed to
approximate the area covered by the HF  cloud, defined as ir times half the width times half the length
of the cloud. Plume areas were not calculated for releases that extended beyond 10 kilometers,
because it is unrealistic to assume that the  wind speed  and direction remain constant beyond the time
it takes the HF plume to pass 10 kilometers. Exhibits 9-5 and 9-6 show that the affected areas range
from 0.1 to 6.5 square kilometers. The area of a cloud is amplified when the extent of the plume
increases.  For example, in comparing scenario 12 to scenario 14, the distance to the IDLH is
doubled, whereas the area is quadrupled.  The plume width varied in these scenarios from about 8 to
12 percent of the plume length. The dispersion data indicated that the location of the maximum
plume width at the IDLH was not always midway between the release location and maximum distance
to IDLH.
                                                                                    Page 151

-------
                                           Exhibit 9-5
                         Modeling Results of Catastrophic Vessel Failures
Scenario Scenarios HGSYSTEM SLAB HGSYSTEM Area of
Number Distance to Distance to Distance to Plume**
IDLH (km) IDLH (km) ERPG-3 (km) (sq. km)
1
2
3
Vessel Rupture-Bulk Storage
D Stability, Empty Vessel
Derailment Anhydrous HF
D Stability, Empty rail car
Derailment 70% Aqueous HF
D Stability, Empty rail car
>10
9.5
1.5
>10
9
1.5
*
5.5
*
-
6.5
0.2
     [Distance to ERPG-3 was not predicted for this scenario.
     "Area Is calculated to nearest 0.1 square kilometer, based on SLAB width and distance to IDLH.
        Comparison of the Models. Overall, the results of the HGSYSTEM and SLAB models paralleled
 each other.  From scenario to scenario, the two models estimated distances that were consistent. In
 all scenarios but one, HGSYSTEM predicted greatest distances to the IDLH than did the SLAB model.
 In Scenario 11, the vessel leak, the SLAB model had a distance to IDLH that was approximately ten
 percent higher than that for HGSYSTEM. This is within the expected uncertainty of the accuracy of
 the results. The slightly  longer distances estimated using HGSYSTEM may be due to the particular
 HF thermodynamic considerations (e.g., polymerization) in the HGSYSTEM model that tend to
 emphasize the dense and cohesive characteristics of the HF cloud.  The HGSYSTEM model also
 incorporates more inputs than the SLAB model, which can help to further describe or account for the
 factors that influence HF dispersion.

        Effects of Atmospheric Stability Category and Wind Speed. Meteorological conditions during a
 release can dramatically  affect the potential hazards. Both F stability (at 1.5 meters per second wind
 speed) and D stability (at 5 meters per second wind speed) were assumed in five release scenarios
 (i.e., hose failure, mitigated hose failure, settler leak-bottom, settler leak-inlet pipe, and pump seal
 failure). The stability input greatly affected the results. Distances estimated for most of the scenarios
 where F stability was assumed to exceed 10 kilometers.  Alternatively, results for scenarios using D
 stability, which indicates  more turbulent, less stable atmospheric conditions, indicated significantly
 shorter distances to the IDLH. Under these conditions, the HF dispersed rapidly.  The distances for
 scenarios with D stability were up to 87 percent less than the distances estimated using F stability
 conditions.

        Some assumptions about meteorological conditions may be overly conservative for the
 purposes of emergency planning.  Because F stability conditions typically exist only at calm, pre-dawn
 hours, however industry recommends that HF unloading/loading operations (scenarios 4 and 5) take
 place during daylight hours when F stability is "rare.60

        Effects of Flow Rate.  The scenarios with the highest initial HF release rates - vessel rupture,
 derailment, hose failure, settler leak-bottom, vessel leak, and settler leak-inlet pipe - resulted in the
 greatest distances to IDLH. Scenarios 15 and 16 (pump seal failure) had the smallest initial flow rate,
 and also had the smallest distance to IDLH. These results  indicate that release rate is a critical factor
 contributing to the distance an HF cloud can travel.  More frequent industry testing (e.g., corrosion or
Page 152

-------
structural testing) on large volume or large flow equipment should help to reduce the potential
process points where large flow releases could occur.  If release rate could be minimized, the
consequences of a release could potentially be reduced.
                                            Exhibit 9-6
                         Modeling Results of a Range of Other Scenarios
HGSYSTEM SLAB HGSYSTEM Area of j
Scenario Distance to Distance to Distance to Plume j
Number Scenario IDLH (km) IDLH (km)
^•^••M
4

5


6

7


8

9

10


11

12


13

14

15

16

I^III^Hil^HBHB^^BH*"*
Hose Failure
F Stability, Empty tank
Hose Failure (mitigated with
automatic shutoff valves)
F Stability, 1 minute release
Hose Failure
D Stability, Empty tank
Hose Failure (mitigated with
automatic shutoff valves)
D Stability, 1 minute release
Settler Leak-Bottom
F Stability
Settler Leak-Bottom
D Stability
Settler Leak-Bottom (mitigated
with water sprays)
D Stability
Vessel Leak
D Stability, Nearly Empty Vessel
Vessel Leak (mitigated with
emergency de-inventory)
D Stability, Nearly Empty Vessel
Settler Leak-Inlet Pipe
F Stability, 20 minute release
Settler Leak-Inlet Pipe
D Stability, 20 minute release
Pump Seal Failure
F Stability, 20 minute release
Pump Seal Failure
D Stability, 20 minute release
••^^^••••B
>10

>10


5

3


>10

6

2.5


3.5

1.5


>10

3

9

1.5

^^mm^^^m^^m
>10

8


4.5

1.5


>10

5.5

1.5


4

1.5


>10

3

7.5

1

ERPG-3 (km) (sq.km) |
>10

6.5


3

*


*

*

*


*

*


*

*

*

*



5.4


1.5

0.1




1.9

0.1


1.4

0.2


_

0.8

3.8

0.1

     "Distance to ERPG-3 was not predicted for this scenario.
     **Area is calculated to nearest 0.1 square kilometer, based on SLAB width and distance to IDLH.
                                                                                           Page 153

-------
         9.7.2  Results from Catastrophic Vessel Failure

         Distances and Areas to Exposure Level.  Exhibit 9-5 shows that the distances to the IDLH for
 the catastrophic vessel failure scenarios examined from releases of anhydrous HF were at least nine
 kilometers. These results are especially noteworthy because the releases were modeled at D stability
 and at a wind speed of 5 meters per second, and not at the worst-case meteorology conditions of F
 stability and a wind speed of 1.5 meters per second.  Even at the stronger wind speed at D stability,
 the distances would still be of serious concern to the public. At F stability, the distances to the IDLH
 from anhydrous HF releases from catastrophic vessel failure would be expected to be much larger
 than 10 kilometers. A sample area of a plume from a derailment involving an HF release is 6.5 square
 kilometers.

        Anhydrous versus Aqueous Releases. Comparing the distances to IDLH from an anhydrous
 versus aqueous release required a modification to the HGSYSTEM model. HGSYSTEM has a
 program to calculate the spill rate and another program to calculate evaporation and dispersion.
 Based on the spill program, the anhydrous release is assumed to aerosolize completely and the 70
 percent aqueous release is assumed to form a pool and begin to evaporate.  However, for the
 aqueous release, the evaporation program assumes that the evaporating pool is anhydrous HF
 instead of 70 percent aqueous HF.  Consequently, to simulate, a 70 percent aqueous HF evaporating
 pool rather than an anhydrous HF  evaporating pool, the evaporation rates generated by HGSYSTEM
 were reduced according to the ratio of the partial pressures of anhydrous  HF and 70 percent aqueous
 HF.

        The results show significant differences between a release from a derailment of a rail car
 containing anhydrous HF versus a release from a rail car  containing 70 percent aqueous HF. Even
 though the amount of liquid spilled from the derailed rail car is similar for both scenarios, the amount
 of HF  released to air is greater in the anhydrous case (65,000 kilograms in scenario 2) than in the 70
 percent aqueous case (3,000 kilograms in scenario 3).  This is due to the fact that HF in anhydrous
 form is assumed to evaporate  completely when exposed to air, while HF in the aqueous phase will
 pool and then will evaporate at the rate of a diluted HF solution.  The area covered by the anhydrous
 HF release in the derailment (scenario 2) is over 32 times greater than the area affected by 70 percent
 aqueous HF release in the derailment (scenario 3). This indicates the greater potential of anhydrous
 HF to  release into air, and affect populations near a release site.  However, the 1.5 kilometer distance
 and 0.2 square kilometer area from the aqueous release could still be cause for concern at some HF
 facilities.

        Comparison with Other Chemicals. Based on results from the HGSYSTEM model, the
 concentrations of HF in a release from a catastrophic failure of a bulk storage vessel remain above 30
 ppm for 30 minutes (IDLH) for more than 10 kilometers. This scenario, although extremely unlikely,
 would  pose a serious hazard to the public surrounding the facility. However, a similar catastrophic
 release of another hazardous substance like chlorine would create a similar hazard.  For the bulk
 vessel rupture (scenario 1), an HGSYSTEM run using chlorine rather than HF indicated that the
 chlorine cloud would reach the chlorine IDLH also at greater than 10 kilometers. Thus, considering
 the consequences  of a catastrophic vessel failure of HF should include the acknowledgment that
 many other hazardous substances, if released catastrophically, may pose similar hazards to the
 public.


        9.7.3 Results from a Range of Other Scenarios

        Distances and Areas to Exposure Level. In most cases, modeling results for a range of other
 release scenarios (Exhibit 9-6) indicated that the distances to the IDLH are  less than those from
Page 154

-------
catastrophic vessel failure. However, the results show that HF would travel over large distances
offsite. Generally, it appears that the rate of release is one of the more important factors influencing
the distance to the exposure level.

       These scenarios indicate a large  range of affected areas, from 0.1 to 5.4 square kilometers.
The differences in consequences due to  atmospheric stability level can be seen by comparing the
mitigated hose failure at F atmospheric stability (scenario 5) and the same mitigated hose failure at D
atmospheric stability (scenario 7). The plume in scenario 5 reached five times the distance reached
by the plume in scenario 7, and about eight times the width, and, as a result, covered approximately
50 times more area.

       Effects of Mitigation. The mitigation systems examined in the modeling either reduced the flow
rate of HF or reduced the duration of the release. The mitigation system employed in scenario 7 (i.e.,
hose failure) was automatic shutoff valves, which stopped the release in one minute. Compared with
the unmitigated case (scenario 6), the distance to the IDLH using the shutoff valves would be reduced
from 5 to 3 kilometers (using HGSYSTEM modeling results and D stability).

       The mitigation system employed  in scenario 10 (settler leak) was a water spray system, which
reduced the effective HF release rate by 90 percent.  Comparison with the unmitigated case (scenario
9) shows a distance reduction from 6 to 2.5 kilometers using HGSYSTEM.  This is consistent with the
general result mentioned above that a reduction in flow would greatly decrease the distance affected.

       The mitigation system employed  in scenario 12 (vessel leak) was an  emergency de-inventory
system which emptied the vessel and stopped the release in three minutes.  Comparing  results for
this case with the unmitigated case (scenario 11) shows a distance reduction from 3.5 to 1.5
kilometers. A shorter duration greatly reduces the distance to the IDLH.

       In all cases, mitigation reduced the distance to IDLH, thereby lessening the potential
consequences of the accident. The magnitude of this reduction in distance ranged from 57 to 73
percent.  It is not the intent to determine  the relative effectiveness of different mitigation strategies.
However, these results indicate a reduction of affected distances by using mitigation systems.
9.8    Sensitivity Analysis

       Several parameters were varied to determine their influence on the modeling results.  The
parameters include stability and wind speed, emission rate, relative humidity, and surface roughness.
The sensitivity analysis was based on either a release from an inlet pipe to an acid settler (scenario
14) or on a release from a hose failure (scenarios 4 and 5).  For the HEGADAS model contained in
HGSYSTEM, the results of the sensitivity analysis are summarized in Exhibits 9-7, 9-8, 9-9 and 9-10.

       9.8.1  Wind Speed and Stability

       Stability conditions are related to ranges of wind speeds. Generally, under more stable
atmospheric conditions and lower wind speeds, higher concentrations are reached downwind.
According to literature sources, downwind concentrations estimated using SLAB and HGSYSTEM
exhibit varying sensitivities to changes in wind speed.  The SLAB model is more sensitive to changes
in wind speed.61 The sensitivity analysis performed on HGSYSTEM shows that the choice of
stability class in combination with wind speed has the  largest affect on the downwind concentration of
HF of any parameter considered. Exhibit 9-7 shows that the dose  downwind for stability class F and
1.5 meters per second wind speed is far worse than the dose downwind for stability class D and 5.0
meters per second wind speed.
                                                                                    Page 755

-------
        9.8.2 Release Rate

        The modeling results presented for HGSYSTEM and SLAB are based on a constant release
 rate. It was assumed that the rate of spill or vapor release remained constant at an initial release rate
 until the vessel was emptied.  However, the actual rate of release from a vessel varies. The release
 rate would decrease over time due to the reduction of pressure in the vessel. SLAB can only handle a
 constant release rate. However, if the models contained in HGSYSTEM are run  separately, it is
 possible to incorporate changes in release rate. The sensitivity analysis presented in Exhibit 9-8
 shows the influence of constant and variable release rates on the dispersion of HF. This influence is
 presented for short (one minute) and long (seven minutes) duration releases from an
 unloading/loading hose.  The exhibit shows that for either short or long durations, the simplifying
 assumption of constant release rates gives more conservative results than variable rates. Also, the
 difference between the distances due to constant and variable release rates are greatest for longer
 duration releases. For this study, it was determined that the simplifying assumption would not
 significantly affect the analysis since the results were being used for comparative purposes rather than
 focusing on the importance of a specific distance.

        9.8.3 Relative Humidity

        Changes in concentration due to changes in relative humidity are a result of thermodynamic
 reactions which alter cloud density.  Relative humidity changes, therefore, alter dispersion by affecting
 both gravity spreading and vertical mixing. Exhibit 9-9 shows a complex relationship between
 concentration and relative humidity. At 10 percent humidity, there is little effect from heat generated
 by reaction between  HF and water.  At 50 percent humidity, the cloud is less dense in the early
 phases of dispersion and, thus, there is less gravity spreading and the plume is narrower.  Exhibit 9-9
 shows that, during the early stages of dispersion, the concentration with 50 percent humidity is close
 to, but slightly less than that for 10 percent.  Because the top surface of the narrow plume is smaller,
 there is a smaller area for mixing to take place, and dilution slows. Thus, after approximately 1,000
 meters downwind, the concentrations at 50 percent humidity become higher than those at 10 percent
 humidity. At 90 percent humidity, the heat of reaction enhances vertical mixing which results in lower
 concentrations.62

        9.8.4 Surface Roughness

        The overall effect of increasing surface  roughness was to enhance the mixing between the
 plume and the environment, thereby decreasing HF concentrations. Exhibit 9-10 shows the effect of
 rural (0.03 meters) and urban (0.5 meters) surface roughness on HF concentration.  The difference in
 concentration remains nearly constant as the cloud travels downwind. An  almost 17-fold increase in
 surface roughness creates only a 1.5 factor decrease in HF average dose.63
Page 156

-------
                               ave. concentration for 30 min. (ppm)
                                               8
                                     g    g
                I

                I
                e

                a
                ~L
01
N
i
>
3 m

•fz
2. ff
to ?»

 I  ?>
                                                                                                 concentration (ppm)
                                                  -L     a
                                                         o
                                           o      o     o
                            -».      o      o      o     o
                            o      o      o      o     o
                            in  i  iiiniii  i i Minn  i ilium  i i nun

-------
                                    Exhibit 9-9
                         Sensitivity Analysis — Relative Humidity
         100000s
                                                 1000
                               downwind distance x (m)
                                    Exhibit 9-10
                        Sensitivity Analysis - Surface Roughness
                                        1000
                              downwind distance x (m)
Page 158

-------
       9.8.5  Combined Effect of Parameter Variation

       To observe the aggregate effects of changes in stability level, relative humidity, and surface
roughness, two variations of scenario 14, inlet pipe to an acid settler, were analyzed. The first
variation was modeled using a combination of worst-case modeling conditions based on the sensitivity
analysis presented above. The conditions included F atmospheric stability, 10 percent relative
humidity, and rural surface roughness (0.03 m).  The second variation was modeled using modeling
conditions that increased HF dispersion, including D atmospheric stability, 90 percent relative
humidity, and urban surface roughness (0.5 m).  All other inputs remained the same.  Modeling results
showed that the combined effect of worst-case conditions on the maximum distance to the IDLH is
significant.  The HF plume in the worst-case modeling conditions of the first scenario traveled greater
than 10 kilometers (calculated at 14 kilometers),  compared to the plume in the second scenario, which
reached slightly less than two kilometers. This seven-fold difference illustrates the dramatic effect on
the resulting consequences.


9.9    Summary  ,

       This chapter examines the potential hazards of HF to the public by modeling a range of
events including worst-case accidental releases.  Computer  models that can predict HF dispersion
can be used to perform consequence analysis for accidental HF releases. Recent research efforts
conducted to better understand HF dispersion have involved field tests and the development of
several models. The capabilities of several dispersion models are discussed, including how the
models address complex issues, such as averaging time and surface roughness. Sixteen HF release
scenarios were selected for modeling, representing catastrophic releases (e.g., vessel rupture,  railcar
derailment) and a range of other release scenarios (e.g., pump  seal failure, hose failure).  Most
scenarios involved releases of anhydrous HF. A few scenarios involved the use of mitigation
measures which reduce the flow or duration of HF releases.  These scenarios were chosen to illustrate
possible accidental release events and the potential effects of mitigation.  Some simplifying
assumptions were made in defining the release scenarios.  It should be emphasized that public risk,
which would involve an analysis of the likelihood of release,  was not examined in this study.

       Two models were selected for use in the HF study-HGSYSTEM and SLAB.  HGSYSTEM is a
complex model that was developed especially for modeling HF  releases.  SLAB is designed to model
dense gases.  Both HGSYSTEM and SLAB make simplifying assumptions to model the complex
release characteristics of HF. Inputs to the models were compiled from sources including studies by
local and state governments, facilities that use HF, accident data, process design data and other
modeling efforts.

       Results were found in terms of maximum distance to IDLH and, for some scenarios, the
maximum distance to ERPG-3.  These results indicate the greatest distance from the release source at
which a person may  be exposed to an IDLH or ERPG-3 dose of HF.  Due to uncertainties in the
modeling assumptions, the models may be inaccurate beyond about 10 kilometers; distances greater
than 10 kilometers are noted as such. Results indicate that an  HF plume can remain largely intact,
and travel substantial distances from the release site. The catastrophic vessel failures of anhydrous
HF resulted in distances at or near 10 kilometers.  The affected distances of the range of other release
scenarios depended largely on release rate, release duration, and atmospheric stability.  In all
scenarios for which mitigation was modeled (e.g., water sprays, emergency de-inventory, automatic
shutoff valves), mitigation significantly reduced the distance  to IDLH, lessening the potential
consequences of the accident.
                                                                                    Page 159

-------
        The modeling analysis provides an indication of the severity of effects and potential doses that
 could result from an accidental HP release.  This consequence analysis, however, is not a prediction
 of what will actually occur in a release.  Actual size and behavior of an HP cloud will depend on site-
 specific conditions of site topography, process operating conditions, weather, mitigation measures,
 and response capability. Facilities that  use HP should  be concerned with possible hazards beyond
 facility boundaries and should estimate  distances using site-specific conditions and analyses.
Page 160

-------
                                        ENDNOTES
1.      Woodward, John, ed., International Conference on Vapor Cloud Modeling, November 2-4, 1987,
       Boston Marriott Cambridge, Cambridge, MA, American Institute of Chemical Engineers, Center
       for Chemical Process Safety, New York, 1987, "Conduct of Anhydrous Hydrofluoric Acid Spill
       Experiments," by Blewitt, D.N., J.F. Yohn,  R.P. Koopman, and T.C. Brown.  (60)

2,      Woodward, John, ed., International Conference on Vapor Cloud Modeling, November 2-4, 1987,
       Boston Marriott Cambridge, Cambridge, MA, American Institute of Chemical Engineers, Center
       for Chemical Process Safety, New York, 1987, "Effectiveness of Water Sprays on Mitigating
       Anhydrous Hydrofluoric Acid Releases," by Blewitt, D.N., J.F. Yohn, R.P. Koopman, T.C. Brown,
       and W.J. Hague. (70)

3.      Woodward, John, ed., International Conference on Vapor Cloud Modeling, November 2-4, 1987,
       Boston Marriott Cambridge, Cambridge, MA, American Institute of Chemical Engineers, Center
       for Chemical Process Safety, NY, 1987, "An Evaluation of SLAB and DEGADIS Heavy Gas
       Dispersion Models Using the HF Spill Test Data," by Blewitt, D.N., J.F. Yohn, and D.L Ermak.
       (130.13)

4.      Koopman, Ronald P., Donald L. Ermak and Stevens T. Chan, "A Review of Recent Field Tests
       and Mathematical Modelling of Atmospheric Dispersion of Large Spills of Denser-Than-Air
       Gases," Atmospheric Environment, Volume 23, Number 4, pp 731-745, 1989. (289.2)

5.      HF Mitigation Water Spray Project, created for a meeting of the American Petroleum Institute
       and provided to the U.S. Environmental Protection Agency, December 10,  1991. (424.34ABC)

6.      Memorandum, Subject: Public Domain Models Which Could be Used to Simulate an Accidental
       Release of HF, From: D.N. Blewitt, To: R.C. Wade, April 1, 1991. (580)

7.      International Conference and Workshop on Modeling and Mitigating the Consequences of
       Accidental Releases of Hazardous Materials, May 20-24, 1991, Fairmont Hotel, New Orleans,
       Louisiana, American Institute of Chemical Engineers, Center for Chemical Process Safety, New
       York, 1991, "HF-DEGADIS Simulation of Dense Gas Dispersion from a Wind Tunnel-Modeled
       Oil Refinery," by Moser, James H., Doug N. Blewitt, Kenneth W. Steinberg, Ronald L. Petersen.
       (350)

8.      Kurata, Susan, and Steve Smith, Final Environmental Assessment for Proposed Rule 1410:
       Hydrogen Fluoride Storage and Use, South Coast Management District, April 2, 1991. (290)

9.      Meeting between U.S. Environmental Protection Agency and Dr. Jerry Havens from the
       University  of Arkansas, April 1992.

10.    Witlox, H.W.M., K. McFarlane, F.J. Rees, and J.S. Puttock, Development and Validation of
       Atmospheric Dispersion Models for Ideal Gases and Hydrogen Fluoride, Part II: HGSYSTEM
       Program User's  Manual, Shell Research Limited, Thornton Research Centre, Chester, England,
       November 1990. (509)

11.    Blewitt, D.N., K.  McFarlane, A. Prothero, J.S. Puttock,  F.J. Rees, P.T. Roberts, and H.W.M.
       Witlox, Development of HFSYSTEM Programs for Hydrogen Fluoride Dispersion Assessment.
       Presented at the 1990  Health and Safety Symposium Organized by AlChE in Orlando, Florida,
       March 18-22, 1990. (57)
                                                                                   Page 161

-------
12.    Woodward, John, ed., "An Evaluation of SLAB and DEGADIS Heavy Gas Dispersion Models
       Using the HF Spill Test Data."

13.    South Coast Air Quality Management District, Review of Dense Gas Modeling with Emphasis
       on Application in Planning for Accidental Releases and Emergency Responses,  Final Report,
       Contract No. C91319, February 26, 1992.  Prepared by Steven R. Hanna, Joseph C. Chang
       and David G. Strimaitis. (441)

14.    Splcer, T.O., and J. Havens, Modeling HF and NH3 Spill Test Data Using Degadis, Department
       of Chemical Engineering, University of Arkansas, Fayetteville, Arkansas, August 1988.  (444)

15.    Woodward, John, ed., "An Evaluation of SLAB and DEGADIS Heavy Gas Dispersion Models
       Using the HF Spill Test Data," p 69.

16.    Woodward, John, ed., "An Evaluation of SLAB and DEGADIS Heavy Gas Dispersion Models
       Using the HF Spill Test Data," p 68.

17.    Ermak, Donald L, User's Manual for SLAB: An Atmospheric Dispersion Model for Denser-than-
       Air Releases,  University of California, Lawrence Livermore National Laboratory, June 1990,
       Document Number: UCRL-MA-105607.   (139.8)

18.    Spicer.T.O.

19.    Woodward, John, ed., "An Evaluation of SLAB and DEGADIS Heavy Gas Dispersion Models
       Using the HF Spill Test Data," p 69.

20.    South Coast Air Quality Management District, Review of Dense Gas Modeling with Emphasis
       on Application in Planning for Accidental Releases and Emergency Responses.

21.    Hanna, Steven R., David G. Strimaitis, and Joseph C. Chang, "Evaluation of Fourteen
       Hazardous  Gas Models with Ammonia and Hydrogen Fluoride Field Data", Journal of
       Hazardous Materials, 1991, vol 26, p. 127-158.

22.    Industry Cooperative HF Mitigation/Assessment Program Ambient Impact Assessment
       Subcommittee,  Development and Validation  of Atmospheric Dispersion Models for Ideal Gases
       and Hydrogen Fluoride, Part I: Technical Reference Manual, Shell Research Limited, Thornton
       Research Centre, Chester, England, November 1990, Document Number TNER.90.015.
       Prepared by K. McFarlane, A. Prothero, J.S. Puttock, P.T. Roberts, and H.W. M. Witlox.
       (328.65)

23.    International Conference and Workshop on Modeling and Mitigating the Consequences of
       Accidental Releases of Hazardous Materials,  "HF-DEGADIS Simulation of Dense Gas
       Dispersion from a Wind Tunnel-Modeled Oil Refinery."

24.    Fthenakis, Vasilis M., Brookhaven National Laboratory, comments from technical review of
       Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992, June 8, 1992. (139.69)

25.    International Conference and Workshop on Modeling and Mitigating the Consequences of
       Accidental Releases of Hazardous Materials, May 20-24,  1991, Fairmont Hotel, New Orleans,
       Louisiana, American Institute of Chemical Engineers, Center for Chemical Process Safety, New
       York, 1991,  "Modeling of Water Spraying of Field Releases of Hydrogen Fluoride," by
       Fthenakis, V.M., K.W. Schatz, and V. Zakkay. (140)
Page 762

-------
26.     Fthenakis, Vasilis M., Brookhaven National Laboratory.

27.     Fthenakis, V.M. and K.W. Schatz, Numerical Simulations of Turbulent Flow Fields Caused by
       Spraying of Water on Large Releases of Hydrogen Fluoride, Biomedical and Environmental
       Assessment Division, Department of Applied Science, Brookhaven National Laboratory,
       Associated Universities, Inc., May 1991. (139.66)

28.     Fthenakis, V. M. and Doug M. Blewitt, "Mitigation of HF  Release: Assessment of the
       Performance of Water Spraying Systems," Presentation at the 7992 AlChE Summer National
       Meeting, Minneapolis, Brookhaven National Laboratory, Upton,  NY, August, 9-12, 1992.
       (139.67)

29     U.S. Environmental Protection Agency, An SAB Report:  Review  of Hydrogen Fluoride Study:
       Report to Congress, Science Advisory Board, Washington, D.C., December 1992. (489.89b)

30.    Hague, William J., Allied-Signal, comments on Hydrogen Fluoride Study, Report to Congress,
       Draft May 8, 1992, June 10, 1992. (153)

31.    Laumer, John and  Harold Lamb, Elf Atochem North America Inc., comments from technical
       review of Hydrogen Fluoride Study, Report to Congress, Draft May 8, 1992, June 4,  1992.
       (290.8)

32.    South Coast Air Quality Management District, Source Terms and Frequency Estimates of
       Selected Accidental Hydrofluoric Acid Release Scenarios in the South Coast Air Basin, El
       Monte, CA, December 1990. Prepared by PLG, Inc. (442.7)

33.    South Coast Air Quality Management District Board, Recommendations on the Potential
       Hazards of Hydrogen Fluoride Transportation, Storage,  and Use, El Monte, CA, April 4, 1990.
        (442.5)

34.     Du Pont Chemicals, Du Font's La Porte, Texas Plant's Disaster Scenarios, Air Permit Submitted
        to TACB-1991, Wilmington, DE, 1991. (139)

35.     Risk Management Plans, HazOp, Safety. (424.34)

36.     ERNS Data Base, U.S. Environmental Protection Agency, U.S. Department of Transportation,
        U.S. Coast Guard, National Response Center,  1987-1991. (139.525)

 37.     U.S. Environmental Protection Agency, U.S. EPA Release Prevention Questionnaires, 1987-
        1988. (490)

 38.     U.S. Environmental Protection Agency, Hydrofluoric Acid Events in Acute Hazards Events
        Database, June  5, 1991.  (200)

 39.     Blewitt, Doug, N., Design of Water Spray Mitigation Systems for Amoco HF Alkylation Units. (56)

 40.     ten Berge, W.F., A. Zwart, and L.M. Appelman, "Concentration-Time Mortality Response
        Relationship of Irritant and Systemically Acting Vapours and Gases," Journal of Hazardous
        Materials, Volume 13,1986, pp 301-309. (54.5)
                                                                                    Page 163

-------
  41.




  42.



  43.




  44.


  45.




  46.

  47.




 48.

 49.

 50.


 51.

 52.

 53.

 54.

 55.

 56.



 57.

 58.
  de Weger, Dik, Chris M. Pietersen, and Paul G.J. Reuzel, "Consequences of Exposure to Toxic
  Gases Following Industrial Disasters," Journal of Loss Prevention in the Process Industries
  Volume 4, July 1991. (502)

  Mudan, Krishna S., Acute  Inhalation Toxicity of Hydrogen Fluoride, Technica Inc., Columbus
  OH, March 1989. (360)

  U.S. Department of Transportation, Study to Modify the Vulnerability Model of the Risk
  Management System, United States Coast Guard, Office of Research and Development,
  Washington, D.C., February 1980, Document Number CG-D-22-80. (489.91)

  "HF Guidelines Change," (Health and Safety Executive Announcement from Jane C. Bugler),
  The Chemical Engineer, June 25, 1992, p 7. (76.3)

  Panketh, Joseph, K. Home, K. Neumann, and S. Keil, Case Studies of Texas Air Control Board:
 Permit Reviews in which Potential for Catastrophic Releases Was Considered, Texas Air Control
 Board, Austin, TX, October 1986. (471)

 South Coast Air Quality  Management District, Source Terms and Frequency Estimates.

 U.S. Environmental Protection Agency, Federal Emergency Management Agency, and U.S.
 Department of Transportation, Technical Guidance for Hazards Analysis:  Emergency Planning
 for Extremely Hazardous Substances, December 1987. (489.95)

 South Coast Air Quality  Management District, Source Terms and Frequency Estimates.

 South Coast Air Quality  Management District, Source Terms and Frequency Estimates.

 South Coast Air Quality  Management District Board, Recommendations on the Potential
 Hazards of Hydrogen Fluoride Transportation, Storage, and Use.

 South Coast Air Quality Management District, Source Terms and Frequency Estimates.

 Du Pont Chemicals, Du Ponfs La Porte,  Texas Plant's Disaster Scenarios.

 Du Pont Chemicals, Du Pom's La Porte,  Texas Plant's Disaster Scenarios.

 Du Pont Chemicals, Du Font's La Porte, Texas Plant's Disaster Scenarios.

 Du Pont Chemicals, Du Font's La Porte, Texas Plant's Disaster Scenarios.

 Risk Management Plans, HAZOP, Safety Evaluations, etc., for various anonymous
 HF/Refining/Alkylation facilities. (424.34E)

 Risk Management Plans, HAZOP, Safety Evaluations, etc.

Woodward, John, ed., International Conference on Vapor Cloud Modeling, November 2-4, 1987,
Boston Marriott Cambridge, Cambridge, MA, American Institute of Chemical Engineers, Center'
for Chemical Process Safety, NY, 1987, "An Evaluation of SU\B and DEGADIS Heavy Gas
Dispersion Models Using the HF Spill Test Data."
Page 164

-------
59.     Industry Cooperative HF Mitigation/Assessment Program Ambient Impact Assessment
       Subcommittee, Development and Validation of Atmospheric Dispersion Models for Ideal Gases
       and Hydrogen Fluoride, Part I: Technical Reference Manual, Shell Research Limited, Thorton
       Research Centre, Chester, England, November 1990, p 7.17-24, Document Number
       TNER.90.015. Prepared by K. McFarlane, A. Prothero, J.S. Puttock, P.T. Roberts, and H.W. M.
       Witlox.  (328.65)

60.     American Petroleum Institute (API), Safe Operation of Hydrofluoric Acid Alkylation Units, API
       Recommended Practice 751, First Edition, June 1992. (10.6)

61.     Guinnup, D.E. and Q.T. Nguyen, "A Sensitivity Study of the Modeling Results from Three
       Dense  Gas Dispersion Models in the Simulation of a Release of Liquefied Methane:  SLAB,
       HEGADAS, and DEGADIS," Seventh Joint Conference on Applications of Air Pollution
       Meteorology with AWMA, New Orleans, LA, January 14-18, 1991.

62.     Industry Cooperative HF Mitigation/Assessment Program Ambient Impact Assessment
       Subcommittee.

63.     Industry Cooperative HF Mitigation/Assessment Program Ambient Impact Assessment
       Subcommittee.
                                                                                   Page 165

-------
                                  This page intentionally left blank.
Page 166

-------
                 10. COMMUNITY AND FACILITY EMERGENCY
                         PREPAREDNESS AND PLANNING
       In the course of visiting facilities to research the question of the hazards associated with the
production and use of HF, some limited information on the interface between the facility and the
surrounding communities was gathered. The information presented in this chapter is, therefore,
anecdotal, and is not intended to be either comprehensive or representative of the U.S. or community
planning and preparedness concerning HF.

10.1   Emergency Preparedness and Planning

       In the event of an HF release, community officials must be informed of the release and be
prepared to take appropriate action (e.g., shelter-in-place or evacuation of the public from affected
areas). However, in the event of an accident involving a release, community officials may not be able
to react quickly or responsibly without prior communication and planning coordination with the facility.
Industry also recognizes the need to plan for emergency response with the community.  Because an
HF  release has the potential to cross the fenceline, facilities handling HF are developing community
outreach activities.

       The HF incident in Texas City, Texas highlighted the need for community and facility
emergency planning and preparedness, especially for HF or similar releases.  Marathon requested
assistance from the Industrial Mutual Aid System (IMAS) and notified the Texas City Emergency
Operations Center (EOC), including the Texas City Police and Fire Departments, to call for an
immediate public evacuation.  A Community Awareness and Emergency Response (CAER) alarm
system had been installed in the facility four months prior to the release.  However, because of a lack
of coordination between the IMAS and the Texas City EOC, the citizens of Texas City were not fully
acquainted with the meaning of the alarms or procedures to follow when the alarms sounded.
Consequently, many people did not understand the sirens' warnings or paid little attention to them.

       The emergency planning and response needs arid interactions between a community and a
facility depend on many factors including the chemicals and processes involved at the facility, the
response capabilities of the community and the facility, and the magnitude and nature of the  potential
threat to the community. To promote community and facility cooperation in developing emergency
response plans to deal with releases of hazardous chemicals, guidance and regulations are provided
by the federal government, by various state or local governments (e.g., the South Coast Air Quality
Management District), and by several industry associations. On the federal level, the Emergency
Planning and Community Right-to-Know Act (also known as SARA Title III) is designed to facilitate the
development of the capability of state and local governments to respond to potential chemical
emergencies through better communication, coordination, and planning with industry. Emphasizing
planning within the local community, SARA Title III mandated the formation of Local Emergency
Planning Committees (LEPCs) comprised of local emergency preparedness officials, industry and
facility officials, and citizens. The primary  purpose of the LEPCs is to develop community emergency
response plans for facilities and transportation routes. Such efforts require the LEPC to maintain
communication with industry about the chemical hazards present at facilities.  The plan outlines the
procedures for communicating with the facility, for transferring  information from the facility, for
mobilizing the emergency response, and for evacuating or sheltering the public.  Some facilities that
handle HF are members of or contribute to LEPCs.  For example, as an active member of the
Ascension Parish LEPC in Louisiana, Allied-Signal provides the LEPC with HF technical information
and resource support.2 In another example, Motorola has assisted the Phoenix Fire Department in
purchasing CAMEO software and equipping fire trucks with Macintosh computers.3  The development
of air dispersion modeling capabilities has been recommended to aid in emergency  planning and
                                                                                  Page 167

-------
 response.4 Such modeling can help to determine the area likely to be affected by a toxic vapor
 cloud and areas that should be evacuated.

        In Rule 1410, the South Coast Air Quality Management District (SCAQMD) requires HF facilities
 in the Los Angeles Basin to coordinate with them to develop systems to transfer HF release
 information quickly in an emergency.  For example, the SCAQMD mandated that the five facilities that
 use anhydrous HF in the Basin link their HF detectors directly to a computer at the SCAQMD to
 identify, monitor, and locate HF releases. This, combined with a dedicated facility phone line to
 SCAQMD, is intended to provide rapid notification of a release. Although the SCAQMD does not have
 response capabilities, this detector arrangement has the effect of involving the SCAQMD in the direct
 information loop concerning HF releases.

 10.2   Industry Programs and Cooperation with the Community

        Industry associations and many individual facilities have been active in attempting to design
 programs to coordinate with community officials and the public.  The Chemical Manufacturers
 Association (CMA) has developed a community-oriented program called the Community Awareness
 Emergency Response (CAER) program. CAER recommends ways for chemical facilities to develop
 working relationships with communities. The following are a few examples of CAER-sponsored
 outreach programs to the community about the hazards of HF:

        >•      Allied-Signal provided to the community a CAER calendar containing a
               map of the facility and emergency routes in the community,
               instructions on how to shelter in place in the event of an HF release, a
               description of the meanings of different emergency sirens, and
               reminders of the calendar dates of siren tests.5

        >-      Champlin Refinery in Humble, Texas involved the community in the
               installation, testing, and inspection of CAER sirens.

        >      The LEPC of La Porte, Texas, posted a highway sign that provides the
               public with a telephone number to call for information on chemical
               hazards in the community. Also in La Porte, Du Pont has installed an
               emergency communication system with the community consisting of a
               dedicated telephone network, emergency broadcast station, and a
               telephone hotline.

        In some industries, facilities have developed mutual aid agreements to respond and assist in
the event of an emergency at any member facility.  CMA has established an HF Mutual Aid Group
made up of emergency response teams with special training, primarily for response to transportation
accidents.  The response teams are activated through CMA's Chemical Transportation Emergency
Center (CHEMTREC), which is a public service established by CMA in 1971 to provide information and
technical assistance to responders during hazardous materials emergencies.6

        Mutual aid agreements may also involve community officials.  For example, the Geismar Area
Mutual Aid group in Ascension Parish, Louisiana, has 24 member chemical companies and provides
much of the funding for materials and training needed by the parish Sheriff's Hazmat Team/
Typically, members of the mutual aid agreements will  also be members of the LEPC.

        Other facilities have developed relationships through direct cooperation with local government
agencies.  For example, the Ultramar Refinery jointly organized and conducted HF release simulations
and drills with the Los Angeles Fire Department. Facilities which have not previously conducted drills
Page 168

-------
can start initially with a table top exercise, and then work up to simulations in cooperation with the
local fire department.

       Because of HP's toxicity, HF producers and shippers have involved hospitals in the
communities' emergency preparedness and response activities. Hospital personnel prepare for
possible emergencies by learning the medical techniques to treat HF exposure injuries, establishing
communication with the facility and community, and participating in mock drills.
HF manufacturers aid in emergency response by providing medical information to medical
professionals and the public during chemical emergencies. CMA recently established  a medical
treatment emergency communication network pilot program in conjunction with the San Francisco Bay
Area Regional Poison Control Center.  In addition, CHEMTREC's 24-hour medical emergency
assistance network provides support to CHEMTREC, CHEMTREC registrants, and medical
professionals for incidents involving exposure to industrial chemicals, including HF.8

        Elf Atochem has a fluorocarbon manufacturing facility located in Wichita,  Kansas that has
conducted drills, exercises, and training with the local government, other facilities, and the
surrounding community in planning for an HF release. The Atochem facility  is adjacent to a facility
owned by Vulcan  Chemical. Because a major HF release could affect personnel in both companies,
Atochem conducts tabletop exercises in conjunction with Vulcan.  Past exercises have simulated an
HF release, associated injuries, off-site monitoring, and medical treatment. To alert facility employees
of an actual emergency, the facility complex is equipped with a warning siren.  Instant alerts have
been installed in homes within a three-mile radius of the complex.  In addition, Atochem has
coordinated with the Sedgewick County Fire Department and the LEPC. Atochem has often assisted
in training the fire department's HazMat team.  In the event of a gas release, the team is capable of
monitoring for dense gas in basements and low lying areas.  Atochem has been well represented on
the LEPC, at one  point serving as LEPC chair. A regional  burn center in the area is also equipped to
deal with victims of HF exposure.9

        Sun Oil in Tulsa, Oklahoma had an HF release in March of 1988. To allay substantial public
fear concerning HF, Sun has conducted an extensive public information campaign.  The LEPC is also
well informed of the hazards of.HF.  Sun trains all local emergency response agencies, including the
fire department, police department, hospital, and emergency medical  units (air and ground)
specifically for an HF release from the facility. Drills are  conducted annually, during which a video of
the 1 988 release is shown. Sun also participates in  exercises with the railroad that transports HF
through Tulsa every week led  by the city  of Tulsa. Smoke is used to simulate the release.
                                                                                10
1 0.3   Ways to Improve Facility Emergency Preparedness and Planning

       Although many facilities have initiated and actively participated in emergency preparedness
and planning activities, Chemical Safety Audits (CSAs) conducted by EPA indicate that the companies
could be doing more. CSAs describe the activities of chemical plants that EPA has reason to believe
may pose risk to the public and recommend ways to improve chemical safety.  Summaries of the
recommendations made by CSAs at two refineries with HF alkylation units, a semiconductor etching
facility, and an aluminum manufacturer are provided below.

       >•      The BP  Oil refinery in Ferndale, Washington, was chosen for a CSA because it
              released hazardous substances nine times in 1989, according to National
              Response Center records.  The audit was conducted March  1 9, 1 990.  The
              refinery  is in an area defined at an earthquake vulnerability ranking of 3.5 out
              of a possible 5 and does not have a forma.1 inspection program for HF storage
              tanks.  BP has conducted emergency drills with the local emergency medical
              services group, but has not scheduled or conducted any exercises with the
              Lummi Tribal Reservation, which is adjacent to the plant and has a population
                                                                                   Page 169

-------
               of 3,200.  The reservation does not have an emergency response plan, and
               the refinery has no public notification capacity in the event of an emergency.
               The CSA recommended that BP install a public notification system and initiate
               an appropriate training program to ensure effective warning to neighboring
               communities.11

        »•      The Motorola semiconductor facility was chosen for a CSA because of its
               location within the densely populated metropolitan area of Phoenix, Arizona,
               and because of the large number and quantity of extremely hazardous
               substances used and stored on site, including HF.  The audit was conducted
               April 9, 1991. Although Motorola has assisted the Phoenix Fire Department in
               purchasing CAMEO software and Macintosh computers, neither CAMEO  nor
               any other hazard analysis method is used by Motorola or the fire department
               for identifying vulnerable zones in the surrounding area in scenarios based on
               the types and quantities of Motorola's chemical inventory.  The CSA
               recommended that Motorola meet with the Phoenix Fire Department and  work
               together with them to run CAMEO prior to an emergency situation.12

        *      EPA chose to audit the Phillips refinery in West Bountiful, Utah, after an
               operator's error resulted in an HF release from the alkylation unit.  The audit
               was conducted May 2, 1989. The facility does not have a representative  on
               the LEPC and has had no contact with this group.  The CSA recommended
               that the facility ensure that it is represented on the LEPC.13

        >      The Columbia Falls Aluminum Company (CFAC) in Columbia Falls, Montana,
               was selected for a CSA because of a release to air of 7,700 pounds of HF.
               The audit was conducted April 30 - May 3, 1991.  The facility is located in a
               potential earthquake and flood zone. Facility emergency plans and training
               sessions have been coordinated with the fire department.  A CFAC
               representative is a member of the LEPC, but at the time of the audit the
               company had not conducted any training or exercises with the LEPC. The
               CSA recommended that the CFAC should consider working more closely with
               local authorities and conduct regular table top exercises to simulate a
               response in the event of a natural disaster.14

Other CSA recommendations have been that facilities initiate training and exercises with LEPCs,
consider using separate siren signals for indicating emergencies and for lunch and quitting time,
review and revise the emergency response plan, and develop an agreement with other local facilities
to coordinate mutual aid.

10.4    Community Efforts to Promote Emergency Preparedness and Planning

        Anecdotal information on community emergency planning and preparedness for HF incidents
has been gathered through contacts with local governments officials, LEPC representatives, and
industry members. The degree of preparedness for a release varies from community to community
depending a number of factors, including the perceived risk of such an incident, the resources
available to respond to it, and the degree of concern demonstrated by the public regarding HF
hazards. Many facilities work closely with local governments to develop emergency plans and
conduct exercises to test response capabilities.

        Delaware County, Pennsylvania, has two HF facilities within its boundaries: the BP refinery and
the General Chemical Corporation. Delaware County Emergency Services officials are well  aware of
the danger posed by HF, have developed a plan to respond to a release, and have conducted table
Page 170

-------
top simulation exercises.  The county developed its plan with the help of an industry advisory council
that brings in experts to discuss the hazards of specific chemicals; including HF.  In addition to the
industry group there is also a citizens advisory group; however, their concerns have focused primarily
on foul odors and soot rather than the threat of an HF emergency.  According to the county's
analysis, the worst-case scenario for an HF emergency would be a rail accident in which a tank car
carrying HF ruptured in a densely populated area.  Releases from the facilities themselves might be
less serious because the facilities are located near a wide stretch of the Delaware River with prevailing
wind direction that would  carry fumes away from populated areas. Officials are confident that because
of their planning, they will be well prepared should an accident occur.15

       The LEPC, whose jurisdiction includes Atochem's main HF production facility in Calvert City,
Kentucky has received funding from industry members and has worked in a close and cooperative
manner with the industries in the area. Although citizens are aware that HF is handled in the area,
there has been little concern directed specifically at HF.  There have been a few citizens voicing
concerns about chemicals in general.

       A table top exercise was conducted to simulate an HF release from the Atochem facility, using
a worst-case release which impacted the nearby GAP facility.  In general, the facility was expected to
address issues that include stopping and mitigating the release, notification of authorities, and the
media; the LEPC coordinated responses external to the facility. Atochem developed the scenario and
made assumptions regarding the quantity released and wind direction.  The LEPC provided situations
to create confusion that Atochem had to address internally in  its response actions. Although the
simulation was considered a low probability event, worst-case incidents are used for planning
purposes.  Industry  personnel, fire,  police, emergency medical, and local citizens participated in the
exercise. In-place sheltering and evacuation were emphasized.16

       There have been some problems, however, in coordinating facility and community emergency
response efforts.  During the site visits EPA conducted to HF facilities to gather information for this
study, many facilities mentioned the difficulty of maintaining or attracting public interest in learning
about the hazards of HF.  Some facilities indicated a sense of frustration that the public is more
concerned about nuisances such as noise and  odors rather than about the hazards of an HF vapor
cloud.  Usually, however, until a release occurs  that directly affects the public in an area, it is difficult
to raise public consciousness regarding chemical hazards.  Also, the residents who live near facilities
are used to living with general chemical hazards and are to some extent unconcerned.  For instance,
both the Allied-Signal HF manufacturing facility in Geismar, Louisiana, and the BP oil refinery in Belle
Chasse, LA, invited the local community to tour their facility (including the HF  processes) and
encouraged them to ask questions.  However, attendance was poor, in part because of the small
population  around the facilities.

       Also, LEPCs charged with developing emergency response plans may be unaware of and
therefore fail to address HF hazards posed to the community. Many LEPCs lack the resources or
manpower to handle the overwhelming workload of developing emergency plans for many hazardous
chemicals.  Consequently, they rely on the best judgment of the facility. Still others may not recognize
the serious hazards posed by HF.          *
                                                                                     Page 171

-------
                                         ENDNOTES
1.     Memorandum, Subject: OSC Report to the National Response Team Major Air Release of
       Hydrofluoric Acid Marathon Petroleum Company Texas City, Galveston County, Texas -
       October 30 to November 1, 1987, From:  Robert M. Ryan, On-Scene Coordinator, U.S.
       Environmental Protection Agency Region IV,  To: National Response Team, March 4, 1988.
       (370)
                                                               "V
2.     Allied-Signal Inc., HF Product Group, Hydrogen Fluoride Production, Geismar Plant,
       Community Relations/Awareness Programs, Geismar, LA, February 1992.  (147.1c)

3.     Motorola, Inc., Chemical Safety Audit Profile, Discrete and Materials, Phoenix, AZ, April 9, 1991.
       (116.3)

4.     Memorandum, Subject: OSC Report.

5.     Allied-Signal Inc., HF Product Group, Hydrogen Fluoride Production, Geismar Plant,
       Community Relations/Awareness Programs, Geismar, LA, February 1992.  (147.1C)

6.     Strickland, Gordon D.,  Chemical Manufacturers Association, comments from technical review
       of Hydrogen Fluoride Study, Report to Congress, Draft  May 8, 1992, June 5,  1992. (466.9)

7.     Allied-Signal Inc., HF Product Group.

8.     Strickland, Gordon D.,  Chemical Manufacturers Association, comments from technical review
       of Hydrogen Fluoride Study, Report to Congress, Draft  May 8, 1992, June 5,  1992. (466.9)

9.     Persona! Communication, Conversation with Larry Masters, Sedgewick County LEPC, Wichita,
       KS, April 17, 1992.  (328.61)

10.    Personal Communication, Conversation with Mike Linville, Tulsa County LEPC, Tulsa County,
       OK, April 14, 1992. (299.95)

11.    Chemical Safety Audit Profile, British Petroleum Company, Ferndale, WA, March 19, 1990.
       (116.2)

12.    Chemical Safety Audit Profile, Motorola, Inc., Discrete and Materials, Phoenix, AZ, April 9, 1991.
       (116.3)

13.    Chemical Safety Audit Profile, Phillips 66 Company Woods Cross Petroleum, West Bountiful,
       UT, May 2, 1989. (116.4)

14.    Chemical Safety Audit Profile, Columbia Falls Aluminum Co., Columbia Falls, MT, April 30,
       1991. (76.8)

15.    Personal Communication, Conversation with George Morgan, Delaware County  Emergency
       Services, Delaware County, PA, April 8, 1992. (343)

16.    Personal Communication, Conversation with Attabelle Willie, Marshall County LEPC, Marshall
       County, KY, April 14, 1992. (506.5)
Page 772

-------
                     11.  FINDINGS AND RECOMMENDATIONS
        This chapter contains the findings and recommendations about the hazards of hydrogen
 fluoride (HF).  They were derived from the technical information gathered about HF, site visits to
 facilities where HF is handled, discussions with a variety of stakeholders, and information from
 technical reviewers.

 11.1   Findings

        The primary purpose of the findings is to characterize the results of the HF study. The
 Agency provided substantial opportunity for comment by stakeholders in this study. Many of the
 findings are simple statements of the facts about HF.  In addition, findings related to the broader
 issues associated with chemical hazards, risks and risk management are included so that the findings
 about HF are  kept in perspective.

        11.1.1 Summary Findings

        HF is  used industrially in large quantities throughout the United States (over 200,000 tons per
 year) and in a great number of applications across a broad range of industries (over 500 facilities).  It
 serves as a major feedstock and source of the fluorine molecule for the production of fluorinated
 compounds.

        An accidental release of HF from one of these industrial facilities could have severe
 consequences.  HF is toxic to humans, flora, and fauna in certain doses and can  be lethal as
 demonstrated by documented workplace accidents.  HF can travel significant distances downwind as
 a dense vapor and aerosol under certain accidental release conditions. Because HF can exist as an
 aerosol, the cloud can contain a substantially greater quantity of the chemical than otherwise would
 be the case. Thus, the potentially high concentration of HF in these dense vapor and aerosol clouds
. could pose a significant threat to the public, especially in those instances where HF is handled at
 facilities located in densely populated areas.  Prompt and specialized medical attention is necessary to
 treat HF exposure properly.

        However, the risk to the public of exposure to HF is a function of both the potential
 consequences and the likelihood of occurrence of  an accidental release; and the  likelihood of an
 accidental release of HF can be kept low if facility owners/operators exercise the general duty and
 responsibility to design, operate, and maintain safe facilities.  In particular, owners/operators can
 achieve an adequate margin of protection both for their workers and the surrounding community by
 assiduously applying existing industry standards and practices, existing regulations, and future
 guidance and regulations  applicable to various classes of hazardous substances in various settings.
 The properties that make HF a potentially serious hazard are found individually or in combination in
 many other industrial chemicals; thus, HF does not require unique precautions. Instead, within each
 of the several  different circumstances in which HF is handled,  an  appropriate combination of general
 and special  precautions should  result in: (1) the safe management of HF and other hazardous
 substances with an emphasis on accident prevention; (2) the preparedness to properly and quickly
 respond to chemical emergencies and to provide specialized medical treatment if  necessary; and (3)
 community understanding of the risks involved.


        11.1.2 Risk of HF to the Public

       This section addresses the findings concerning the risks of HF  to the public. Consequences
 and likelihood are individually addressed in subsequent sections.
                                                                                    Page 173

-------
•      Risks from a major accidental HF release may vary from facility to facility depending on site-
       specific factors such as the density and distribution of nearby populations and the quality of
       process safety management and risk management practiced at the facility.  Some facilities
       present greater risk than others.

•      Since risk is a product of both consequences and likelihood, judgements of risk should not be
       made solely on the  basis of the analysis of consequences. Thus, risk reduction must take
       both consequence and likelihood of occurrence into account. Actions taken based on
       consequence analysis alone, without consideration of likelihood of occurrence, could lead to
       increased risk or to  transfer of risk. For example, installation of equipment intended to reduce
       accident hazards may require additional connections (e.g., welds or gaskets).  These
       connections represent additional locations where failure could occur.

•      Similarly, the risks associated with an alternative chemical or process  must be assessed and
       compared to HF process risks before any decision about the alternative can be made. Thus,
       the risks of sulfuric acid alkylation technology in petroleum refineries must be assessed and
       compared to the risks of HF alkylation technology before any decision is made to require a
       switch from one technology to another.  Such an assessment must include hazards and risks
       associated with sulfuric acid manufacture, transportation, use, and regeneration. Although the
       uses and hazards of sulfuric acid were not  analyzed in this study, some issues were identified:

               Although sulfuric acid has a much higher boiling point than HF and is not likely to
               vaporize or to generate an aerosol  upon accidental release at normal alkylation
               conditions, other situations and scenarios where sulfuric acid  could be vaporized or
               generate sulfur dioxide or sulfur trioxide releases (such as in the acid regeneration
               process) must be considered in a comprehensive assessment.

               Sulfuric acid alkylation involves significantly greater transportation requirements and
               associated transportation hazards where regeneration capacity is not available on site
               or via pipeline due to the much larger acid supply requirements.

               Conversion  from one alkylation technology to another requires consideration of factors
               such as differing equipment, differing catalyst performances, and the potential physical
               space limitations in some refineries.

       11.1.3  Consequences of an HF Release

       This section addresses the consequences of a range of accidental HF release scenarios,
including worst case scenarios as required by the Clean Air Act Amendments. A fundamental part of
risk assessment is an  assessment of the consequences of accidental releases.

       Physical/Chemical  Properties and Hazards

       Consequence assessment requires an understanding of a number of  factors including the
physical/chemical properties and hazards of the chemical and its behavior upon release. The
following sections address these issues.

•      The physical/chemical properties of HF are complex but are not unique; several other
       industrial chemicals exhibit similar properties and hazards.

               HF is toxic and is mandated for inclusion in the Clean Air Act  section 112(r)(3) list. It
               is also an extremely hazardous substance (EHS) under the Emergency Planning and
               Community-Right-to-Know Act (EPCRA).
Page 174

-------
       Possible immediate acute effects of HF exposure depend on the dose and exposure
       route (i.e., inhalation, dermal contact, ingestion) and can be as severe as death or
       pulmonary edema or as minor as mild skin irritation or no effect at all.

       As an acid, HF can be very corrosive to common metals in the dilute, aqueous form.
       The use of anhydrous HF typically  does not require exotic metallurgy when moisture
       levels are properly controlled.

       Anhydrous HF has a boiling point of 67°F and will vaporize and become airborne if
       released above this temperature at atmospheric pressure. Under certain conditions,
       and particularly if released at temperatures below its boiling point, HF will form a liquid
       pool with vaporization at a much slower rate than at temperatures above the boiling   .
       point.

       HF vapor may exhibit dense gas behavior when released although it has a molecular
       weight and density less than that of air in its dissociated state.  When concentrated
       HF is released, the HF molecules in the vapor may be in an associated state (bound
       together as molecular groups) with a density greater than air. As the vapor entrains
       surrounding air, the molecular groups tend to dissociate to individual molecules of HF.
       Dissociation cools the plume which increases the density of the plume.  The property
       of density greater than air presents a potential hazard because dense plumes
       resulting from a release tend to move at ground level where exposure is most  likely  to
       occur.

       Substances liquefied under pressure at temperatures above their atmospheric boiling
       point may initially form aerosols (mists of liquid droplets) when released. The  aerosol
       vaporizes as the plume travels downwind.  Anhydrous HF may be handled under
       these conditions (chlorine and ammonia are other examples).  Cooling of the vapor
       also occurs as a result of expansion during release. The hazard presented  by an
       aerosol is the increased mass rate of release to atmosphere presented  by the liquid
       droplets as compared to a gaseous vapor.  The higher mass rate of release to
       atmosphere can generate greater downwind impacts.

       HF is soluble in water and heat is generated as HF dissolves in water.  The  buoyancy
       of an HF vapor cloud is increased  by the thermal effects of interaction between HF
       and moisture in the air. This property directionally counteracts the density effects of
       HF dissociation.

       Anhydrous HF poses a greater atmospheric release hazard than aqueous solutions of
       HF.  Even at ambient temperatures above the boiling point of HF, aqueous solutions of
       HF form liquid pools when released although the solutions may fume to a degree
       determined by the temperature and concentration of HF.   HF in aqueous solution will
       vaporize at a much slower rate than anhydrous HF.

Acute HF exposure (inhalation, dermal contact, or ingestion) requires prompt and specialized
first-aid and follow-up medical treatment. Some of these procedures are unique for exposures
to HF.

Actual and Potential Impacts on the  Public

HF in its  anhydrous and aqueous forms is used industrially in large quantities and by  a large
number of facilities.  HF is the primary source of fluorine for fluorinated chemicals and it is
                                                                             Page 175

-------
        used in the manufacture of a wide variety of products.  Products include refrigerants,
        electronics, gasoline, detergents, and drugs.

               U.S. annual capacity for HF production was 206,000 tons in 1992.

               Over 500 facilities reported handling HF in the 1990 Toxic Release Inventory. This
               number represents only a fraction of the total number of facilities in which HF is
               handled.

        Facilities that handle HF are located in both urban and rural areas.  Populations around
        facilities handling more than 10,000 pounds of HF can range from 0 to 24,000 within 1 mile
        and can range from 0 to 550,000 within 5 miles of such facilities.

        If accidentally released to the air in sufficient quantities and *at sufficient rates,  HF presents a
        potential off-site threat to the life and health of the exposed public.

               Actual impacts on the public in the U.S. from accidental HF releases have been rare
               and have been limited to injury (ranging from mild irritation to effects requiring hospital
               observation). No off-site fatalities have been known to result from an HF release.

               Two events were documented in this study where members of the public were injured
               as a result  of exposure to HF.  In one incident, a construction accident in a shutdown
               refinery alkylation unit resulted  in 1,037 persons reporting to hospitals  with
               approximately 100  admitted for treatment. In the second incident, a leaking rail car
               resulted in  71 injuries among the public.

               12 events were documented in this study  where the public was  evacuated as a result
               of an accidental HF release.  Some of these evacuations were precautionary
               measures.

        A variety of worst-case accidental release scenarios show that HF could result in doses
        equivalent to the NIOSH IDLH  beyond 10 km from the point of release.  The AIHA ERPG-3
        may also be exceeded for significant distances. The worst-case scenario is useful to facilities
        and to the community surrounding facilities to gain an understanding of the potential
        magnitude of severe situations and should be taken into account along  with more probable
        scenarios when setting priorities for community emergency planning.  Note, however, that the
        worst-case is designed to generate the maximum  impact off-site and is considered to be
        extremely unlikely.  The accidental release is based on situations such as catastrophic vessel
        failure or other scenarios where release rates are very high. These  scenarios may also
        include worst-case  meteorological and dispersion  conditions.  These situations and conditions
        do not take into account a variety of factors that can significantly alter the outcome of the
        downwind impacts.

               The worst-case scenario does not take into account the  role of process safety
               management in reducing the probability of loss of containment.

               The worst-case scenario does not take into account passive mitigation (such as diked
               areas or spill containment) or active mitigation (such as water spray and rapid de-
               inventory systems)  that can  reduce the amount released into the air.

               The cloud is assumed to travel  over smooth terrain. Actual  terrain may be complex
               and obstacles may be present that affect the path and dispersion of the cloud.
Page 176

-------
              Worst-case meteorological conditions may not persist for long time periods or over
              long distances.  Dispersion at more typical meteorological conditions significantly
              reduces the magnitude of the impact.

              The dose that is actually received is uncertain and may be reduced or avoided by
              sheltering-in-place or evacuation.

       An assessment of a range of accident scenarios that are judged to be more likely to occur
       than catastrophic vessel failure or other scenarios with very  high release rates shows a wide
       range of impacts that are highly sensitive to scenario conditions.  Large, uncontrolled releases
       (e.g., transfer hose failures, pipe ruptures) appear to lead to significant downwind impacts,
       while small, quickly controlled and mitigated releases (e.g., leaks of small magnitude or short
       duration) show little or no off-site impact. These results vary widely depending on the rate of
       release, the speed at which the release is controlled and mitigated, the  dispersion conditions,
       and the actual dose received by the public.  Although these results may be judged to be
       imprecise and very uncertain predictions of what might actually occur in a release, they are
       extremely useful for comparison across scenarios to demonstrate the significant value of
       management to prevent large releases, and rapid control and mitigation of all accidental
       releases.

       Atmospheric Dispersion Modeling

       Atmospheric dispersion models are used to predict the downwind consequences of accidental
releases. Models provide a way to estimate mathematically the dispersion behavior of a released
substance in the atmosphere.  Some are relatively simple and easy to use while others are complex
and sophisticated.

•      Because HF demonstrates complex behavior upon accidental release (e.g., association,
       dissociation, reaction with water, aerosol formation), accurate dispersion modeling requires
       sophisticated modeling techniques that can account for this behavior. Even with the most
       sophisticated models, results from modeling efforts are only gross approximations of actual
       conditions.  Precise prediction of downwind concentration from an actual  release is unlikely for
       reasons such as:

              Assumptions about release rate may not reflect what actually occurs  in a real event;

              The effects of surface roughness and obstacles in the path of the cloud that influence
              dispersion are  not precisely known although the models used assume greater
              dispersion as a result of complex terrain; and

              Variability of  meteorological conditions  and exposure level affect the result.

       Dispersion models including SLAB and HGSYSTEM have been found to provide data
       consistent with HF cloud concentrations measured from controlled field tests of HF within a
       factor of five and  sometimes within a factor of two.  This difference indicates  a lack of
       precision.  However, model results are useful for general predictions, development of process
       safety management efforts, and for comparative assessments between scenarios.

 •      The calculated results from the same scenario  using different models are slightly different.
       This reflects the differing degree to which models account for the complex factors which affect
        releases and atmospheric dispersion of released materials.
                                                                                      Page 177

-------
                    Results are extremely sensitive to assumed release rates (which relate to the accident
                    scenario) and assumed atmospheric conditions (e.g., wind speed, atmospheric stability).
                    Atmospheric conditions such as F stability may not occur for long periods of time or for long
                    distances.

                    The behavior of dense vapor plumes of HF, or other hazardous materials, the detection and
                    mitigation of accidental releases, and the effects of exposure to toxic substances are
                    continuing subjects of research interest and activity.

                          The influence of complex terrain (the presence of obstacles such as facility and off-site
                          structures) has not been adequately determined. It is possible that obstacles and
                          certain terrain could inhibit dispersion rather than increase it.

                          The levels used to predict the onset of toxic effects are uncertain.  Research continues
                          to determine appropriate acute exposure levels for emergency planning.

                    11.1.4 Likelihood of an HF Accident

                   The likelihood of an accidental  release can be reduced through appropriate control measures.

                    Prevention of Accidental Releases

                   Accidental releases of HF can be prevented by application of process safety management
                    principles.  The following are among the ways that these principles are adopted:

                          Under the Clean Air Act Amendments, industry has a responsibility to identify hazards,
                          take the actions necessary to prevent chemical accidents, and to take action to
                          mitigate accidents in the event they do occur.

                          OSHA has promulgated a process safety management standard that requires facilities
                          to implement process safety management programs for chemicals including HF to
                          protect workers from chemical accidents. These same measures can also prevent
                          chemical accidents that might affect the public.

                          Under the Clean Air Act Amendments, EPA must promulgate rules that require facilities
                          handling HF to implement a risk management plan designed to prevent chemical
                          accidents that adversely affect the public.

                          Voluntary initiatives (e.g., codes, standards, recommended practices) such as the API
                          recommended practices for process safety management and HF alkylation, CMA's
                          Responsible Care program, and the adoption of standards more stringent than
                          regulatory requirements for transport vessels have been implemented by some
                          facilities.

                          Four states (New Jersey, Delaware, Nevada, and California) have adopted accident
                          prevention requirements for many chemicals including HF.

                   Research and development efforts are currently underway with one goal  of modifying
                   petroleum refinery HF alkylation systems such that accidental  releases do not present a
                   potential threat to the public.

                          Additives to reduce the  volatility of liquid HF are being developed. Limited testing has
                          been conducted.
_
           Page 178

-------
              Development of a solid alkylation catalyst is underway and is also in the pilot plant
              phase.

       Detection and Mitigation

       If an accidental release occurs, both modeling and actual experience demonstrate that rapid
detection of a release and mitigation or control of the release can minimize the potential
consequences.

       Several technologies to detect HF releases exist. The concept has been included in South
       Coast Air Quality Management District (SCAQMD) regulations for petroleum refineries and
       refrigerant production facilities in southern California and endorsed as a recommended
       practice by the API for refinery HF alkylation units.

       Mitigation technologies have been field tested, endorsed as recommended practices by the
       API for HF alkylation units in petroleum refineries, and required as an interim measure by the
       SCAQMD for certain petroleum refineries and refrigerant production facilities.  Properly
       designed and  operated water spray mitigation systems have been shown in field tests to
       significantly reduce the amount  of HF that travels downwind. Known technologies include:

              Water spray and  deluge systems at the perimeter of an HF unit or particular
              equipment.
              Rapid transfer systems to remove the contents of a leaking vessel.
              Remotely or locally controlled valve systems to segregate or to isolate leaking parts of
              a process.
              Remotely- and locally- controlled equipment to shut down or isolate affected
              equipment.

        In spite of the availability of detection and mitigation measures, all facilities have not uniformly
        adopted such measures.  In addition, the reliability of such equipment and the site-specific
        conditions must be considered  before particular detection or mitigation systems are adopted
        or implemented.

        Preparedness and Response

        Comprehensive risk management must include planning and preparation for, and response to,
 chemical emergencies.  Successful chemical accident prevention, preparedness, and response
 programs for HF and  all other hazardous substances require the active participation of all
 stakeholders (e.g., workers, community, first responders, and industry).

        Facilities handling HF are subject to the emergency planning requirements of the Emergency
        Planning and Community Right-to-Know Act (EPCRA).

        Not all Local Emergency Planning  Committees (LEPCs), HF facilities, or the public surrounding
        facilities handling bulk anhydrous HF or along anhydrous HF transportation routes are aware
        of, have planned for, and are prepared to deal with an HF emergency. The public is generally
        not aware of the hazards and risks existing at facilities that handle hazardous substances,
        including HF. In cases where this is true, the public would not  know the proper protective
        actions to take in the event of a release.


               This finding is based on contact with a limited number of LEPCs in areas where
                facilities handling HF are located.
                                                                                      Page 179

-------
                Some, but not all facilities and LEPCs, have determined vulnerable zones, performed
                planning, and conducted table-top exercises and drills associated with an HF disaster.

                Contact with a limited number of facilities indicates that few facilities have attempted to
                communicate risk to surrounding residents.

  •       Many facilities claim to be prepared to treat cases of acute exposure to HF or to arrange for
         such treatment.

                A limited number of facilities have indicated that they have developed in-house
                capability and/or worked with  local hospitals and clinics to ensure the capability to
                treat acute HF exposure.

  •       The accidental release prevention provisions of the Clean Air Act Amendments will require
         facilities handling HF above threshold  quantities to implement an emergency response
         program.


  11.2    Recommendations

        The Clean Air Act Amendments require that the EPA make recommendations to the Congress
 for the reduction of the hazards of HF in industrial and commercial applications, if appropriate. This
 section contains recommendations to the Congress and to industry regarding the reduction of such
 hazards. Recommendations for further EPA efforts are also included. These recommendations are
 based on the findings about HF hazards and the technical information gathered about HF during the
 course of the study.

        11.2.1  General

 •      The EPA does not recommend legislative action from the Congress at this time to reduce the
        hazards associated with HF. The regulations already promulgated, and being developed,
        under the authorities provided to EPA  in  CERCLA, EPCRA, and the accidental release
        prevention provisions of the CAM, and to OSHA in the process safety management
        provisions of the CAAA, provide a good framework for the prevention of accidental chemical
        releases and preparedness in the event that they occur.

               EPA should continue its Chemical Safety Audit program and conduct audits at HF
               facilities that have not been audited. Previously audited sites should be contacted to
               learn what improvements have been made, if any, since the initial audit.  EPA should
               track implementation of current and future industry standards and recommended
               practices at HF facilities to determine which industry sectors comply with the
               standards.  EPA should consider outreach specifically directed at non-participating
               sectors.

               EPA should investigate any chemical accidents associated with HF that cause or have
               the potential to cause public impacts in order to determine the root cause of such
               accidents. This information can be used to determine whether current practices are
               adequate.  Such investigations should be coordinated with the OSHA to encompass
               worker safety issues.

               EPA should continue to investigate the need for additional rulemaking under the CAAA
               to require implementation of certain prevention, detection,  monitoring, and mitigation
               efforts at facilities  where extremely hazardous substances  (such as HF) could generate
Page 180

-------
       vapor clouds and aerosols and travel off-site. The level of voluntary industry initiatives
       and degree of participation, the results of Chemical Safety Audits, and accident history
       should be taken into account.

EPA should encourage dialogue on the implications of the findings of this study for industry,
emergency response organizations, and the public.  Information should also be shared via
domestic and international presentations on the safety and research issues associated with
HF.

       EPA will distribute this study widely to facilities, States, and localities where anhydrous
       HF is used. The EPA has already distributed an advisory to LEPCs to alert them to
       the hazards and process safety management issues associated with HF.

       EPA will share this study with other countries, particularly Mexico and Canada, for their
       use related to HF handling in their country and for border-related emergency issues.

HF producers should  practice product stewardship and should assure that customers are
safely handling HF, are trained on the hazards of HF, and are taking steps to prevent
accidental releases and to promptly respond and mitigate releases that do occur.

11.2.2 Facility and LEPC Coordination

Facilities which handle hazardous substances that could form dense vapor clouds or aerosols
if accidentally released, such as  anhydrous HF, should work closely with their LEPC to prevent
accidents and to be prepared to respond to such accidents.

        Facilities should identify and thoroughly understand the hazards and conditions that
        can lead to accidental releases and the potential impacts on the public. These
        hazards and potential impacts should be communicated to the LEPC.

        EPA should work with fluorocarbon producers, HF manufacturers, and petroleum
        refineries that use anhydrous HF in populated  areas and their associated LEPCs to
        offer assistance for prevention, preparedness, and response.

        All facilities that handle HF and the LEPC for that area should conduct drills and
        exercises with workers, the  community, first responders and others to test mitigation,
        response and medical treatment for a simulated major HF accident.  All facilities
        handling HF should have training programs and procedures in place for HF
        emergencies.

 Facilities that handle  HF should  actively conduct outreach efforts to ensure that the community
 is aware of the hazards of HF, that  protective measures are in place to prevent public health
 impacts, and that proper actions will be taken during an emergency.  Such outreach should
 be  conducted through the LEPCs.

        All facilities that handle anhydrous HF should be able to rapidly detect, mitigate, and
         respond to accidental releases in order to minimize the  consequences (e.g., through
         detection, monitoring, mitigation, and alert or alarm systems). Site-specific risk factors
         should be taken into account.

         EPA should support the development and continued refinement of release detection
         and mitigation systems for  hazardous substances, such as HF,  in .order to improve
         their reliability and effectiveness.
                                                                               Page 181

-------
                All facilities that handle any form of HF should have proper medical treatment supplies
                and trained personnel available and should ensure that first responders, hospitals, and
                clinics in the area are prepared to treat HF exposure.

         EPA should coordinate with industry and others to determine which other toxic substances
         generate aerosols at certain conditions upon accidental release and to communicate this
         hazard information to users and LEPCs.

         11.2.3  Research and Further Studies

         Further study on the acute exposure levels of HF that result in irreversible health effects or
         lethality in humans should be conducted in order to improve emergency planning tools such
         as atmospheric dispersion models.

         EPA should monitor alkylation catalysis and HF additive research for potential process safety
         improvements.

         With EPA participation, industry should modify software for dense gas atmospheric dispersion
         models such as HGSYSTEM for HF to make them more user-friendly and more broadly
         disseminated to facility owner/operators.

         Further research on the effects of surface roughness and obstacles on dense-gas dispersion
         behavior should be conducted to determine their influences on toxic substance concentrations
         in a  dispersing vapor cloud. The Liquefied Gaseous Fuels Spill Test Facility should be used
         for spill tests to assist in this research.

         EPA should continue to study the issues surrounding worst-case releases, their consequences
         and the likelihood of worst-case or other significant releases for extremely hazardous
         substances and the role and relationship of these issues to prevention, preparedness, and
         response;
Page 182

-------
                                      APPENDIX]
                  SUMMARY NOTES FROM HF ROUNDTABLE
Introduction

       Under the Clean Air Act Amendments of 1990, the Environmental Protection Agency (EPA) is
required to "complete a study of the industrial and commercial applications of hydrofluoric acid and to
examine the potential hazards of hydrofluoric acid and the uses of hydrofluoric acid in industrial and
commercial applications to public health and the environment considering  a range of events including
worst-case accidental releases and shall make recommendations to the Congress for the reduction of
such hazards, if appropriate." The study must be submitted to Congress by November 15, 1992.
                         t
       EPA believes that the study must reflect input from those individuals and organizations with a
"stake" or interest in its outcome. Such stakeholders include environmental groups, labor, industry,
trade associations, professional societies, and state and federal government agencies.  Consequently,
EPA organized a "Roundtable"  meeting with individuals representing these interests. The goals of the
Roundtable were to:

       >      discuss  key issues raised by EPA in its initial effort to begin the study;

       c      solicit input on additional issues important to stakeholders;

       >      solicit input on EPA's approach to the study and the ways to address critical issues;
              and

       ••      solicit input to the issues and background information on hydrogen fluoride (HF) and
              to establish peer reviewers for the study.

       Each meeting attendee briefly discussed his or her view of the purpose of the study, what the
study should attempt to accomplish and how. The issues raised by each  attendee were documented
and discussed along with issues previously raised. Listed below are the issues and the results that
were generated at the meeting.
 General Results
               EPA believes the Roundtable meeting was quite successful. Some unresolved
               questions about release scenarios, consequences, probability, and risk may need to
               be addressed in future meetings with stakeholders.

               EPA needs to communicate the real issues associated with HF to dispel
               misconceptions and to place the hazards of HF in the context of other hazardous
               materials.

               Participants indicated that the study should also account for international efforts with
               respect to HF and not just U.S. interests.

               Additional stakeholders should be contacted and included in the study such as the
               Federal Emergency Management Agency (FEMA), the Nuclear Regulatory Commission
               (NRC), the  Department of Transportation (DOT), the Department of Energy (DOE),
                                                                                    Page 1-1

-------
                State Emergency Response Commissions (SERCs), Local Emergency Planning
                Committees (LEPCs), City Managers, and Air Quality Management Districts. The
                public and workers also should be considered stakeholders.

 Study Approach Results

         Feedback from stakeholders included:

         »•       All agreed that while discussion of HF hazards should be addressed in the context of
                other hazardous substances, the study should not get bogged down in general
                hazardous chemical issues such that the focus on HF is lost.

         >       Participants had conflicting viewpoints about how the report should address
                alternatives to HF, such as sulfuric acid.  Some believed that sulfuric acid should be
                examined in detail whereas others supported the HF-only approach.  EPA concluded
                to address all of the safety and handling  issues involving sulfuric acid, but to present
                them in a summary.

         >       Participants pointed out that HF has  many industrial applications - some  more
                hazardous than others. For example, uses of HF include etching glass, making
                Pharmaceuticals, and making integrated circuits.  All agreed that while the Clean Air
               Act requires EPA to address all concentrations and states of HF and all uses and
                issues associated with HF, certain "cuts" could be made using criteria. For example,
                aqueous solutions of HF should be discussed in the report but their importance
               should be discounted because scientific evidence indicates little catastrophic potential
               from aqueous HF.

        »•      EPA should not attempt to perform risk assessments itself because of time and
               resource limits.  Instead, EPA should use the results of assessments already
               performed by industry.  In addition, participants indicated that the model facility
               approach for assessment of release scenarios, consequences, and risks is not feasible
               because each site is unique. The model facility concept, however, is useful in
               describing successful release prevention practices and process safety management.

        >      Participants indicated that consequence analyses are critical in characterizing the
               potential impact from accidents at a facility and in examining how hazards are
               handled. EPA should review these consequence analyses of credible worst-case and
               near worst-case scenarios.

        >      EPA should not include a cost/benefit analysis in the Report to Congress,  however,
               EPA should discuss the important role of cost/benefit analysis in making decisions
               about risk reduction alternatives.

        »•      Participants indicated that the study should stress the importance of both worker
               safety and prevention.

        >      EPA needs to examine management practices and facility design in characterizing the
               safe production or use of HF.
Page 1-2

-------
Future Steps
              Obtain information on various HF industry segments from the Roundtable participants.

              Obtain information on current research areas such as the toxicology of HF, accidents
              involving HF, and the transportation and storage of HF.
                                                                                       Page 1-3

-------
                                   This page intentionally left blank.
Page 1-4

-------
                                     APPENDIX II
  SUMMARY OF COMMENTS FROM TECHNICAL REVIEWERS ON DRAFT
                          HYDROGEN FLUORIDE STUDY
       EPA developed a draft Hydrogen Fluoride Study report (May 8, 1992) that was sent to
representatives of environmental groups, labor, industry, trade associations, professional societies,
academia, and state and local government agencies for technical review. This draft did not include
findings and recommendations.  The intent of the review process was to ensure that technical
information in  EPA's report is complete and accurate and to provide EPA with an adequate technical
basis for developing findings and making recommendations to Congress. About 100 draft review
copies of the report were mailed out and forty-four reviewers provided EPA with information,
recommendations for revisions to the document,  and general comments. The reviewers who provided
comment are listed in Exhibit 1 following the text of this document.

       EPA appreciates the efforts of the reviewers, who provided much useful information and many
helpful comments.  The Agency has considered these comments and has revised the Hydrogen
Fluoride (HF) Study based on many of the comments provided during the technical review. This
appendix discusses comments made on the study and EPA's revisions based on the comments.  The
discussion is organized by chapter of the HF Study.  Major technical comments are noted; however,
this document is not intended to be a complete compilation of all comments and revisions. EPA
thanks reviewers who suggested editorial changes or pointed out typographical errors.  Corrections
have been made to the study based on these comments, but they are not specifically noted in this
document.

Chapter 1 - Introduction

       Several reviewers recommended that a clearer distinction be made, in the introduction and
throughout the HF Study, between aqueous and anhydrous HF. EPA agrees that this distinction
should be clearly made and has revised the study, including the Introduction,  accordingly.

       EPA has eliminated the list of questions included in the Introduction, based on comments by
reviewers. These questions were related to thought-provoking issues brought up at the stakeholders
meeting;  however, they do not accurately reflect the topics covered in EPA's report. The Introduction
now emphasizes those issues that are the focus of the report.

       Several reviewers provided useful information about the Goldfish studies of HF releases; EPA
also has incorporated this information into the Introduction.

Chapter 2 - Properties and Hazards of HF

       2.1 Description of Physical and Chemical Properties

       Based on information from reviewers, EPA has revised its discussion of the corrosive
properties of HF; EPA has emphasized the distinction between aqueous and anhydrous HF in this
revised discussion, noting that aqueous, rather than anhydrous, HF is corrosive to a number of
common industrial materials. The use of carbon steels for aqueous HF in concentrations of 70
percent or greater and chlorobutyl rubber-lined equipment for aqueous HF up to 70 percent
concentration is mentioned in the document, as recommended by reviewers.
                                                                                 Page 11-1

-------
       The discussion of the reactivity of HF has been modified to note HP's similarity to other acids.
In addition, based on comments by reviewers, the discussion of adding water to HF has been
expanded to include information on use of water in large excess as a mitigation agent.

       2.2 Health Hazards

       Several reviewers provided additional information or clarification concerning the health hazards
of HF. Data provided on HF concentrations that cause eye irritation and breathing difficulty have been
incorporated into the document.

       EPA has modified its description of the Galveston study of health effects from an HF release,
based on comments by reviewers. The text has been corrected to note that subjects were interviewed
once after the release as part of the exposure study and again for the symptom and  disease
prevalence study. In addition, EPA has included a brief discussion of questions  raised by reviewers
regarding the validity of the Galveston study, including the lack  of a  medical baseline, disagreement
over definitions, and questions about causes of symptoms.

       Several reviewers questioned the presentation  of Immediately Dangerous to Life and Health
(IDLH) levels and Emergency Response Planning Guidelines (ERPGs) for HF and other substances.
EPA has clarified the discussion of these guideline levels and has changed the exhibit from a bar
graph display to a table, as suggested by reviewers.

        Comments were provided on the discussion of probit equations in Chapter 2; detailed
comments were provided on Appendix III on probtts. These comments were very helpful and were
used as a basis for revisions. EPA has noted in the text that the same uncertainties that apply to
probit analysis would apply to any analytical methods that use toxicity data for HF or other toxic
substances.

       EPA has modified its discussion of skin  and eye contact with HF based on comments by
reviewers, noting that HF should be diluted and rinsed from the skin, followed by additional treatment.
The effect of exposure to HF vapor is discussed.

       The discussion of recommended medical treatment for HF exposure has been modified to
incorporate several comments.  The draft document erroneously stated that quaternary ammonium
compounds should be injected; as pointed out by reviewers, quaternary ammonium compound
solutions are used to treat exposed skin by soaking. EPA has incorporated into  the document
additional details regarding treatment with calcium gluconate; e.g., treatment can involve gel
application of calcium gluconate or injection of a 5  percent solution.

       2.3 Environmental Hazards

       As recommended by reviewers, EPA has noted that the effect of HF on aquatic and terrestrial
life depends on exposure concentration. The addition  of fluoride to drinking water to prevent tooth
decay is noted.

       2.4 Release Characteristics

       Based on comments by reviewers, the discussion of the behavior of HF upon release has
been modified and clarified.  It is noted that HF  spilled  as a liquid  will evaporate at a  rate that depends
on release temperature and atmospheric conditions, and as the liquid pool evaporates, the  pool
temperature will drop, causing a decrease in evaporation rate.
Page 11-2

-------
       The description of the Goldfish test series has also been modified based on comments.  The
text now notes that the conditions of the tests approximated HF alkylation unit operating parameters
and that,  upon release, about 20 percent of the HF vaporized and about 80 percent formed HF/water
vapor aerosol. Additional references have been cited for the Goldfish tests, as recommended by
reviewers.

Chapter 3 - Characterization of Hydrogen Fluoride Industry

       3.1 Production of HF

       Reviewers provided several corrections and additions to data on HF production and on
producers identified in the document.  These changes,  including the deletion of references to Allied-
Signal's Canadian plant,  now closed, the addition of Industrias Quimicas de Mexico, and updates to
international HF production capacity data, have been incorporated.

       Several reviewers commented that data for U.S. imports of HF from Kenya and China reported
in the document actually include fluorspar; however, EPA has been  unable to verify this  comment with
the U.S. Department of Commerce and, therefore, has only noted it  in the document.

       3.2 Uses of HF

       Data on HF facilities reporting to EPA's Toxic Release Inventory (TRI) have been updated to
1990, the most up-to-date data available.  The percentages of total HF represented by various end
uses have been updated to 1990, based on data provided by reviewers. The discussion of HF as an
alkylation catalyst has also been revised to include updated data from reviewers on quantities of HF
consumed and percent of alkylate produced  using HF.

       3.3  Market Outlook

       The discussion of market  outlook for  HF has been extensively revised, based on comments by
reviewers. The Montreal Protocol and London Amendments phasing out CFCs and their effect on the
HF market are discussed.  Predictions provided by reviewers, indicating a  decline in HF consumption
in the U.S. from 1991 to  1996, are included in the discussion.

Chapter  4 - Regulations and Initiatives

       4.1  U.S. Federal Regulation of Hydrogen Fluoride

       The section discussing OSHA's Process Safety Management Standard has been expanded to
include more details about the  requirements  of this standard.

       EPA has expanded the section on DOT regulations to include the  DOT Response  Guides for
anhydrous and aqueous HF, as recommended by reviewers. EPA also has modified the discussion of
DOT labeling and shipping requirements based on  information provided by DOT, noting that
"Corrosive 8" placards are required. In addition, HF meets the definition of a poison gas because of its
vapor pressure at ambient temperature and its toxicity, and is a poisonous by inhalation material.

       4.2  U.S. State and Local Regulations

       EPA has added  a discussion of the Texas Air Control Board's permitting  program  to control
emissions of air pollutants, as recommended by reviewers.
                                                                                    Page 11-3

-------
        EPA has substantially revised the discussion of regulation of HF by the South Coast Air
 Quality Management District (SCAQMD).  It is noted that the SCAQMD rule is in litigation and
 implementation has been suspended, as pointed out by reviewers. The direct comparison of HF and
 H2SO4 alkylation processes has been removed from this section because EPA decided it was
 Inappropriate. EPA has included in this section a discussion of issues related to switching from HF to
 H2SO4 as an alkylation catalyst.

        Based on information provided by reviewers, EPA has corrected inaccuracies in the discussion
 of the suit by the City of Torrance, CA, against Mobil Oil Corporation.

        4.3  International Efforts

        EPA has added information provided by reviewers to the discussion in the subsection on
 Great Britain concerning the development and use of risk-based criteria for determination of planning
 decisions regarding major hazard facilities by the U.K.'s Health and Safety Executive (HSE).
 Information also is included about the "Dangerous Toxic Load' (DTL) for HF published by the HSE
 under the Control of Major Accident Hazards (CIMAH) regulations.

        The subsection on the Netherlands has been expanded to include a short discussion of
 increased safety in facilities through improved design and management and inventory reduction,
 based on comments by reviewers.

        EPA has added a description of an HF modeling system developed by ELF to the subsection
 on France.

        The description of Canada's Life-Cycle Management of Toxic Chemicals has been slightly
 revised based on comments.

         The subsection on other international efforts now notes that Western Australia; Victoria,  New
 South Wales; and Hong Kong have risk-based criteria in place for facilities that handle hazardous
 materials, based on information provided by reviewers.

        EPA has added information on several additional international efforts, as recommended by
 reviewers. Subsections have been added on the Guiding Principles for Chemical Accident Prevention,
 Preparedness and  Response, developed by the Organisation for Economic Co-operation and
 Development (OECD) and Awareness and Preparedness for Emergencies at the Local Level (APELL),
 an  initiative sponsored by the Industry and Environment Office (IEO) of the United Nations
 Environment Programme (UNEP), in cooperation with the CMA and the Conseil Europeen des
 Federations de I'lndustrie Chimique (CEFIC).

 Chapter 5 - Process Descriptions of Hydrogen Fluoride Industry

        5.1  HF Manufacture

        The  description of HF manufacturing and the flow chart showing the process have been
 modified for greater accuracy and completeness based on information provided by reviewers.

        5.2  Transportation  and Storage

        Several reviewers noted that air and hydrocarbon gases are not used in the compressed  gas
 method for unloading HF. EPA has deleted references to use of gases other than nitrogen for this
 purpose. EPA also has modified the range reported for bulk storage tank capacities for HF based on
 comments by reviewers.
Page 11-4

-------
       5.3 Fluorocarbon Production (title changed from "Chlorofluorocarbon Production" as per
comments received from fluorocarbon manufacturers)

       EPA has modified the description of production processes, and the flowchart illustrating these
processes, based on comments from fluorocarbon manufacturers. Additional information on the
production of HCFCs and HFCs, provided by reviewers, has been included.

       5.4 Alkylate Production for Gasoline

       Information provided by reviewers on the importance of alkylate for aviation gasoline has been
incorporated into this section.  The descriptions of the HF alkylation processes have been modified
based on comments by reviewers.

       5.5 Uranium Processing

       Reviewers provided useful information on the chemistry and processes involved in the use of
HF in uranium processing, including descriptions of two different processes. EPA has incorporated
this additional information into the document.

       5.6 Aluminum Fluoride and Aluminum Manufacturing

       Information provided by reviewers on the reactions and processes involved in aluminum
production has been incorporated into this section to provide a clearer and more accurate description.

       5.7 Electronics Manufacturing

       Corrections and additional information provided by reviewers on concentrations used, silicon
etching, and process details have been incorporated into this section.

       5.8 Chemical Derivatives Manufacturing

       Information provided by a reviewer has been added to this section describing the production
of fluoroaromatics by diazotization of substituted anilines  with sodium nitrite in anhydrous HF.

       5.9 Processes Using Aqueous HF

       No substantive comments were received on this section.

       5.10  Dissolving Ores for Production of Tantalum and Columbium (Niobium) Metals
('Niobium* added to title for clarification, based on reviewers' comments)

       No substantive comments were received on this section.

       5.11  Linear Alkylbenzene Production

       A few minor changes were made to clarify this section, based on comments (e.g., LAB for
linear alkylbenzene, rather than Lab).

       5.12  Pharmaceutical Production

       EPA added this new section to the document, based on information provided by a
pharmaceutical producer. This addition helps to provide  a more complete overview of industries using
HF.
                                                                                    Page 11-5

-------
 Chapter 6 - Hazards and Industry Practices for Processes Involving Hydrogen Fluoride

        6.1 General Hazards

        EPA has made minor modifications to this section for clarification; e.g., editorial changes to
 improve the wording.

        6.2 General Industry Practices

        Reviewers provided several suggestions for improving this section.  EPA has added more
 information on the OSHA Process Safety Management Standard. Based on comments by reviewers,
 EPA has noted in the text the complexity of hazard evaluation and its interpretation.  The discussion of
 Quantitative Risk Analysis (QRA) has been modified to include additional discussion of its strengths
 and uncertainties.

        The section on industry-wide standards has been updated to include the API Recommended
 Practice 751, Safe Operation of Hydrofluoric Acid Alkylation Units. The descriptions of the National
 Petroleum Refiners Association and the Chemical Manufacturers Association Hydrogen Fluoride Panel
 have been modified for greater accuracy and completeness, based on comments by reviewers.

        The discussion of corrosion has been modified based on comments by reviewers; the use of
 special carbon steels and other materials is discussed, and materials used for aqueous HF solutions
 are also noted.  Potential cracking of welds in carbon steel pressure vessels and prevention of such
 cracking is discussed.

        Based on comments from reviewers, EPA has  corrected its description  of shell and tube heat
 exchangers in HF service.

        6.3  Specific Industry Hazards and Practices

        EPA has made several changes to this section, based on information provided by reviewers.
 The HF production method  used by Du Pont is described as an internally heated reactor method, and
 it is noted that Du Pont uses a Texas Nuclear Alloy Analyzer to verify composition of alloys.
 information  on HCFC and HFC production has been added to the subsection on fluorocarbon
 production (previously called chlorofluorocarbon production), and additional information provided by
 reviewers on equipment failure has been included. It is also noted that the overall reaction in CFC
 production is endothermic. not exothermic.

        In the subsection on alkylate production, EPA has added information provided by reviewers on
 the use of isolation valves, double seals in the acid circulation pump in the UOP process, and the
 lower pressures used in the Phillips process.

        Reviewers also provided additional information for the subsection on transportation and
 storage, including information on transfer procedures and transportation equipment.

        The subsection on other uses was modified to  include a discussion  of the production of
 electronic aqueous HF, based on information provided by reviewers.

        6.4 Research Efforts to Modify or Substitute  HF In Alkylation

        Based on comments by reviewers, EPA has noted in the document research efforts by UOP
 and Texaco, and by Phillips Petroleum.
Page 11-6

-------
Chapter 7 - Industry Practices to Detect and Mitigate Hvdroaen Fluoride Releases

       7.1 General Industry Practices to Detect HF Releases

       Based on comments by reviewers, a number of modifications have been made to this section.
In the subsection on visual observation, it is noted that HF clouds are similar in appearance to steam
clouds.

       Information on thermal imaging systems, provided by reviewers, has been incorporated into
the subsection on detector equipment and systems. The use of hydrocarbon detectors at alkylation
unit sites is also noted.  The discussion of multi-detector systems has been modified, based on
information from reviewers.

       7.2 HF Detectors Used by Specific Industries

       EPA has corrected information relating to HF detection at specific facilities, based on
comments by reviewers.

       7.3 General Industry Practices to Mitigate HF Releases

       Several reviewers noted that diking may not be very useful for dealing with spills of liquid HF
because of its volatility; EPA has noted in the document that HF vaporizes rapidly. The subsection on
secondary containment has been revised to reflect this.

       The discussion of water spray systems has been  esctensively modified, based on information
provided by reviewers.  The text includes a description of the water spray test program, the  Hawk Test
series, sponsored and conducted by the Industry Cooperative HF Mitigation/Assessment Program
(ICHMAP). Development of the HFSPRAY model for assessment of water spray systems is also
mentioned.  The description of water spray systems has been clarified; stationary water streams, water
monitors, and water deluge systems are described. SCAQMD requirements are noted.

       7.4  HF Mitigation Systems Used by Specific Industries

       The discussion of HF mitigation systems at specific facilities has been corrected, based on
comments by reviewers.  It is noted that additional refineries have developed or are  planning water
spray mitigation systems, and the description of Amoco's water spray system has been modified
based on comments.

Chapter 8 - Characterization of Hvdroaen Fluoride Accidents

        8.1  Examples of Major Accidents

        EPA has noted in the description of the Marathon accident that the release included flashing
vapor and droplets and took place over 44 hours and modified the discussion as suggested based on
the pollution report by EPA's On-Scene Coordinator, which also includes OSHA notices of violations.
The description of the Mobil accident has been modified based on comments, and EPA has expanded
the description of the uranium hexafluoride accident at Sequoyah Fuels Corporation (subsidiary of
 Kerr-McGee), based on information provided by reviewers.

        8.2 Analysis of HF Accident Databases

        The analysis of reported incidents has been modified to distinguish, to the extent possible,
 incidents involving anhydrous HF from those involving aqueous HF.
                                                                                    Page H-7

-------
        8.3 Overview of HF Accident Data

        The discussion of transportation incident data has been expanded to note that almost five
 billion pounds of anhydrous HF have been shipped over the past ten years with few releases and no
 fatalities or serious injuries due to HF, as pointed out by reviewers.  Incidents involving loading and
 unloading have been separated from incidents that occurred in transit, where possible.

 Chapter 9 - Modeling  Hydrogen Fluoride Releases

        9.2 Models for Hydrogen Fluoride Releases

        Based on comments, EPA has decided the discussion of sulfuric acid spill tests is
 inappropriate to a report on HF; therefore, this discussion has been deleted.

        The discussion of HF thermodynamics has been expanded, based on comments  by reviewers,
 to include more information about the thermodynamic effects that can occur over time in an HF
 release (e.g., cooling of the cloud by aerosol evaporation).

        EPA has incorporated information from reviewers into its description of available models and
 comparisons with spill tests. The discussion of SLAB and DEGADIS has been expanded to reference
 additional reports comparing model results to the Goldfish tests. The aerosol evaporation model
 added to SLAB is described briefly.  Difficulties in interpretation of the DEGADIS results are discussed.
 A 1992 study comparing results of DEGADIS, SLAB, and other models to field test data is cited.
 Recent studies are cited showing that the latest version of HFSPRAY has improved capabilities for
 modeling water spray mitigation.

        9.4  Worst-Case Accident Scenarios

        This section has been expanded to include both worst-case scenarios, as mandated by
 Congress, and credible or reasonable worst-case scenarios for the types of facilities that handle HF.
 As recommended by reviewers, it is noted that caution must be used when reviewing consequence
 analysis results.

        Reviewers provided extensive comments on EPA's modeling efforts. EPA is currently
 conducting additional modeling using HGSYSTEM, taking into account these comments. Such issues
 as averaging time and toxic dose will be addressed.  EPA also is modifying the scenarios  presented in
 the document based on comments by reviewers.

        EPA has removed the discussion and modeling results using the ALOHA model as per
 comments received from reviewers and the SAB.  EPA is including additional discussion and modeling
 results using the SLAB and DEGADIS models.

 Chapter 10  - Community and Facility Emergency Preparedness and Planning

       EPA has added an introductory paragraph to this chapter, explaining that the information is
 anecdotal and  not intended to be comprehensive. Descriptions of specific industry/community
 programs have been  corrected based on comments by reviewers. CMA's medical treatment
 emergency communication network pilot program and CHEMTREC are noted. EPA has added a new
 subsection describing EPA's Chemical Safety Audit program and discussing the recommendations
 made in Chemical Safety Audits of several facilities handling HF.
Page 11-8

-------
Appendices

       EPA has modified several of the appendices to the HF document based on comments by
reviewers, as described below.

       Appendix V - Overview of Probit Equations

       Probit equation coefficients in this appendix have been corrected based on comments, and
the appendix has been modified to include additional probiit equations and results for several different
exposure times. The discussion of probits  has been modified to reflect information provided by
reviewers.

       Appendix VII - Facilities Reporting to TRI for Hydrogen Fluoride

       This appendix has been updated by replacing 1989 data with 1990 data from TRI.

       Appendix VIII - U.S. Producers of Fluorocarbons and  of Other Chemicals Manufactured
with Hydrogen Fluoride or Chlorofluorocarbons

       Information on producers in this appendix has been corrected as appropriate (e.g., by
changes to the list of types of fluorocarbons produced) based on comments by reviewers.

       Appendix IX - U.S. and Canadian  Petroleum Refineries with HF Alkylation Units

       This appendix has been updated based on information provided by reviewers.

       Appendix XI - Containers for Transportation of HF

       Corrections have been  made to this appendix based on information provided by reviewers
 (e.g., corrections to the description of tank car specifications).
                                                                                    Page 11-9

-------
                                              EXHIBIT 11-1
                    Reviewers Who Commented on Draft Hydrogen Fluoride Report
                        Organization

  Aluminum Association
  3M
  Lawrence Uvermore National Laboratory
  Mallinckrodt Specialty Chemicals Company
  U.S. Department of the Interior, Bureau of Mines
  Wayne State University
  SRI International
  IBM
  Hoechst Celanese
  Techntca Inc.
  MetroHeatth
  National Petroleum Refiners Association (NPRA)
  U.S. Environmental Protection Agency, Office of Pollution
  Prevention and Toxics
  U.S. Environmental Protection Agency - Region V
  Marathon
  Rohm & Haas Company
  University of Louisville
  ELF Atochem North America
  Du Pont Chemicals
  BPOil
  Chemical Manufacturers Association (CMA)
  Ultramar Inc.
  Mobil Oil Corporation
  Vista Chemical Company
  International Union of Electronic, Electrical, Salaried, Machine
  and Furniture Workers; Occupational Health Foundation
  Working Group on Community Right-To-Know
  Brookhaven National Laboratory
  South Coast Air Quality Management District
  U.S. Department of Commerce, National Oceanic and
 Atmospheric Administration  (NOAA)
 American Petroleum  Institute (API) HF Alkylation Committee

 U.S. Environmental Protection Agency, Office of Air and
 Radiation
 Allied-Signal, Inc.
 Allied-Signal, Inc. (modeling); and EPA Science Advisory
 Board, invited specialty expert
 Workers Board in British Columbia
 Energy Safety Council
 Kerr-McGee Corporation
 Wacker Siltronic Corporation
 Environment Canada
 Phillips Petroleum Company
                   Individual

   Roy Carwile
   C.S. Chow
   Ron Koopman
,   Thomas H. McFadden
   William L Miller
   Daniel A. Growl
   Ray Will
   David Gunnarson
   James L Paul
   Krishna Mudan
   Steven Borron
   Norbert Dee
   Scott Prothero

   Glenn Cekus
  Robert Mason
  Stanley J. Schecter
  George C. Rodgers, Jr.
  John Laumer and Harold Lamb
  Carolyn S. Seringer
  Charles E. Fryman
  Gordon D. Strickland, F.S. Leiva
  Diane Sinclair
  T.D. Cole
  F.E. Linstead
  Scott Schneider

  Paul Orum
  Vasilis Fthenakis
  David Yeh
  Roy Overstreet

  Jeff Morris, Fina Oil and Chemical Company;
  Richard M. Gustafson, Texaco; Doug N. Blewitt
  Jose Fernandez

  William Hague and Robert Pratt
  William Hague

  Ian Solomon
  Edward V. Badolato and Keith M.  Karnofsky
 Steven Emerson
 Thomas C. McCue
 John Shrives
 Barbara Price and Jeffrey M. Reamy
Page 11-10

-------
                                           EXHIBIT 11-1
                 Reviewers Who Commented on Draft Hydrogen Fluoride Report
                                            (Continued)
                      Organization

Texas Air Control Board
STRATCO, Inc.
U.S. Department of Transportation
HHS-CDC
Friends of the Earth; and EPA Science Advisory Board, invited
specialty expert
Air Products and Chemicals, Inc.
Bob Puschinsky, Inc., Alkylation Technologies
University of Arkansas; and EPA Science Advisory Board,
invited specialty expert
U.S. Environmental Protection Agency, Office of Pollution
Prevention and Toxics
EPA Science Advisory Board, Chairman; and Union Carbide
Corporation
EPA Science Advisory Board, member; and University of
Michigan
EPA Science Advisory Board, member; and Michigan
Department of Natural Resources
EPA Science Advisory Board, member; and Oregon
Department of Energy
EPA Science Advisory Board, member; and Pilko & Associates,
Inc.
EPA Science Advisory Board, member; and GeoTrans, Inc.
EPA Science Advisory Board, member; and Electric Power
Research Institute
EPA Science Advisory Board, member; and University of
 Pittsburgh
 EPA Science Advisory Board, member; and GEI Consultants,
 Inc.
 EPA Science Advisory Board, member; and Stanford University
 EPA Science Advisory Board, member; and Energy and
 Environmental Research Corporation
 EPA Science Advisory Board, member; and Corporation on
 Resource Recovery and the Environment
 EPA Science Advisory Board, member; and Rice University
 EPA Science Advisory Board, advisory staff
 EPA Science Advisory Board, Staff Director
                                                                         Individual
Sam Crowther
Kenneth R. Masters
James O'Steen
Mark McClanahan
Dr. Fred Millar

Robert W. Ormsby
Bob Puschinsky
Dr. Jerry Havens

Gail Froiman

Richard Conway

Dr.  Linda Abriola

Dr.  George Carpenter

Christin Ervin

Dr.  Wayne Kachel

Dr.  James Mercer
Dr. Ishwar Murarka

Dr. Frederick Pohland

Dr. Robert Pojasek

 Dr. Paul Roberts
 Dr. William Seeker

 Dr. Walter Shaub

 Dr. Herb Ward
 Dr. Jack Kooyoomjian
 Dr.  Donald Barnes
                                                                                           Page 11-11

-------
                                   This page intentionally left blank.
Page 11-12

-------
                                 APPENDIX III
              SUMMARY OF ORAL AND WRITTEN COMMENTS
        FROM PUBLIC MEETING ON HYDROGEN FLUORIDE STUDY
                                 Jujy 12, 1993
            EPA Auditorium, 401 M Street, SW, Washington, D.C.
      On July 12, 1993, a public meeting was held to present and to discuss preliminary findings of
the Hydrogen Fluoride Study. The preliminary findings, which were used to develop
recommendations, were distributed to stakeholders prior to the meeting. Fourteen of the attendees
presented their comments orally, and 15 interested parties submitted written comments. The
Appendix contains a summary of comments and is not a verbatim transcript. Exhibit 111-1 lists the
reviewers who provided written comments.
                                                                        Page 111-1

-------
 III.1    Summary of Oral Comments

 Elaine Davies-EPA

        Introduction and Background. The stakeholders' meeting held on October 17, 1991, focused
 on issues to be included in the report. Comments were incorporated into the report, and additional
 revisions were made per EPA's Science Advisory Board comments.  A court-ordered deadline of
 September 30, 1993, has been set for the final report. The purpose of this public meeting is to receive
 technical comments on the findings, and relevant technical information which will be used to develop
 recommendations.

 Ed Freedman-EPA

        The meeting will end with a wrap-up.  Significant changes to the report include expanded
 modeling input descriptions, and an expanded number of scenarios.  Scenarios now include
 catastrophic vessel failure and others.  Sensitivity analysis was done for certain assumptions.
 Consequence analysis is based on dose rather than concentration.

               Major points of the findings:

 •       Manufacture and  use of HF present some  risk to the public (depends on potential
        consequence and likelihood of occurrence).
 •       Dispersion modeling of large scale catastrophic releases resulted in impacts at distances
        greater than 10 km.
 •       Actual off-s'rte releases are rare, no deaths to the public have occurred.
 •       A number of controls, and  many regulations, are in  place;  owner/operator must comply with
        the general duty clause to  operate facilities safely. These  controls should decrease the
        likelihood of accidents.
 •       Emergency planners have  limited preparation for HF accidents.

 Barry Weissman-Director of Regulatory Affairs for Ausimont USA

        His company is a plastics and fluoropolymer manufacturer in NJ, employing 97 people.
 Anhydrous HF has been used since 1989. Its use was registered with NJDEPE under the Toxic
 Catastrophe Prevention Act regulation issued in June 1988.  Risk Management Plans (BMP) have
 been developed, and risk assessments have been performed.  Ausimont follows manufacturer's
 recommendations in developing RMP/operations plans, and conducts preventive maintenance and
 yearly ultrasonic tank thickness inspections. Fully  automated loading/unloading, and fully automated
 monitoring of vapors is being installed. The facility has fire monitors, and water curtains.  Training is
 conducted for  contractors, and an employee emergency response team. Access to areas with
 specified hazardous chemicals is limited.  Ausimont has on-site wastewater treatment. Since 1989,
 there have been ten injuries and five releases.  The largest release was 174 pounds, which was kept
 on-site through the use of water spray mitigation devices.

       Safety  costs money up-front, but saves money if it works.  The speaker suggests keeping
 OSHA In mind, and asks that new regulations  not be developed.  He also suggests computer
 compatible forms, and simple record keeping.

 Jeff Morris-American Petroleum Institute (API)

       The speaker compliments EPA on the  involvement of the stakeholders in the process of
 developing the HF report; the stakeholders have been able to make an impact.  He asks that this
 involvement be continued.
Page 111-2

-------
       Generally, the findings are balanced and objective.  The speaker would like to reinforce two
points. First, he suggested that EPA focus on how process safety is managed; handling is important;
techniques are critical (i.e., API Recommended Practices 750, 751). Second, he suggested that EPA
should recognize that these practices and the application of techniques are site specific.

Fred Millar-Friends of the Earth (FOE)

       Given the limited time remaining, the speaker suggested that EPA look at Process Safety
Management's (PSM) track record. The speaker has been looking at companies' involvement of
employees in HazOps procedures, etc. He has interviewed workers-- their comments have varied.
The speaker suggested that EPA examine if OSHA is ensuring compliance with PSM.

Geoff Kaiser-Science Applications International Corporation

       The.speaker has provided technical advice on HF in hazops, risk assessments, and PSM.  The
comments in the findings are sensible and wise. They emphasize that risks are reduced if facilities
comply with the general duty clause, etc.  The speaker has three points:

>      Concerning risk being minimized by facility compliance with regulations regarding HF, 'or any
       other hazardous substances:' HF should  be placed in perspective with other chemicals.
       Emphasis on existing regulations  is helpful.  Regulations provide a broad objective, but should
       allow varying ways to implement.
»•      Judgments of risk should not be based on analysis of consequences alone, likelihood of
       occurrence must also be considered.  Judgments should be made in context of risk.
K      The speaker does not believe that fatalities are possible beyond 10 km. His work has never
       found releases that went this far, especially not in urban settings.

William J. Hague-Allied-Signal

       The speaker has several comments regarding the volatility of HF above its boiling point.
Pressure must also be sufficient for HF to vaporize upon release.  Allied-Signal has done experiments
where HF autorefrigerates and a liquid pool forms at temperatures above the boiling point.

       The speaker has comments regarding the increased potential dose posed by aerosol.
Aerosol in a release is vaporized as the plume progresses,  at approximately 10-20 meters. This
increases potential dose because a greater source term is created.

       The speaker notes a correction to the findings: heat of reaction is counteracted by
dissociation, not by polymerization.

       The speaker is concerned that we seem to tie worst-case scenario to worst-case
consequences.  EPA needs to look at maximum worst-case risks to public, and must look at likelihood
of release. Emergency response should focus on events with some probability.  A worst-case
determination should address high probability, a combination of consequence and probability.
 [Craig Matthiessen asked for suggested approaches]  The speaker suggests a matrix as presented in
the Technical Guidance for Hazards Analysis  document (also known as the Green Book). EPA should
 use the judgment of several people and focus on  mitigation for reasonable cases that fit into the
matrix.

 Fred Millar-FOE

 >     The speaker stated that the study should  be provided along with the findings.
                                                                                    Page 111-3

-------
 >•      The statement that HF is an essential compound goes beyond what should be said. Other
        compounds can be alternatives.
 >•      The comment that a framework of additional future controls is being implemented is an
        overstatement.  Industry has not been sufficiently surveyed.
 >•      Compliance with "spirit" of existing regulations-this language should not be used, industry
        wants specific  regulations.
 >      South Coast Air Quality Management District (SCAQMD) has not been given enough
        attention--EPA  needs to evaluate further the effectiveness of SCAQMD's programs.
 >•      The speaker suggests that the discussion of mitigation field tests is misleading.  He would like
        to see information on the engineering uncertainty of water spray effectiveness.  There is a lot
        of skepticism on real operating units. The speaker wants more information in the report on
        who is actually using these techniques. He would also like to see the results of any reliability
        tests.
 >•      The speaker appreciates that information from safety audits was included, and wants more
        emphasis on this.  He would like to review the study.
 »•      The speaker would like to have access to documents on New Jersey's risk management rule.

 William Hague-Allied-Signal

        A lot has happened on water spray mitigation technology.  The original tests were in a closed
 chamber but correlate with the earlier 1986 and 1987 tests.  Although a 40-1 water to vapor ratio
 achieved 90% effectiveness in an operating tunnel, this tunnel is not optimal. 20-1 may also be just as
 effective.  Work is ongoing.

 Jerrv Havens-Univers'rty of Arkansas

        This is an important study because it will set precedents with respect to consideration of
 hazardous chemicals in commerce  and how to consider  risk of chemicals and their use. It is the first
 time a particular chemical has been singled out.  The speaker emphasizes that it is incorrect to base
 analysis on  probability or consequence separately; there is no logical, rational basis for evaluating risk
 to public without considering both.  There is a danger in  addressing consequence and  risk  together,
 the two can't be mixed.  When people talk about likelihood, the importance of consequence is
 dismissed because of improbability. The importance of consequences of highly improbably accidents
 should not be dismissed; one should be fully aware of the potential  severity of accidents.

        The speaker has two specific comments on the findings. First, the statement about aerosols
 posing a greater hazard does not mention a principal reason for concern-that the cloud will be much
 heavier than air.  Second, the issue of complex terrain continues to be researched.  Consequence
 assessment presents a lot of information, all over idealized flat terrain.  The question is how to use
 those tools to assess consequences in urban setting. EPA should include some mention of the fact
 that we do not know enough about the influence of terrain on consequences.

 Charles Barrett-Oil. Chemical, and Atomic Workers International Union (OCAWILO

       The  speaker's main concern is the importance placed on OSHA's PSM standard.  This report
 gives the impression it is going well. OCAWIU's impression is that it is behind schedule. The
 importance of worker involvement needs to be emphasized.  Unless worker involvement provisions are
 adhered to, we won't see what's going on with compliance. There is no specific requirement for
 training workers on the full  nature of PSM standard.  Unless training encompasses the entire standard,
 workers will be lost as aides to enforcement. A company often hand picks the people to be in charge
 of these areas and does not always make the best choice. The training of contractors in no way
 equals that which workers receive. The speaker asks that workers be surveyed on problems of
 contractor knowledge/training.
Page 111-4

-------
       Very few companies are involving workers.  We need to look at PSM in more detail. Labor,
environment, and communities need to analyze PSM. EPA should look at PSM plans in more detail,
and have an oversight committee on compliance with PSM. The report appears to say that PSM is
working, and that the standard is being enforced.

Ed Freedman-EPA

       The speaker asked if these comments have been shared with OSHA. [answer: yes] The HF
report says success is conditional on PSM being applied properly.  EPA has not found that everybody
is in compliance with regulatory standards.

Henry Jones-OCAWIU-BP Refinery

       It is important to have worker involvement in PSM. OSHA often dismisses health and safety
concerns. Workers are needed to oversee PSM because OSHA will not.  Workers tend to overlook
problems and cannot follow through. They fear for retribution, or fear that plants will be closed down.
Workers need to follow up on oversight. At BP, employee involvement has improved safety. Workers
are holding 8 hour safety sessions. PSM cannot work without worker involvement, even then it will be
a struggle.

       Statistics show contractors have more accidents than plant workers per man-hour.  At BP,
training for contractors is done by contractors; they should have more contact with local workers.
Loss of mechanical integrity is perhaps the single most frequent cause of releases and accidents.  At
BP, they have a full-time inspector for HF.  Companies should have properly trained people to do the
job.
The speaker states that workers are beginning to be asked to report problems.  He asks that LEPCs
be given  more funding.

Vince Morroni-OCAWIU member

       The speaker, who has been a Chevron alkylation unit operator for 17 years, reaffirms the
importance of worker involvement.  Equipment integrity is often compromised by testing of equipment
by operators, rather than trained technicians. Operation of water mitigation systems is not sufficiently
tested.

Craig Matthiessen & Ed Freedman-EPA

     The report is scheduled to be finalized by Sept. 30.  EPA will not have time to distribute a new
draft. The next step is to incorporate comments into the findings,  and use findings to develop
recommendations.  The report will go through the Agency approval process.

Jerry Havens-University of Arkansas

       The speaker suggests holding another stakeholder review. The report and findings need to
be reviewed together.  The review process is procedurally weak.

William Hague-Allied-Signal

       A previous EPA technical report was full of problems.  Could an addendum be issued if
technical inaccuracies exist in HF report?
                                                                                   Page 111-5

-------
 Craig Matthiessen-EPA

        EPA would consider having an addendum/follow-up report.  International feedback is
 expected.

 Jerry Havens-University of Arkansas

        Congress should be made aware that an addendum/response may possibly follow.

 Craig Matthiessen-EPA

        The speaker notes that this is a good point.

 Fred Miilar-F.O.E.

        France (SNPE) has done the only other HF study.  Where is it?

 Craig Matthiessen--EPA

        A Canadian life-cycle analysis HF study is also being done.  The speaker thinks the SNPE
 study is on-going.

 Bill Hilller-M.W. Kellogg Co.

        The speaker has no basic disagreement with broad findings.  He suggests an addition to [the
 former] section 9.2 on substitution and modification, regarding M.W, Kellogg's development of an
 alkylation process (per written comments).

 Mike Moosemlller-Technica

        The speaker reiterated that PSM is involved in a lot of HF discussion.  PSM involves
 engineering systems. A lot of money is being spent.  The speaker would like EPA to take a broader
 perspective.  EPA should look at management practices, inspection systems, contractors, and
 mechanical integrity. Flexibility is needed to allow people to come to the most effective realization of
 their limited resources.

 Jeff Morris-API

        The speaker almost entirely supports the OCAWIU speakers, and recognizes that enlightened
 management and employees are needed. We are changing the way we think about and operate
 refineries;  some are farther along than others. Business and workers need to work together; need to
 protect jobs, environment, safety of workers and public.  PSM is the proper way to be successful in
 this business; API endorses PSM techniques. It is important to be cognizant of realities in moving
 from where we are to where we want to be-difficult balancing act.

       The speaker is concerned that the worst-case scenario discussion in the report could be taken
 out of context. He reserves the right to address the worst-case discussion in the report. The speaker
 registers concern on not being able to see the report.
Page HI-6

-------
Fred Millar-FOE

       OSHA does not have written model plans of action, nor does Chemical Manufacturers
Association or the American Petroleum Institute.  What good is the program without a model? There's
a risk that PSM is not being taken seriously if people don't have a model and no measures exist on
the early effectiveness of PSM.

       Did the HF study look at whether facilities have catastrophic insurance? Are there small, "mom
and pop" facilities that may go out of business because they are not adequately insured?

Craig Matthiessen--EPA

       That topic was not examined in this study.  Another EPA study looked at the role of insurance
companies.

Coriolana Simon--EPA Chemical Accident Prevention staff

       There were no conclusions from that study; the results were contradictory.  It was clear that
the insurance industry does not want to be a surrogate regulator.

Chuck Galloway-Chevron East Coast

       The speaker notes that Chevron will respond in writing to answer the OCAWIU member's
concerns regarding the water mitigation system that is being installed at an alkylation unit.

Chuck Barbell-Refractories Institute

       The speaker expresses his concern about this regulatory process.  EPA has gone back and
forth on what is required. Industry doesn't have a chance to do what they've planned.  This report will
go to Congress,  and they will make laws based on it.

Craig Matthiessen--EPA

       The speaker assures the commenter that this is not a regulatory process.  Any regulation that
would  be under other sections of CAA, would require a proposed rule, public comment, etc. EPA
realized the importance early on to develop a technically accurate  report. There are no hidden
regulations in the report's recommendations.  There is a possibility that Congress may  make decisions
based on the report, therefore EPA had to be very careful.

        EPA wants all issues documented, wanted to have dialogue, and review of the  report.  EPA
would  like to have more time.  EPA wants to get the report out, and put any comments, changes in an
addendum. The speaker wants all written comments on the preliminary findings by COB Monday
(today).

        The speaker has heard that there are concerns about PSM, and will try to get a better
understanding of these issues. Worst-case is also a hot issue.  EPA shares the concerns expressed,
and hopes to consider all these issues, and provide a sensible and best path forward on Risk
Management Plan.
                                                                                   Page 111-7

-------
 111.2    Summary of Written Comments

 Allied-Signal-William J. Hague

        The commenter clarifies that the dense gas behavior of HF is due to temperature gradients in
 the plume, and to the presence of aerosol. He also notes that the limitation of the LEPCs in planning
 for transportation incidents is not unique to HF.

 American Petroleum Institute-C.J. Krambuhl

        API supports the findings that 1) the risks of HF release are site-specific and can be mitigated
 by PSM practices, and 2) further knowledge may be gained with respect to alternatives to HF acid.
 The commenter emphasized that HF be considered in perspective with other hazardous substances.
 He proposes using the term "associated" rather than "polymerized" to describe the behavior of HF
 molecules. EPA should  emphasize the usefulness of  modeling for PSM efforts.

 Ausimont. USA-Barry Weissman

        This reviewer's comments were made in the oral presentation summarized above. Written
 attachments included the requirements of New Jersey's risk management plan (RMP), injuries/releases
 at the facility due to HF,  and risk assessments and hazard analysis reports required by NJ's Toxic
 Catastrophe Prevention Act.

 BP Oil-Chuck Fryman

        The commenter suggests that publishing the entire draft report,  including findings and
 recommendations, will ensure that all stakeholders' comments have been considered.  He would like
 to see more emphasis on the prevention aspects of HF risk management. The commenter wants EPA
 to emphasize that no deaths to the  public have occurred due to HF exposure in the more than 50-
 year operating history of HF alkylation plants. EPA should mention the work underway to develop HF
 aerosol inhibitors, as well as the costs associated with conversion from HF to sulfuric acid.  EPA
 should emphasize HP's hazards relative to those  of other dangerous substances like chlorine and
 ammonia. The commenter noted that site-specific factors must be evaluated to determine the
 appropriateness of installing specific detection and mitigation measures.

 Chemical Manufacturers  Association (CMAV-Joe J. Mayhew

        CMA notes that little is gained by the development and application of consequence analysis
 which stacks simplistic assumptions. CMA emphasizes that the likelihood of occurrence must be tied
 to consequence analysis. The commenter is concerned that a revised draft was not made available to
 reviewers with the findings. He made several suggestions advocating the use of site-specific and
 technical information in analyzing worst-case scenarios, especially as these related to the RMP
 rulemaking. The commenter suggests that EPA allow the use of realistic and relevant levels of
 concern in consequence analysis.

 Chevron-Don E. Tormev

       The commenter responded to Vince Morroni of OCAWIU's concerns regarding equipment
 operation at a Chevron Phillips alkylation unit. He described the use, installation, and testing of HF
 detectors and a water spray system at the refinery.
Page 111-8

-------
Hoechst-Celanese-James L. Paul Ph.D.

       The commenter strongly supports CMA's comments dealing with worst-case scenario and
implications for EPA's forthcoming RMP rulemaking.

Lawrence Livermore National Laboratory-Ronald P. Koopman

       The commenter found no errors or problems with the preliminary findings.

Marathon Oil Company-Ned F. Seppi

       The commenter asked for clarification of the reference in the findings of a "framework for
additional future controls." He suggested that EPA mention the greater transportation requirements
associated with sulfuric acid alkylation, due to higher acid consumption, if regeneration is not available
on-site.  The commenter suggests changing the term "polymerize" to "aerosolize" with respect to HF
molecule behavior.  He requests clarification of "aqueous solutions" in terms of weight percent HF.

M.W. Kellogq-William J. Hillier

       The commenter provided a description of a "next generation alkylation process" under
development by Kellogg and Topsoe for commercial use, which could be appropriately included in the
Substitution and Modification section of the final report.
                                      K.
Oil. Chemical, and Atomic Workers International Union (OCAWIU)-Charles Barrett

       OCAWIU encourages EPA to recommend incorporation of preventive measures of adequate
spacing between alkylation units along with continuous gas detection systems. Other written
comments were reflected in the oral presentation, summarized in section 111.1.

Phillips Petroleum Company-Barbara J. Price

       The commenter expressed  support for the findings, and for the decision to evaluate HF from a
risk management view point rather  than simply hazard avoidance. She asks that EPA emphasize that
HF is like many other extremely hazardous substances. The commenter suggests that EPA include in
the findings a section on the valuable contributions of HF products to the U.S. manufacturing industry
and the public.  She questions the  appropriateness and necessity of using the term "severe" to
describe the results of a possible HF release. The  commenter emphasizes that factors determining
the movement and  potential hazards of an HF vapor plume are complex, and time- and site-specific.
She requests that EPA give a low-range estimate, along with the high-range estimates presented, of
populations located near facilities which handle HF. The commenter emphasizes the usefulness of
model results in the development of PSM  efforts. She asks that EPA note that HF containment failures
with off-site consequences are extremely rare. The commenter asks that EPA revise its conclusion
that not all facilities have adopted detection and mitigation measures, and instead state that  not all
facilities have undertaken analyses to determine what measures may be necessary.

Science Applications International Corporation-Geoff Kaiser

       The commenter clarified that under certain  conditions HF may form liquid droplets which can
become airborne.  He wanted clarification regarding the impact of release consequences and the
basis of priorities for emergency planning. Regarding water spray mitigation systems, he added that
the cited 90% reduction in HF flow  has been achieved for 40:1 waterHF ratios, but that higher
reductions have occurred for higher ratios.
                                                                                   Page HI-9

-------
 South Coast Air Quality Management District-David Yen. Ph.D.

        The commenter requests that EPA disclose the concentration level assumed for the "severe
 impacts" cited in the findings. With respect to stability level, EPA should clarify what is meant by "long
 periods of time," and note that F stability may not be worst-case in all. situations.  The commenter asks
 that EPA note that releases of sulfuric acid under alkylation conditions have generated a small amount
 of aerosol (not greater than 7 percent).

 STRATCO-Kenneth R. Masters

        The commenter suggests that all references to sulfuric acid be deleted from the Preliminary
 Findings and the final report.  STFIATCO notes that the transportation requirements for sulfuric
 alkylation do not translate into greater transportation hazards. EPA should not comment on
 conversion from an HF alkylation unit to a sulfuric acid unit when EPA has chosen not to include a
 direct comparison of sulfuric acid and HF in the final report.
Page 111-10

-------
                                       EXHIBIT 111-1
        Reviewers Who Submitted Written Comments on the Preliminary Findings
                     Organization

Allied-Signal
American Petroleum Institute
Ausimont, USA
BPOjl
Chemical Manufacturers Association
Chevron
Hoechst-Celanese
Lawrence Livermore National Laboratory
Marathon Oil Company
M.W. Kellogg Company
Oil, Chemical & Atomic Workers International Union
Phillips Petroleum Company
Science Applications International Company
South Coast Air Quality Management District
STRATCO
                                                                        Individual
William J. Hague
C.J. Krambuhl
Barry Weissman
Chuck Fryman
Joe J. Mayhew
Don E. Tormey
Jarnes L Paul,  Ph.D.
Ronald P. Koopman
Ned F. Seppi
William J. Hillier
Charles Barrett
Barbara J. Price
Geoff Kaiser
David Yeh, Ph.D.
Kenneth R. Masters
                                                                                     Page II 1-11

-------
                                   This page intentionally left blank.
Page 111-12

-------
                                     APPENDIX IV
                EXPOSURE LEVELS FOR  HYDROGEN FLUORIDE
       A number of regulatory and guideline exposure levels have been developed for HF and other
toxic substances. Some of these levels are discussed below. Exhibit IV-1  (next page) presents a
comparison of some exposure levels for HF with those for other toxic substances.  A method for
estimating reference exposure levels for HF, and the levels derived based on that method, are also
discussed in this appendix, following the exhibit.

»-      IDLH. The Immediately Dangerous to Life or Health (IDLH) level, developed by.the National
       Institute for Occupational Safety and Health (NIOSH), represents the maximum concentration
       from which one could escape within 30 minutes without any escape-impairing symptoms or
       any irreversible health effects.1

>      EEGL. The National Research Council (NRC) of the National Academy of Sciences (NAS) has
       developed the Emergency Exposure Guidance Level (EEGL), defined as a concentration of a
       substance in air judged to be acceptable for the performance of specific tasks by military
       personnel during emergency conditions, usually lasting one hour.  The EEGLs are based
       primarily on acute toxicity. The NRC has also developed Short-Term Public Exposure
       Guidance Levels (SPEGLs), defined as ceiling concentrations for a single, unpredicted short-
       term exposure to the public, for a few chemicals; however, no EEGL or SPEGL has  been
       developed for HF.

*      ERPG.  Emergency Response Planning Guidelines (ERPGs)  have been developed for a limited
       number of chemicals by the American Industrial Hygiene Association (AIHA). The ERPGs are
       based on primarily on acute toxicity data and possible long-term effects from short-term
       exposure.

               The  ERPG-1 is defined as the concentration below which nearly all people
               could be exposed for one hour with only mild, transient adverse health effects
               or an objectionable odor.

               The  ERPG-2 is the concentration below which nearly all people could be
               exposed for  one hour without irreversible or other serious health effects or
               symptoms that would impair their ability to take protective action.

               The  ERPG-3 is defined as the maximum concentration in air below which
               nearly all people could be exposed for one hour without life-threatening health
               effects.

 »•      OSHA PEL. The Occupational Safety and Health Administration (OSHA) has developed
       Permissible  Exposure Limits (PELs), enforceable by law, for worker exposure to substances
       listed as air  contaminants (29 CFR Part 1910).  For most listed chemicals, eight-hour time
       weighted average (TWA) concentrations that cannot be exceeded in an eight hour work day,
       are specified.  In addition, for some chemicals, OSHA has developed Short-Term Exposure
       Limits (STELs), 15-minute time weighted  averages that cannot be exceeded at any time during
       a work  day,  and ceilings, levels that cannot be exceeded at any time during the work day.
       The PELs are developed for the protection  of worker health,  and are based on consideration
       of health effects and economic and technological feasibility.
                                                                                  Page IV-1

-------
                        EXHIBIT IV-1
Comparison of Regulatory and Guideline Exposure Levels for HF and
                    Other Toxic Substances
CAS No. Chemical Name
7664-39-3 Hydrogen Fluoride
7664-41-7 Ammonia
7782-50-5 Chlorine
7647-01-0 Hydrogen Chloride
75-44-5 Phosgene
7446-09-5 Sulfur Dioxide
7664-93-9 Sulfuric Acid
Listed as mg/m3 rather than ppm.
NIOSH
IDLH
(30 minutes)
30 ppm
(25 mg/m3)
500 ppm
(348 mg/m3)
30 ppm
(87 mg/m3)
100 ppm
(149 mg/m3)
2 ppm
(8 mg/m3)
100 ppm
(262 mg/m3)
20 ppm
(80 mg/m3*)
NAS
EEGL
(1 hour)


3 ppm
(9 mg/m3)
20 ppm
(30 mg/m3)
0.2 ppm
(0.8 mg/m3)
10 ppm
(26 mg/m3)
0.25 ppm
(1 mg/m3*)
AIHA
ERPG-3
(1 hour)
50 ppm
(41 mg/m3)
1000 ppm
(695 mg/m3)
20 ppm
(58 mg/m3)
100 ppm
(149 mg/m3)
1 ppm
(4 mg/m3)
15 ppm
(39 mg/m3)
7 ppm
(30 mg/m3 )
OSHA PEL
(8-hr TWA except
as noted)
3 ppm
(2.6 mg/m3)
STELSSppm
(27 mg/m3)
0.5 ppm
(1.5 mg/m3)
STEL: 1 ppm
(3 mg/m3)
Ceiling: 5 ppm
(7 mg/m3)
0.1 ppm
(0.4 mg/m3)
2 ppm
(5 mg/m3)
STEL: 5 ppm
(10 mg/m3)
0.25 ppm
(1 mg/m3*)
Sources: National Institute of Occupational Safety and Health
National Academy of Sciences
American Industrial Hygiene Association
Occupational Safety and Health Administration

-------
       Reference Exposure Levels Developed Using a "Benchmark Dose" Approach. Researchers
at the California Environmental Protection Agency2 have developed reference exposure levels (RELs)
for HF for protection of the public. The exposure levels were developed from existing toxicological
data on animals and humans. Log-probit extrapolation of available sets of concentration-response
data was employed to estimate a Practical Threshold (PT), defined as a "benchmark dose" estimated
to produce one percent response, specific to each set of data. A set of data-specific factors were
developed to account for species sensitivity, response severity, and slope differences among the
available concentration-response data, and RELs were estimated from each data set. The "best" REL
estimates were chosen based on the relative reliability of the test data used to derive the RELs.  Using
this method the one-hour reference exposure level to protect the public against any irritation from a
routine emission (REL-1) is 0.7 ppm and the level to protect against severe irritation from a once-in-a-
lifetime release (REL-2) is 2 ppm.  These levels of exposure are not industry or government standards.
                                                                                      Page IV-3

-------
  1.
  2.
                                  ENDNOTE

U.S. Department of Health and Human Services, NIOSH Pocket Guide to Chemical Hazards,
Centers for Disease Control, National Institute for Occupational Safety and Health
Washington, D.C., June 1990.  (362)


Alexeeff, George V., David C. Lewis, and Nancy L. Ragle, "Estimation of Potential Health
Effects from Acute Exposure to Hydrogen Fluoride Using a 'Benchmark Dose' Approach " Risk
Analysis, February 1993, p 63-69.
Page IV-4

-------
                                       APPENDIX V
                        OVERVIEW OF PROBIT EQUATIONS
       This appendix presents an overview of probit equations and their application to releases of
HF. Several probit equations developed for HF are presented.  EPA has not reviewed the data or
methodology used to develop these equations and is not endorsing a particular equation or method.
This appendix is presented for information purposes only.

       Probit (probability unjt) functions are used to determine statistical probability of an effect.  The
probit is a unit of measurement of statistical probability based on deviation from the mean of a normal
distribution.  A probit function takes the following form:1

                                      Pr  = a  + b Iog0V
where:         Pr is a measure of the percent of the vulnerable resource affected;
               V is a function of the factor that causes injury or damage to the vulnerable resource;
               a is a location parameter; and
               b is a slope parameter.

        Probit functions are applicable to various types of incidents, including fires and explosions as
well as toxic releases. They can be used to develop a probability distribution of consequences for
such incidents.  Probit functions can be used to quantify the number of fatalities that are likely to
occur from a given exposure to a toxic chemical in cases where there is information on dose-response
relationships. Toxicity probit equations are applicable to cases of non-linear as well as linear dose-
response relationships. Toxicity probit equations can  be used in conjunction with dispersion models
that take exposure duration into account to predict fatality levels at various locations based on the
results of dispersion modeling. The probit function for a toxic exposure is a logarithmic expression of
the form:2

                                    Pr  =  a + b log-(C"t)
 where:        a, b, and n are constants;
               C = concentration; and
               t = exposure time.

        The term C"t is the "toxic load," which provides a measure of the effect of exposure to the
 chemical as a function of concentration and duration of exposure. The toxic effect considered in
 probit analysis is lethality. An exponent (n) of one indicates that the effect of exposure to the chemical
 is directly related to concentration times duration of exposure; i.e., exposure to a concentration of 10
 ppm for 30 minutes would have the same effect as exposure to 300 ppm for one minute.

        The probit function is used along with a standard probit table to relate chemical concentration
 and duration of exposure to the  estimated percentages of people affected or estimated number of
 fatalities. A standard probit table is shown in Exhibit V-1.
                                                                                      Page V-1

-------
                                            EXHIBIT V-1
                                            Problt Table
             Probits are the three digit numbers in the table. Percents are read along the top
             and side margin of the table.  The vertical column of percents gives the decade; the
             horizontal column gives the unit. The table entry appearing in the row of the
             decade value and the column of the unit value is the probit corresponding to that
             percent. The last two rows in the table provide a finer reading for very  high percent,
             from 99.0 to 99.9.  The second to last row is the tenths of percent to be added to
             99%. The last row consists of the corresponding probits.

     Source:   Eisenberg, et al., Vulnerability Model: A Simulation System for Assessing Damage Resulting from Marine
             Spills, Springfield, VA, NTIS, 1975.
Page V-2

-------
       Several probit equations have been reported for HF. The values of a, b, and n derived for four
different probit equations3'4'5'6 are presented in Exhibit V-2.

       The results of the probit analysis vary depending on the coefficients used in the equation.
Exhibit V-3 shows the concentrations calculated to produce one percent, 10 percent, and 50 percent
fatality for a five-minute exposure, based on probits calculated using the coefficients presented in
Exhibit V-2 for four different probit equations and the probit table in Exhibit V-1. Exhibits V-4 and V-5
show the concentrations calculated to lead to the same fatality levels for 30-minute and 60-minute
exposure durations.  The highest and lowest concentrations estimated to cause one percent fatality
differ by a factor of nearly 12 for an exposure time of five minutes, while the highest and lowest
concentrations estimated to cause 50 percent fatality for any of the exposure times considered differ
by a factor of about six. The de Weger and ten Berge probit equations use similar coefficients (see
Exhibit V-2) and give generally similar results. The Mudan equation and the Perry and Articola
equation both assume a linear relationship between exposure time and effect of exposure (i.e., n=1);
however, the coefficient a for the Mudan equation is nearly twice as large as that for the Perry and
Articola equation. Use of the Mudan equation results in consistently higher concentrations than the
Perry and Articola equation.

       Data from animal experiments are usually used to derive probit equations, which are based on
lines of best fit to experimental data.  Animal experiments are usually done on rats or mice, but other
animal species may also be used.  The variation  in toxic effects  between different species may be
substantial. In addition, there is no definitive correlation between human and animal responses, which
may vary widely depending on the substance tested.

        Probits exhibit difficulties similar to those encountered in more generic attempts at modelling
human toxic responses; e.g., lack of data and difficulty in extrapolation of animal data to humans.
Additionally, care must be taken in applying a probit relationship;  a common mistake is extrapolating
the probit relationship outside of the original data.7 Probit analysis is weakest when applied to
prediction of toxic loads that would affect only small percentages of the population;8 hazard
assessments often require such predictions.

        The IDLH level of 30 ppm, intended to represent the level from which people could escape
within 30 minutes without suffering escape-impairing symptoms or irreversible health effects, is
considerably lower (by factors of about six to 40) than the concentrations estimated from the four
probit equations for one percent fatality from a 30-minute exposure (see Exhibit V-4).  This result is
consistent with the definition of the IDLH.  The ERPG-3 value of 50 ppm, intended to represent the
level to which nearly all people could be  exposed for one hour without life-threatening effects,  is
relatively close to the lowest concentration (83 ppm) calculated  to cause one percent  fatality from a
60-minute exposure, based on the Perry  and Articola probit equation (see Exhibit V-5). The three
other probit equations give higher concentration levels for one percent fatality from a 60-minute
exposure (2.5 to 12 times the ERPG-3).
                                                                                       Page V-3

-------
                                                   EXHIBIT V-2
                                    Coefficients for Four Problt Equations
Source of Probit
Equation
Mudan
Perry and Articola
de Weger
ten Berge
a
-48.33
-25.8689
-8.4
-7.35 •
b
4.853
3.3545
1
0.71
n
1
1
1.5
2
Units
ppm
ppm
mg/m3
mg/m3
        Sources: Mudan, Krishna S., Acute Inhalation Toxicity of Hydrogen Fluoride, AlChE Summer Annual Meeting,
                 Philadelphia, August 24-36, 1989.
                 Perry, W.W., and W.P. Articola, Study to Modify the Vulnerability Model of the Risk Management System,
                 Prepared for U.S. Department of Transportation, U.S. Coast Guard, Washington, D.C., 1980.
                 de Weger, Dik, Chris M. Pietersen, and Paul G.J. Reuzel, 'Consequences of Exposure to Toxic Gases
                 Following Industrial Disasters," Journal of Loss Prevention in the Process Industries, Volume 4, July
                 1991.
                 ten Berge, W.F., A. Zwart, and LM. Appelman, "Concentration-Time Mortality Response Relationship of
                 Irritant and Systemically Acting Vapours and Gases," Journal of Hazardous Materials, Volume 13, 1986
                 pp 301-309.
                                                  EXHIBIT V-3
                 Results Based on Several Probit Equations for Five-Minute Exposures
Source of Probit
Equation
Mudan
Perry and Articola
de Weger
ten Berge
Concentration for
1% Fatality
ppm
7,328
990
672
636
mg/m3
5,982
809
548
519
Concentration for
10% Fatality
ppm
9,099
1,355
1,353
1,331
mg/m3
7,428
1,106
1,104
1,087
Concentration for
50% Fatality
ppm
11,845
1,984
3,176
3,279
mg/m3
9,669 I
1,620
2,592
2,677
     Sources: Mudan, Krishna S., Acute Inhalation Toxicity of Hydrogen Fluoride, AlChE Summer Annual Meeting,
              Philadelphia, August 24-26, 1989.
              Perry, W.W., and W.P. Articola, Study to Modify the Vulnerability Model of the Risk Management System,
              Prepared for U.S. Department of Transportation, U.S. Coast Guard, Washington, D.C., 1980.
              de Weger, Dik, Chris M. Pietersen, and Paul G.J. Reuzel, "Consequences of Exposure to Toxic Gases
              Following Industrial Disasters," Journal of Loss Prevention in the Process Industries, Volume 4, July 1991.
              ten Berge, W.F., A. Zwart, and L.M. Appelman, "Concentration-Time Mortality Response Relationship of
              Irritant and Systemically Acting Vapours and Gases," Journal of Hazardous Materials, Volume 13 1986  pp
              301-309.
Page V-4

-------
                                            EXHIBIT V-4
            Results Based on Several Probit Equations for 30-Minute Exposures
Source of Probit
Equation
Mudan
Perry and Articola
de Weger
ten Berge
Concentration for
1% Fatality
ppm
1,221
165
203
259
mg/m3
997
135
166
212
Concentration for
10% Fatality
ppm
1,516
226
410
543
mg/m3
1,238
184
334
444
Concentration for
50% Fatality
ppm
1,974
331
961
1,338
mg/m3
1,612
270
785
1,093
Sources:  Mudan, Krishna S., Acute Inhalation Toxicity of Hydrogen Fluoride, AlChE Summer Annual Meeting,
         Philadelphia, August 24-26, 1989.
         Perry, W.W., and W.P. Articola, Study to Modify the Vulnerability Model of the Risk Management System,
         Prepared for U.S. Department of Transportation, U.S. Coast Guard, Washington, D.C., 1980.
         de Weger, Dik, Chris M. Pietersen, and Paul G.J. Reuzel, "Consequences of Exposure to Toxic Gases
         Following Industrial Disasters," Journal of Loss Prevention in the Process Industries, Volume 4, July 1991.
         ten Berge, W.F., A. Zwart, and L.M. Appelman, "Concentration-Time Mortality Response Relationship of
         Irritant and Systemically Acting Vapours and Gases," Journal of Hazardous Materials, Volume 13, 1986, pp
         301-309.
                                            EXHIBIT V-5
             Results Based on Several Probit Equations for 60-MInute Exposures
Source of Probit
Equation
Mudan
Perry and Articola
de Weger
ten Berge
Concentration for
1% Fatality
ppm
611
83
128
183
mg/m3
499
67
105
150
Concentration for
10% Fatality
ppm
758
113
258
384
mg/m3
619
92
211
314
Concentration for
50% Fatality
ppm
987
165
606
946
mg/m3
806
135
495
773
Sources: Mudan, Krishna S., Acute Inhalation Toxicity of Hydrogen Fluoride, AlChE Summer Annual Meeting,
         Philadelphia, August 24-26, 1989.
         Perry, W.W., and W.P. Articola, Study to Modify the Vulnerability Model of the Risk Management System,
         Prepared for U.S. Department of Transportation, U.S. Coast Guard, Washington, D.C., 1980.
         de Weger, Dik, Chris M. Pietersen, and Paul G.J. Reuzel, "Consequences of Exposure to Toxic Gases
         Following Industrial Disasters," Journal of Loss Prevention in the Process Industries, Volume 4, July 1991.
         ten Berge, W.F., A. Zwart, and L.M. Appelman, "Concentration-Time Mortality Response Relationship of
         Irritant and Systemically Acting Vapours and Gases," Journal of Hazardous Materials, Volume 13, 1986, pp
         301-309.
                                                                                                 Page V-5

-------
                                         ENDNOTES

 1.      Eisenberg, Norman A., Cornelius J. Lynch, Roger J. Breeding, Vulnerability Model, A Simulation
        System for Assessing Damage Resulting From Marine Spills, Enviro Control, Incorporated,
        Rockville, MD, June 1975, Document Number AD-A015245. Prepared for U.S. Coast Guard,
        Washington, D.C. (139.513)

 2.      American Institute of Chemical Engineers, Guidelines for Chemical Process Quantitative Risk
        Analysis, Center for Chemical Process Safety, New York, 1989. (10.46)

 3.      Mudan, Krishna S., Acute Inhalation Toxicity of Hydrogen Fluoride, AlChE Summer Annual
        Meeting, Philadelphia, August 24-26, 1989. (360)

 4.      Perry, W.W., and W.P. Articola, Study to Modify the Vulnerability Model of the Risk Management
        System, Prepared for U.S. Department of Transportation, U.S. Coast Guard, Washington, D.C.,
        1980, No. CG-D-22-80. (489.91)

 5.      de Weger, Dik, Chris M. Pietersen, and Paul G.J. Reuzel, "Consequences of Exposure to Toxic
        Gases Following Industrial Disasters," Journal of Loss Prevention in the Process Industries,
        Volume 4, July 1991, p 272.  (502)

 6,      ten Berge, W.F., A. Zwart, and LM. Appelman, "Concentration-Time Mortality Response
        Relationship of Irritant and Systemically Acting Vapours and Gases," Journal of Hazardous
        Materials, Volume 13, 1986, pp 301. (54.5)

 7.      Gustafson, Richard M., Texaco Inc., comments from technical review of Hydrogen Fluoride
        Study, Report to Congress, Draft May 8, 1992, May 29,  1992. (344, Appendix 1)

 8,      Chikhliwala, E.D., "Extensions of Consequence Analysis," 6th SAFER/TRACE Users Meeting,
        San Antonio, TX, September 1991. (121)
Page V-6

-------
                            APPENDIX VI!
           EXAMPLE OF MEDICAL TREATMENT GUIDELINES
              FROM A HYDROGEN FLUORIDE PRODUCER
     This appendix presents medical and first aid treatments for exposure to HF recommended by
Allied-Signal.
                                                              Page VI-1

-------
i.'v
.ecommended Medical Treatmeri:

:'or Hydrofluoric Acid Exposere
                            Hied
                            Signal

-------
Hydrofluoric  Acid
Quick Reference Chart
    'FOLD OUT HERE5
   This Booklet describes the special First Aid and. Medical Treatment measures
   necessary following exposure to or injury from HYDROFLUORIC ACID.

                   However, it must be emphasized that

                      PREVENTION
               of exposure or injury must be the primary goal.

   Preventive measures include making sure that:

     1.  Everyone who handles or uses HF is aware of its properties and
        dangers.
     2.  Everyone handling or using HF is trained in proper handling and safety
        precautions.
     3.  All appropriate engineering controls are in place, are maintained, are
        functioning properly, and are utilized.
     4.  Everyone who handles or uses HF has available, knows how to use,
        and is required to use appropriate safety and personal protective
        equipment.
     5.  Arrangements are made ahead of time to provide first aid or medical
        treatment measures if necessary.

   If additional information is necessary., you should write to:

        Technical Service Manager - Hydrofluoric Acid
        Allied-Signal Inc.
        P.O. Box 1053
        101 Columbia Road
        Morristown, New Jersey 07962-1053

-------
';,-*".${ menl of Hydrofluoric Acid Exposure1


Kirsl Aid
• IRtii /Eill>






Medical
Treatment













Skin Burns
Cone. HF Dilute HF
Water Water
Wash Wash
THEN THEN
ZEPfflRAN® A ZEPfflRAN® A
0.13% soaks 0.13% soaks
OR OR
Calcium Gluconate Calcium Gluconate
2.5% Gel 2.5% Gel
Debride Debride
THEN THEN
Continue Continue
Soaks or Soaks or
Calcium Gluconate Calcium Gluconate
2.5% Gel 2.5% Gel
OR OR
Calcium Gluconate Calcium Gluconate
5% Injection2 5% Injection2
AND Systemic
Observe For/ And Effects3
Treat Unlikely
Systemic
Effects3
(especially if >25sq. in.)






















Eye Exposure
A11HF
Water
Wash
THEN
1% Calcium
Gluconate
Irrigation


Topical
Tetracaine
Hydrochloride
THEN
1% Calcium
Gluconate
Irrigation
AND
Consult
Ophthalmologist




















Inhalation
Cone. HF
Oxygen
AND
2.5% Calcium
Gluconate by
Nebulizer



Observe
AND
Treat
Dilute HF
Oxygen
AND
2.5% Calcium
Gluconate by
Nebulizer



Observe
(Serious
Effects
Bronchoconstriction, Unlikely)
Pulmonary Edema
and Systemic
Effects3






NOTE: In addition to the usual medical history.
the physician will find it helpful to obtain the
following information: concentration of
Hydrofluoric Acid, date and time of exposure,
duration of exposure, bow exposure occurred,
body parts exposed/affected, first aid measures

1) This is a brief summary of Fust Aid and Treatment measures. The text of the booklet "RECOMMENDED to dilute Hydrofluoric Acid solutions or low
MEDICAL TREATMENT TOR HYDROFLUORIC ACID EXPOSURE" must be consulted for more com- ^^^^a^^M^^S^
plete information. exposure.
2) 5% Calcium gluconate injections must be used if the gel does not significantly relieve pain in 30-40
minutes. Injections may also be used as the primary treatment, especially for larger and/or deeper burns.
3) Systemic effects include hypocalcemia, hypomagnesemia, hyperkalemia, cardiac arrhythmias, and * Registered Trademark, Wmthrop
altered pulmonary henwdynaraics. TREATMENT includes cardiac monitoring, monitoring serum Laboratories, New York, NY 10016.
calcium, magnesium, and electrolytes; administration of intravenous calcium gluconate, correcting n Registered Trademark, Smart
magnesium and electrolyte imbalance, and dialysis. Phaniuceuacals.Wiimuigton, DE 19897























Ingestion
A1IHF
Do Not Induce Vomiting
Milk or
Water
THEN
Milk of
Magnesia
or
MYLANTA®8
Lavage with
Limewater
AND
Treat
Systemic Effects3




•.


















additional refc charts or information on rties st eandhandT
or medical treatment for hydrofluoric acid, contact:
The HF Products Group
Allied-Signal Inc.
P.O. Box 1053
Morristown, NJ 07962-1053

In the event of an emergency with this product, call the 24-hour Allied-Signal
emergency telephone number: (201) 455-2000.

£age»ttimcoiinn»hgpottl*
-------
Treatment of HF Exposure: Quick Reference            Inside Cover

Table of Contents		,	1

Introduction	2

Acute Toxicity	2
  Skin Contact	2
  Systemic Toxicity	3
  Eye Contact	3
  Inhalation	3
  Ingestion	4

Chronic Toxicity	4

First Aid Treatment For Hydrofluoric Acid Burns		4
  Skin Contact	•	4
  Eye Contact	5
  Inhalation	6
  Ingestion	6


Medical Treatment For Hydrofluoric Acid Burns	6
  Burns of the Skin - General	6
    Quaternary Ammonium Compounds	6
    Calcium Gluconate Gel	7
    Calcium Gluconate Injections	7
    Burns of the Fingers and Nails	7
    Intra-arterial Calcium Infusion	8
    Additional Measures	8
  Systemic Absorption and Metabolic Effects	8
  Eye Injuries	9
  Inhalation Injuries	9
  Ingestion Injuries	10

 References	•	H
 Appendix-First Aid and Medical Supplies	13
 Notes	-	I4

-------
 Introduction
 Hydrofluoric acid is a very strong inorganic acid.
 Both  anhydrous  hydrofluoric  acid   (hydrogen
 fluoride) and its solutions are clear, colorless liquids.
 When exposed to  air,  concentrated  solutions and
 anhydrous hydrofluoric acid produce pungent fumes
 which are  especially dangerous. Unless  heated,
 dilute concentrations of hydrofluoric acid in water do
 not produce vapors.

 NOTE: Persons  unfamiliar with  hydrofluoric acid
 often mistake it or confuse it with hydrochloric acid.
 Although hydrofluoric acid (HF)  and hydrochloric
 acid (HC1) sound similar, the toxicity of these two
 acids is very different. To decrease or avoid confusion,
 we recommend that  HYDROFLUORIC ACID, and
 HYDROGEN FLUORIDE be referred to as "HF".

 HF is primarily an industrial raw material. It is used
 in stainless steel manufacturing, iron and steel foun-
 dries, metal finishing, aluminum manufacturing, in-
 organic  and  organic  chemical  manufacturing,
 petroleum refining,  mineral processing, glassmaking
 and electronic components manufacturing. It is also
 used in certain industrial and consumer cleaning com-
 pounds. However, its use in consumer products is
 discouraged because of the  hazards  described
 herein.

 Most non-industrial burns are caused by dilute con-
 centrations of HF. Most of the HF used in the elec-
 tronics  industry  is  less than  50%  concentration.
 However, many industrial uses  of HF involve con-
 centrated (50-100%) HF.
Skin Contact
Hydrofluoric acid (HF) can cause serious, painful
bums of the skin. Specialized first aid and medical
treatment is required. Burns larger than 25 square
inches (160 square cm) may result in serious
systemic toxicity.

Hydrofluoric acid is a highly corrosive acid which can
severely burn skin, eyes, and mucous membranes. The
vapors from anhydrous hydrofluoric acid or its con-
centrated solutions can also burn these tissues.

Hydrofluoric acid is similar to other acids in that the
initial extent of a burn depends on the concentration,
the temperature and the duration of contact with the
acid. Hydrofluoric acid differs,  however, from other
WARNING: BURNS WITH CONCENTRATED HF
ARE USUALLY VERY SERIOUS, WITH  THE
POTENTIAL  FOR SIGNIFICANT  COMPLICA
TIONS  DUE TO FLUORIDE TOXICITY. CON
CENTRATED HF, LIQUID  OR  VAPOR, MAY
CAUSE  SEVERE  BURNS,  METABOLIC IM
BALANCES, PULMONARY EDEMA AND LIFE
THREATENING  CARDIAC  ARRYTHMIAS
EVEN MODERATE EXPOSURES TO CONCEN-
TRATED HF MAY RAPIDLY PROGRESS TO~A
FATALITY IF LEFT UNTREATED.

Allied-Signal is  the  world's  leading supplier  of
hydrofluoric acid. The recommended medical pro-
cedures described below are based on many years ex-
perience  in dealing with  the unique hazards of this
product,  in addition  to  a review  of the  medical
literature. Every  effort must be made to prevent ex-
posure to hydrofluoric acid. If exposure does occur
the specialized procedures which follow are recom-
mended to avoid  the very serious consequences that
might otherwise occur.

Because  the medical  treatment of  hydrofluoric
acid exposure is so specialized and differs from the
treatment of other inorganic acid exposures, not all
physicians may be aware of appropriate treatment
measures. It is recommended that HF users make ar-
rangements ahead  of time with local medical resources
to be sure that  users are familiar with first aid
measures and that professional personnel are familiar
with the toxicity of HF and the treatment of HF ex-
posure.   This  would  include,  at a  minimum,
thoroughly  reviewing this booklet and making sure
that treatment facilities and supplies are available.
acids because the fluoride ion readily penetrates the
skin, causing destruction of deep tissue  layers in-
cluding bone. Unlike other acids which are rapidly
neutralized, this process may continue for days.

Strong  acid  concentrations (over 50%),  and par-
ticularly anhydrous HF (AHF or 100% HF), usually
cause immediate, severe, burning pain and a whitish
discoloration of the skin which usually proceeds to
blister formation. Exposure to HF vapors can also
result in similar burns.

In contrast to the immediate effects of concentrated
HF,  the  effects  of  contact  with  more  dilute
hydrofluoric  acid or its vapors may be delayed, and
this is one of the problems with the recognition of
some HF burns. Contact with  acid concentrations in

-------
the 20% to 50% range may not produce clinical signs
or symptoms for one to eight hours. With concentra-
tions less than 20%, the latent period may be up to
twenty-four  hours. HF concentrations as low as 2% j
may cause  symptoms if the contact time is  long
enough (1).
HF skin burns are accompanied by severe, throbbing
pain which is thought  to be due to irritation of nerve
endings by increased levels of potassium ions enter-
ing the extracellular  space to compensate for the
reduced levels of calcium ions,  which have  been
bound to the fluoride. Thus, relief of pain is an im-
portant guide to the success of treatment.
The usual initial signs of an HF burn are redness,
edema, and blistering. With more concentrated acids,
a blanched  white area appears.  The  fluoride ion
penetrates the upper layers of the skin. A thick granular
exudate may form under blisters due to liquifaction
necrosis. In rare (and  untreated) cases, there may be
penetration to underlying bone with decalcification.
 HF burns require immediate and specialized first
 aid and medical treatment (2,3,4,5,6) differing from
 the treatment of other chemical burns. If untreated or
 improperly treated, permanent damage, disability or
 death may result (7). If, however, the burns  are
 promptly and properly recognized and managed, the
 results of treatment are generally favorable.

 Treatment is directed toward tying up the fluoride ion
 to prevent tissue  destruction. High molecular weight
 quaternary   ammonium  compounds,  e.g.
 ZEPHIRAN® (benzalkonium chloride), are used as
 soaking agents* (8,9,10). Calcium gluconate as a gel
 or ointment  can be applied locally,  and calcium
 gluconate solution may be injected (subcutaneously,
 intravenously, or intra-arterially), inhaled, or used as
 an irrigant (3,11,12,13,14).

 Speed is of the  essence. Delays in first aid care or
 medical treatment or improper medical treatment will
 likely result in greater damage or may, in some cases,
 result in a fatal outcome.


 Systemic  Toxicity
 One of the most serious consequences of severe ex-
 posure to HF by any route is the marked lowering of
 serum calcium (hypocalcemia)  and  other  metabolic

  *Quaternary ammonium compounds which have proven clinically
  successful  in treating  HF burns include  ZEPHIRAN18 (ben-
  zalkonium  chloride) and HYAMINE® 1622 (benzethonium
  chloride). Because it is available hi the U.S. as a non-prescription
  drug, ZEPHIRAN® is recommended.
changes, which may result in a fatal outcome if not
recognized and treated. Hypocalcemia should be con-
sidered a possible risk in all instances of inhalation or
ingestion, and whenever skin burns  exceed 25
square inches, (160 square  centimeters).  Serum
magnesium may also be lowered, and elevations in
serum potassium have been reported to further com-
plicate the metabolic imbalances which will need to
be monitored and corrected (15,16,17).  High levels
of fluorides have been noted both in the blood and
body organs. Hemodialysis has been reported to be
effective therapy for cases of severe systemic intox-
ication (18,19). Treatment for shock may also be re-
quired as for other severe injuries.

Other effects reported from fluoride exposure include
coagulation defects and inhibition of a number of en-
zymes,  including  preglycolytic   enzymes,
phosphatases and cholinesterase. The results of this
enzyme  inhibition  include  inhibition  of  cellular
glucose phosphorylation and subsequent glycolysis,
inhibition of respiration, and increased sensitivity of
cholinergic mechanisms to acetyl cholinesterase (20).

While  hypocalcemia has  been  traditionally con-
 sidered the major systemic effect of severe poisoning
with  HF,  it  is  apparent that hypbmagnesemia,
 hyperkalemia, the cardiodepressing and vasodilating
 effects  of fluoride,  and  effects  on  pulmonary
 hemodynamics and systemic capacitance vessels, in-
 cluding an increase hi pulmonary vascular resistance,
 all play a role in systemic toxicity. Although some of
 these effects  have  been .described, the implications
 for therapeutic measures have not been well defined
 (21,22,23).                        '
E
   ye Contact
 Hydrofluoric acid can cause severe eye burns, with
 destruction or opacification of the cornea. Blindness
 may result from severe or untreated exposures. Im-
 mediate first aid and specialized medical care is re-
 quired (3,12,).


 Inhalation
 Hydrofluoric acid fumes  may cause. bronchospasm
 and/or  acute pulmonary  edema.  Acute symptoms
 may  include  coughing,  choking, chest tightness,
 chills, fever and cyanosis.  Most reported  fatalities
 from HF  exposures  have been due  to severe
 pulmonary  edema (coupled with systemic  toxicity)
 that did not respond to medical treatment.
 Burns  from vapors   or  liquid  contact  to  the
 oropharyngeal mucosa or upper airway may cause
  severe   swelling  to  the  point  of  requiring a

-------
 tracheostomy. It is recommended that all patients
 with such exposures be hospitalized for observation.
 Because of the strong irritant nature of hydrofluoric
 acid, an individual inhaling HF vapors or fumes will
 usually experience upper  respiratory injury, with
 mucous membrane  irritation and  inflammation  as
 well as cough. All individuals suspected of inhalation
 should  be observed for pulmonary  effects.' This
 would include those individuals with significant up-
 per  respiratory irritation,  broncho-constriction by
 pulmonary auscultation or spirometry, and any in-
 dividual with HF exposure to the head, chest or neck
 areas.  It has been  reported that pulmonary edema
 may be delayed for several hours and even up to two
 days. If there is no initial upper respiratory irritation,
 significant inhalation exposure can generally be ruled
 out.
 The  Permissible  Exposure Limit (PEL)  set by the
 U.S. Occupational Safety and Health Administration
 (OSHA) is a ceiling level of 3 ppm and a 15 minute
 short term exposure limit (STEL) of 6 ppm (24). The
 National Institute for Occupational Safety and Health


 Chronic Toxicitv
 ——————^————^—_—_	«/—
Chronic toxicity from overexposure to fluoride ion
has been  reported to  result in  tooth mottling in
children, bone fluorosis and sometimes osteosclerosis
in adults and children.
Skeletal fluorosis is known to be associated with ex-
cessive exposure to fluoride compounds. Cases of
skeletal fluorosis have been reported in populations
exposed to naturally occurring drinking water  con-
taining greater than 10 ppm of fluoride ion and in in-
dividuals exposed to.high levels of fluoride contain-
ing  dusts.  However, skeletal fluorosis has not been
reported as a consequence of hydrofluoric acid ex-
posure.
Fluorides  are not  carcinogenic, and have not  been

First  Aid  Treatment

For Hydrofluoric  Adh
 (NIOSH) has established a level that is immediately
 dangerous to life and health (IDLH) at 30 ppm (25). The
 American  Industrial  Hygiene  Association  has
 published Emergency Response Planning Guidelines
 setting 50, ppm as the maximum level below which
 nearly all individuals could be exposed for one hour
 without experiencing or developing life-threatening
 health effects (ERPG-3), 20 ppm as the  maximum
 level below which nearly all individuals could be ex-
 posed for one hour without developing irreversible
 health effects or symptoms which would impair tak-
 ing protective action (ERPG-2), and 5 ppm as the
 maximum level below which nearly all individuals
 could be exposed up to one hour without experiencing
 other  than mild, transient  adverse  health effects
 (ERPG-1) (26).

 Ingestion
 If hydrofluoric acid is ingested, severe burns to the
 mouth, esophagus and stomach may occur. Severe
 systemic effects usually also occur. Ingestion of even
 small amounts of dilute HF have resulted in death
 (27).
reported to cause male or female reproductive ef-
fects. Fluoride  exposures should be  kept  below
recommended levels to assure no adverse effects to
the developing fetal skeletal system.

Monitoring of urine for fluorides is an  accepted
method of determining exposure (28). Urine fluoride
levels above  3  mg/liter  at  the  beginning of a
workshift, or above 10 mg/liter  at the end of a
workshift, may  indicate excessive  absorption of
fluoride. It should be noted that fluorides are often
present in significant amounts in persons not occupa-
tionally  exposed (because of dietary sources of
fluoride such as tea), and that the urine fluoride deter-
mination is not specific for HF (29, 30).
In  Case of  Contact  or Suspected  Contact  with
Hydrofluoric Acid:

    Skin  Contact
     1.  Move  victim  immediately under  safety
        shower or other water source and flush af-
        fected area thoroughly with large amounts
        of cool running water. Speed in washing
        off the acid is of primary importance.
    2.  Remove  all  contaminated clothing while
        flushing with water.
    3.  Rinse with large amounts of cool running
        water. If 0.13% ZEPHIRAN® solution or
        2.5% calcium gluconate gel are available,
        the rinsing may be limited to 5 minutes,
        with the soaks or gel applied as soon as
        the rinsing is stopped. If ZEPHIRAN® or
        calcium gluconate gel are not available,
        rinsing must continue until medical treat-
        ment is rendered.
     4.  While  the victim is being rinsed  with
        water,  someone should alert first aid or
        medical personnel and arrange for subse-
        quent treatment.

-------
5.  Immediately after thorough washing, use
    one of the measures below:

    a .  Begin soaking the affected areas in
        0.13% ZEPHIRAN® solution.
        If immersion is not practical, towels
        should be  soaked with iced 0.13%
        ZEPHIRAN® solution  and used as
        compresses for the burned area. Com-
        presses should be changed every two
        to four minutes.
        Do not use ZEPHIRAN® solution for
        burns of the eyes. Exercise  caution
        when  using  ZEPHIRAN® solution
        near the eyes as it is an eye irritant.
        ZEPHIRAN® soaks  or  compresses
        should  be  continued  until pain  is
        relieved or until more definitive care is
        rendered (see below).
    b . Start  massaging   2.5%   calcium
        gluconate gel into the burn site.
        Apply  gel  every  15  minutes and
        rub hi continuously until pain and/or
        redness disappear or more definitive
        care is given (see below).
        It is advisable for the individual apply-
        ing the calcium gluconate gel to wear
        surgical gloves to prevent a possible
         secondary HF burn.

  6.   After treatment of burned areas is begun,
      the victim should be examined to ensure
      there are no other burned areas which have
      been overlooked.


  7.   Arrange to have the victim seen by a physi-
      cian. During transportation to a medical
      facility or while waiting for a physician to
      see the victim, continue the ZEPHIRAN®
      soaks or compresses or continue massaging
      calcium gluconate gel. In many situations,
      particularly  for minor  burns covering a
      small  skin area or  for bums caused by
      dilute  HF, continued treatment with soaks
      or gel may be effective as the sole type of
      medical  care. All  persons with extensive
      burns  or burns with significant blister for-
      mation or with the appearance of whitish or
      dead skin need to be seen by a physician.


  8.   The physician may advise continuation of
      ZEPHIRAN® soaks  or calcium gluconate
      gel.

      a .  If the  physician  advises  continued
          treatment with ZEPHIRAN® soaks or
          compresses, the soaks or compresses
          are usually required for 2 to 4 hours.
      Significant relief of pain should be
      noted within the first 30 minutes.  If
      this does not occur, the victim must be
      seen  by   a  physician  and  more
      definitive care instituted. If the pain is
      substantially  relieved, continue the
      treatment  for  two hours. After that
      time,  discontinue  treatment  and
      observe for the recurrence of pain. If
      pain recurs, continue soaks or com-
      presses for an additional two  hours.
      Soaking for six hours is sometimes
      needed. (Note: Because prolonged im-
      mersion in the ice bath may result hi
      discomfort, relief may be obtained by
      removing the part from the bath every
      ten minutes for a minute or so and then
      reimmersing it. After the initial 30-60
      minutes of treatment, less ice  can  be
      used so the bath is cool rather than
      cold.)

   b. Calcium gluconate gel may be used for
   several hours  or  even  repeated over  a
   period of a few days. However, if signifi-
   cant  relief of pain does not occur within
   30-40 minutes, more definitive treatment
   will be required. For small bums, or burns
   of uje face, ears,  and  near mucous mem-
   branes, calcium gluconate gel may be very
   useful. The  gel is  applied every fifteen
   minutes and massaged into the burned area.
   This is continued until relief is obtained or
   further medical care is available.

9.  Seek medical attention as soon as possible
    for all burns regardless of how minor they
    appear initially.
Eye Contact
1.   Immediately flush the eyes for at least 15
    minutes with large amounts of gently flow-
    ing water. Hold the eyelids open and away
    from the eye  during  irrigation  to allow
    thorough flushing of the eyes. Do not use
    the ZEPHIRAN® solutions described for
    skin treatment. If 1% calcium gluconate
    solution  is available,  washing  may  be
    limited to 5 minutes, after which the 1%
    calcium gluconate solution should be used
    repeatedly  to  irrigate the  eye using  a
    syringe.

2.   Take the victim to a doctor, preferably an
    eye specialist, as  soon as  possible. Ice
    water compresses should be applied to the
    eyes while transporting the victim to the
    doctor.

-------
        3.  If a physician is not immediately available,
            apply one or two drops of 0.5% PONTO-
            CAINE® hydrochloride solution or other
            aqueous, topical ophthalmic anesthetic and
            continue irrigation. Use no other medica-
            tions unless instructed to do so by a physi-
            cian. Rubbing of the eyes is to be avoided.

        For Inhalation of Vapors

        1.  Immediately move victim to fresh air and
            get medical attention.
        2.  Keep victim warm, quiet and comfortable.
       •3.  If breathing has stopped,  start artificial
            respiration at once. Make sure mouth and
            throat are free of foreign  material.
        4.  Oxygen should be administered as soon as
            possible by a trained individual.  Continue
            oxygen  while awaiting medical attention
            unless instructed otherwise by a physician.
        5.  A nebulized solution  of 2.5% calcium
            gluconate may be administered with oxy-
            gen by inhalation.
       6.  Do not give stimulants unless instructed to
          do so by a physician.
       7.  The victim should be examined by a physi-
          cian and held under observation for at least
          a 24 hour period.
       8.  Vapor exposures  can cause  skin  and
          mucous  membrane  burns  as  well  as
          damage to pulmonary tissue. Vapor burns
          to the skin are treated the same as liquid
          HF burns.

      If Ada'  is Ingested

      1.  Have the victim  drink large amounts of
         water as quickly  as possible to dilute the
         acid. Do not induce vomiting. Do not give
         emetics or baking soda. Never give any-
         thing by mouth to an unconscious person.
      2.  Give several glasses of milk or several
         ounces of milk of magnesia, MAALOX®
         MYLANTA®   etc.  The  calcium  or
         magnesium in these compounds may act as
         an antidote.
      3.  Get immediate  medical attention.
  Medical   Treatment
  For Hydrofluoric  Acid Burns
  Burns of the Skin - General

 Bums from dilute acid are difficult to  distinguish
 from other chemical burns and usually appear as
 areas of erythema. However, they may progress, if
 not treated, to areas of blistering, necrosis or ulcera-
 tion. Bums from more concentrated acid  have  a
 rather  characteristic  appearance and present  as
 severely  reddened, swollen  areas  with blanched,
 whitish regions which rapidly progress to blistering
 and necrosis.  A thick granular exudate usually ap-
 pears under these blisters which requires debridement
 and removal.


 Hydrofluoric acid burns cause  extreme  pain. The
 pain is thought to result from nerve ending irritation
 due to  increased levels of potassium ions  in ex-
 tracellular spaces to  compensate for the reduced
 levels of calcium ions which have been bound by the
 fluoride. Relief of pain is an excellent indication of
 the success of  treatment and,  therefore,  local
 anesthetics should be avoided.

 Many different types of therapies have been sug-
 gested for HF burns. The aim of all  treatment is to
 chemically sequester the fluoride ion  and to prevent
 extensive, deep-tissue destruction (31).

After treatment of burned areas is begun, the victim
should be carefully examined to insure there are no
other burned areas which may have been overlooked.
 Quaternary Ammonium Compounds

 Most HF burns can be satisfactorily treated by im-
 mersion of the burned part in an iced, aqueous (or
 aqueous-alcohol) solution of a quaternary ammonium
 compound. Two solutions have been clinically suc-
 cessful,  0.13%  ZEPHIRAN® (benzalkonium
 chloride) or 0.2% HYAMINE® 1622 (benzethonium
 chloride). Because of its availability  as  a non-
 prescription drug, ZEPHIRAN® is recommended in
 the United States.

 The solutions should be cooled with ice cubes. (Shaved
 or crushed ice may cause excessive cooling, with the
 danger of frostbite.)

 If immersion in the solution is not practical, soaked
 compresses of the same iced solution should be ap-
 plied to the burned area.  The immersion or com-
 presses should be used for at least two hours. Com-
 presses should be changed or soaked with additional
 solutions approximately every  two to four minutes.

 If blisters are present, they should be opened and
 drained and necrotic tissue should be debrided prior
 to use of ZEPHIRAN® immersion or compresses.

 Prolonged immersion  in  the  iced ZEPHIRAN®
 bath may result in discomfort due to excess chilling;
 relief may be obtained by removing the part from the
 bath every ten to fifteen minutes for a few minutes
and then reimmersing it.  After the initial 30-60
minutes of treatment, less ice can be used so the bath
is cool rather than cold.

-------
The success of this treatment is indicated by relief of
the severe  pain hi the burned area. If pain recurs
when the treatment is stopped at the end of the first
two  hours, immersion or compresses should be
resumed for an additional two hours. A total of four
to six hours  immersion or use  of compresses of
ZEPHIRAN® is usually required for the treatment of
most burns. No further treatment will be required in
many instances. The  use of iced quaternary am-
monium compound solutions offer several advan-
tages:
      a. reduction of local pain
      b. possible slowing of the rate of tissue destruc-
        tion
      c. possible  slowing  of  the  passage  of the
        fluoride  ion   into  tissues and  into  the
        bloodstream

Large burns,  significant burns  due to concen-
trated HF, or burns with delayed treatment will
probably require the use of calcium gluconate in-
jections in   addition  to  or   instead   of  the
ZEPHIRAN® soaks.

Quaternary ammonium compounds should not be
used  for burns on the face, ears or other sensitive
areas due to their irritating nature. It is preferable to
use calcium gluconate gel or calcium gluconate injec-
tion in these areas.

 Calcium Gluconate Gel
Calcium gluconate gel, consisting of 2.5%  USP
calcium gluconate in a surgical water soluble lubri-
cant, is widely used for first aid and/or primary treat-
ment of HF burns of the skin. The gel is convenient to
cany and can be used to initially treat small burns that
might occur away from medical care. The gel is used
by massaging it promptly and repeatedly into the
burned area,  until pain  is  relieved.  If possible,
surgical gloves should be worn during initial applica-
tion of the gel, so the person providing treatment will
not receive a secondary burn. This treatment can be
started without waiting for medical direction.

If used as the only method of treatment, liberal quan-
tities of calcium gluconate gel must be massaged into
the burned area intermittently  for several hours.
Relief of pain can be used to assess the efficacy of this
treatment.  If good relief of pain is not obtained after
30-40 minutes, alternate methods of treatment such as
calcium gluconate injections or ZEPHIRAN® soaks
should be considered. The gel may have to be used
4-6 times daily for 3 to 4 days.

 The gel is especially useful for burns on the face, par-
 ticularly near the mouth and eyes  or on the ears.  It
may be convenient to use the gel for very small burns
where the victim can easily apply and massage the gel
himself. Use of the gel may be more convenient for
dilute acid burns such as occur with commercial pro-
ducts like rust removers, aluminum cleaners or etch-
ing solutions.

Calcium  Gluconate Injections
After first aid measures have been taken, injection of
calcium gluconate solution is indicated as the primary
medical treatment for large burns (over 25  square
inches). For smaller burns, if ZEPHIRAN® soaks or
calcium gluconate gel do not promptly result hi relief
of pain, injection of calcium gluconate solution is in-
dicated. Injection  of calcium gluconate solution may
also be indicated  for burns hi which treatment has
been delayed. The physician should inject sterile 5%
aqueous calcium gluconate beneath, around and into
the burned area. Calcium gluconate is packaged as a
10% solution, and must be diluted 50-50 (equal parts)
with normal saline. (Note: DO NOT USE calcium
chloride, which is corrosive and may result hi addi-
tional damage.)

If subcutaneous calcium gluconate injections are used,
the amount  injected initially is small and should not
exceed 0.5 cc per square centimeter of affected skin
surface.  The injections should not distort the  ap-
pearance of the skin. A  small-gauge needle  (27-30
gauge) should be used, and the burned area should be
injected through multiple sites. The patient can usual-
ly advise when the pain stops, and this is an indicator
of adequate  treatment. Multiple injections hi skin that
has compromised integrity may increase the risk of
infection, and the use of antibiotic creams such as
SILVADENE® (silver   sulfadiazine)   or
GARAMYCIN® (gentamicin sulfate  cream) should
be considered  following  such  treatment.  Local
anesthetics should not be used since they mask pain
relief which is an important indication of adequacy of
treatment.

Some physicians  prefer using calcium gluconate in-
jections initially as the primary treatment, instead of
using  quaternary  ammonium  compound  soaks or
compresses or using calcium gluconate gel. Injections
often are not necessary when there has been early and
adequate treatment with soaks or gel.
 Burns Of The Fingers And Nails
 Burns of the fingers often create special problems hi
 treatment.  Finger and toe nails permit penetration of
 fluoride ions but prevent soaks or gels from being ef-
 fective.  It  may be necessary on occasion to split or
 even remove nails to allow the topical methods of

-------
 treatment to be effective. One author has cautioned
 that removal of the nail should rarely be necessary in
 the case of dilute HF acid (less than 10%) burns (32).
 The treating physician must consider the morbidity
 associated with removal of the nail versus the need to
 treat the HF exposure.

 If immersion in ZEPHIRAN® solution is started im-
 mediately, it may be possible to avoid removing the
 nail. Sometimes better penetration under the nail can
 be successfully accomplished by splitting the nail or
 by drilling several burr holes in the nail using a large
 gauge needle or a nail drill. If calcium gluconate in-
jection is used as treatment, the nail may still need to
 be split or removed. Some authorities recommend the
 use of general anesthesia or a regional nerve block,
 rather than local anesthesia, to remove the nail so that
 pain relief by calcium gluconate injections may be used
 as an indicator of effective treatment. When using
 calcium gluconate injections hi the digits, care must
be taken to inject the solution slowly so as to avoid
any compromise to the circulation in these areas.

 Intra-arterial  Calcium Infusion
 Reports in the literature  have described the use of
intra-arterial injection or infusion of dilute  calcium
 gluconate solutions to treat HF burns of the hand and
digits which do not respond to other methods, either
due to inadequate or improper treatment, or  in cases
where treatment  has been greatly delayed.  The
method is described as follows:

    "A long catheter was inserted percutaneously in-
    to the radial artery using standard aseptic tech-
    nique. Intra-arterial catheter placement was con-
    firmed by pressure transducer and oscilloscope.
    If the burn involved only the thumb, index, or
    long fingers, the catheter was advanced only a
    few  centimeters proximally in  preparation for
    digital subtraction arteriography. If the burn in-
    volved the ring or small fingers, the catheter was
    advanced  proximally  into the  brachial artery
    because  access to the  ulnar  circulation  was
    necessary.
    Following  satisfactory placement of the arterial
    catheter, we performed digital  subtraction  ar-
    teriography on all patients in our series to  identify
    the origin of vascular supply to digits involved.
    Once the tip of the arterial  catheter was in the
    desired location, a dilute preparation of calcium
    [gluconate] (10 ml of a 10% solution mixed in 40
    to 50 ml 5% dextrose) was infused with a pump
    apparatus into the catheter over four hours. We
    generally have  used calcium gluconate....  Each
   patient was observed closely during the infusion
   period for progression of symptoms and potential
     complications of the procedure,  such as altera-
     tions of distal vascular supply.
     Following  the four-hour infusion, the arterial
     catheter  was maintained in place in the usual
     manner while the patient underwent an observa-
     tion period. If typical HF pain returned within
     four hours,  a  second calcium infusion was
     repeated until the patient was pain free four hours
     following completion of the calcium infusion
     (13)"

 This method, although  rather involved, should be
 considered in selected cases, especially where inade-
 quate or delayed treatment has occurred.

 Additional Measures
 Where  blistering  and/or  necrosis  occur,  early
 debridement may facilitate healing.

 In instances of extensive burns, skin grafting has oc-
 casionally been required, but the need for this treat-
 ment should be markedly reduced by immediate and
 aggressive primary treatment.

 Follow-up care requires monitoring to prevent secon-
 dary infections, and the use of antibiotic creams such
 as  SILVADENE®or GARAMYCIN® has  proven
 effective. HF burns may heal slowly, but if properly
 treated most heal with little or no scarring in 14 to 28
 days.


Systemic Absorption  and
Metabolic  Effects
Significant amounts of fluoride ion may be absorbed
by skin  contact,  inhalation, or by  ingestion.  If
systemic   absorption   of  fluoride  occurs,
hypocalcemia, hypomagnesemia  and  hyperkalemia
may also occur.  All of these parameters need to be
monitored and appropriate- therapeutic measures in-
stituted. The patient should be observed for clinical
signs of hypocalcemia following ingestion or inhala-
tion or following extensive burns greater than 25
square inches. Serum calcium determinations must be
performed immediately and periodically to monitor
and treat hypocalcemia. Severe lowering of serum
calcium levels can occur within one to two hours even
with HF burns covering less than 2.5% of body sur-
face area (7). Continuous EKG monitoring to observe
prolongation of the Q-T interval may be useful to
detect early changes in serum calcium, although pro-
found hypocalcemia following HF exposure has been
reported in the absence of EKG changes.

The fall in serum calcium may occur precipitously
following HF exposure. In two reported cases of ex-
posure to anhydrous HF, the serum calcium fell to

-------
levels around 3 milliequivalents per liter (mEq/L)
[normal = 8.8 -  10.3 mEq/L] within one to three
hours of exposure (7,32).
If necessary, aqueous  calcium gluconate may be
given intravenously. Calcium  gluconate as a 10%
solution must be given slowly  since excess calcium
can  produce  vagal bradycardia,  ventricular  ar-
rhythmias  and  ventricular fibrillation. The  IV
calcium gluconate should be repeated until serum
calcium levels  return to,  and remain  at,  normal
levels. In one fatal case, 280 mEq of calcium over
four hours was not sufficient to correct the profound
hypocalcemia (7). Without additional measures such
as renal dialysis, it may not be possible to correct ex-
treme hypocalcemia.
Serum magnesium levels should also be monitored
and magnesium loss should be replaced intravenously
if indicated. Serum potassium must also be carefully
monitored. Significant elevations of serum potassium
have been noted in cases of fluoride toxicity and also
in laboratory studies. Hyperkalemia has also been im-
plicated as a causative factor in cardiovascular col-
lapse. The use of quinidine may be helpful in preven-
ting this serious complication (19).

Renal dialysis with fluoride free  water, in conjunc-
tion with other treatments mentioned, should be con-
sidered in all cases of serious burns and may need to
be repeated if indicated (18,19).  Serum  fluoride
levels should be monitored. Normal plasma fluoride
levels may differ because of various methodologies
and analytical techniques. The decision to use dialysis
should be based on the clinical condition of the pa-
tient, including the serum levels of fluoride, calcium
and potassium.
Primary excision  has been recommended by some
practitioners as a method of reducing systemic ab-
sorption of fluoride (33). While this could in some in-
stances be life saving, it is a rather drastic measure. It
is likely that renal dialysis could be used to effectively
treat systemic toxicity  and  would  not result in
disfigurement, disability, or morbidity which could
be associated with primary excision.
 Eye  Injuries

 HF can cause severe eye bums which, if not properly
 treated,  may result  in scarring and blindness.. The
 prognosis is not good if first aid treatment is delayed
 or inadequate. After first aid treatment (see FIRST
 AID section) the following medical treatment may be
 provided:
For minor exposures with very dilute HF, the follow-
ing treatment has been successful:

     Mix 10 ml of 10%. calcium gluconate with 100
     ml of normal saline to give approximately a 1 %
     calcium gluconate solution. With a syringe, ir-
     rigate the eye intermittently for a period of 15
     to 30 minutes or until relief of pain occurs.

With more serious HI' eye burns, good results have
been reported with the following procedure:
     Mix 50 ml of 10% calcium gluconate with 500
     ml of normal saline to give approximately a 1 %
     calcium gluconate solution. Using an eye clamp
     and IV infusion set under local anesthetic eye
     drops,  instill the solution  over a period of one
     to two  hours. More prolonged use of the solu-
     tion could  possibly  damage the cornea.  A
     MORGAN THERAPEUTIC LENS® connected
     to an IV line may be a simpler method of infus-
     ing  the calcium gluconate solution.  This treat-
     ment has been  reported to result in reversal of
     corneal edema and  to prevent permanent eye
     damage or loss of vision.  Consultation with an
      ophthalmologist to consider the use of steroids,
      antibiotics or  additional  treatment  is  recom-
      mended.

Inhalation Injuries

Patients with inhalation exposures should also be
observed for  signs of systemic absorption and
•fluoride toxicity.

Exposure to hydrofluoric acid fumes can cause acute
 respiratory irritation, bronchospasm, and/or pulmonary
edema. Medical personnel should also be alert to the
possibility of  development of pulmonary  edema
when extensive burns of the face, neck or chest have
occurred.

The victim should be removed from exposure and ad-
ministered 100%  oxygen immediately. The use of
2.5% aqueous calcium gluconate given by nebulizer
with 100%  oxygen, or  with  intermittent positive
pressure, has been recommended. Theoretically this
 should reduce toxicity and damage from the fluoride
 ion and should be seriously considered in cases of in-
halation exposure.

 Burns of the oral mucosa or upper airway may cause
 severe swelling and necessitate a tracheostomy. It is,
 therefore, recommended that all such patients be ad-
 mitted to a hospital for observation.

 Because  inhalation  of HF may be  associated with
 significant bronchospasm, inhaled, oral or parenteral
 bronchodilators should be administered as necessary.

-------
10
                   Pulmonary function testing may be helpful in assess-
                   ing the degree and progress of pulmonary injury.

                   Specific measures may be needed to treat pulmonary
                   edema.  High doses of parenteral steroids  may be
                   needed along with the administration of appropriate
                   diuretics. Caution should be taken not to administer
                   excessive  fluid.  Hemoconcentration  may  require
                   treatment  by  phlebotomy.  The  management of
                   pulmonary edema may result in renal failure due to
                   reduced fluid volume, and this may be another indica-
                   tion for renal dialysis.

                   If  it is necessary to  relieve anxiety,  use  general
                   measures and do not use sedatives which could cause
                   central nervous system depression or  hypoventila-
                   tion. Although  right heart failure is uncommon in
                   chemically-induced pulmonary edema, monitoring of
                   pulmonary pressure, arterial pressure, and central
                   venous pressure may be indicated.

                   Secondary infections must be treated. It is preferable
                   to start antibiotics at the first signs of infection such
                   as fever or tachycardia. Periodic blood cultures may
                   be advisable. Prophylactic use of antibiotics is not ad-
                   vised.

                   Ingestion Injuries
                   After first aid is completed (drinking several glasses
                   of  water followed by two glasses of milk or two
                   ounces of milk of magnesia, MYLANTA®, or other
                   calcium or  magnesium  containing antacids),  the
                   stomach may be lavaged with lime water. The-Levin
                   tube must be passed with care to prevent perforation.
                   Treatment is the same as for ingestion of other strong
                   acids.  Systemic toxicity is  very likely  to occur and
                   may require aggressive treatmenT

-------
References
1. Derelenko, M.J., et al: Acute Dermal Toxicity of Dilute Hydrofluoric
Acid. J. Toxicol-Cut and Ocular Toxicol 4:73-85, 1985.
2. MacKinnon,  M.A.:  Hydrofluoric Acid Burns. Dermatologic Clinics
6:67-74, January, 1988.
3. Trevino, M.A.: Treatment of Severe Hydrofluoric Acid Exposures. J. Oc-
cup. Med. 25:861-3, December, 1983.
4. Edelman, P.: Hydrofluoric Acid Burns. State of the Art Rev. Occup. Med.
1:89-103, 1986.
5. Upfal, M. and Doyle,  C.: Medical Management of Hydrofluoric Acid Ex-
posure. J. Occup. Med.  32: 726-31, August, 1990.
6. Caravati, E.M.: Acute Hydrofluoric Acid Exposure. Am. J. Ernerg. Med.
6: 143-50, March, 1988.
7. .Tepperman, P.B.: Fatality Due to Acute Systemic Fluoride Poisoning
Following a Hydrofluoric Acid Skin Burn. J. Occup. Med. 22: 691-2, Oc-
tober, 1980.
8. Wetherhold, J.M.: Treatment of Hydrofluoric  Acid Burns. J. Occup.
Med. 7:193-5, May, 1965.
9. MacKinnon, M.A.: Treatment of Hydrofluoric Acid Burns (letter). J. Oc-
cup. Med. 28:804, September, 1986.
10. Reinhardt, C.F.: Hydrofluoric Acid Burn Treatment. Am. Ind. Hyg.
Assoc.  J. 27:166-71, 1966.
11. Browne, T.D.: The Treatment of Hydrofluoric Acid Burns. I. Soc. Oc-
cup. Med. 24:80-9, July, 1974.
12. Rose L. and Trevino, M.A.: Further Evaluation of Hydrofluoric Acid
Burns of the Eye. J. Occup. Med. 26:483-4, July, 1984.
13. Vance, M.V.: Digital Hydrofluoric Acid Burns: Treatment with Intra-
arterial Calcium Infusion. Ann. Emerg. Med. 15:59-65, August, 1986.
14. Davanzo, F.  et al:  Hydrofluoric Acid Intoxication: A New Therapy.
Med. Lav. 78:333-6, 1987.
15. Mclvor, M.E.:  Delayed Fatal Hyperkalemia in a Patient with Acute
Fluoride Intoxicaiton. Ann. Emerg. Med.  16:1166-7. October, 1987.
16. Cummings, C. and Mclvor, M.E.: Fluoride Induced Hyperkalemia: The
Role of CA2 Dependent  K Channels. Am. J. Emerg. Med. 16:1-3, January,
1988.
17.  Boink,  A.M.,  et  al: An Investigation  of  the Pathophysiological
Mechanisms of Hydrofluoric  Acid Intoxication  in Rats and  Pigs: Interim
Report. RIVM Report 318802 002, August, 1990.
18. Picazo, H.C.:  Fluoride Removal from the Blood by  Hemodialysis.
Presented at HF Industry Safety Seminar, August, 1990.
19. Mclvor, M.E.:  Acute Fluoride Toxicity: Pathophysiology and Manage-
ment. Drug Saf. 5:79-85, 1990.
20. Sodium Fluoride. Hazardous Substances Data Bank, National Library of
Medicine, 1990.
21. Gaugl, J.F. and Woolridge,  B.: Cardiopulmonary Response to Sodium
Fluoride Infusion in the  Dog. J.  Toxicol Environ. Health  11:765-82, 1983.
22. Strubelt, O., et al:  The Pathophysiological Profile of the Acute Car-
diovascular Toxicity of Sodium Fluoride. Toxicology 24:313-23, 1982.

-------
12
       References Continued
      23. Braun, C.L.J.: RIVM/CTEF study on HF. Personal communication,
      1990.
      24. Air Contaminants; Final Rule. Department of Labor (OSHA) Fed. Reg.
      54:2465, January 19, 1989.
      25. NIOSH Pocket Guide to Chemical Hazards. DHHS (NIOSH) Publication
      No. 90-117, p 126, June, 1990.
      26.  Emergency  Response  Planning Guidelines:  Hydrogen  Fluoride.
      American Industrial Hygiene Assn., October, 1988.
      27. Manoguerra, A.S.: Fatal Poisoning from Acute Hydrofluoric Acid Inges-
      tion. Am. J. Emerg. Med. 4:362-363, 1986.
      28. Kono, K., et al: Urinary Fluoride Monitoring of Industrial Hydrofluoric
      Acid Exposure. Environ. Res. 42:415-20, April, 1987.
      29.1991-1992 Threshold Limit Values for Chemical Substances and Physical
      Agents and Biological Exposure Indices. American Conference of Govern-
      mental Industrial Hygienists, p 23, Cincinnati, Ohio, 1991.
      30. Documentation of the Biological Exposure Indices. American Conference
      of Governmental  Industrial Hygienists, pp BE197-101, Cincinnati, Ohio,
      1988.
      31. Dunn, B.J., et al: An Assessment of the Efficacy of Selected Treatment
      for HF  Dermal  Burns. Presented  at Society of  Toxicology  Meeting,
      February, 1990.
      32. Roberts, J.R. and Merigian, K.S.: Acute Hydrofluoric Acid Exposure
      (letter). Am. J. Emerg. Med. 7:125-6, January,  1988.
      33.  Buckingham,  P.M.:  Surgery:  A   Radical  Approach  to  Severe
      Hydrofluoric Acid Burns. J. Occup. Med. 30:873-4, November, 1988.

-------
                                                                                                     13
Appendix

First Aid and
Medical Supplies
The following supplies should be maintained in a
dispensary or first aid station near hydrofluoric acid
handling and storage areas:

    1. ZEPHIRAN® solution*
      a .    For soaks  and compresses,  3 to 4
            gallons of 0.13% water solution (1:750)
            of  ZEPHIRAN® (benzalkonium
            chloride). The  1:750 (0.13%) solution
            can be purchased as a non-prescription
            drug in gallon containers.
            This solution should be obtained in ad-
            vance and  replaced annually.  It  is
            recommended mat the solution be stored
            in properly  labelled light-resistant con-
            tainers.

      b .   Ice cubes (not crushed or shaved ice).
      c .   Assorted basins (for immersing burned
            areas in ZEPHIRAN® solution).
      d .   Towels (for use as wet compresses).

    2. Calcium gluconate gel
      Calcium  gluconate  gel  (2.5%   calcium
      gluconate in a water soluble base) may require
      a prescription from a physician. It may be for-
      mulated  by a  pharmacist  by combining 2.5
      grams of calcium gluconate USP  in 100 ml
      K-Y® Jelly  (3.2 grams per 4 ounce tube). It
      may also be made by mixing one ampule of
       10%  calcium gluconate solution for  each 1.5
       ounces of K-Y® Jelly (about 30 ml per 4
       ounce tube). Although this makes a somewhat
       "soupy" mixture, it has  the advantage that
       the ingredients may be stored separately until
       needed,  and Shelf life is less of a concern.


    3. Aqueous calcium gluconate,  10% USP. 10 cc
       ampules
       a .    To make calcium gluconate gel, or
       b .    To mix with sterile saline for eye irriga-
             tion (5 ampules 10% calcium gluconate
             per 500 cc sterile normal saline for a 1 %
             solution), or

       *In addition to ZEPHIRAN®, HYAMINE®  1622 has also
       been used successfully to  treat HF burns. Because of its
       availability as a non-prescription drug,  ZEPHIRAN* is
       recommended. Additional  information  concerning
       ZEPHIRAN® or HYAMINE® 1622 solutions can be ob-
       tained by writing to the addresses listed on page 14.
      c .   To mix with sterile saline for administra-
           tion with oxygen by nebulization (10 cc
           10% calcium gluconate in 30 cc sterile
           saline for a 2.5% solution), or
      d .   To  be  administered  by  a physician.
           When  injected subcutaneously,  10%
           calcium gluconate must be diluted half
           and half with normal saline to produce a
           5% solution
   4. Sterile 0.9% saline
      a.   50  cc vials to dilute  10% calcium
           gluconate to 5% for injection.
      b.   500  cc  IV to  dilute  10% calcium
           gluconate to 1% for eye irrigation.
   5. 0.5% PONTOCAINE®  (tetracauie  hydro-
      chloride)   solution  to   counteract
      blepharospasm and facilitate eye irrigation.
   6. Medical oxygen.
   7. Nebulizer,  to  administer  2.5%  calcium
      gluconate with oxygen.
   8. Beta adrenergic bronchodilators and steroids
      for inhalation.
   9. Surgical gloves.
   10. Syringes and needles (27-30 gauge).
   11. MORGAN THERAPEUTIC LENS®
The FIRST AID AND MEDICAL TREATMENTS
AND SUPPLIES recommended in this booklet are
based on information reported in the medical literature
and the personal experience of physicians with Allied-
Signal, Inc. It should be noted that there are no medica-
tions in the U.S. for which the specific indication is the
treatment of hydrofluoric acid bums. The physican has
the dilemma of using prescription drugs  in  a'non-
approved manner, or of using substances which are not
approved drugs but which have been proven effective
for medical treatment. Given the choice between recom-
mending effective treatment, or recommending the use
of only drags which are approved, we have chosen to
recommend the effective treatment.

ZEPHIRAN® (benzalkonium chloride) is available in
the U.S. as a non-prescription drug. It is a surface ac-
tive agent:  sold for use as a disinfectant. It is available
in a 7:750 aqueous solution, a 17% concentrate, and a
tinted tincture. The concentrated 17% solution must
be dilutee!. The tinted tincture is not recommended to
treat HF exposures.

HYAMINE® 1622 (benzethonium chloride) has been
used in veterinary medicine as an antiseptic for
wounds and infections, but it is not available as a

-------
14
                 Appendix Continued
                drug. Care should be taken that HYAMINE® 1622 is
                used, not  HYAMINE® with other numeric  or
                alphanumeric modifiers.

                CALCIUM GLUCONATE GEL is  not available
                commercially  in  the  United States as  a phar-
                maceutical agent.  It is hoped that a commercial
                preparation of calcium gluconate gel will be approved
                by the U.S. Food and Drug Administration,  and
                become available hi the U.S. as it now is in Canada,
                Great Britain, and other countries. At the present
                tune in the U.S., however, calcium gluconate gel re-
                quires a prescription either for the compounded gel or
                for the 10% solution. The shelf life is uncertain, but it
                is recommended that calcium gluconate gel be replac-
                ed on an annual basis.

                CALCIUM GLUCONATE INJECTION, USP (one
                gram in 10 ml, 10% solution)  is labelled for in-
                travenous use only. Experience has shown that when
                diluted to  5%  with normal  saline,  and  used  as
                described hi this booklet, it is a  safe and effective
                treatment for HF  skin  exposure. When diluted to
                2.5 % and used as described, it is safe for nebulization
                and inhalation, and when diluted to 1.0% and used as
                described, it is safe for eye irrigation.
                Notes:
                GARAMYCDSf® is a Registered Trademark of Scher-
                ing Corporation, Kenilworth, NJ 07033
                HYAMINE® 1622  is a  Registered Trademark of
                Lonza, Inc., Fairlawn, NJ 07410

                K-Y® Jelly is a Registered Trademark of Johnson &
                Johnson Products, Inc., New Brunswick, NJ 08903

                MAALOX®is  a Registered Trademark of Rorer
                Pharmaceutical Corporation, Fort Washington PA
                19034
                MORGAN THERAPEUTIC LENS® is a Registered
                Trademark of MorTan, Inc., Missoula, MT 59806
                MYLANTA® is a Registered  Trademark of Stuart
                Pharmaceuticals, Wilmington, DE  19897
                PONTOCAINE® is a Registered Trademark of Win-
                throp Laboratories, New York, NY 10016
                SILVADENE®  is a Registered Trademark of Marion
                Laboratories, Kansas City, MO 64137
                ZEPHIRAN® is a Registered Trademark  of  Win-
               throp Laboratories, New York, NY 10016

-------
                                                                                                      15
                                  Notes
FOR ADDITIONAL INFORMATION

For additional information on properties, storage and handling, or medical treatment for
hydrofluoric acid, contact:

           Technical Service Manager—Hydrofluoric Acid
           Allied-Signal, Inc.
           P.O. Box 1053
           101 Columbia Road
           Morristown, NJ 07962-1053

In the event of an emergency with this product, call the 24-hour Allied-Signal emergency
telephone number: (201) 455-2000.

All statements, information, and data given herein are believed to be accurate and reliable but are
presented without guaranty, warranty, or responsibility of any kind, express or implied. Statements or
suggestions concerning possible use of our products are made without representation or warranty that
any such use is free of patent infringement and are not recommendations to infringe any patent. The user
should not assume that all medical and first aid measures are indicated or that other measures may not be
required.

-------
                                 This page intentionally left blank.
Page Vl-20

-------
                    APPENDIX VII
FACILITIES REPORTING TO TRI FOR HYDROGEN FLUORIDE, 1990
FACILITY
3M
D & H CHEMICALS INC.
REYNOLDS METALS CO. ALLOYS PLANT
REYNOLDS METALS CO. SHEFFIELD PLANT
TELEDYNE SC
CHEM-FAB
GREAT LAKES CHEMICAL CORP. SOUTH PLANT
AMERICAN NATIONAL CAN CO.
CHEM RESEARCH CO. INC.
DOLPHIN INC.
GENERAL SEMICONDUCTOR INDUSTRIES INC.
INTEL CORP. CHANDLER CAMPUS
KERLEY AGING.
MICROCHIP TECHNOLOGY INC.
MOTOROLA INC.
MOTOROLA INC. DMTG
OLIN HUNT SPECIALTY PRODUCTS INC.
ADVANCED MICRO DEVICES INC.
AEROCHEM INC.
CITY
DECATUR
NEW BROCKTON
MUSCLE SHOALS
SHEFFIELD
HUNTSVILLE
HOT SPRINGS
EL DORADO
PHOENIX
PHOENIX
PHOENIX
TEMPE
CHANDLER
PHOENIX
CHANDLER
MESA
PHOENIX
CHANDLER
SANTA CLARA
ADELANTO
STATE
AL
AL
AL
AL
AL
AR
AR
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
CA
CA
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
100,000-999,000
10,000-99,999
10,000-99,999
10,000-99,999
10,000-99,999
10,000-99,999
1,000,000-9,999,999
10,000-99,999
100-999
100-999
1,000-9,999
10,000-99,999
100,000-999,000
1,000-9,999
100-999
10,000-99,999
10,000-99,999
1,000-9,999
100,000-999,000
QUANTITY RELEASED
OR TRANSFERRED
(POUNDS PER YEAR)
86,900
0
0
250
299
255
210
250
750
250
250
750
5
260
9,350
5,801
1
626
9,977
SIC
CODE
2821
2842
3353
3479
3399
3728
2869
3411
3471
3324
3674
3674
2873
3674
3674
3674
2819
3674
3728
                                                             Page VII-1

-------
FACILITY
AEROCHEM INC.
ALCOA COMPOSITES INC. ASTECH DIV.
ALLIED-SIGNAL INC.
ALLOYS CLEANING INC.
AMERICAN NATIONAL CAN CO.
AMERICAN NATIONAL CAN CO.
AMERICAN NATIONAL CAN CO.
ANALOG DEVICES INC. BOURNS CO.
APPLIED SOLAR ENERGY CORP.
CARPENTER TECHNOLOGY CORP. SPECIAL
PRODUCTS DIV.
CASPIAN INC.
CHEM-TRONICS
CHEMTECH INDUSTRIES INC.
CROWN BEVERAGE PACKAGING
CROWN BEVERAGE PACKAGING
CYPRESS SEMICONDUCTOR CORP.
DOUGLAS AIRCRAFT CO.
DOUGLAS AIRCRAFT CO.
DOW CHEMICAL CO.
DU PONT ANTIOCH ANTIOCH
EG&G KT AEROFAB
CITY
ORANGE
SANTA ANA
EL SEGUNDO
LOS ANGELES
CHATSWORTH
FAIRFIELD
LOS ANGELES
SANTA CLARA
CITY OF INDUSTRY
EL CAJON
SAN DIEGO
EL CAJON
GARDENA
UNION CITY
VAN NUYS
SAN JOSE
LONG BEACH
TORRANCE
PITTSBURG
ANTIOCH
EL CAJON
STATE
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
100-999
1,000-9,999
100,000-999,000
10,000-99,999
10,000-99,999
10,000-99,999
10,000-99,999
1,000-9,999
1,000-9,999
10,000-99,999
1,000-9,999
10,000-99,999
10,000-99,999
10,000-99,999
10,000-99,999
100-999
10,000-99,999
10,000-99,999
100,000-999,000
1,000,000-9,999,999
0-99
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
1,500
11,076
516
10
250
250
250
2,270
250
0
255
1,551
250
0
0
2,753
2,455
6,605
4
600
250
SIC
CODE
3728
3471
2819
3479
3411
3411
3411
3674
3674
3499
3728
3724
2899
3411
3411
3674
3721
3728
2879
2816
3728
Page vii-2

-------
FACILITY
EXAR CORP.
EXSIL INC.
FLO-KEM INC.
GENERAL CHEMICAL CORP.
GOLDEN WEST REFINING CO.
HENKEL CORP. PARKER+AMCHEM
HEWLETT PACKARD CO. SAN JOSE SITE
HEXFET AMERICA
INTEGRATED DEVICE TECHNOLOGY
INTEL CORP. FAB 1/D2
INTEL CORP. FAB 3
INTERNATIONAL LIGHT METALS CORP,
INTERNATIONAL MICROELECTRONIC PRODUCTS
INTERNATIONAL RECTIFIER CORP.
J. R. SIMPLOT CO.
LINEAR TECHNOLOGY CORP.
LSI LOGIC CORP.
METAL CONTAINER CORP. OF CA OF CALIFORNIA
MICROSEMI CORP.
MOBIL OIL CORP. TORRANCE REFINERY
NATIONAL SEMICONDUCTOR CORP.
CITY
SUNNYVALE
SAN JOSE
COMPTON
PITTSBURG
SANTA FE SPRINGS
FREMONT
SAN JOSE
TEMECULA
SALINAS
SANTA CLARA
LIVERMORE
LOS ANGELES
SAN JOSE
EL SEGUNDO
LATHROP
MILPITAS
SANTA CLARA
CARSON
SANTA ANA
TORRANCE
SANTA CLARA
STATE
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
100-999
100-999
1,000-9,999
10,000,000-49,999,999
100,000-999,000
10,000-99,999
1,000-9,999
1,000-9,999
1,000-9,999
1,000-9,999
100-999
1,000-9,999
100-999
1,000-9,999
1,000-9,999
100-999
1,000-9,999
1,000-9,999
1,000-9,999
10,000-99,999
10,000-99,999
QUANTITY RELEASED
OR TRANSFERRED
(POUNDS PER YEAR)
86
250
300
1,340
448
10
237
250
511
5,194
7,573
24,067
18,966
Q
1,095
15
14,016
47
0
250
750
SIC
CODE
3674
3679
2842
2819
2911
2899
3674
3674
3674
3674
3674
3354
3674
3674
2873
3674
3674
3411
3674
2911
3679
Page VII-3

-------
-=—"-"•"• —-' - -••- 	 	 	 — 	 •- '•-.•!• 'LJL-UJU, 	 1. 	 	 ._•"!.„. _-.— —
	 	 FACILITY
••^ 	 ^^^ 	 LL'I 	 	 :|f|— —
NEC ELECTRONICS INC.
NORTHERN TELECOM ELECTRONICS INC.
POWERINE OIL CO.
PRECISION METAL PRODUCTS INC.
PRECISION SPECIALTY METALS INC.
REYNOLDS METALS CO.
REYNOLDS METALS CO. TORRANCE CAN PLANT
ROCKWELL INTERNATIONAL
ROHR INDUSTRIES INC.
SEEQ TECHNOLOGY INC.
SEMTECH CORP.
SIGNETICS CORP.
SILICONIX INC.
SONY MFG. CO. OF AMERICA
SPECTROLAB INC.
SURFACE TREATMENT & INSPECTION INC. (ST&I)
TITECH INTERNATIONAL INC.
TRW LSI PRODUCTS INC.
ULTRAMAR INC.
VLSI TECHNOLOGY INC.
XICOR INC.
CITY
ROSEV1LLE
SAN DIEGO
SANTA FE SPRINGS
EL CAJON
LOS ANGELES
HAYWARD
TORRANCE
NEWPORT BEACH
CHULA VISTA
SAN JOSE
NEWBURY PARK
SUNNYVALE
SANTA CLARA
SAN DIEGO
SYLMAR
PARAMOUNT
POMONA
SAN DIEGO
WILMINGTON
SAN JOSE
MILPITAS
STATE
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
,,_J 	 J !. 	 1
10,000-99,999
100-999
10,000-99,999
1,000-9,999
100-999
1,000-9,999
1,000-9,999
1,000-9,999
1,000-9,999
100-999
1,000-9,999
1,000-9,999
1,000-9,999
1,000-9,999
1,000-9,999
1,000-9,999
1,000-9,999
100-999
10,000-99,999
100-999
1,000-9,999
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
	
250
5
5
500
1,700
11
60
255
5
255
14,750
89
750
487
0
0
18
5,330
153
10
17,600
- - —
SIC
CODE
-
3674
3674
2911
3399
3398
3411
3411
3674
3728
3674
3674
3674
3674
3671
3674
3470
3369
3674
2911
3674
3674
Page VII-4

-------
FACILITY
ATMEL CORP.
COORS BREWING CO.
METAL CONTAINER CORP. (WND)
NCR MEPD FORT COLLINS
NCR MICROELECTRONIC PRODUCTS DIV.
ALLEGHENY LUDLUM CORP.
AMERICAN NATIONAL CAN CO.
FELDSPAR CORP.
PRATT & WHITNEY
SIKORSKY AIRCRAFT STRATFORD
UNO NAVAL PRODUCTS
WYMAN GORDON INVESTMENT CASTINGS
GENERAL CHEMICAL CORP. DELAWARE VALLEY
WORKS
AMERICAN NATIONAL CAN CO.
AT&T MICROELECTRONICS
AVESTA SANDVIK TUBE INC.
CONSOLIDATED MINERALS INC.
FLORIDA TILE INDUSTRIES INC. FLORIDA TILE DIV.
HARRIS CORP. GOVERNMENT SYSTEMS SECTOR
HARRIS CORP. SEMICONDUCTOR PALM BAY
CITY
COLORADO
SPRINGS
GOLDEN
WINDSOR
FORT COLLINS
COLORADO
SPRINGS
WALLINGFORD
DANBURY
MIDDLETOWN
EAST HARTFORD
STRATFORD
UNCASVILLE
GROTON
CLAYMONT
JACKSONVILLE
ORLANDO
WILDWOOD
PLANT CITY
LAKELAND
PALM BAY
PALM BAY
STATE
CO
CO
CO
CO
CO
CT
CT
CT
CT
CT
CT
CT
DE
FL
FL
FL
FL
FL
FL
FL
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
1,000-9,999
10,000-99,999
1,000-9,999
1,000-9,999
1,000-9,999
10,000-99,999
10,000-99,999
10,000-99,999
10,000-99,999
10,000-99,999
1.000-9.999
100-999
1,000,000-9,899,999
10,000-99,999
1,000-9,999
10,000-99,999
10,000-99,999
0-99
0-99
10,000-99,999
QUANTITY RELEASED
OR TRANSFERRED
(POUNDS PER YEAR)
505
501
10
1,407
147
4,555
250
0
1,340
10,868
19,500
1,250
2,090
250
76
7,840
11,700
27,560
5
22,635
SIC
CODE
3674
2082
3411
3674
3674
3316
3411
3299
3724
3721
3559
3324
2819
3411
3674
3317
2874
3253
3669
3674
Page VII-5

-------
r ir- TT-I rimmmir -.-,' 	 	 ~-~- 	 __ .„!'". .. 	 	 	 _ . ....mim.^ 	 ,, 	 ..,,„. 	
FACILITY
-"" 	 	 - 	 ~- 	 	 	 - -— -m- - ~. 	 -.. 	 ,!.• 	 _
METAL CONTAINER CORP.
REYNOLDS METALS CO. TAMPA CAN PLANT
ABC COMPOUNDING CO. INC.
AMERICAN NATIONAL CAN CO.
CROWN BEVERAGE PACKAGING
CROWN BEVERAGE PACKAGING
CROWN CORK & SEAL CO. INC.
FARMERS FAVORITE FERTILIZER INC.
HERAEUS AMERSIL INC.
MEARL CORP. SFM DIV.
MILLER BREWING CO. MOULTRIE CONTAINER PLANT
NIAGARA NATIONAL CORP.
NOVAMAX TECHNOLOGIES (U.S.) INC.
OXFORD CHEMICALS INC.
PRATT & WHITNEY
TRANSCHEM INDUSTRIES INC.
TRENT TUBE DIV.
ZEP MFG. CO.
ROCKWELL INTERNATIONAL CORP.
WHINK PRODUCTS CO.
AMERICAN MICROSYSTEMS INC.
======
CITY
=====
JACKSONVILLE
TAMPA
MORROW
FOREST PARK
ATLANTA
PERRY
ATLANTA
MOULTRIE
BUFORD
HARTWELL
MOULTRIE
ATLANTA
ATLANTA
CHAMBLEE
COLUMBUS
EAST POINT
CARROLLTON
ATLANTA
CEDAR RAPIDS
ELDORA
POCATELLO
l-r^.,,,1, =
STATE
FL
FL
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
IA
IA
ID
=====:
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
!=-.!?•:•'" ..1 ' 	 _..L !'...' - JIMI ,_
1,000-9,999
1,000-9,999
1,000-9,999
10,000-99,999
10,000-99,999
10,000-99,999
1,000-9,999
100,000-999,000
1,000-9,999

1,000-9,999
10,000-99,999
10,000-99,999
10,000-99,999
10,000-99,999
1,000-9,999
10,000-99,999
10,000-99,999
1,000-9,999
10,000-99,999
1,000-9,999
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
85
26
260
250
0
0
0
250
170
40
0
500
0
500
250
101
325
0
1,005
15
750
SIC
CODE
—
3411
3411
2842
3411
3411
3411
3411
2874
3295
3295
3411
2841
2899
2842
3724
2842
3317
2841
3669
2842
3674
Page Vll-6

-------
FACILITY
MICRON TECHNOLOGY INC.
MONSANTO CO.
ZILOG INC.
3M
ALLIED-SIGNAL INC.
ALLIED-SIGNAL INC. DANVILLE WORKS
ALUMAX MILL PRODUCTS INC.
AMERICAN NATIONAL CAN CO.
AMOCO PETROLEUM ADDITIVES CO.
BEECO MFG.
C. J. SAPORITO PLATING CO.
CHEMICAL-WAYS CORP.
CHEMTECH INDUSTRIES INC.
CLARK OIL & REFINING CORP. BLUE ISLAND
CLARK OIL & REFINING CORP. WOOD RIVE
CORAL INTERNATIONAL INC.
MARATHON OIL CO.
MOBIL JOLIET REFINING CORP.
NATIONAL INTERCHEM CORP.
PRECOAT METALS
UNO-VEN CO. CHICAGO REFINERY
CITY
BOISE
SODA SPRINGS
NAMPA
CORDOVA
METROPOLIS
DANVILLE
MORRIS
CHICAGO
WOOD RIVER
CHICAGO
CICERO
LAKE BLUFF
EAST SAINT LOUIS
BLUE ISLAND
HARTFORD
WAUKEGAN
ROBINSON
JOLIET
CHICAGO
CHICAGO
LEMONT
STATE
ID
ID
ID
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
10,000-99,999
100-999
1,000-9,999
100,000-999,000
100,000-999,000
100,000-999,000
1,000-9,999
10,000-99,999
1,000-9,999
1,000-9,999
1,000-9,999
1,000-8,989
1,000,000-9,999,999
100,000-999,000
100,000-999,000
10,000-99,999
10,000-99,999
100,000-999,000
10,000-99,999
1,000-9,999
100,000-999,000
QUANTITY RELEASED
OR TRANSFERRED
(POUNDS PER YEAR)
250
86,000
1,006
81,800
4,375
434
0
250
780
1,500
525
5
7,100
5
5
505
1,000
27,394
0
0
7,826
SIC
CODE
3674
2819
3674
2821
2819
2869
3353
3411
2869
3993
3471
2842
2819
2911
2911
2841
2911
2911
2842
3479
2911
Page VII-7

-------
FACILITY
ZENITH ELECTRONICS CORP. RAULAND DIV.
AEROFORGE CORP.
ALCOA WARRICK OPERATIONS
ALLEGHENY LUDLUM CORP.
AVESTA INC.
CIRCLE-PROSCO INC.
DELCO ELECTRONICS CORP. BYPASS
FORD ELECTRONICS & REFRIGERATION CORP.
HAYNES INTERNATIONAL INC.
INDIANA FARM BUREAU CO-OP ASSN. INC. MT.
VERNON REFINERY
MARATHON OIL CO.
SLATER STEELS FORT WAYNE SP ALLOYS DIV.
THOMSON CONSUMER ELECTRONICS
WORLD WIDE CHEMICALS
ANODIZING INC.
ATOCHEM N.A. WICHITA FACILITY
BOEING WICHITA
COASTAL DERBY REFINING CO.
COASTAL DERBY REFINING CO.
FARMLAND INDUSTRIES INC.
NATIONAL COOPERATIVE REFINERY ASSOCIATION
—fJlilJii-L^^JMiji.....^^ 	 	 in .1-j.i .... ,, ......
CITY
MELROSE PARK
MUNCIE
NEWBURGH
NEW CASTLE
NEW CASTLE
BLOOMINGTON
KOKOMO
CONNERSVILLE
KOKOMO
MOUNT VERNON
INDIANAPOLIS
FORT WAYNE
MARION
INDIANAPOLIS
FORT SCOTT
WICHITA
WICHITA
EL DORADO
WICHITA
COFFEYVILLE
MC PHERSON
— —
STATE
IL
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
KS
KS
KS
KS
KS
KS
KS
^i^-iTTi-riSBSSBS 	 !' "'mS™ PI-I TiV-nra-aiBS
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
m^^Hli^s^^^im^^^^Iimu^^^^^HUfmmesm
10,000-99,939
1,000-9,998
1,000-9,999
10,000-89,999
10,000-99,999
1,000-9,999
100,000-999,000
10,000-99,999
10,000-99,999
100,000-999,000
10,000-99,999
10,000-99,999
10,000-99,999
1,000-9,999
0-99
1,000,000-9,999,999
10,000-99,999
10,000-99,999
10,000-99,999
100,000-999,000
100,000-999,000
L.,,,L, , , „,_.,.,,.._.,.„, 	 ""--r— ---._-J«A
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
3,822
5
918,540
17,155
87,000
750
1,000
750
7,010
1,500
2,750
15,450
10
4,155
5
67,650
10
250
250
3,300
48,259
SIC
CODE
=
3671
3463
3334
3312
3312
2899
3469
3714
3356
2911
2911
3312
3671
2842
3471
2813
3728
2911
2911
2911
2911
Page VII-8

-------
FACILITY
PROSOCO INC.
TEXACO REFINING & MARKETING INC. EL DORADO
PLANT
TOTAL PETROLEUM INC.
A. 0. SMITH CORP. PROTECTIVE COATINGS DIV.
ALCAN INGOT DIV. SEBREE ALUMINUM PLANT
ASHLAND PETROLEUM CO. CATLETTSBURG REFINERY
ATOCHEM NORTH AMERICA INC.
COMMONWEALTH ALUMINUM
DU PONT LOUISVILLE PLANT LOUISVILLE WORKS
FLORIDA TILE INDUSTRIES INC. FLORIDA TILE DIV.
GE CO. KENTUCKY GLASS PLANT
MALLINCKRODT SPECIALTY CHEMICALS CO.
NATIONAL-SOUTHWIRE ALUMINUM CO.
PHILIPS LIGHTING CO.
AGRICO CHEMICAL CO. DIV. OF FREEPORT MCMORAN
ALLIED-SIGNAL INC. BATON ROUGE SOUTH
ALLIED-SIGNAL INC. GEISMAR PLANT
BP OIL CO.
HAYNES INTERNATIONL INC.
LAROCHE CHEMICALS INC.
MARATHON OIL CO. LOUISIANA REFINING DIV.
CITY
KANSAS CITY
EL DORADO
ARKANSAS CITY
FLORENCE
HENDERSON
ASHLAND
CALVERT CITY
LEWISPORT
LOUISVILLE
LAWRENCEBURG
LEXINGTON
PARIS
HAWESVILLE
DANVILLE
UNCLE SAM
BATON ROUGE
GEISMAR
BELLE CHASSE
ARCADIA
GRAMERCY
GARYVILLE
STATE
KS
KS
KS
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
LA
LA
LA
LA
LA
LA
LA
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
1,000-9,999
100,000-999,000
10,000-99,999
0-99
0-99
100,000-999,000
1,000,000-9,999,999
100-999
1,000,000-9,999,999
0-99
10,000-99,999
100,000-999,000
0-99
10,000-99,999
10,000-99,999
100,000-999,000
1,000,000-9,999,999
1,000,000-9,999,999
10,000-99,999
100,000-999,000
100,000-999,000
QUANTITY RELEASED
OR TRANSFERRED
(POUNDS PER YEAR)
250
43,950
15,900
575
102,830
500
9,500
76,208
859
123,400
2,000
84,305
473,040
31,968
5
500
14,500
2,340
95
45
61,960
SIC
CODE
2899
2911
2911
2899
3334
2911
2869
3355
2822
3253
3229
2869
3334
3229
2874
2869
2819
2911
3356
2812
2911
Page VII-9

-------
	 FACILITY
••^ •••^^^-•^^•••••^^^-••^^••••M^M^^MaMMaiJ^M.iia^ii^MM^aMi^^^^^SSSE
MOBIL OIL CORP. CHALMETTE REFINERY
MURPHY OIL USA INC. MERAUX REFINERY
PLACID REFINING CO.
VISTA CHEMICAL CO. LAKE CHARLES CHEMICAL
COMPLEX
ALLEGRO MICRO SYSTEMS INC.
ANALOG DEVICES SEMICONDUCTOR
AT&T
CROWN CORK & SEAL CO. INC.
DIGITAL EQUIPMENT CORP.
GTE PRODUCTS CORP. QUARTZ PLANT
NORTON CO.
POLY-METAL FINISHING INC.
TELEDYNE RODNEY METALS
TEXAS INSTRUMENTS INC.
UNITRODE CORP.
WYMAN-GORDON CO.
WYMAN-GORDON CO.
BALTIMORE SPECIALTY STEELS CORP.
BETHLEHEM STEEL CORP. SPARROWS POINT PLANT
CROWN BEVERAGE PACKAGING
EASTALCO ALUMINUM CO.
"~~7 — — — • '— 	 	 . . • . ..ii'iiFiiiiijiiui-^T
CITY
CHALMETTE
MERAUX
PORT ALLEN
WESTLAKE
WORCESTER
WILMINGTON
NORTH ANDOVER
LAWRENCE
HUDSON
IPSWICH
WORCESTER
SPRINGFIELD
NEW BEDFORD
ATTLEBORO
WATERTOWN
NORTH GRAFTON
WORCESTER
BALTIMORE
SPARROWS POINT
BALTIMORE
FREDERICK
=
STATE
— i =
LA
LA
LA
LA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MD
MD
MD
MD
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
^^^S^t^^^^^Hmmoii^^^f^mf^^^^^^ffMm'
1,000,000-9,999,999
100,000-999,000
100,000-999,000
1,000,000-9,999,999
1,000-9,999
1,000-9,999
1,000-9,999
1,000-9,999
1,000-9,999

1,000-9,999
1,000-9,999
1,000-9,999
10,000-99,999
1,000-9,999
10,000-99,999
10,000-99,999
10,000-99,999
100,000-999,000
1,000-9,999
10,000-99,999
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
iaa^__i^__: 	 ^^__^ 	 i 	 jn
1,000
750
343
1,152
1,642
1,250
255
250
40
99,290
13,325
500
49,500
39
250
134,650
35,500
255
4,400
0
64,000
SIC
CODE
— • • —
2911
2911
2911
2869
3674
3674
3661
3411
3674
3229
3291
3471
3316
3341
3674
3462
3462
3312
3312
3411
3334
Page VII-10

-------
FACILITY
EASTERN STAINLESS CORP.
NATIONAL SEMICONDUCTOR CORP.
ACUSTAR INC. MCGRAW GLASS DIV.
ANCOTECH INC.
DOW CORNING CORP.
DU PONT MONTAGUE WORKS
GMC PONTIAC EAST ASSEMBLY
HOWMET CORP. PLANT 5
J & L SPECIALTY PRODUCTS CORP.
JET DIE/BARNES GROUP INC.
NIPPONDENSO MANUFACTURING USA INC.
PARKER & AMCHEM HENKEL CORP.
PARKER+AMCHEM
TOTAL PETROLEUM INC. ALMA REFINERY
UPJOHN CO. PRODUCTION FACILITY
3M CHEMOLITE CENTER
AMERICAN NATIONAL CAN CO.
ASHLAND PETROLEUM CO. ST. PAUL PARK REFINERY
CROWN BEVERAGE PACKAGING
FREMONT INDUSTRIES INC.
AMERICAN NATIONAL CAN CO.
CITY
BALTIMORE
SOUTH PORTLAND
DETROIT
DEARBORN
HIGHTS
MIDLAND
MONTAGUE
PONTIAC
WHITEHALL
DETROIT
LANSING
BATTLE CREEK
WARREN
MORENCI
ALMA
PORTAGE
COTTAGE GROVE
SAINT PAUL
SAINT PAUL PARK
SAINT PAUL
SHAKOPEE
PEVELY
STATE
MD
ME
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
MN
MN
MN
MN
MN
MO
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
100,000-999,000
10,000-99,999
1,000-9,999
1,000-9,999
100,000-999,000
1,000,000-9,999,999
1,000-9,999
10,000-99,999
10,000-99,999
1,000-9,999
1,000-9,999
100,000-999,000
10,000-99,999
100,000-999,000
10,000-99,999
100,000-999,000
10,000-99,999
10,000-99,999
10,000-99,999
1,000-9,999
10,000-99,999
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
23,291
37,120
140
43,260
2,055
1,334
500
59,034
250
2,250
250
3,780
500
5,500
0
5,323
250
250
0
0
250
SIC
CODE
3312
3674
3231
3356
2869
2869
3711
3369
3312
3490
3714
2899
2899
2911
2834
2899
3411
2911
3411
2842
3411
Page VII-11

-------
FACILITY
AT&T MICROELECTRONICS KANSAS CITY WORKS
BRIGGS & STRATTON CORP.
COMBUSTION ENGINEERING NFM HEMATITE
HITCHINER MFG. CO. INC.
KO MFG. INC.
MALLINCKRODT SPECIALTY CHEMICALS CO.
MCDONNELL DOUGLAS CORP.
MEMO ST. PETERS PLANT
METAL CONTAINER CORP.
NORANDA ALUMINUM INC.
REYNOLDS METALS CO. K.C. CAN PLANT
TRADCO INC.
WILLERT HOME PRODUCTS
AMERICAN NATIONAL CAN CO.
CROWN CORK & SEAL CO. INC.
CENEX REFINERY
COLUMBIA FALLS ALUMINUM CO.
CONOCO BILLINGS REFINERY
EXXON BILLINGS REFINERY BILLINGS REFINERY
RHONE-POULENC BASIC CHEMICALS
ALCOA BADIN WORKS
CITY
LEES SUMMIT
POPLAR BLUFF
HEMATITE
O'FALLON
SPRINGFIELD
SAINT LOUIS
SAINT LOUIS
O'FALLON
ARNOLD
NEW MADRID
KANSAS CITY
WASHINGTON
SAINT LOUIS
OLIVE BRANCH
BATESVILLE
LAUREL
COLUMBIA FALLS
BILLINGS
BILLINGS
SILVER BOW
BADIN
STATE
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MS
MS
MT
MT
MT
MT
MT
NC
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
1,000-9,999
1,000-9,999
100-999
1,000-9,999

100,000-999,000
10,000-99,999
10,000-99,999
10,000-99,999
0-99
10,000-99,999
10,000-99,999
100-999
10,000-99,999
10,000-99,999
100,000-999,000
1,000-9,999
100,000-999,000
10,000-99,999
100-999
0-99
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
230
255
7,200
10
10
5,132
21,400
2,000
84
277,044
29
521
0
250
750
19,500
372,100
8,400
40,900
95,300
185,644
SIC
CODE
3678
3519
2819
3369
2841
2869
3721
3674
3411
3334
3411
3356
2879
3411
3411
2911
3334
2911
2911
2819
3334

-------
FACILITY
ARROCHEM INC.
CYPRUS FOOTE MINERAL CO.
GE CO.
MILLER BREWING CO.
MITSUBISHI SEMICONDUCTOR AMERICA INC.
PPG INDUSTRIES INC.
REYNOLDS METALS CO. SALISBURY CAN PLANT
STROH BREWERY CO.
TELEDYNE ALLVAC
TEXASGULF INC. PHOSPHATE OPERATIONS
AMOCO OIL CO.
REiNKE MFG. CO. INC.
UNITRODE INTEGRATED CIRCUITS CORP.
ALLIED-SIGNAL INC. ELIZABETH
AMERICAN NATIONAL CAN CO.
AMERICAN NATIONAL CAN CO.
AUSIMONT USA INC.
CERAGRAPHIC INC.
CP CHEMICALS INC.
CROWN CORK & SEAL CO. INC.
DU PONT CHAMBERS WORKS CHAMBERS WORKS
CITY
MOUNT HOLLY
KINGS MOUNTAIN
WILMINGTON
REIDSVILLE
DURHAM
SHELBY
SALISBURY
WINSTON-SALEM
MONROE
AURORA
MANDAN
DESHLER
MERRIMACK
ELIZABETH
MONMOUTH
JUNCTIO
PISCATAWAY
THOROFARE
HACKENSACK
SEWAREN
NORTH BERGEN
DEEPWATER
STATE
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
ND
NE
NH
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
100,000-999,000
10,000-99,999
10,000-99,999
1,000-9,999
1,000-9,999
1,000-9,999
1,000-9,999
10,000-99,999
1,000-9,999
10,000-99,999
10,000-99,999
Q-99
1,000-9,999
10,000-93,999
10,000-99,999
10,000-99,999
1,000,000-9,999,999
1,000-9,999
10,000-99,999
1,000-9,999
100,000-999,000
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
5
CT
2,100
0
5
255
31
0
0
1,550
3,978
255
255
45
250
250
113
29,011
500
0
3,172
SIC
CODE
2841
2819
2819
3411
3674
3229
3411
3411
3356
2874
2911
3523
3674
2889
3411
3411
2821
3200
2819
3411
2865
Page VII-13

-------
FACILITY
ESSEX CHEMICAL CORP.
ICI AMERICAS INC.
J.T. BAKER INC.
KRAMER CHEMICALS INC. DELKAY DIV.
MOBIL OIL CORP.
PROSOCO INC.
SHIELDALLOY METALLURGICAL CORP.
SWEPCO TUBE CORP.
GIANT REFINING CO. CINIZA
INTEL CORP.
NAVAJO REFINING CO.
SIGNETICS CO.
KENNAMETAL INC. NEVADA REFINERY
AL TECH SPECIALTY STEEL CORP.
AL TECH SPECIALTY STEEL CORP.
ALCOA
ALLIED-SIGNAL INC.
AMPHENOL CORP. BCO
CARBORUNDUM CO.
CORNING INC.
DU PONT NIAGRA FALLS
CITY
PAULSBORO
BAYONNE
PHILLIPSBURG
CAMDEN
PAULSBORO
SOUTH PLAINFIELD
NEWFIELD
CLIFTON
GALLUP
RIO RANCHO
ARTESIA
ALBUQUERQUE
FALLON
DUNKIRK
WATERVLJET
MASSENA
BUFFALO
SIDNEY
NIAGARA FALLS
CORNING
NIAGARA FALLS
STATE
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NM
NM
NM
NM
NV
NY
NY
NY
NY
NY
NY
NY
NY
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
100,000-999,000
100-999
10,000-99,999
10,000-99,999
100,000-999,000
1,000-9,999
0-99
10,000-99,999
10,000-99,999
1,000-9,999
10,000-99,999
1,000-9,999
10,000-99,999
1,000-9,999
10,000-99,999
100,000-999,000
1,000-9,999
1,000-9,999
1,000-9,999
10,000-99,999
1,000-9,999
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
1,510
255
75
500
250
250
250
59,725
4,199
1,005
1,790
4,400
1,250
260
3,800
114,578
140
3,505
505
25,219
368
SIC
CODE
2819
2821
2819
2819
2911
2899
3312
3498
2911
3674
2911
3674

3315
3312
3334
2869
3678
3297
3231
2812
Page Vll-14

-------
FACILITY
GCF INC.
GMC HARRISON RADIATOR DIV.
GRUMMAN AEROSPACE CORP.
IBM EAST FISHKILL FACILITY
IMAGING & SENSING TECHNOLOGY DIV.
METAL CONTAINER CORP. NWB
MILLER BREWING CO. CONTAINER DIV.
OCCIDENTAL CHEMICAL CORP. NIAGARA PLANT
PHILIPS LIGHTING CO.
REYNOLDS METALS CO.
REYNOLDS METALS CO. WALLKILL CAN PLANT
SPECIAL METALS CORP.
TOSHIBA DISPLAY DEVICES INC.
UTICA CORP.
WESTINGHOUSE ELECTRIC CORP. HORSEHEADS
OPERATIONS
AIRFOIL FORGING TEXTRON INC.
ALCAN ROLLED PRODUCTS CO.
ALCOA
AMERICAN MATSUSHITA ELECTRONICS CORP.
AMERICAN NATIONAL CAN CO.
CITY
BUFFALO
LOCKPORT
BETHPAGE
HOPEWELL
JUNCTIO
HORSEHEADS
NEW WINDSOR
FULTON
NIAGARA FALLS
BATH
MASSENA
MIDDLETOWN
NEW HARTFORD
HORSEHEADS
WHITESBORO
HORSEHEADS
EUCLID
WARREN
CLEVELAND
TROY
FREMONT
STATE
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
OH
OH
OH
OH
OH
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)

1,000-9,999
100,000-999,000
10,000-99,999
1,000-9,999
1,000-9,999
10,000-99,999
1,000,000-9,999,999
10,000-99,999
100-999
1,000-9,999
1,000-9,999
1,000-9,999
1,000-9,999
0-99
1,000-9,999
10,000-99,999
10,000-99,999
1,000-9,999
10,000-99,999
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
10
255
42,398
450
1,500
14
0
2,057
1,750
88,000
34
14,050
1,863
56,050
10
81,090
10
40,565
5,955
250
SIC
CODE
3471
3714
3721
3674
3663
3411
3411
2812
3641
3334
3411
3313
3671
3724
3699
3724
3354
3463
3672
3411
Page VII-15

-------
FACILITY
AMERICAN NATIONAL CAN CO.
AMERIMARK BUILDING PRODUCTS INC.
ARMCO ADVANCED MATERIALS CO.
ASHLAND PETROLEUM CO. CANTON REFINERY
ASTRO METALLURGICAL INC.
BETZ LABORATORIES INC. NEW PHILADELPHIA
BRUSH WELLMAN INC.
CINCINNATI SEMICONDUCTOR INC.
COLD METAL PRODUCTS CO. INC.
COSHOCTON STAINLESS DIV.
CROWN BEVERAGE PACKAGING
ENGELHARD CORP.
ENGELHARD CORP.
EPCO EXTRUSION PAINTING CO.
GE CO. LIGHTING WILLOUGHBY QUARTZ PLANT
GE CO. SUPERABRASIVES
GMC DELCO MORAINE NDH DIV. SOUTH
HARRIS SEMICONDUCTOR INTERNATIONAL INC.
J & L SPECIALTY PRODUCTS CORP.
LW STEEL CO. INC. CLEVELAND TUBULAR PLANT
LUCERNE PRODUCTS INC.
CITY
WHITEHOUSE
GNADENHLJTTEN
ZANESVILLE
CANTON
WOOSTER
NEW
PHILADELPHIA
ELMORE
MAINEVILLE
YOUNGSTOWN
COSHOCTON
CINCINNATI
CLEVELAND
ELYRIA
BOARDMAN
WILLOUGHBY
WORTHINGTON
DAYTON
FINDLAY
LOUISVILLE
CLEVELAND
HUDSON
STATE
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
10,000-99,999
1,000-9,999
10,000-99,999
10,000-99,999
1,000-9,999
10,000-99,999
100,000-999,000
1,000-9,999
10,000-99,999
10,000-99,999
10,000-99,999
100,000-999,000
10,000-99,999
1,000-9,999

100,000-999,000
1,000-9,999
10,000-99,999
100,000-999,000
1,000-9,999
100-999
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
250
260
854,435
500
30,900
523
5
2,950
14,335
3,055
0
885
255
0
750
521
255
326
250
8,212
4,305
SIC
CODE
3411
3499
3312
2911
3499
2899
3339
3644
3316
3316
3411
2819
2819
3354
3229
3291
3714
3674
3312
3312
3699
P««« Mil -1£S
ayt? vii- t\j

-------
FACILITY
MAN-GILL CHEMICAL CO.
METAL CONTAINER CORP.
OI-NEG TV PRODUCTS INC.
ORMET CORP.
PHILIPS DISPLAY COMPONENTS CO.
REPUBLIC ENGINEERED STEELS INC. MASSILLON CFB
RMI TITANIUM CO. NILES PLANT
SAWYER RESEARCH PRODUCTS INC.
SUPERIOR TUBE CO.
TUNGSTEN PRODUCTS PLANT
ZIRCOA INC.
AMERICAN NATIONAL CAN CO.
CONOCO PONCA CITY REFINERY
KERR-MCGEE REFINING CORP.
MCDONNELL DOUGLAS TULSA
OZARK MAHONING CO. FORMERLY OZARK-MAHONING
CO.
SEQUOYAH FUELS CORP.
SUN REFINING & MARKETING CO.
TOTAL PETROLEUM INC.
FUJITSU MICROELECTRONICS INC.
HEWLETT PACKARD CO.
CITY
CLEVELAND
COLUMBUS
COLUMBUS
HANNIBAL
OTTAWA.
MASSILLON
NILES
EASTLAKE
WAPAKONETA
EUCLID
SOLON
OKLAHOMA CITY
PONCA CITY
WYNNEWOOD
TULSA
TULSA
GORE
TULSA
ARDMORE
GRESHAM
CORVALLIS
STATE
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OK
OK
OK
OK
OK
OK
OK
OK
OR
OR
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
1,000-9,999
10,000-99,999
10,000-99,999

10,000-99,999
10,000-99,999
10,000-99,999
10,000-99,999
10,000-99,999
1,000-9,999
10,000-99,999
10,000-99,999
100,000-999,000
100,000-999,000
10,000-99,999
10,000-99,999
100,000-999,000
100,000-999,000
1,000,000-9,999,999
1,000-9,999
100-999
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
5
84
86,010
295,159
10,250
18,325
44,036
26,525
3,500
500
255
250
16,000
2,000
45,150
1,450
20,814
4,060
1,364
864
305
SIC
CODE
2899
3411
3229
3334
3672
3316
3356
3679
3317
3399
3297
3411
2911
2911
3721
2819
2819
2911
2911
3670
3674
Page VII-17

-------
FACILITY
INTEL CORP.
NORTHWEST ALUMINUM CO. INC.
REYNOLDS METALS CO.
SILTEC SILICON
TELEDYNE WAH CHANG ALBANY
TITANIUM BUSINESS OPERATIONS
WACKER SILTRONIC CORP.
AIR PRODUCTS & CHEMICALS INC.
ALEX C. FERGUSSON INC.
ALLEGHENY LUDLUM CORP.
ALLEGHENY LUDLUM CORP.
ALLEGHENY LUDLUM CORP. LEECHBURG WORKS
ALLEGRO MICROSYSTEMS INC.
AMERICAN NATIONAL CAN CO.
AMERICAN PLATING INC.
ARMCO ADVANCED MATERIALS CO.
ARMCO ADVANCED MATERIALS CO.
ASHLAND CHEMICAL INC.
AT&T MICROELECTRONICS MICROELECTRONICS
AT&T MICROELECTRONICS MICROELECTRONICS
BP OIL CO. MARCUS HOOK REFINERY
CITY
ALOHA
THE DALLES
TROUTDALE
SALEM
ALBANY
MILWAUKIE
PORTLAND
TAMAQUA
FRAZER
BRACKENRIDGE
VANDERGRIFT
LEECHBURG
WILLOW GROVE
FOGELSVILLE
ZELIENOPLE
BUTLER
BUTLER
EASTON
ALLENTOWN
READING
TRAINER
STATE
OR
OR
OR
OR
OR
OR
OR
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
100-999
100-999
100,000-999,000
1,000-9,999
10,000-99,999
10,000-99,999
1,000-9,999
100,000-999,000
1,000-9,999
10,000-99,999
10,000-99,999
100,000-999,000
1,000-9,999
10,000-99,999
100-999
100,000-999,000
10,000-99,999
100,000-999,000
1,000-9,999
10,000-99,999
1,000,000-9,999,999
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
520
19,000
189,000
605
4,500
1,139
5,093
122
505
378,955
79,555
71,455
6,477
250
2,200
73,580
12,250
500
2,800
64,250
1,200
SIC
CODE
3674
3334
3334
3674
3339
3369
3674
2813
2842
3312
3312
3312
3674
3411
3471
3312
3312
2819
3674
3674
2911
Page VII-18

-------
FACILITY
BULK CHEMICALS INC.
BULLEN COMPANIES
CABOT CORP.
CARPENTER TECHNOLOGY CORP.
CHEVRON USA
CORCO CHEMICAL CORP.
CORNING INC.
CROWN CORK & SEAL CO. INC.
DYNAMET INC.
GECO.
GRINNELL CORP. MFG. DIV.
GTE PRODUCTS CORP.
GTE PRODUCTS CORP. CHEM & MET DIV.
HANDY & HARMAN TUBE CO.
HARRIS SEMICONDUCTOR INTERNATIONAL INC.
HEINTZ CORP.
J & L SPECIALTY PRODUCTS CORP.
JESSOP STEEL CO.
LENOX CRYSTAL
LUKENS STEEL CO.
MOLYCORP INC.
CITY
MOHRSVILLE
FOLCROFT
BOYERTOWN
READING
PHILADELPHIA
FAIRLESS HILLS
CHARLEROI
PHILADELPHIA
WASHINGTON
MALVERN
COLUMBIA
WELLSBORO
TOWANDA
NORRISTOWN
MOUNTAIN TOP
PHILADELPHIA
MIDLAND
WASHINGTON
MOUNT PLEASANT
COATESVILLE
YORK
STATE
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
10,000-99,999
1,000-9,999
1,000,000-9,999,999
100,000-999,000
100,000-999,000
10,000-99,999
0-99
1,000-9,999
1,000-9,999
1,000-9,999
1,000-9,999
10,000-99,999
1,000-9,999
100-999
1,000-9,999
1,000-9,999
100,000-999,000
10,000-99,999
1,000-9,999
10,000-99,999
1,000-9,999
QUANTITY RELEASED
OR TRANSFERRED
(POUNDS PER YEAR)
0
1,000
31,157
27,808
255
500
66,660
0
43,880
255
687
9,450
0
21,259
480
250
150,250
44,500
477
151,466
40
SIC
CODE
2899
2842
3339
3312
2911
2819
3229
3411
3356
3613
3322
3229
3339
3317
3674
3724
3312
3312
3229
3312
2819
Page VII-19

-------
FACILITY
- - -
NF & M INTERNATIONAL INC.
OLJN HUNT SPECIALTY PRODUCTS INC.
PITTSBURGH FLATROLL CO. STEEL DIV.
POWEREX INC.
SANDVIK STEEL CO.
SHARON STEEL CORP. DAMASCUS TUBE DIV. PLANT
#1
SUPERIOR TUBE CO.
TELEDYNE COLUMBIA-SUMMERILL
THOMSON CONSUMER ELECTRONICS INC.
WASHINGTON STEEL CORP.
WESTINGHOUSE ELECTRIC CORP.
AMERICAN NATIONAL CAN CO.
ANAQUEST CARIBE INC.
CROWN CORK DE PUERTO RICO INC.
SCHERING INDUSTRIAL DEVELOPMENT CORP.
CHERRY SEMICONDUCTOR CORP.
GTE PRODUCTS CORP.
ALUMAX OF SOUTH CAROLINA
AMERICAN NATIONAL CAN CO.
CAROLINA METALS INC.
HALOCARBON PRODUCTS CORP.
=====:
CITY
========
MONACA
NAZARETH
PITTSBURGH
YOUNGWOOD
CLARKS SUMMIT
GREENVILLE
COLLEGEVILLE
SCOTTDALE
SCRANTON
WASHINGTON
BLAIRSVILLE
CATANO
GUAYAMA
CAROLINA
MANATI
EAST GREENWICH
CENTRAL FALLS
GOOSE CREEK
BISHOPVILLE
BARNWELL
NORTH AUGUSTA
STATE
=======
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PR
PR
PR
PR
Rl
Rl
SC
sc
SC
sc
	
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
— •^•^^•MMBlSlB^^^M^^^^.^^^,^^
1,000-9,999
100,000-999,000
10,000-99,999
1,000-9,999
10,000-99,999
1,000-9,999
100,000-999,000
1,000-9,999
1,000-9,999
10,000-99,999
1,000-9,999
10,000-99,999
10,000-99,999
100-999
1,000-9,999
1,000-9,999
10,000-99,999
0-99
10,000-99,999
1,000-9,999
10,000-99,999
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
—
255
1,198
500
250
620
41,543
125,907
22,500
255
500
19,100
250
10
250
500
250
16,100
78,917
250
111
817
=
SIC
CODE
==
3399
2819
3312
3674
3317
3317
3841
3317
3672
3316
3356
3411
2834
3411
2833
3674
3229
3334
3411
3399
2869
Page VII-20

-------
FACILITY
MEMO ELECTRONIC MATERIALS INC.
SPARTANBURG STEEL PRODUCTS INC.
WESTINGHOUSE COMMERCIAL NUCLEAR FUEL DIV.
ALCOA
BRISTOL METALS INC.
MAPCO PETROLEUM INC.
MIDLAB INC.
OCCIDENTAL CHEMICAL CORP.
OXFORD TENNELEC/NUCLEUS INC.
STAUFFER CHEMICAL CO. FURNACE PLANT
TIMET INC.
USDOE Y-12 PLANT
ACI CHEMICALS INC.
ADVANCED MICRO DEVICES INC.
ALCOA POINT COMFORT OPERATIONS
ALCOA ROCKDALE WORKS
AMERICAN NATIONAL CAN CO.
AMOCO CHEMICAL CO. TEXAS CITY PLANT B
AMOCO OIL CO. TEXAS CITY REFINERY
ASHLAND CHEMICAL E & LP
CELANESE ENGINEERING RESINS INC.
CITY
MOORE
SPARTANBURG
COLUMBIA
ALCOA
BRISTOL
MEMPHIS
SWEETWATER
COLUMBIA
OAK RIDGE
MOUNT PLEASANT
MORRISTOWN
OAK RIDGE
LANCASTER
AUSTIN
POINT COMFORT
ROCKDALE
HOUSTON
TEXAS CITY
TEXAS CITY
DALLAS
BISHOP
STATE
SC
sc
SC
TN
TN
TN
TN
TN
TN
TN
TN
TN
TX
TX
TX
TX
TX
TX
TX
TX
TX
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
10,000-99,999
10,000-99,999
10,000-99,999
1,000-9,999
10,000-99,999
10,000-99,999
10,000-99,999
100-999
100-999
0-99

1,000-9,999
1,000-9,999
10,000-99,999
1,000-9,999
0-99
10,000-99,999
10,000-99,999
1,000,000-9,999,999
1,000-9,999
10,000-99,999
QUANTITY RELEASED
OR TRANSFERRED
(POUNDS PER YEAR)
2,751
1,005
260
577,848
9,300
41
10
198,423
6,770
16,250
500
3,146
10
140,500
3,550
1,305,634
250
483
2,400
21
8,840
SIC
CODE
3674
3411
3219
3334
3317
2911
2842
2819
3810
2819
3356
3499
2841
3674
2819
3334
3411
2865
2911
2819
2869
Page VI 1-21

-------
I 	 FACILITY
===- -^— — —
CHAMPUN REFINING & CHEMICALS INC.
CHEMICAL DYNAMICS INC.
CHEVRON USA INC. PORT ARTHUR REFINERY
I COASTAL REFINING & MARKETING INC.
I COOPER INDUSTRIES CAMERON FORGED PRODUCTS
1 DIV.
CROWN BEVERAGE PACKAGING
CROWN BEVERAGE PACKAGING
CROWN CENTRAL PETROLEUM HOUSTON REFINERY
CROWN CORK & SEAL CO. INC.
CROWN CORK & SEAL CO. INC.
CYPRESS SEMICONDUCTOR TEXAS INC.
DALLAS SEMICONDUCTOR CORP.
DIAMOND SHAMROCK REFINING & MARKETING CO
THREE RIVERS
DU PONT CORPUS CHRISTI PLANT CORPUS CHRISTI
PLANT
DU PONT LA PORTE PLANT LA PORTE PLANT
FINA OIL & CHEMICAL CO.
GENERAL DYNAMICS FT. WORTH DIV. USAF PLANT 4
GOODYEAR TIRE & RUBBER CO. BEAUMONT
CHEMICAL PLANT
HITACHI SEMICONDUCTOR AMERICA INC.


CITY
CORPUS CHRISTI
WEATHERFORD
PORT ARTHUR
CORPUS CHRISTI
HOUSTON
LONGVIEW
SUGAR LAND
PASADENA
ABILENE
CONROE
ROUND ROCK
DALLAS
THREE RIVERS
INGLESIDE
LA PORTE
BIG SPRING
FORT WORTH
BEAUMONT
IRVING

»«^«^«Bmi^^«^«
STATE
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX

^•^^•^^i^
— —
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
=^^^=^--
100,000-999,000
10,000-99,999
100,000-999,000
10,000-99,999
10,000-99,999
10,000-99,999
10,000-99,999
100,000-999,000
1,000-9,999
100-999
1,000-9,999
100-999
100,000-999,000
1,000,000-9,999,999
10,000,000-49,999,999
1,000-9,999
10,000-99,999
10,000-99,999
1,000-9,999

•— — — — -^«
i ' •• 	 • " ' :=
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
•SSS^Sfi^BHSi^^^tl^^^^^jj^^^^^B^fr^S
0
2,000
11,250
0
7,500
0
0
1,486
0
0
10,184
54
4,700
845
2,931
4,840
1,400
140
2,823


SIC
CODE I
' 	 "—» -
2911
3471
2911
2911
3462
3411
3411
2911
3411
3411
3674
3674
2911
2869
2819
2911
3721
2822
3674
II

Page VII-22

-------
FACILITY
KOCH REFINING CO.
MANNINGTON CERAMIC CO.
MARATHON PETROLEUM CO.
MILLER BREWING CO. FORT WORTH CONTAINER DIV.
MONSANTO CO.
MOTOROLA INC.
NATIONAL SEMICONDUCTOR CORP.
PEARL CONTAINER CO.
PHIBRO REFINING INC.
PHILLIPS 66 CO.
PHILLIPS 66 CO.
PRECOAT METALS
SGS-THOMSON MICROELECTRONICS INC.
SHELL OIL CO. ODESSA REFINERY
SONY MICROELECTRONICS CORP.
SOUTHWESTERN REFINING CO. INC. INC.
STROH CONTAINER CO.
TEXAS INSTRUMENTS INC.
TEXAS INSTRUMENTS INC.
TEXAS INSTRUMENTS INC.
TEXAS INSTRUMENTS INC. SHERMAN
CITY
CORPUS CHRISTI
MOUNT VERNON
TEXAS CITY
FORT WORTH
ALVIN
AUSTIN
ARLINGTON
SAN ANTONIO
TEXAS CITY
BORGER
SWEENY
HOUSTON
CARROLLTON
ODESSA
SAN ANTONIO
CORPUS CHRISTI
LONGVIEW
DALLAS
LUBBOCK
STAFFORD
SHERMAN
STATE
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
100,000-999,000
0-99
100,000-999,000
1,000-9,999
100,000-999,000
10,000-99,999
10,000-99,999
1,000-9,999
100,000-999,000
1,000,000-9,999,999
100,000-999,000
1,000-9,999
1,000-9,999
100,000-999,000
100-999
100,000-999,000
10,000-99,999
100,000-999,000
10,000-99,999
10,000-99,999
1,000-9,999
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
59
27,440
19,250
5
6,670
76,170
590
0
0
4,327
3,300
0
260
870
505
500
0
3,300
1,000
35,500
3,375
SIC
CODE
2911
3253
2911
3411
2869
3674
3674
3411
2911
2911
2911
3479
3674
2911
3674
2911
3411
3674
3674
3674
3674
Page VII-23

-------
FACILITY
="— " 	 "- '- 	 " 	 !J1 wTnr""" 	 _---'- 	 ^- -
VALERO REFINING CO.
VIRGINIA KMP CORP.
VLSI TECHNOLOGY INC.
BIG WEST OIL CO.
CHEVRON USA INC. SALT LAKE REFINERY
FLAMECO (BLDG.#2)
INTERSTATE BRICK CO.
NATIONAL SEMICONDUCTOR
PHILLIPS 66 CO. WOODS CROSS REFINERY
SIGNETICS CO.
WESTERN ZIRCONIUM
BABCOCK & WILCOX CO. MT. ATHOS FACILITY
BALL PACKAGING PRODUCTS GROUP MCD
HOWMET CORP. HAMPTON CASTING DIV.
TEXASGULF INC. SALTVILLE OPERATIONS
GECO.
IBM CORP.
ALCOA WENATCHEE WORKS
AMERICAN NATIONAL CAN CO.
BOEING COMMERCIAL AIRPLANES FABRICATION
BOEING DEFENSE & SPACE GROUP PLANT II
CITY
CORPUS CHRIST!
DALLAS
SAN ANTONIO
NORTH SALT LAKE
SALT LAKE CITY
OGDEN
WEST JORDAN
WEST JORDAN
WOODS CROSS
OREM
OGDEN
LYNCHBURG
WILLIAMSBURG
HAMPTON
SALTVILLE
NORTH
CLARENDON
ESSEX JUNCTION
WENATCHEE
KENT
AUBURN
SEATTLE
STATE
=
TX
TX
TX
LO-
UT
UT
UT
UT
UT
UT
UT
VA
VA
VA
VA
VT
VT
WA
WA
WA
WA
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
100,000-999,000
10,000-99,999
100-999
10,000-99,999
100,000-999,000
1,000-9,999
0-99
1,000-9,999
100,000-999,000
10,000-99,999
10,000-99,999
10,000-99,999
1,000-9,999
10,000-99,999
0-99
10,000-99,999
100,000-999,000
0-99
10,000-99,999
1,000-9,999
1,000-9,999
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
515
500
255
2,636
9,450
22,192
11,580
1,831
4,700
1,805
500
6
10
4,815
1,209
8,112
2,750
223,300
250
22,700
7,430
SIC
CODE
- - • —
2911
3585
3674
2911
2911
3728
3251
3674
2911
3674
3356
3443
3411
3324
2874
3724
3674
3334
3411
3728
3728
Page VII-24

-------
FACILITY
BP OIL CO. FERNDALE REFINERY
COLUMBIA ALUMINUM CORP.
CROWN BEVERAGE PACKAGING
INTALCO ALUMINUM CORP.
KAISER ALUMINUM & CHEMICAL CORP.
KAISER ALUMINUM & CHEMICAL MEAD WORKS
NATIONAL SEMICONDUCTOR CORP.
REYNOLDS METALS CO. REDUCTION PLANT
REYNOLDS METALS CO. SEATTLE CAN PLANT
SANDVIK SPECIAL METALS CORP.
SEH AMERICA INC.
VANALCO INC.
WESTERN PNEUMATIC TUBE CO.
BRIGGS & STRATTON CORP.
CHEMICAL PACKAGING CORP.
CROWN BEVERAGE PACKAGING
CROWN CORK & SEAL CO. INC.
KOHLER CO.
LADISH CO. INC.
METAL CONTAINER CORP. FTA
MILLER BREWING CO. MILWAUKEE CONTAINER PLANT
CITY
FERNDALE
GOLDENDALE
OLYMPIA
FERNDALE
TACOMA
MEAD
PUYALLUP
LONGVIEW
KENT
FINLEY
VANCOUVER
VANCOUVER
KIRKLAND
WAUWATOSA
MILWAUKEE
LA CROSSE
GLENDALE
KOHLER
CUDAHY
FORT ATKINSON
MILWAUKEE
STATE
WA
WA
WA
WA
WA
WA
WA
WA
WA
WA
WA
WA
WA
Wl
Wl
Wl
Wl
Wl
Wl
Wl
Wl
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
100,000-999,000
0-99
10,000-99,999
0-99
0-99
0-99
1,000-9,999
0-99
10,000-99,999
10,000-99,999
10,000-99,999
0-S9
1,000-9,999
1,000-9,999
1,000-9,999
10,000-99,999
10,000-99,999
1,000-9,999
100,000-999,000
1,000-9,999
1,000-9,999
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
8,000
89,000
0
62,345
28,200
372,000
1,480
260,255
40
500
21,179
469,403
14,800
141
515
0
0
18,030
500
10
5
SIC
CODE
2911
3334
3411
3334
3334
3334
3674
3334
3411
3356
3674
3334
3498
3519
2842
3411
3411
3261
3462
3411
3411
Page VII-25

-------
FACILITY
MURPHY OIL USA INC. SUPERIOR REFINERY
NORTHERN ENGRAVING CORP.
TRENT TUBE
U.S. CHROME CORP. OF WISCONSIN
3M
DU PONT WASHINGTON WORKS
EAGLE CONVEX GLASS SPECIALTY CO.
INCO ALLOYS INTERNATIONAL INC.
RAVENSWOOD ALUMINUM CORP.
CHEVRON CHEMICAL CO.
CROWN CORK & SEAL CO. INC.
FRONTIER REFINING INC.
CITY
SUPERIOR
SPARTA
EAST TROY
FOND DU LAC
CHARLES TOWN
WASHINGTON
CLARKSBURG
HUNTINGTON
RAVENSWOOD
ROCK SPRINGS
WORLAND
CHEYENNE
STATE
Wl
Wl
Wl
Wl
wv
wv
wv
wv
wv
WY
WY
WY
MAXIMUM
QUANTITY ON SITE
(RANGE, POUNDS)
10,000-99,999
1,000-9,999
10,000-99,999
1,000-9,999
1,000-9,999
0-99
10,000-99,999
10,000-99,999
0-99
0-99
1,000-9,999
1,000-9,999
Total Quantity Released:
QUANTITY RELEASED
OR TRANSFERRED*
(POUNDS PER YEAR)
0
340
5,263
10
5
43,021
10
2,494
81,126
7,700
0
0
12,658,031
SIC
CODE
2911
3479
3317
3471
3555
2821

3356
3334
2874
3411
2911

  Total of annual quantities reported as fugitive or non-point air emissions, stack or point air emissions, discharges to receiving streams or
water bodies, underground injection on-site, releases to land on-site, discharges to POTW, and other transfers in waste to off-site locations.


Source:  Toxic Release Inventory, 1990

Note:  TRI reports exclude all non-manufacturing facilities and those manufacturers with fewer than 10 employees. The threshold for
       reporting for manufacturing or processing a TRI listed chemical was 25,000 pounds and the threshold for reporting for otherwise
       using the chemical is 10,000 pounds.
Page VII-26

-------
                               APPENDIX VIII
        U.S. PRODUCERS OF FLUOROCARBONS AND OF OTHER
     CHEMICALS MANUFACTURED WITH HYDROGEN FLUORIDE OR
                        CHLOROFLUOROCARBONS
      Exhibit VIII-1 of this appendix presents U.S. producers of fluorocarbons, with their location and
production capacity. Exhibit VIII-2 lists some other chemicals made directly or indirectly from HF,
manufacturers of these chemicals, and the source of the fluorine (i.e., HF or an HF product). The
chemicals in Exhibit VIII-2 are examples of chemicals produced from HF; there are many other
chemicals, including nearly all fluorine containing chemicals that are made from HF or HF products.
                                                                     Page VIII-1

-------

-------
                                                         EXHIBIT VIII-1
                                                U.S. Producers of Fluorocarbons
            Company

Allied-Signal Inc.
  Engineered Materials Sector
Atochem North America, Inc.
  Fluorine Chemicals Division
 Ausimont USA, Inc.


 Du Pont Chemicals
   Fluorochemicals
           Plant Site
Baton Rouge, LA
Danville, IL
El Segundo, CA
Calvert City, KY
Wichita, KS
Thorofare, NJ


Antioch, CA
Corpus Christi, TX
Louisville, KY
Montague, Ml
Deepwater, NJ
 Halocarbon Products Corporation     North Augusta, SC
 La Roche Chemicals, Inc.
Gramercy, LA
                                   Total
 Annual Capacity
(Millions of Pounds)

       315
       240
                                        1,360
        Type of Fluorocarbons

Trichlorofluoromethane (CFC-11)
Dichlorodifluoromethane (CFC-12)
Chlorodifluoromethane (CFC-22)
Trichlorotrifluoroethane (CFC-113)
Dichlorotetrafluoroethane (CFC-114)

Trichorofluoromethane (CFC-11)
Dichlorodifluoromethane (CFC-12)
Chlorodifluoromethane (CFC-22)
Dichlorofluoroethane (HCFC-1416)
Dichlorod'rfluoroethane (HCFC-1426)

Dichlorofluoroethane (HCFC-1416)
Dichlorodifluoroethane (HCFC-1426)

Trichlorofluoromethane (CFC-11)
Dichlorodifluoromethane (CFC-12)
Chlorodifluoromethane (HCFC-22)
Chlorotrifluoromethane (CFC-23)
Trichlorotrifluoroethane (CFC-113)
Dichlorotetrafluoroethane (CFC-114)
Chloropentafluoroethane (CFC-115)
Hexafluoroethane (HFC-116)
Tetrafluoroethane (HFC-134a)
Difluoroethane (HFC-152a)

Dichlorotrifluoroethane (HCFC-123)
Tetraf luoroethane (HFC-134a)

Trichlorofluoromethane (CFC-11)
Dichlorodifluoromethane (CFC-12)
Chlorodifluoromethane (CFC-22)
Note:  Capacities are SRI estimates.  The types of fluorocarbons listed exclude halons, azeotropes, and any products produced in pilot plants
only.

Source:        SRI International, 1991 Directory of Chemical Producers, United States.
              Seringer, Carolyn S., Du Pont Chemicals, comments from technical review of Hydrogen Fluoride Study Report to Congress, Draft
              May 8, 1992, June 5, 1992. (436.4)

-------
                                                            EXHIBIT Vlll-2
                             Other Chemicals Manufactured Using HP or Based on Another HF Product
                                                (Examples Onty - Not a Complete List)
  Chemical

  Bromochlorodifluoromethane (Halon 1211)
  (Produced by bromlnatlon of fluorocarbon)

  Bromotrifluoromethane (Halon 1301)
  (Produced by bromlnatlon of fluorocarbon)

  Fluoboric acid
  (Produced using 70% HF solution)
 Fluosulfonic acid
 Vinyl Fluoride
Use

Fire extinguishers


Fire extinguishers


Manufacture of fluoborate salts used
in metal processing and as catalysts
Preparation of boron trifluoride,
catalyst
Monomer for plastics
Company                      Plant Site

Great Lakes Chemical Corp.      El Dorado, AK


Great Lakes Chemical Corp.      El Dorado, AK


Atochem North America          Tulsa, OK
 Chemical Specialties Div.

Chemtech Industries, Inc.         St. Louis, MO
 Fluoride Manufacturing Div.
  Harstan Div.

Englehard Corp.                 Cleveland, OH
 Catalysts and Chemicals Div.

Fidelity Chemical Products Corp.   Newark, NJ

Johnson Matthey, Inc.            Danvers, MA
 Alfa Products

New Hampshire Oak             Claymont, DE
 General Chemical Corp.

Philipp Brothers Chemicals        Sewaren, NJ
 C.P. Chemicals, Inc., sub.

DuPont Co.                     La Porte, TX
 DuPont Chemicals

New Hampshire Oak             Claymont, DE
 General Chemical Corp.

Sigma-Aldrich Corp.              Milwaukee, Wl
 Aldrich Chemical Co., sub.

DuPont Co.                     Louisville, KY
 Fluorochemicals
Sources:    SRI International, 1991 Directory of Chemical Producers, United States of America, 1991.
           Kirk-Othmer Encyclopedia of Chemical Technology, Volumes 10 and 11, 1980.
           Seringer, Carolyn S., Du Pont Chemicals, comments from technical review of Hydrogen Fluoride Study Report to Congress Draft May 8 1992
           Junes, 1992(436.4).

-------
                                APPENDIX IX
     U.S. AND CANADIAN PETROLEUM REFINERIES WITH HYDROGEN
                       FLUORIDE ALKYLATEON UNITS
      Exhibit IX-1 of this appendix presents U.S. petroleum refineries that use HF as an alkylation
catalyst. The exhibit shows the name and location of the resfinery and also indicates whether the
Phillips or UOP process is used. Canadian refineries with HF alkylation units are also listed.
                                                                        Page IX-1

-------
                                        EXHIBIT IX-1
                         Petroleum Refineries with HF Alkylation Units
    Petroleum Refiners
    Amoco Oil Co.
    Amoco Oil Co.
    Ashland Petroleum
    Ashland Petroleum
    Ashland Petroleum
    BP Oil, Inc.
    BP Oil, Inc.
    BP Oil, Inc.
    Cenex Refinery
    Champlin Refining
    Chevron, USA
    Chevron, USA
    Chevron, USA
    Clark Oil & Refinery Corp.
    Clark Oil & Refinery Corp.
    Coastal Refinery & Marketing
    Coastal Refinery & Marketing
    Coastal Refining & Marketing
    Conoco
    Conoco
    Crown Central Petroleum
    Diamond Shamrock
    Exxon Company USA
    Farmland Industries
    Fina Oil and Chemical
    Flying J Petroleum Co.
    Frontier Oil and Refining
    Giant Refining Company
    Golden West Refining
    Hill Petroleum
    Indiana Farm Bureau Coop.
    Kerr-McGee Refining Corp.
    Koch Refining
    MAPCO Petroleum
    Marathon Petroleum
    Marathon Petroleum
    Marathon Petroleum
U.S. Refineries

  Plant Site                 Process

  Texas City, TX            Phillips
  Mandan, ND              Phillips
  Catlettsburg, KY           UOP
  Canton, OH               UOP
  St. Paul Park,  MN          UOP
  Marcus Hook, PA          UOP
  Belle Chase, LA           Phillips
  Ferndale, WA              Phillips
  Laurel, MT                UOP
  Corpus Christi, TX         UOP
  Salt Lake City, UT          UOP
  Philadelphia, PA           Phillips
  Port Arthur, TX            Phillips
  Blue Island, IL            UOP
  Hartford, IL               UOP
  El Dorado, KS            Phillips
  Wichita, KS               UOP
  Corpus Christi, TX         UOP
  Ponca City, OK            Phillips
  Billings, MT               Phillips
  Houston, TX               Phillips
  Three Rivers, TX           UOP
  Billings, MT               UOP
  Coffeyville, KS            UOP
  Big Spring, TX            UOP
  North Salt Lake City, UT    Phillips
  Cheyenne, WY            UOP
  Gallup, NM               UOP
  Santa Fe Springs, CA       UOP
  Texas City, TX            Phillips
  Mt. Vernon, IN            UOP
  Wynnewood, OK           Phillips
  Corpus Christi, TX          UOP
  Memphis, TN              UOP
  Robinson, IL               UOP
  Indianapolis, IN            UOP
  Texas City, TX            Phillips
Page IX-2

-------
                              EXHIBIT IX-1 (Continued)
                     Petroleum Refineries with HF Alkylatlon Units

                                  I. U.S. Refineries
Petroleum Refiners
Marathon Petroleum
Mobil Oil
Mobil Oil
Mobil Oil
Mobil Oil
Murphy Oil U.S.A.
Murphy Oil U.S.A.
National Cooperative Refiners Assoc.
Navajo Refining
Phillips 66 Company
Phillips 66 Company
Phillips 66 Company
Placid Refining
Powerine Oil
Shell Oil
Southwestern Refining Corp.
Sun Refining Co.
Texaco Marketing & Refining
Total Petroleum
Total Petroleum
Total Petroleum
Ultramar, USA Inc.
Uno-Ven Corp.
Valero Refining Co
Wyoming Refining Co.
       Plant Site
       Garyville, LA
       Chalmette, LA
       Joliet, IL
       Paulsboro, NJ
       Torrance, CA
       Superior, Wl
       Meraux, LA
       McPherson, KS
       Artesia, NM
       Borger, TX
       Sweeny, TX
       Woods Cross, UT
       Port Allen, LA
       Santa Fe Springs, CA
       Odessa, TX
       Corpus Christi, TX
       Tulsa, OK
       El Dorado, KS
       Ardmore, OK
       Arkansas City, KS
       Alma, Ml
       Wilmington, CA
       Lemont, IL
       Corpus Christi, TX
       New Castle, WY
 Petroleum Refiners
 Esso Petroleum Canada
 Esso Petroleum Canada
 Petro-Canada Products
 Petro-Canada Products
 Shell of Canada
 Sunoco Inc.
II.  Canadian Refineries

       Plant Site
       Sarnia, Ontario
       Edmonton, Alberta
       Edmonton, Alberta
       Taylor, British Columbia
       Montreal, Quebec
       Sarnia, Ontario
Process

Phillips
Phillips
Phillips
Phillips
UOP
UOP
UOP
UOP
UOP
Phillips
Phillips
Phillips
UOP
UOP
Phillips
UOP
UOP
Phillips
UOP
UOP
UOP
UOP
Phillips
Phillips
Phillips
 Process
 UOP
 UOP
 UOP
 Phillips
 UOP
 Phillips
Source: American Petroleum Institute, Submission to EPA, April 22, 1991.
                                                                                Page IX-3

-------
                                  This page intentionally left blank.
Page IX-4

-------
                                    APPENDIX X
                    RULE 1410: SCAQMD REGULATION ON
                  HYDROGEN FLUORIDE STORAGE AND USE
Introduction

       The South Coast Air Quality Management District (SCAQMD) adopted Rule 1410 on April 5,
1991.  It represents the only direct regulation of HF in the United States. The three major sections in
Rule 1410:  Phase Out; Interim Control Measures; and Reporting and Storage/Usage Inventory
Requirements are discussed below.


Phase Out

       In accordance with the HF phase out schedule, refineries must cease use of HF on or before
January 1, 1998 and fluorocarbon production facilities on or before January 1, 1999 unless the HF is
contained in a mixture which in a serious, near worst-case accidental release, will not result in
atmospheric concentrations equal to or greater than 20 parts per million (ppm) for five minutes and
120 ppm for one minute at or beyond the facility boundary. These concentrations were based on the
SCAQMD analysis and study of irritation thresholds for HF and based on extrapolation from existing
toxicrty standards (i.e., IDLH, OSHA).

       The SCAQMD conducted studies before determining that anhydrous HF should be phased
out. These studies are described in the March 19,1991 "Supporting Document for the Proposed Rule
1410." Based on a computer model and scenarios chosen by the SCAQMD, a release of anhydrous
HF was deemed to pose unacceptable risk to the  public. However, for facilities using  HF contained in
a mixture, SCAQMD allows the facility to conduct a computer modeling run to determine if a near
worst-case accidental release will result in atmospheric concentrations below the exemption cutoffs
mentioned above. Facilities have to take a certain approach approved by the SCAQMD for their
modeling.  Facilities must use the Dense Gas Dispersion (DEGADIS) computer model  used by
SCAQMD or another approved model for calculating concentration and exposure, with SCAQMD-
determined input parameters for surface roughness, worst-case meteorological conditions, reference
height for wind speed, relative humidity, and ambient temperature.  In addition, the SCAQMD defines
the near worst-case scenario as a release from a two inch diameter pipe failure which was based on
actual failure rate data, and SCAQMD's judgment  of the worst-case release in view of the Marathon
event. The facility itself is responsible for running the models and for choosing model source terms
such  as release rate and rate duration. These source terms must be approved by the SCAQMD.


interim Controls

        In order for a facility that is subject to the  phase out to continue to use HF until the 1998
deadline takes effect, the facility must install safety equipment and implement procedures required by
the SCAQMD to reduce the risk of HF. On or after January 1, 1992, facilities must:

        >       Maintain a facility-specific minimum HF inventory (the maximum amount allowed on
               site);

        >       Maintain HF-sensitive paint for leak detection on all valves and flanges for pipes and
               vessels handling HF;
                                                                                  Page X-1

-------
         >•      Maintain emergency isolation valves operated by remote switches in the control room
                or in appropriately safe locations which are accessible during HF releases;

         >•      Operate and maintain automatic HF detection and alarm systems in all HF areas and
                use an approved on-site remote terminal unit (RTU) that is linked to the SCAQMD and
                capable of linkage with the local city and county fire departments;

         >•      Maintain safety devices and procedures to neutralize accidental releases;

         >      Maintain control room detectable audible and visual alarm systems to eyewashes and
                safety showers in all areas where HF is present;

         *•      Maintain direct supervision of all maintenance and technical support personnel and
                laborers when they work within the HF unit boundaries or on equipment directly
                related to the operation of the HF unit and at such times maintain individuals trained in
                HF safety, that have authority to act, at or about the HF unit;

         >      Administer job-specific safety training for all maintenance and technical support
                personnel and require that contractors do the same and provide written and walk
                through performance examinations and maintain records of such training; and,

         >•      Ensure that all HF loading and unloading operations are administered within the
                presence of a facility-trained operator.

 After January 1,1993, the owner or operator of each facility must:

        >       Maintain containment systems;

        >•       Maintain facility-specific automated evacuation systems (i.e., rapid vessel de-inventory);

        *       Maintain facility-specific automated water spray systems,  or an SCAQMD-approved
               alternative, designed to achieve an HF removal efficiency of 90 percent or equivalent
               in the HF areas; and

        >•      Ensure facility-specific seismic upgrade of support structures for all HF-related process
               equipment as specified in the 1988 Uniform Building Code Section 2312.

        Facilities must also develop a plan each  year describing the specific steps it will take to
 comply with specific risk reduction measures and submit the plan to the SCAQMD.  In addition, by
 January 1, 1995, the owner or operator of any facility must submit a compliance plan to the SCAQMD.


 Reporting and Storage/Usage Inventory Requirements

        After July 1, 1991, an owner or operator must report to the SCAQMD any HF release that
 results in exposed persons requiring medical treatment at an off-site facility, evacuation of any portion
 of the facility premises, or aerosol HF transport beyond the facility property boundaries. Such  a report
 must be made within one hour of the time the release is known, or reasonably should have been
 known to any employer, officer, or agent of the owner or operator.  The report must include:

        »•      The name and specific location of the facility;

        >      Identification and title of the notifier;
Page X-2

-------
      >       Cause and extent of the release, including approximate amount, concentration, and
              the area affected;

      >.       Specific location of the release and equipment involved in the release;

      *       Any and all measures taken to mitigate or stop the release, including repairs;

      *       A complete description, to the extent known, and number of any injuries or fatalities;
              and

       *       Names of other agencies notified of the release, the time of notification, and the name
              of the person notified.

       In addition, the owner or operator must submit a follow-up written  report with all the
information presented at the time  of the initial notification and any other related inf°rm2°"n ^r
January 1 1992, the owner or operator of an HF facility must upon the alarm of a HF senso at the
facility/notify the SCAQMD within  fifteen minutes of the alarm and  provide the same information
outlined above.

       Bv July 1  1992  and July  1 of each subsequent year, all facilities must also submit HF storage
and usage report's describing the quantities stored and used during the previous calendar year. Such
an inventory shall include:

       *      The name of the company, telephone and address and company identification and
               applicable equipment permit numbers, as administered by the SCAQMD;

       ».      The name and title of the person conducting the inventory;

       *      The name and address of the manufacturer or distributor of the HF;

       ».      A brief description of the process and/or equipment using HF;

       »      The total annual quantity of HF received in gallons, the size and frequency of
               deliveries, and the mode of transport;

        *      The concentration of HF for each piece of process equipment and as received and in
               storage;

        ^      The total quantity of HF used annually per specified process in gallons; and

        ^      A description of the type of storage and the maximum and average quantities at any
               one time, of HF in possession or control of the owner or  operator of the facility in fixed
               or mobile storage containers both on-site and at other locations within the SCAQMD.

        Rule 1410 also provides for three exemptions from this reporting requirement. It exempts
 facilities that do not store, transport or use anhydrous HF  and facilities that si«re, transport o,-use
 aqueous HF exclusively in solutions in concentrations less than or equal to 50 percent by weigh tit
 also exempts any facility that stores, transports, or uses less than or equal to one gallon of anhydrous
 HF at any one time.1
                                                                                      Page X-3

-------
                                         ENDNOTES
  1.     Hydrogen Fluoride Storage and Use, Rule 1410, South Coast Management District, CA, April 5,
             I • ^&Ouy
PageX-4

-------
                                     APPENDIX XI
     CONTAINERS FOR TRANSPORTATION OF HYDROGEN FLUORIDE
   The transportation of HF is regulated by the DOT under the Hazardous Materials Transportation
Act (HMTA) and the associated Hazardous Materials Regulations (HMR).1'2  The containers and
specifications for transport of HF are determined for anhydrous HF, aqueous HF with concentration
greater than 60%, and aqueous HF with concentration less than 60%. The specifications for these
strengths of HF are outlined in Exhibit XI-1.

Rail Tank Cars

   According to DOT regulations, anhydrous HF must be shipped in DOT specified rail tanks
(specification numbers 105, 112, or 114) with test pressures of 300 psi or greater.  The commonly
used 105A300W model rail tanks have a capacity range of approximately 4,000 to  16,000 gallons.
Exhibit XI-2 shows a typical class 105A300W car. The other classes used for anhydrous hydrogen
fluoride are similar in appearance and design.  Specifications associated with three tank cars of the
class 105A300W are given in Exhibit XI-3.3

   For anhydrous HF, the only opening permitted in the railway tanks is a single manway located in
the center at the top.  Five valves are mounted inside the dome cover, four are angle valves used for
the connections to pump HF and the fifth is the safety relief valve used to release HF gas in the event
of tank overpressurization.4

   Aqueous HF may be shipped in DOT rail tank models 103, 104, 105, 109, 111, 112, 114, or 115.
Tank cars of class 112S400W, as an example, generally have capacities of 4,500 gallons to 8,000
gallons, although larger capacities are possible. Several railway tank cars that may be used in the
transportation of hydrofluoric acid are described in Exhibit XI-4.

   As a standard for industry, the safety relief valve is of the spring-loaded type and is usually
combined with a frangible (or bursting) disc.  Frangible disks are used because the continuous
corrosive action of HF on the relief valve may cause the relief valve to fail. Railway tankers may be
rubber-lined for solutions of up to 40 percent aqueous hydrofluoric acid.6

Motor Vehicle Tank Cars            <

   According to DOT regulations, both anhydrous and aqueous HF transported via motor vehicle
must be transported in tank trucks  of specification MC304, MC307, MC331, MC310, MC311, MC312,
DOT 407, and DOT 412. (See Exhibit XI-5.)

   Aqueous hydrofluoric acid is shipped by tank motor vehicles with steel tanks with capacities up to
4200 gal.  Similar to  railway cars, these highway tankers are unloaded from the top.7

Cylinders

    Cylinders holding about 400 pounds may be used to transport small volumes  of anhydrous
hydrofluoric acid.  DOT specification cylinders 3, 3A, 3AAA, 3B, 3C, 3E, 4, 4A, 25,  or 38 may be used.
Specifications 4B, 4BA, 4BW, or 4C cylinders may be used if they  are not brazed.  Exhibit XI-6
describes the 4B and 4BA cylinders. Laboratories are the primary users or these types of small
cylinders. Cylinders may also be used to transport aqueous hydrofluoric acid.
                                                                                  Page XI-1

-------
                                                                                                EXHIBIT XH
                                                                        D«pirtm»nt of Trtntportitlon Regulitlont for the TomporUtlon of HF



Anhydrous
HF

















Greater
»tr*ngth
HF)












Cargo Unlu


Specification MC 304, MC307,
MC331 cargo tank motor
vehicles; and

MC3IO, MC311,MC312,
DOT 407, and DOT 41 2 cargo
tank motor vehicles with tank
design pressure of at least
17Z4 kPa (25 psi) are
authorized.









Specification M306 and DOT
406 cargo tanks, and DOT 57
portable tanks are not
authorized.
Specification MC 304, MC307,
MC331 cargo tank motor
vehicles; and MC310, MC 31 1,
MC 312, DOT 407, and DOT
412 cargo tank motor vehicles
with tank design pressure of at
least 172.4 kPa (25 psi} are
authorized.









puactn


Tank cars must be marked with the nama of the cargo,

The only tank cars authorized am class DOT 105, 1 12,
and 1 1 4 tank car tanks with ft test pressure of 2069 kPa
(300 psi) or greater.

Each tank must have a minimum shelf thickness of 10
mm (.384 Inch} with mild steel with at least 5.0 mm
(.197 Inch) lead lining.


Riveted tank car tanks are not authorized.







Class DOT 103, 104, 105. 109, 111, 112, 114. or 115
tank car tanks; and Class 106 or 1 10 muttl unit tank car
tanks are authorized.
Riveted tank car tanks and AAR208 tank car tanks are
prohibited.
Tank cars must be marked with the name of the cargo.

Tanks must be made of steel that Is rubber lined or
unlined. Unlined tanks must be passtvated before
being placed In service. If unlined tanks are washed
out with water, they must be repassfvated prior to return
to service. Cargo in unlined tanks must be Inhibited so
that the corrosive effect on steel is not greater than that
of hydrofluoric acid of 65 percent concentration.

Each tank shatJ have a minimum shell thickness of 8.0
mm (.315 inch) mild steel

Gauging devices are required on Class DOT 103, 104,
and 111 tank car tanks.



Veswlitorag*

Must be stowed *on deck* on a
cargo vowel, but Is prohibited
on a passenger vessel.

Must be stored •dev of (Mag
quarters' and •separated from*
foodstuffs.












Must be stowed 'on deck1 on a
cargo vessel, but Is prohibited
on a passenger vessel.
Stow 'dear of living quarters'
and "separated from' foodstuffs.
Keep as cool as reasonably
practicable.
Stow 'away from' sources or
heal










General

Cylinders: SpedffcaUon 3, 3A,
3AA, 3B. 3C, 3E. 4. 4A. 25, OT38
cylinder*; or

Specification 48, 4BA, 4BW, or
4C cylinders, H they are not
brazed.

Filling density must not exceed
85 percent of the water weight
capacity of the cylinder.

Cylinders used exclusively In this
service may, In lieu of a periodic
hydrostatic retest, be given a
complete external visual
Inspection, Such Inspections
shall be mode on cylinders
cleaned to bare metal.
Steel packaglngs must be
corroston-resfetant or have
protection against corrosion.
Packaglngs must be protected
with non-metallic linings
Impervious to the cargo or have a
suitable corrosion allowance.

Glass materials of construction
are not authorized for any part of
a packaging which is normally In
contact with the hazardous
material.
Aluminum construction materials
are not authorized for any part of
a packaging which Is normally In
contact with the hazardous
material.




Hon-Wk Pwduging ContefcMf*
CoMfefettiOf
Outer
WA


















For combination packaglngs, K
plastic Inner packaglngs are
used, they must be packed In
tightly closed metal receptacles
before packing In outer
packaglngs.
Steel drum: 1A1 or 1A2
Aluminum drum: 1B1 or 1B2
Metal drum other than steel or
aluminum: 1N1 or 1N2
Plywood drum: ID
Fiber drum: 1G
Plastfc drum: 1H1 or 1H2
Steel Jerrican: 3A1 or 3A2
Plastic jerrican: 3H1 or 3H2
Steel box: 4A1 or 4A2
Aluminum box: 481 or 482
Natural wood box: 4C1 or4C2
Plywood box:4D
Reconstituted wood box; 4F
Fiberboard box: 4G
Expanded plastfc box: 4H1
Solid plastic box: 4H2
l
Inner
N/A


















Qlassor
earthenware
receptacles
Plastfc receptacles
Metal receptacles
Glass ampoules












Stogie

N/A


















Except for transportation by passenger aircraft,
the following single packagings are authorized:
Steel drum: 1A1 or 1A2
Aluminum drum: 1B1 or 1B2
Metal drum other than steel or aluminum: 1N1 or
1N2
Plastic drum: 1H1 or 1H2
Steel jerrican: 3A1 or 3A2
Plastic Jerrican: 3H1 or 3H2
Plastic receptacle In steel, aluminum, fiber or
plastic drum:6HA1, eHB1.flHQ1.6HH
Plastic receptacle In steel, aluminum, wooden,
plywood or fiberboard box: 6HA2, 6HB2, 6HC.
6HD2, or 6HG2.
Glass, porcelain or stoneware in steel aluminum
or fiber drum: 6PA1, 6PB1,or6PG1.
Glass, porcelain or stoneware in steel, aluminum
or fiberboard box: 6PA2. 6PB2. 6PC or 6PG2
Glass, porcelain or stoneware in solid or
expanded plastic packaging! 6PH1 or 6PH2,
Cylinders, specification, as prescribed for any
compressed gas, except for Specifications 8
and - 3HT.
Alrc»A/R**o«l


Forbidden


















For transportation In one
package by passenger-
carrying aircraft or
passenger-carrying rail car:
0.5 L
For transportation by cargo
aircraft: 2.5 L











XI-2

-------
                                                                                                                                 EXHIBIT XI-1
                                                                                                Department of Transportation Regulations for the Transportation of HF
Lee* than
60%
•bength
(•qiMotn
HF)
                       Cargo tanks
Specification MC 304, MC307,
MC331 cargo tank motor
vehicles; and MC310, MC 311,
MO 312, DOT 407, and DOT
412 cargo tank motor vehicles
with tank design pressure of at
toast 172.4 kPa (25 psl) are
authorized.
Tank cars must be marked with the name of the cargo.

Each tank shall have a minimum shell thickness of 8.0
mm (.315 Inch) mild steel

Class DOT 103, 104, "105,109, 111,112, 114, or 115
tank car tanks: and Class 106 or 110 multi unit tank oar
tanks are authorized.
                                             Gauging devices are required on Class DOT 103,104,
                                             and 111 tank car tanks.

                                             Riveted tank car tanks are not authorized.
—


••••••
>f the cargo.
mess of 8.0
14, or 115
unit tank car
DT 103, 104,






Vetwlatonge

•••••••••••••••
Material must be stowed 'on
deck1 on a cargo vessel, but Is
prohibited on a passenger
vessel.
Stow 'clear of living quarters'
and "separated from* foodstuffs.
Keep as cool as reasonably
practicable.
Stow 'away from" sources or
heat




Non-bulk Pack
General

Packaglngs must be protected
witti non-metallic linings
Impervious to the cargo or have a
suitable corrosion allowance.
Steel packaglngs must be
corrosion-resistant or have
protection against corrosion.
Glass materials of construction
are not authorized for any part of
a packaging which Is normally In
contact with the hazardous
material.
Aluminum construction materials
are not authorized for any part of
a packaging which Is normally In
contact «Hi ins h=:d=--
materU.



=====
aging Container*
ComUruUon
outer
For combination packaglngs, If
plastic Inner packaglngs are
used, they must be packed In
tightly closed metal receptacles
before packing In outer
packagings.
Steel drum: 1A1 or 1A2
Aluminum drum: 1B1 or 182
Metal drum other than steel or
aluminum: 1N1 or 1N2
Plywood drum: 1D
Fiber drum: 1G
Plastic drum: 1H1 or 1H2
Wooden barrel: 2C2
Steel jerrican: 3A1 or 3A2
Plastic Jerrican: 3H1 or 3H2
Steel box: 4A1 or 4A2
A!ii.T.!ri»m box: 4B1 or 4R2
Natural wood box: 4C1 or 4C2
Reconstituted wood box: 4F
Fiberboard box: 4G
Expanded plastic box: 4H1
Solid plastic box: 4H2
Inner
fmf^aa^a
Glass or
earthenware
receptacles
Plastic receptacles
Metal receptacles
Glass ampoules






Single
s==a=Ba=a=s=saaaefa
Except for transportation by passenger aircraft,
the following single packaglngs are authorized:
Steel drum: 1A1 or 1A2
Aluminum drum: 1B1 or 1B2
Metal drum other than steel or aluminum: 1N1 or
Plastic drum: 1H1 or 1H2
Wooden barrel: 2C1
Steel Jerrican: 3A1 or 3A2
Plastic jerrican: 3H1 of 3H2
Plastic receptacle In steel, aluminum, fiber or
plastic drum BHA1, 6HB1, 6HG1, 6HH.
Plastic receptacle In steel, aluminum, wooden,
plywood orfiberboard box : 6HA2, 6HB2. 6HC.
6HD2, or 6HG2.
Glass, porcelain or stoneware In steel aluminum
or fiber drum: 6PA1, SPB1, orepst.
Glass, porcelain or stoneware In steel, aluminum
or ftbeiboardboxePA2, 6PB2, 6PCorePG2.
Glass, porcelain or stoneware In solid or
expanded plastic packaging: 6PH1 or 6PH2.
Plastic receptacle In plywood drum: eHDI


Aircraft/Railroad
transportation maximum

E>J>JHMI*^BMM
For transportation In ona
package by passenger-
carrying aircraft or
passenger-carrying rail car:
1L
For transportation by cargo
aircraft: 30 L




                                                                                                                                                                                                                                                                       XI-3

-------
                                        EXHIBIT XI-2
                          Typical Class 105A300W Tank Car
                                                           Discharge Outlets
                                          Safety Valve




                            Detail of top unloading arrangement
                                                 Loading Platform
                               Illustration of tank car layout
Page XJ-4

-------
                                     EXHIBIT XI-3
            Typical Railway Tank Car Specifications - Class 105A300W
Description
Overall
Nominal Capacity
Car weight - empty
Car weight - max.
Tank Car Size (Imp. gal.)
9,000

9000 gal.
66,800 Ib.
177,000 Ib.
21,000

21 ,000 gal
90,000 !b.
1 84000 Ib.
28,000

28,000 gal
11 2,000 Ib.
263000 Ib.
Tank
Material
Thickness
Inside Diameter
Test Pressure
Burst Pressure
Steel
11/16 in.
88 in.
300 psi.
750 psi.
Steel
11/16 in.
95 in.
300 psi.
750 psi.
Steel
11/16 in.
120 in.
300 psi.
750 psi.
Source:  Environment Canada Environmental Protection Service, Hydrogen Fluoride and Hydrofluoric Acid,
        Environmental and Technical Information for Problem Spills, Technical Services Branch, Ottawa,
        Ontario, July 1984. (220)
                                                                                       Page X/-5

-------
                                                 EXHIBIT XI-4
                                      Railway Tank Car Specifications
DOT Specification Number
103AW
105A100W
105A300W
111A100W2
111A100W4
111A100W5
112A400W
112S400W
114A400W
Description
Steel tank with dome. Insulation optional. Safety valve
set at 2070 kPa (300 psi). Bottom outlet and washout
prohibited. Test pressure:
Steel tank with dome. Insulation required. Safety valve
set at 517 kPa (75 psi). Bottom outlet and washout
prohibited. Test pressure: 100 psi.
Steel tank with dome. Insulation required. Safety valve
set at 1550 kPa (225 psi). Bottom outlet and washout
prohibited. Test pressure: 300 psi.
Steel tank without dome. Insulation optional. Safety
valve set at 517 kPa (75 psi). Test pressure: 100 psi.
Steel tank without dome. Insulation required. Safety
valve set at 517 kPa (75 psi). Bottom outlet and
washout prohibited. Test pressure: 1 00 psi.
Rubber-lined steel tank without dome. Insulation
optional. Safety vent, burst at 413 kPa (60 psi). Bottom
outlet and washout prohibited. Test pressure: 100 psi.
Steel tank with dome. Insulation not used. Safety valve
set at 2070 kPa (300 psi). Bottom outlet and washout
prohibited. Test pressure: 400 psi.
Special permit tank car. Same as 112A400W, except
no insulation used and equipped with head shield. Test
pressure: 400 psi.
Steel tank with dome. Insulation not used. Safety valve
set at 2070 kPa (300 psi). Bottom outlet and washout
optional. Test pressure 400 psi.
          Source:  Environment Canada Environmental Protection Service, Hydrogen Fluoride and Hydrofluoric Acid,
                   Environmental and Technical Information for Problem Spills, Technical Services Branch, Ottawa,
                   Ontario, July 1984. (220)
Page XI-6

-------
                                 EXHIBIT Xi-5
                       Motor Vehicle Tank Specifications
       DOT Specification
            Number
      MC312
                 Description
Steel butt-welded tank.  Design and construct
in accordance with American Society for
Mechanical Engineers (ASME) Code when
unloading by pressure in excess of 103 kPA
(15psi).  Gauging device is not required.  Top
and/or bottom discharge outlet.  Minimum one
pressure relief device per compartment as
required by ASME Code.  One minimum 15 in.
diameter manhole per compartment. Bottom
washout optional.
Source: Environment Canada Environmental Protection Service, Hydrogen Fluoride and Hydrofluoric Acid,
       Environmental and Technical Information for Problem Spills, Technical Services Branch, Ottawa,
       Ontario, July 1984. (220)
                                  EXHIBIT XI-6
                             Cylinder Specifications
   DOT Specification Number
   4B
   4BA
                     Description
    Welded and brazed steel cylinders. Service
    pressure 1035 to 3450 kPa (150 to 500 psi).
    Capability must not exceed 1000 Ib. water.
    Welded or brazed steel cylinders made of
    definitely prescribed steels.  Service pressure
    1550 to 3450 kPa (150 to 500 psi). Capacity
    must not exceed 1000 Ib. water.
Source:  Environment Canada Environmental Protection Service, Hydrogen Fluoride and Hydrofluoric Acid,
        Environmental and Technical Information for Problem Spills, Technical Services Branch, Ottawa,
        Ontario, July 1984. (220)
                                                                               Page XI-7

-------
 Polyethylene carboys

     Reagent grade hydrofluoric acid (49%) is often supplied in polyethylene bottles. These carboys
 can generally hold 30 or 55 gal.8  Polyethylene containers (specification 2SL) with a steel overpack
 (Specification 6D) are frequently employed.  The container has a 260 Ib. capacity and is designed for
 one-way service, but is returned for disposal9.
Page XI-8

-------
                                        ENDNOTES
1.   "Performance-Oriented Packaging Standards; Changes to Classification, Hazard Communication,
    Packaging and Handling Requirements Based on U.N. Standards and Agency Initiative, Final
    Rule," 49 CFR Part 107, et. a/., Department of Transportation Research and Special Programs
    Administration. (139.9)

2.   "Hazardous Materials Regulations," 49 CFR 171-180. (156.5)

3.   Environment Canada Environmental Protection Service, Hydrogen Fluoride and Hydrofluoric Acid,
    Environmental and Technical Information for Problem Sp///s, Technical Services Branch, Ottawa,
    Ontario, July  1984. (220)

4.   Environment Canada Environmental Protection Service.

5.   Environment Canada Environmental Protection Service.

6.   Chemical Industries Association, Guide to Safe Practice in the Use and Handling of Hydrogen
    Fluoride, London, England, 1988. (147)

7.   Environment Canada Environmental Protection Service.

8.   Mark, Herman F., Donald F. Othmer, Charles G. Overberger, and Glenn T. Seaborg, eds., Kirk-
    Othmer Encyclopedia of Chemical Technology, Third Edition, Volume 10, John Wiley and Sons,
    New York, 1978. (285)

9.   Environment Canada Environmental Protection Service.
                                                                                   Page XI-9

-------
                                  This page intentionally left blank.
Page XI-10

-------
                                     APPENDIX XII
            DATA BASE SOURCES FOR ACCIDENT INFORMATION
ARIP Database

       The Accidental Release Information Program (ARIP) database, maintained by EPA, contains
about 2,200 records of chemical accidental release events that have occurred since October, 1986.
The purpose of the ARIP database is to collect more detailed information on the causes of accidents,
the prevention efforts already in use at the facility to prevent chemical accidents, and the changes to
prevent a reoccurrence.  ARIP was started because no other database in operation contains such
information.

       Facilities are required under CERCLA to report releases of CERCLA hazardous substances
when such releases exceed a reportable quantity (RQ). These releases are reported to the  National
Response Center (NRC) and shared with EPA's Emergency Release Notification System (ERNS; see
below). EPA periodically screens the ERNS database to find events that meet one or more  of these
triggers:
              The quantity released was above a certain multiple of the RQ;
              The release resulted in a death or injury;
              The release was one in a trend of frequent releases from the same facility; or
              The release involved an extremely hazardous substance (EHS) as listed by EPA under
              section 302 or SARA/EPCRA.
       EPA then sends a detailed questionnaire to the facility that reported the release event that
meets these criteria. When the questionnaire is returned, the data is encoded into the database.

       In addition to routine information such as the chemical released, the amount lost, the media
affected (air, water, etc.) and consequences of the release (deaths, injuries, evacuations), ARIP
captures unique details such as the duration of the release, circumstances leading up to the release,
whether a hazard evaluation has been conducted for the process where the release occurred and
changes instituted or planned to prevent the release from ireoccurring.  Note however, that ARIP is not
statistically representative of all industry. ARIP is designed to capture events involving CERCLA or
extremely hazardous substances with more severe consequences. It does not contain events
associated with flammable or petroleum products.

AHE Database

       The Acute Hazardous Events (AHE) database was developed by EPA to provide an historical
perspective on the magnitude  of chemical accidents in the United States following the Bhopal, India
disaster. The database contains about 6,200 records that represents information on roughly 11,000
incidents that occurred between 1982 and 1986.  Data on the events was collected from a variety of
sources including the United Press International (UPI), Associated Press (AP), 26 daily newspapers,
EPA Region VII office files, six  offices of five state governments and from spill reports to the National
Response Center (NRC, see below).  The data collection v/as only intended to provide a "snapshot" of
the number of chemical accidents occurring at fixed facilities versus transportation, fire and explosion
events versus toxic releases, and the degree of deaths, injuries, evacuations and environmental
damage associated with these kinds of incidents in the United States. The data has not been
thoroughly verified and caution should  be used when interpreting certain findings.
                                                                                  Page XII-1

-------
 ERNS

        The Emergency Response Notification System (ERNS) is maintained by EPA from data
 reported to the National Response Center (NRC); the U.S. Coast Guard, or directly to EPA regional
 offices. A facility is required by law to report releases of more than a reportable quantity (RQ) of a
 CERCLA hazardous substance to the NRC, the Coast Guard or to EPA regional offices.  Reports to
 the NRC are used primarily to determine if a federal response team is needed to assist with the
 response to the incident. EPA then compiles the data into ERNS and shares it with EPA Regional
 Offices. Much of the data in ERNS is very early information during the life of an emergency and data
 is often changed later when more accurate details are available and the system often contains
 duplicate records.  EPA is continually updating information to delete duplicates and verify reported
 data.

 HMIS

        The Hazardous Materials Information System (HMIS) is a database containing information  on
 chemical releases from transportation incidents. Information includes carrier name, shipper name,
 mode of transport, release due to vehicular accident or derailment, number of containers shipped,
 reason(s) for failure of containers, amount of material released, and consequences such as the
 number of deaths or injuries and the number of individuals evacuated.  HMIS identifies transportation
 incidents involving HF and can indicate those deaths and injuries directly attributed to the HF release
 and those attributed to the physical impact of the accident (e.g. collision).
Page Xll-2

-------
                              APPENDIX Xill
         DESCRIPTIONS OF HYDROGEN FLUORIDE ACCIDENTS
      This appendix presents general descriptions of hydrogen fluoride accidents from the ARIP,
AHE, ERNS, and HMIS databases and other sources.
                                                                   Page XIII-1

-------

-------
                            EXHIBIT XIII-1
General Description of Hydrogen Fluoride Accidents From the ARIP Database
DATE
10-31-86
12-19-86
2-26-87
6-16-87
8-3-87
10-30-87
10-4-88
10-18-88
12-20-88
1-9-89
FACILITY (TYPE)*
Exxon, Billings Refinery
(R)
ARMCO, Inc. (SS)
J&L Specialty Products
(SS)
Mobil Oil, Paulsboro
Refinery (R)
Ashland Chemical (C)
Marathon Petroleum (R)
J&L Specialty Products
(SS)
Phillips 66 Company (R)
Philips Display
Components (GE)
Mobil Joliet Refinery (R)
STATE
MT
PA
PA
NJ
PA
TX
OH
UT
OH
IL
QUANTITY
RELEASED
	 *-*=
2,000 Ibs
1,590 Ibs
9,450 Ibs
200 Ibs
500 Ibs
53,200 Ibs
3,418 Ibs
1,600 Ibs
150 Ibs
124 Ibs
MEDIA
AFFECTED
1,000 Ibs to air,
1,000 Ibs to sw
sw
9,282 Ibs to air,
168 Ibs to sw
air
air
air, sw, land
land
air, sewer to
treatment facility
land
air
CONCENTRATION*"
anhydrous
70% aqueous
70% aqueous
40-67% aqueous
-92% aqueous
anhydrous
70% aqueous
anhydrous
25% aqueous
-70 ppm aqueous
RELEASE
EVENT
vapor release
spill
NPDES
excursion
vapor release
vapor release
vapor release
spill
vapor release
spill
vapor release
MIGRATION OFF
SITE
750 Ibs to air
1,590 Ibs to sw
168 Ibs to sw
200 Ibs to air
500 Ibs to air
air, sw, land
no
yes
no
124 Ibs to air
CAUSE/DESCRIPTION
piping failure in alkylation
unit due to equipment
allure
failure of welded fitting
due to corrosion in
concrete containment
area
equipment failure in
storage vessel 	
bull plug failure in piping
of HF alkylation unit
rupture disk on storage
tank failed due to unusual
stress due to steel vent
pipe extension
crane dropped a
convection unit onto an
HF storage vessel
faulty piping on process
vessel caused leak
employee opened valve
on a line under pressure
with HF acid
closed valve caused
overpressurization in
head tank and failure of
top gasket .
increased water levels
doubled rate of acid
soluble oil production and
increased amount of HF
sent to process heater

-------
                       EXHIBIT Xlll-1 (continued)
General Description of Hydrogen Fluoride Accident* From the ARIP Database
DATE
1-17-89
1-29-89
6-12-89
6-27-89
7-24-89
8-9-89
8-11-89
8-30-89
9-12-89
12-22-89
FACILITY (TYPE)*
3M(C)
Sikorsky Aircraft (A)
Hughes Aircraft (EL)
Great Lakes Chemical (C)
Learjet, Inc. (A)
Carpenter Technology
(SS)
General Chemical
Corporation (C)
Kennametal Nevada
Refinery (U)
Columbia Falls Aluminum
(AL)
Allied-Signal (C)
STATE
AL
CT
CA
AR
KS
PA
CA
NV
MT
IL
QUANTITY
RELEASED
800 IDS
115 Ibs
1 Ib
1,320 Ibs
unknown
795 Ibs
4,000 Ibs
174 Ibs
7,700 Ibs
810 Ibs
MEDIA
AFFECTED
air, treatment
air, land
land
air

land, air
sewer to
treatment facility
land, air
air
200 Ibs to sw, 10
Ibs to land, 600
Ibs to sewer to
treatment facility
' ""-""'in' "in ' i,iroesa[aag^joa»^B
CONCENTRATION"*
75% aqueous
2-3 oz/gal aqueous
70% aqueous
anhydrous
< 1% aqueous
70% aqueous
49% aqueous
aqueous
anhydrous
36% aqueous
RELEASE
EVENT
	 	 -•• " " ' '~~*
vapor release
spill
spill
vapor release
spill
spill, vapor
release
spill
spill
vapor release
spill
MIGRATION OFF
SITE

air
no
no
no
no
no
no
7,700 Ibs to air
200 Ibs to sw
CAUSE/DESCRIPTION
==^==^==^=
HF leaked through control
valve during repairs
pipe fitting in the
discharge pump
developed a leak and HF
was released into
secondary containment
bottle of HF fell from a
pallet and was broken by
a forklift
pressure gauge on
storage vessel failed due
to corrosion
process vessel leaked
storage tank leaked
failed containment area,
equipment failure,
operator error
flanged joint on acid
pipeline between storage
tank and process line
failed
flow of absorptive alumina
for aluminum processing
was cut off due to
operator error
extremely cold weather
caused freezing of tank
vent line and resulted in
tank overpressure and
tank failure

-------
                       EXHIBIT Xlll-1 (continued)
General Description of Hydrogen Fluoride Accidents From the ARIP Database
DATE

12-23-89
4-27-90


5-26-90
6-26-90
7-6-90
7-27-90




7-30-90

8-6-90

9'5-90


9-27-90


FACILITY (TYPE)*

Atochem (CFC)
BP Oil (R)


Texas Instruments (EL)
Intel (EL)
Chemical Dynamics (T)**
Atochem North America,
Inc (C)




General Chemical
Corporation (C)

Flowline (SS)

McDonnell Douglas (A)


General Chemical
Corporation (C)

STATE

KY
PA


TX
CA
TX
NJ




CA

PA

CA


CA


QUANTITY
RELEASED
200 Ibs
333 IbS


188 Ibs
800 Ibs
6,812 Ibs
500 Ibs




300 Ibs

852 Ibs

15 Ibs


350 Ibs


MEDIA
AFFECTED
air
150 Ibs to air, 183
Ibs to sewer to
treatment facility
air
land
water
sw




land

water, land

air


land


CONCENTRATION"*

anhydrous
90% aqueous


49% aqueous
2% aqueous
aqueous
832 mg/l aqueous




49% aqueous

aqueous

70% aqueous


49% aqueous


RELEASE
EVENT
vapor release
vapor release


vapor release
spill
spill
spill




spill

spill

vapor release


spill


MIGRATION OFF
SITE




188 Ibs to air
no

500 Ibs to sw




no



no


no


CAUSE/DESCRIPTION

sulfuric acid froze, HF kiln
vented via seal
gasket failure on a sight
glass assembly

spill pit indicator failure
caused leak
open valve caused
overflow onto floor
2" pipe coupling failed,
leaked waste etchant mix
open valve caused
wastewater to siphon
through a pump to the
effluent line and
discharge into the
Delaware river
spill during drum transfer,
RCRA sump drain system
failed and allowed waste
to sntsr effluent system
leak of spent HF (waste
pickle liquor) from storage
tank
residue in drum left in the
sun volatilized causing HF
to vent through a bung
hole
valves on process vessel
failed causing a sewer
line leak in the packaging
area

-------
                                                                         EXHIBIT XIII-1 (continued)
                                                General Description of Hydrogen Fluoride Accidents From the ARIP Database
DATE
10-8-90
11-14-80
3-15-91
FACILITY (TYPE)*
Chevron (R)
Alcoa (AL)
Roadway Express (T)**
STATE
UT
AR
QA
QUANTITY
RELEASED
225 Ibs
576 Ibs
199 Ibs
MEDIA
AFFECTED
air
air, vegetation
land
CONCENTRATION*"
anhydrous
anhydrous
aqueous
RELEASE
EVENT
vapor release
vapor release
spill
MIGRATION OFF
SITE



CAUSE/DESCRIPTION
contractor ruptured valve
during repairs
accidental release from
production (probably
AlFg)
drum punctured with
Source: ARIP Database, U.S. Environmental Protection Agency, 1986 to 1991.  Data as of January 29,199^ (45)

       * Facility Type
       (A)     Aircraft manufacture
       (AL)    Aluminum manufacture
       (C)     Chemical manufacturing, including the production of HF
       (CFC)  Chlorofluorocarbon manufacturing
       (CL)    HF used in cleaning
       (EL)    Manufacture of electronics, including semi-conductors
       (GE)   Glass etching, including TV cathode ray tube production
       (R)     Oil refineries
       (SS)    Stainless steel manufacture
       (T)     Transportation, includes loading and unloading, equipment failures during transport, and package failures during transport
       (U)     Undetermined

       ** ARIP designates these incidents as having occurred at a fixed facility, but as the HF was in transit at the time, it has been designated here as (T)
       *** Some data on concentration (aqueous or anhydrous) was provided by Allied-Signal.1
       Letter, Segregation of Databases for Aqueous and Anhydrous HF, From: William J. Hague, Allied-Signal, To: Craig Matthiessen, U.S. Environmental Protection Agency, August 18,
       1992.  (293.3)

-------
                            EXHIBIT XIII-2
General Descriptions of Hydrogen Fluoride Accidents From the AHE Database
DATE
6-1-80
4-12-83
4-29-83
9-7-83
9-9-83
10-10-83
10-17-83
11-10-83
6-21-84
6-25-84
8-24-84
11-10-84
2-5-85
3-3-85
3-3-85
5-7-85
5-14-85
6-10-85
7-5-85
8-27-85
3-6-86
FACILITY (TYPE)*
City Chemical, Co. (C)
Alchem-tron, Inc. (C)
Kerr-McGee Corp. (T)
Chevron Chemical (R)
Harshaw Chemical (T)
New Jersey Warehouse (T)
Teff Freight Lines (T)
Southern Pacific Railroad (T)
Chemical Waste
Management (T)
Allied Chemical (C)
Sid Harvey, Co. (U)
Garrett Truck Lines (T)
Conrail Railroad (T)
McDonnell Douglas (A)
unknown (T)
Inland Container (T)
unknown (T)
Marathon Petroleum (R)
SHMAS Contracting (CL)
unknown (T)
Chevron Chemical (T)
STATE
NJ
OH
OK
UT
OH
NJ
CA
CA
NC
LA
MO
CA
IN
MO
CO
MO
OH
TX
NJ
CA
TX
QUANTITY
RELEASED
412 Ibs
unknown
unknown
unknown
unknown
unknown
488 Ibs
unknown
unknown
unknown
unknown
188 Ibs
unknown
150 Ibs
unknown
413 Ibs
unknown
100 Ibs
unknown
unknown
300 Ibs
MEDIA
AFFECTED
air, land
air
air
air
air
air, water, land
air, land
air
land
air
air, water

air
air, land
air
on-site
air
air
air, land
air
air
CONCEN-
TRATION"
aqueous
anhydrous
anhydrous
anhydrous
anhydrous
anhydrous
anhydrous
anhydrous
aqueous
anhydrous
aqueous
anhydrous
anhydrous
70% aqueous
anhydrous
aqueous
anhydrous
anhydrous
aqueous
anhydrous
anhydrous
RELEASE
EVENT
spill
vapor release
vapor release
vapor release
vapor release
vapor release
spill
spill
spill
vapor release
spill
spill
vapor release
spill
vapor release
spill
vapor release
vapor release
spill
spill
vapor release .
CAUSE/DESCRIPTION
leaking drum
explosion from enclosed vapors in
tank
valve failed in truck
operator opened flanges on process
line before it was cleared
train

leak from truck at freight company
cap off car on train, leaking from
manway
tank/truck rubber lining failed

plastic drums leaking
truck
rail car cracked due to accumulate
fatigue
pipeline rupture
worker opened tanker dome on train
forklift punctured drum
venting from valve

cleaning building with acid compound
open container inside truck
relief valve failed on railroad car

-------
                                                                         EXHIBIT Xlll-2 (continued)
                                                 General Description of Hydrogen Fluoride Accidents From the AHE Database
DATE
9-12-86
9-30-86
10-4-86
10-31-86
12-5-86
10-30-87
FACILITY (TYPE)*
Conway Central Express (T)
United Parcel Service (T)
Allied/Signal Corp. (C)
Exxon USA (R)
May Trucking (T)
Marathon Petroleum (R)
STATE
Ml
CA
LA
MT
NV
TX
QUANTITY
RELEASED
412 Ibs
unknown
2,750 Ibs
2,000 Ibs
262 Ibs
270,000 Ibs
MEDIA
AFFECTED
air, land
air, land
air
air
land
air, land
CONCEN-
TRATION"
aqueous
aqueous
anhydrous
anhydrous
aqueous
anhydrous
RELEASE
EVENT
spill
spill
vapor release
vapor release
spill
vapor release
CAUSE/DESCRIPTION
plastic drum failed on truck
plastic bottle leaked in truck
power failure led to bypass of
scrubber system

barrels leaked in truck accident
falling equipment sheared pipe from
tank
Source: AHE Database, U.S. Environmental Protection Agency, 1960-1987.  Data as of June 5,1991 (05)

       * Facility Type

       (A)     Aircraft manufacture
       (AL)    Aluminum manufacture
       (C)     Chemical manufacturing, including the production of HF
       (CFC)  Chlorofluorocarbon manufacturing
       (CL)    HF used in cleaning
       (EL)    Manufacture of electronics, including semi-conductors
       (GE)    Glass etching, including TV cathode ray tube production
       (R)     Oil refineries
       (SS)    Stainless steel manufacture
       (T)     Transportation, includes loading and unloading, equipment failures during transport, and package failures during transport.
       (U)     Undetermined

       ** Some data on concentration (aqueous or anhydrous) was provided by Allied-Signal.2
       Letter, Segregation of Databases for Aqueous and Anhydrous HF, From: William J. Hague, Allied-Signal, To: Craig Matthiessen, U.S. Environmental Protection Agency  August 18
       1992.  (293.3)                                                                                                                                  a   y,    a     -

-------
                            EXHIBIT XIII-3
General Descriptions of Hydrogen Fluoride Accidents From the ERNS Database
DATE
12-19-86
1-19-87
2-3-87
2-5-87
2-26-87
3-8-87
4-2-87
4-6-87
4-11-87
5-6-87
6-16-87
6-25-87
6-25-87
6-26-87
7-24-87
8-3-87
8-7-87
8-19-87
10-1-87
10-20-87
10-30-87
11-13-87
11-29-87
12-18-87
FACILITY (TYPE)*
Armco (SS)
Intel (EL)
Allied-Signal (CFC)
Allied (CFC)
J&L Specialty Products
Allied Chemicals (C)
Fairchild Semiconductors (EL)
Du Pont (C)
Allied (T)
Du Pont (C)
Mobil (R)
Allegheny Ludlum (T)
Ozark Mahonig (C)
Chem-Security Systems (T)
Moreland (U)
Ashland Chemical (C)
A & B Industrial (T)
Monsanto (C)
Matlack Trucking (T)
Mercury Stainless (SS)
Marathon (R)
Delco Electronics (EL)
Mobil (R)
Allied-Signal (C)
STATE
PA
CA
IL
LA
PA
NJ
CA
NJ
Ml
TX
NJ
PA
OK
OR
NY
PA
Ml
MO
CA
OH
TX
IN
CA
IL
QUANTITY RELEASED
1,590 Ibs"
5 IDS
171 Ibs
100 IDS
9,450 Ibs"
unknown
500 Ibs
90 Ibs
63 Ibs
105 Ibs
200 Ibs**
135 Ibs
3 Ibs
8 Ibs
unknown
500 Ibs**
180 Ibs
200 Ibs
unknown
500 Ibs
53,200 Ibs**
90 Ibs
unknown
1,800 Ibs
CONCENTRATION***
70% aqueous
aqueous
anhydrous
anhydrous
aqueous
anhydrous
aqueous
anhydrous
70% aqueous
anhydrous
anhydrous
70% aqueous
anhydrous
anhydrous
anhydrous
anhydrous
60% aqueous
aqueous
anhydrous
70% aqueous
anhydrous
aqueous
anhydrous
1% aqueous
CAUSE/DESCRIPTION
leak on storage tank lining
operator error, spill from waste tank
pump failed into contained area
leak in fluorocarbons plant
acid tank spill lost from containment
unknown
2 acid drums reacted, cause unknown •
leak in one-ton cylinder
leak from truck
reactor seal failure
line failure in alkylation unit
hose failed during truck unloading
fire in unit, unknown cause
truck valve malfunction
leaks from two corroded cylinders
anhydrous storage leak
drum leak in truck
waste water sewer leak
truck valve broke, release to air
pump hose prime line valve rupture
crane accident broke pipes on tank
pump failed causing tank overflow
pipe broke during repairs, alkylation unit,
gasket leak in transfer piping

-------
                        EXHIBIT XIII-3 (continued)
General Description of Hydrogen Fluoride Accidents From the ERNS Database
DATE
1-4^88
1-27-88
2-3-88
3-6-88
5-4-88
5-20-88
5-28-88
7-7-88
7-12-88
8-4-88
8-5-88
8-21-88
8-22-88
8-22-88
8-26-88
8-27-88
8-27-88
8-28-88
9-11-88
10-5-88
10-13-88
10-18-88
10-20-88
FACILITY (TYPE)*
Classic Glass (GE)
Phillips ECG (EL)
Soco-Wester Chem (C)
Amoco Oil (R)
Chevron (R)
Matlack Trucking (I)
GE Solid State (EL)
Aratex Services (U)
Mobil Oil (R)
General Chemical (C)
Matlack (T)
Norfolk Southern Railroad (T)
Balero Refining (R)
Pennwalt (T)
Delco Moraine (T)
Shipper (I)
unknown (T)
Allied-Signal (CFC)
Reynolds Metals (T)
J & L Specialty Products (SS)
Chevron (R)
Phillips 66 (R)
Metellics (U)
STATE
NC
OH
CA
TX
PA
MO
PA
CA
NJ
CA
MO
VA
TX
KY
OH
PR
PR
LA

OH
TX
TX
CA
QUANTITY RELEASED
450 Ibs
unknown
250 Ibs
220 Ibs
1,300 Ibs
90 Ibs
1,800 Ibs
unknown
unknown
127 Ibs
unknown
9 Ibs
>RQ
580 Ibs
27 Ibs
279 Ibs
270 Ibs
150 Ibs
2,250 Ibs
3,418 Ibs"
unknown
1,600 Ibs"
unknown
CONCENTRATION*"
aqueous
25% aqueous
70% aqueous
anhydrous
anhydrous
anhydrous
aqueous
anhydrous
anhydrous
anhydrous
23% aqueous
aqueous
anhydrous
70% aqueous
aqueous
anhydrous
anhydrous
anhydrous
anhydrous
aqueous
anhydrous
anhydrous
anhydrous
CAUSE/DESCRIPTION
drain from building runs waste etchant solution
valve corrosion
top of drum blew off
circulation line, nipple failed on alkylation unit
alkylation unit heater stack, unknown cause
rupture disk failed on trailer
leak in storage tank wall
HF mixed with Clg
propane and small amount of HF leaked
line parted on discharge of pump, 200 gallon
reactor
sump plate hole cover on trailer leaked
gallon jug on train leaked
alkylation unit leak
leak occurred on loading line to truck
residue in an empty drum spilled
container leaked on ship
2 drums fell off pallet on loading dock and leaked
hole in catalyst stripper
container failure
piping system on raw acid storage tank leaked
three alarm fire, alkylate or other
bleeder valve on acid tank opened
incompatible with chemical mix

-------
                        EXHIBIT XIII-3 (continued)
General Description of Hydrogen Fluoride Accidents From the ERNS Database
DATE
12-15-88
12-20-88
1-11-89
2-6-89
3-18-89
4-18-89
4-24-89
5-12-89
5-23-89
5-30-89
6-23-89
6-30-89
7-19-89
7-24-89
8-9-89
8-12-89
8-30-89
9-8-89
3-8-90
3-15-90
4-11-90
4-20-90
4-28-90
FACILITY (TYPE)*
Florida Marine Chem (C)
Philips Display (GE)
Safeway (U)
Hi-Pure Chemical (C)
Vista Chemical (C)
Rogers Cartage (T)
MEMC Electronic Materials (EL)
unknown (T)
Pennwalt (C)
Ashland Petroleum (R)
Mobil Oil (R)
Ripley FD (T)
Westinghouse (SS)
Learjet (A)
Carpenter Technology (SS)
General Chemical (C)
Kennametal (U)
Yellow Freight (I)
Darivain (U)
Roadway Express (T)
Mobil Oil (R)
Golden West Refinery (R)
unknown (U)
STATE
FL
OH
CA
PA
LA
IL
SC

NJ
MN
CA
TN
UT
KS
PA
CA
NV
KS
TX
GA
LA
CA
CA
QUANTITY RELEASED
unknown
150 Ibs**
unknown
unknown
unknown
225 Ibs
750 Ibs
45 Ibs
unknown
10 Ibs
9 Ibs
unknown
unknown
unknown**
795 Ibs**
4,000 Ibs**
174 Ibs**
495 Ibs
unknown
100 Ibs
150 Ibs
2 Ibs
unknown
CONCENTRATION***
aqueous
25% aqueous
anhydrous
aqueous
anhydrous
anhydrous
aqueous
aqueous
anhydrous
anhydrous
anhydrous
aqueous
aqueous
aqueous
70% aqueous
49% aqueous
aqueous
aqueous
anhydrous
44% aqueous
anhydrous
anhydrous
aqueous
CAUSE/DESCRIPTION
solution used to clean boat hulls spilled into water
operator error, tank overflow from misvalving
unknown
storage tank had leak in liner
acid regenerator column flange leak
tank truck rupture disk cracked
HF with nitric acid and acetic acid spilled from
mixing area
truck wreck caused drums to leak
pipelines at plant leaking
alkylation unit valve failure
unknown
300 gallon drum leaking behind facility
pickling tank drain line failure
processing tank, 6,000 gallon liner failure
10,000 gallon tank began leaking at couplings
58,000 Ibs storage tank leak
pipe leaked at flange
etohant solution ate through drum
storage tank relief valve failure
30 gallon drum puncture
HF alkylation unit pump seal failure
seal rupture on valve
acid bath left overnight vented

-------
                       EXHIBIT XIII-3 (continued)
General Description of Hydrogen Fluoride Accidents From the ERNS Database
DATE
5-26-90
6-12-90
6-26-90
7-2-90
7-6-90
7-6-90
7-31-90
8-6-90
8-7-90
8-15-90
9-18-90
9-19-90
9-27-90
10-3-90
10-8-90
10-25-90
11-12-90
11-14-90
11-29-90
12-4-90
12-5-90
2-5-91
FACILITY (TYPE)*
Texas Instruments (EL)
AERO Chem (T)
Intel (EL)
Merlin Specialty (U)
Teledyne (U)
Chemical Dynamics (T)
General Chemical (C)
Flowline (SS)
BP Oil (R)
Marathon Petroleum (R)
Local Fairchild (A)
Intel (EL)
General Chemical (C)
Pratt & Whitney (A)
Chevron (R)
General Electric (EL)
Cifco Industries (U)
Alcoa (AL)
Ashland (T)
Atochem (T)
Timet (U)
unknown (U)
STATE
TX
CA
CA
MD
OR
TX
CA
PA
LA
LA
CA
CA
CA
CT
UT
KY
OH
AR
OH
KY
OH
CA
QUANTITY RELEASED
188 Ibs"
19lbs
800 Ibs"
900 Ibs
unknown
6,812 Ibs"
350 Ibs"
852 Ibs"
21 Ibs
1 Ib
30 Ibs
82,125 Ibs
350 Ibs"
4,500 Ibs
225 Ibs"
150 Ibs
1,800 Ibs
576 Ibs"
9 Ibs
200 Ibs
9,000 Ibs
9 Ibs
CONCENTRATION"*
49% aqueous
70% aqueous
2% aqueous
anhydrous
aqueous
aqueous
aqueous
aqueous
anhydrous
anhydrous
aqueous
aqueous
aqueous
aqueous
anhydrous
aqueous
aqueous
anhydrous
anhydrous
anhydrous
aqueous
anhydrous
CAUSE/DESCRIPTION
9,000 gallon liner cracked
storage tank vent leaked during delivery
spill was contained and neutralized
storage tank, unknown cause
system transfer line failed, unknown cause
2-inch pipe coupling failed
drum transfer, sump basin malfunction, spill
leak of spent HF from storage tank
thermowell leak
leaking line in fractionator, valve shut
holding tank overflow, spill to floor
contained and neutralized
drain line valve failed
nitric acid and HF mixed due to failed check valve
in waste line
contractor ruptured valve during repairs
6' drain line broke
HF and nitric acid spilled when valve broke on
storage tank
accidental release from production (AlFg)
valve left open on truck, spill on road
rail car overfilled
hole in bottom of tank, unknown cause
unknown

-------
                                                                        EXHIBIT XIII-3 (continued)
                                                General Description of Hydrogen Fluoride Accidents From the ERNS Database
DATE
5-6-91
7-7-91
7-17-91
9-9-91
10-12-91
FACILITY (TYPE)*
Ausimot (CFC)
Formosa Plastics (C)
ARMCO steel (SS)
Kelly Industrial (T)
unknown (T)
STATE
NJ
TX
PA
MA

QUANTITY RELEASED
200 IDS
unknown
unknown
1lb
9 IDS
CONCENTRATION***
anhydrous
anhydrous
aqueous
aqueous
anhydrous
CAUSE/DESCRIPTION
transfer pipe between two tanks failed due to
corrosion
30 minute release to air
spill into TP to creek
vandals opened spigot on drum
spilled on pavement
Source: ERNS Database, U.S. Environmental Protection Agency, U.S. Department of Transportation, U.S. Coast Guard, National Response Center, 1987-1991.  Data as of November 10,
1991. (139.525)

       * Facility Type

       (A)     Aircraft manufacture
       (AL)    Aluminum manufacture
       (C)     Chemical manufacturing, including the production of HF
       (CFC)  Chlorofluorocarbon manufacturing
       (CL)    HF used in cleaning
       (EL)    Manufacture of electronics, including semi-conductors
       (GE)   Glass etching, including TV cathode ray tube production
       (R)     Oil refineries
       (SS)    Stainless steel manufacture
       (T)     Transportation, includes loading and unloading, equipment failures during transport, and package failures during transport.
       (U)     Undetermined

       ** Release quantities have been updated with facility-verified ARIP information.
       *** Some data on concentration (aqueous or anhydrous) was provided by Allied-Signal.3
   3   Letter, Segregation of Databases for Aqueous and Anhydrous HF, From:  William J. Hague, Allied-Signal, To: Craig Matthiessen, U.S. Environmental Protection Agency,  August 18,
       1992. (293.3)

-------
                             EXHIBIT XIII-4
General Descriptions of Hydrogen Flurolde Accidents From the HMIS Database
DATE
5-2-80
1-12-81
1-14-81
6-29-81
2-18-84
2-4-85
3-28-85
5-14-85
6-9-85
9-26-85
11-27-85
6-25-87
1-19-89
4-18-89
FACILITY (TYPE)*
United Parcel Service,
Inc. (I)
Ryder Truck Lines, Inc.
0)
Ryder Truck Lines, Inc.
m
APA Transport Corp. (T)
Illinois Central Gulf
Railroad (T)
Consolidated Rail
Corporation (T)
Matlack Inc. (T)
Baltimore & Ohio Railroad
Co. (T)
Southern Pacific (T)
Matlack Inc. (T)
Illinois Central Gulf
Railroad (T)
Bendix Corp. (T)
ABF Freight System, Inc.
CO
Rogers Cartage Co. (T)
STATE
CT
LA
FL
NJ
LA
IN
CA
OH
AR
CA
LA
PA
OR
CA
QUANTITY
RELEASED
unknown
8 Ibs
40 Ibs
8 Ibs
unknown
6,400 Ibs
unknown
unknown
0.48 Ibs
16 Ibs
160 Ibs
800 Ibs
200 Ibs
200 Ibs
MEDIA
AFFECTED





air

air






CONCEN-
TRATION"
anhydrous
anhydrous
aqueous
aqueous
anhydrous
anhydrous
anhydrous
anhydrous
anhydrous
anhydrous
anhydrous
70% aqueous
aqueous
aqueous
RELEASE
EVENT
spill
spill
spill
spill
spill
vapor release
vapor release
vapor release
spill
spill
spill
spill
spill
spill
MIGRATION OFF
SITE





vapor cloud
traveled 2.5 miles








CAUSE/DESCRIPTION
dropped cylinder
valve failure on cylinder
bottom of a plastic drum failed
dropped plastic bottle
loose valve on a tank car
leak occurred when weld
failed on pressurized tank car
because of continued
transportation of a tank car
that was reported to be
leaking
packaging failed on cargo
tank
loose valve caused venting on
tank car
derailment caused loose valve
on pressurized tank car
ruptured piping on cargo tank
packaging failed on tank car
packaging failed on tank
trailer
loose valve on drum
defective valve on cargo tank

-------
                                                                       EXHIBIT XIII-4 (continued)
                                               General Description of Hydrogen Fluoride Accidents From the HMIS Database
DATE
5-15-89
7-21-89
1-26-90
1-26-90
8-27-90
FACILITY (TYPE)*
Markair (T)
Consolidated Freightways
0)
Union Pacific Railroad
Co. (T)
United Parcel Service,
Inc. (T)
Enroserv Midwest,
STATE
AK
IN
AR
CA
OH
QUANTITY
RELEASED
8 IDS
Bibs
8 Ibs
0.32 Ibs
16 Ibs
MEDIA
AFFECTED





CONCEN-
TRATION**
aqueous
aqueous
anhydrous
aqueous
aqueous
RELEASE
EVENT
spill
spill
vapor release
spill
spill
MIGRATION OFF
SITE





CAUSE/DESCRIPTION
loose valve on jug
loose valve and improper
blocking on plastic jug
loose valve on tank car
dropped and ruptured closure
of plastic bottle
corrosion on the bottom of the
inner liner of a metal drum
Source: Hazardous Materials Information System (HMIS), Department of Transportation, January 1,1980 - December 31,1990. Data as of November 11,

       * Facility Type

       (A)    Aircraft manufacture
       (AL)   Aluminum manufacture
       (C)    Chemical manufacturing, including the production of HF
       (CFC)  Chlorofluorocarbon manufacturing
       (CL)   HF used in cleaning
       (EL)   Manufacture of electronics, including semi-conductors
       (GE)   Siass etching, tneiuuing TV c-aihode ray tube production
       (R)    Oil refineries
       (SS)   Stainless steel manufacture
       (T)    Transportation, includes loading and unloading, equipment failures during transport, and package failures during transport.
       (U)    Undetermined

       ** Some data on concentration (aqueous or anhydrous) was provided by Allied-Signal.4
1991. (155)
    4   Letter, Segregation of Databases for Aqueous and Anhydrous HF, From:  William J. Hague, Allied-Signal, To:  Craig Matthiessen, U.S. Environmental Protection Agency, August 18,
        1992.  (293.3)

-------
                                                                        EXHIBIT XIII-5
                                            General Descriptions of Hydrogen Fluoride Accidents From Other Sources
DATE
6-5-78
6-5-85
1-4-86
11-24-87
1-13-90
FACILITY (TYPE)*
Phillips Petroleum
Company (R)
Allied Corporation
(C)
Kerr-McGee (R)
Mobil Torrance
Refinery (R)
Powerine Oil
Company (R)
STATE
KS
LA
OK
CA
CA
QUANTITY
RELEASED

10,000 Ibs
29,500 Ibs
100 Ibs

MEDIA
AFFECTED
air

air
air

CONCEN-
TRATION"
anhydrous
anhydrous
anhydrous
anhydrous
anhydrous
RELEASE
EVENT
vapor release
vapor release
vapor release
spill
vapor release
MIGRATION
OFF SITE


air, land
no

CAUSE/DESCRIPTION
operator opened valve without checking
pressure of the vessel causing acid
vapors to back up into the gas purge
line and release from tips of pilot light
and drain valve
premature rupture disk failure at normal
operating pressures and human failure
to detect a missing pressure gauge
assembly caused release
operator heated overfilled cylinder, in
violation of standard operating
procedures, resulting in a cylinder failure
and HF release
due to malfunctioning monitoring devices
and human error, HF overflowed into a
tank and mixed with another chemical,
causing an explosion
corroded pipe holding HF ruptured as
employees were trying to replace it
SOURCE
Newspaper
Accident
Report2
Newspaper
Newspaper
Newspaper
5
* Facility Type

(A)    Aircraft manufacture
(AL)   Aluminum manufacture
(C)    Chemical manufacturing, including the production of HF
(CFC)  Chlorofluorocarbon manufacturing
(CL)   HF used in cleaning
(EL)   Manufacture of electronics, including semi-conductors
(GE)   Glass etching, including TV cathode ray tube production
(R)    Oil refineries
(88)   Stainless steel manufacture
(T)    Transportation, includes loading and unloading, equipment failures during transport, and package failures during transport.
(U)    Undetermined
** Some data on concentration (aqueous or anhydrous) was provided by Allied-Signal.

-------
                                                                              ENDNOTES
1.      The Kansan, 'Acid Leak Under Lens at Phillips," June 7,1978. (03)
2.      Letter, From: E.G. Calamari, Allied-Signal, Baton Rouge, LA, To: Gustave Von Bonungen, LA Department of Environmental Quality, Baton Rouge, LA, June 12,1985. (294b)
3.      Hazardous Materials Intelligence Report, 'Leak of Toxic Gas at OK Uranium Processing Plant,' January 10,1986, p 1. (520)
4.      Stein, George, "Mobil Refinery Explosion Laid to Human Error," Los Angeles Times, February 11,1990, p B1. (463.5)
5.      Daunt, Tina, "Officials Probe Chemical Leak That Injured 7 at Refinery," Los Angeles Times, February 11,1990, p J1. (134)
                                                                                                                                                              Page XIII-16

-------

-------
                                APPENDIX XIV
                  DEATHS, INJURIES, OR EVACUATIONS
             CAUSED BY HYDROGEN FLUORIDE ACCIDENTS
      This appendix presents information on deaths, injuries, or evacuations caused by hydrogen
fluoride accidents listed in the ARIP, AHE, ERNS, and HMIS data bases and other sources.
                                  EXHIBIT XIV-1
Dea
DATE
12-19-86
10-30-87
1-17-89
6-12-89
7-24-89
3-15-91
ths, Injuries,
FACILITY
Armco, Inc.
Marathon
Petroleum
3M
Hughes
Aircraft
Learjet,
Inc.
Roadway
Express
or Evacual
Liste
STATE
PA
TX
AL
CA
KS
GA
ions Caused by 1
d in the ARIP Dat
EVACUATION
no
-5,800
general public
no
no
no
no
Hydrogen Flu
abase
ON-SITE
INJURIES
1
0
1
1
2
1
loride Accidei
OFF-SITE
INJURIES
0
1,037
0
0
0
0
its
DEATHS
0
0
0
0
0
0
         Source:  ARIP Database, U.S. Environmental Protection Agency, 1986 to 1991. Data as of January 29, 1992.
               (45)
                                                                       Page XIV-1

-------
                                                  EXHIBIT XIV-2
Deaths
DATE
4-12-83
10-10-83
10-17-83
2-5-85
3-3-85
3-3-85
5-14-85
7-5-85
8-27-85
9-12-86
10/30/87
>, Injuries, or Evacual
Liste
FACILITY
Alchem-tron, Inc.
New Jersey
Warehouse
Teff Freight Lines
Conrail Railroad
McDonnell
Douglas
unknown
unknown
SHMAS
Contracting
unknown
Conway Central
Express
Marathon
Petroleum
tions Cauj
sd in the fi
STATE
OH
NJ
CA
IN
MO
CO
OH
NJ
CA
Ml
TX
sed by Hydrogen
iHE Database
EVACUATION
unknown
150
evacuated
100 residents
evacuated
yes
no
no
yes
no
unknown
no
-5,800*
residents
evacuated
Fluoride Ac<
INJURIES
1
0
0
0
0
1
0
1
2
1
1,037*
:idents
DEATHS |
0
0
0
0 I
2 I
0 I
I
o I
0 I
0 1
0 I
            Source:  AHEData Base, U.S. Environmental Protection Agency, 1960-1987. Data as of June 5, 1991. (05)




            *        Number has been updated with facility-verified ARIP information.
Page XIV-2

-------
                                 EXHIBIT XIV-3
 Deaths, Injuries, or Evacuations Caused by Hydrogen Fluoride Accidents
                        Listed in the ERNS Database
DATE
6-25-87
10-30-87
8-22-88
10-18-88
10-20-88
1-11-89
6-23-89
7-24-89
9-8-89
10-3-90
10-8-90
FACILITY
Ozark
Mahonig
Marathon
Petroleum
Balero
Refining
Phillips 66
Metellics
Safeway
Mobil Oil
Learjet
Yellow
Freight
Pratt &
Whittney
Chevron
STATE
OK
TX
TX
UT
CA
CA
CA
KS
KS
CT
UT
EVACUATION
yes
yes
yes
no
yes
no
no
13
yes
no
no
INJURIES
(ON- OR
OFF-SITE)
1
-1000
0
2
0
1
1
0
6
4
1
DEATHS
0
0
0
0
0
0
0
0
0
0
0
Source:  ERNS Data Base, U.S.
        Coast Guard, National
Environmental Protection Agency, U.S. Department of Transportation, U.S.
Response Center, 1987-1991. Data as of November 10, 1991. (139.525)
                                                                                Page XIV-3

-------
                                               EXHIBIT XIV-4
Deaths, Injuries, or Evacuations Caused by Hydrogen Fluoride Accidents
Listed in the HMIS Database
DATE

2-4-85
3-28-85
5-14-85

6-25-87
5-15-89
FACILITY

Consolidated
Rail Corp
Matlack Inc.
Baltimore &
Ohio
Railroad Co.
Bendix Corp.
Markair
STATE

IN
CA
OH

PA
AK
EVACUATION

1,500
evacuated in
1.1 mi2 area
no
yes

no
no
ON-SITE
INJURIES
4 response
personnel
6
0

1
1
OFF-SITE
INJURIES
71 others
0
0

unknown
unknown
DEATHS

0
0
0

0
0
            Source:  Hazardous Materials Information System (HMIS), Department of Transportation, January 1,  1980 •
                    December 31, 1990. Data as of November 11, 1991. (155)
Page XIV-4

-------
                                              EXHIBIT XIV-5
Deaths, Injuries, or Evacuations Caused by Hydrogen Fluoride Accidents
Listed in Newspapers and Accident Reports
DATE

6-5-781


6-S-852

1-4-863

1-24-874


1-13-905


FACILITY

Phillips
Petroleum
Company
Allied
Corporation
Kerr-McGee

Mobil
Torrance
Refinery
Powerine
Oil
Company
STATE

KS


LA

OK

CA


CA


EVACUATION

unknown


facility
employees
facility
employees
unknown


unknown


ON-SITE
INJURIES
34


2

100

3


7


OFF-SITE
INJURIES
0


0

0

0


0


DEATHS

0


0

1

0


0


1.        "Acid Leak Under Lens at Phillips," The Kansan, June 7, 1978. (03)

2.        Letter, From: E.G. Calamari, Allied-Signal, Baton Rouge, LA, To: Gustave Von Bonungen, LA Department of
         Environmental Quality, Baton Rouge, LA, June 12, 1985. (294 b)

3.        "Leak of Toxic Gas at OK Uranium Processing Plant," Hazardous Materials Intelligence Report, January 10,
         1986, p 1. (520)

4.        Cole, T.D., Mobil Oil Corporation, comments from technical review of Hydrogen Fluoride Study Report to
         Congress, Draft May 8, 1992, June 4, 1992. (126)

5.        Daunt, Tina, "Officials  Probe Chemical Leak That Injured 7 at Refinery," Los Angeles Times, February 11,1990,
         p J1. (134)
                                                                                                   Page XIV-5

-------
                                  This page intentionally left blank.
Page XIV-6

-------
                                      APPENDIX XV
                        POPULATION CHARACTERIZATION
       Populations within a one-mile and five-mile radius of selected, facilities that produce, use, or
handle anhydrous HF were estimated by use of 1990 census data1. These facilities and estimated
populations around them are shown in Exhibit XV-1.  Facilities examined include HF producers,
fluorocarbon manufacturers, petroleum refineries with HF alkylation units, and facilities that reported
inventories of 100,000 pounds or greater in the 1990 Toxic Release Inventory (TRI) database.
Population estimates were developed for facilities on this list if further review indicated that they use
anhydrous HF rather than HF in solution. The location of each facility was determined by using
latitude and longitude coordinates reported in the TRI database.  Using the facility location as the
center, circles with one-mile and five-mile radii were defined.  Populations were based on census
blocks with centroids (the approximate center points of census blocks) within the radii. For example, if
the centroid of a census block is within the one-mile or five-mile radius, the entire population of the
census block is included in the estimate even though a portion of the population is located outside
the radius. On the other hand, the population of a census block with a centroid located outside the
radius would not be counted within the population estimate even though a portion of the census block
is physically located within the radius.  Thus, the populations represented in Exhibit XV-1 are
approximations.  Populations are estimated to the nearest thousand; populations estimated as less
than 500 are reported as "< 1,000."
                                                                                   Page XV-1

-------
EXHIBIT XV-1
Population Characterization
FACILITY NAME
3M
Great Lakes Chemical Corp. - South Plant
Kerley AG Products Inc.
Allied-Signal Inc.
Dow Chemical Co.
Du Pont
General Chemical Corp.
Golden West Refining Co.
Mobil Oil Corp.
Powerine Oil Co.
Ultramar Inc.
General Chemical Corp.
3M
LOCATION
Decatur, AL
El Dorado, AR
Phoenix, AZ
El Segundo, CA
Pittsburgh, CA
Antioch, CA
Pittsburgh, CA
Santa Fe Springs, CA
Torrance, CA
Santa Fe Springs, CA
Wilmington, CA
Claymont, DE
Cordova, IL
COUNTY
Morgan
Limestone
Union
Maricopa
Los Angeles
Contra Costa
Contra Costa
Contra Costa
Los Angeles
Los Angeles
Los Angeles
Los Angeles
New Castle (DE)
Salem (NJ)
Delaware (PA)
Clinton (IA)
Rock Island
Whiteside
POPULATION WITHIN RADIUS
1 MILE 5 MILE
<1,000
0
0
3,000
Undetermined
< 1,000
8,000
2,000
14,000
9,000
12,000
1,000
2,000
0
0
1,000
0
0
21,000
1,000
4,000
191,000

105,000
69,000
56,000
411,000
550,000
522,000
436,000
56,000
3,000
75,000
6,000
1,000
1,000

Page XV-2

-------
EXHIBIT XV-1
Population Characterization (continued)
FACILITY NAME
Allied-Signal Inc. - Danville Works
Allied-Signal Inc.
Chemtech Industries Inc.
Clark Oil & Refining Corp.
Clark Oil & Refining Corp.
Marathon Petroleum Co.
Mobil Joliet Refining Corp.
Uno-Ven Co.
Indiana Farm Bureau Co-op. Assn. Inc.
Marathon Oil Co.
Atochem
Coastal Derby Refining Co.
Coastal Derby Refining Co.
LOCATION
Danville, IL
Metropolis, IL
East Saint Louis, IL
Hartford, IL
Blue Island, IL
Robinson, IL
Joliet, IL
Lemont, IL
Mount Vernon, IN
Indianapolis, IN
Wichita, KS
Wichita, KS
El Dorado, KS
COUNTY
Vermilion (IL)
Vermillion (IN)
Massac
McCracken (KY)
St. Clair
Madison
Cook
Crawford
Will
Will
Cook
DuPage
Posey
Marion
Hamilton
Boone
Sedgewick
Sedgewick
Butler
POPULATION
1 MILE
2,000
0
1,000
0
5,000
0
10,000
0
0
0
0
0
4,000
3,000
0
0
0
9,000
0
WITHIN RADIUS
SMILE
38,000
1,000
9,000
2,000
98,000
23,000
380,000
9,000
12,000
69,000
10,000
1,000
10,000
61,000
4,000
10,000
34,000
212,000
14,000

PageW-3

-------
EXHIBIT XV-1
Population Characterization (continued)
FACILITY NAME
Farmland Industries Inc.
National Cooperative Refinery Association
Texaco Refining & Marketing Inc.
El Dorado Plant
Total Petroleum Inc.
Ashland Petroleum Co.
Atochem North America Inc.
Du Pont Louisville Works
Allied-Signal Inc. Geismar Plant
Allied-Signal Inc. Baton Rouge South
BP Oil Co.
Laroche Chemicals Inc.
LOCATION
Coffeyville, KS
McPherson, KS
El Dorado, KS
Arkansas City, KS
Ashland, KY
Calvert City, KY
Louisville, KY
Geismar, LA
Baton Rouge, LA
Belle Chasse, LA
Gramercy, LA
COUNTY
Montgomery
Nowata (OK)
McPherson
Butler
Cowley
Boyd
Lawrence (OH)
Marshall
Livingston
Jefferson
Floyd (IN)
Iberville
Ascension
East Baton Rouge
West Baton Rouge
Plaquemines
St. James
St. John the Baptist
POPULATION
1 MILE
2,000
0
4,000
3,000
1,000
< 1,000
0
0
0
5,000
0
0
0
1,000
0
0
0
0
WITHIN RADIUS
SMILE
15,000
1,000
13,000
13,000
16,000
6,000
7,000
4,000
2,000
186,000
8,000
4,000
3,000
127,000
11,000
0
10,000
6,000
Marathon Petroleum Co.
Garyville, LA
St. John the Baptist
2,000
15,000
Page XV-4

-------
EXHIBIT XV-1
Population Characterization (continued)
POPULATION WITHIN RADIUS
FACILITY NAME
Mobil Oil Corp. Chalmette Refinery



Murphy Oil USA Inc. Meraux Refinery



Placid Refining Co.

Vista Chemical Co. -
Lake Charles Chemical Co.
Dow Corning Corp.


Du Pont Montague Works
Total Petroleum Inc.
3M

Ashland Petroleum Co.
St. Paul Park Refinery

Mallinckrodt Specialty Chemicals Co.


LOCATION
Chalmette, LA



Meraux, LA



Port Allen, LA

Westlake, LA

Midland, Ml


Montague, Ml
Alma, Ml
Cottage Grove, MN

Saint Paul Park, MN


Saint Louis, MO


COUNTY
St. Bernard
Jefferson
Plaquemines
Orleans
St. Bernard
Jefferson
Plaquemines
Orleans
West Baton Rouge
East Baton Rouge
Calcasieu

Bay
Midland
Saginaw
Muskegon
Gratiot
Washington
Dakota
Dakota
Washington
Ramsey
St. Louis
St. Clair (IL)
Madison (111)
1 MILE
4,000
0
0
2,000
3,000
0
0
0
< 1,000
0
3,000

0
0
0
1,000
2,000
1,000
0
3,000
1,000
0
Undetermined
0
0
SMILE
56,000
34,000
1,000
95,000
59,000
8,000
1,000
50,000
11,000
99,000
51,000

2,000
2,000
1,000
9,000
18,000
25,000
12,000
49,000
34,000
2,000

32,000
26,000
Page XV-5

-------
EXHIBIT XV-1
Population Characterization (continued)
FACILITY NAME
Cenex Refinery
Conoco Billings Refinery
Exxon Co. USA - Billings Refinery
Amoco Oil Co.
Ausimont USA Inc.
Du Pont Chambers Works
Mobil Oil Corp.
Giant Refining Co. - Ciniza
Navajo Refining Co.
Alcoa
Occidental Chemical Corp. - Niagara Plant
LOCATION
Laurel, MT
Billings, MT
Billings, MT
Mandan, ND
Thorofare, NJ
Deepwater, NJ
Paulsboro, NJ
Gallup, NM
Artesia, NM
Massena, NY
Niagara Falls, NY
COUNTY
Yellowstone
Yellowstone
Yellowstone
Morton
Burleigh
Gloucester
Philadelphia (PA)
Delaware (PA)
Salem
New Castle (DE)
Gloucester
Delaware (PA)
Philadelphia (PA)
McKinley
Eddy
St. Lawrence
Niagara
Erie
POPULATION
1 MILE
0
3,000
1,000
3,000
0
1,000
0
0
1,000
1,000
3,000
0
0
0
2,000
0
4,000
0
WITHIN RADIUS
SMILE
8,000
77,000
36,000
15,000
15,000
79,000
32,000
11,000
25,000
71,000
34,000
65,000
15,000
1,000
12,000
14,000
80,000
10,000
Ashland Petroleum Co. Canton Refinery
Canton, OH
Stark
2,000
129,000
Page XV-6

-------
FACILITY NAME
Conoco Ponca City Ref.
Kerr-McGee Refinery Corp.
Sequoyah Fuels Corp.
Sun Refining & Marketing Co.
Total Petroleum Inc.
Reynolds Metal Co. (produce)
Air Products & Chemicals Inc.
Ashland Chemical Inc.
BP Oil Co. Marcus Hook Refinery
Chevron USA
Mapco Petroleum Inc.
Amoco Oil Co. Texas City Refinery
EXHIBIT XV-1
Population Characterization (continued)
LOCATION COUNTY
Ponca City, OK Kay
Osage
Wynnewood, OK
Gore, OK
Tulsa, OK
Ardmore, OK
Troutdale, OR
Tamaqua, PA
Easton, PA
Trainer, PA
Philadelphia, PA
Memphis, TN
Texas City, TX
Garvin
Murray
Sequoyah
Tulsa
Osage
Carter
Multnomah
Schuylkill
Northampton
Bucks
Warren (NJ)
Delaware
Glouster (NJ)
Philadelphia
Delaware
Gloucester (PA)
Camden (NJ)
Shelby
Crittendon (AR)
Galveston
POPULATION
1 MILE
5,000
0
< 1,000
0
0
2,000
0
0
0
0
5,000
0
0
4,000
0
< 1,000
0
0
0
3,000
0
3,000
WITHIN RADIUS
SMILE
29,000
2,000
4,000
< 1,000
0
121,000
6,000
0
61,000
15,000
76,000
1,000
20,000
115,000
5,000
381,000
105,000
35,000
6,000
142,000
2,000
54,000
Page XV-7

-------
EXHIBIT XV-1
Population Characterization (continued)
FACILITY NAME
Champlin Refining & Chemicals Inc.
Chevron USA Inc. Port Arthur Refinery
Coastal Refining & Marketing, Inc.
Crown Central Petroleum Houston Refinery
Diamond Shamrock Refining & Marketing Co.
Three Rivers
Du Pont Corpus Christ! Plant
Du Pont La Porte Plant
Fina Oil & Chemical Co.
Koch Refining Co.
Marathon Petroleum Co.
Monsanto Co.
Phibro Refining Inc.
Phillips 66 Co.
Phillips 66 Co. Sweeny Complex
Shell Oil Co. Odessa Refinery
LOCATION
Corpus Christi, TX
Port Arthur, TX
Corpus Christi, TX
Pasadena, TX
Three Rivers, TX
Ingleside, TX
La Porte, TX
Big Spring, TX
Corpus Christi, TX
Texas City, TX
Alvin, TX
Texas City, TX
Borger, TX
Sweeny, TX
Odessa, TX
COUNTY
Nueces
Jefferson
Nueces
Harris
Live Oak
San Patricio
Harris
Howard
Nueces
Galveston
Brazoria

Hutchinson
Brazoria
Ector
POPULATION WITHIN RADIUS
1 MILE SMILE
'5,000 97,000
0
24,000
2,000
1,000
0
0
1,000
0
1,000
Undetermined
Undetermined
0
0
0
30,000
186,000
220,000
2,000
16,000
59,000
18,000
21,000
52,000


13,000
5,000
55,000
Page Xv-S

-------
FACILITY NAME

Southwestern Refining Co.

Valero Refining Co.

Chevron USA Inc.


Phillips 66 Co.


BP Oil Co.

Murphy Oil USA


Frontier Refining Inc.
                                                      EXHIBIT XV-1
                                           Population Characterization (continued)
LOCATION

Corpus Christi, TX

Corpus Christi, TX

Salt Lake City, UT


Woods Cross, UT


Ferndale, WA

Superior, Wl


Cheyenne. WY
                                                                                          POPULATION WITHIN RADIUS
COUNTY
Nueces
Nueces
Salt Lake
Davis
Davis
Salt Lake
Whatcom
Douglas
St. Louis (MN)
1 MILE
6,000
0
0
0
9,000
0
0
1,000
0
SMILE
103,000
35,000
0
0
66,000
1,000
7,000
28,000
2,000
Laramie
8,000
64,000
                                                                                                                 Page XV-9

-------
                                        ENDNOTES
  1.
"1990 Census of Population and Housing - Summary Tape Files," U.S. Department of
Commerce, Bureau of the Census, Data User Services Division, Washington, D.C.
Page XV-10

-------
                                    APPENDIX XVI
                    DESCRIPTION OF HGSYSTEM AND SLAB
       HGSYSTEM and SLAB are computer modeling systems which predict the dispersion of
accidental releases of hazardous substances.  The following describes the modeling systems, the
inputs needed and the data outputs obtained.

HGSYSTEM

       HGSYSTEM is a system of models developed for the prediction of the dispersion of accidental
releases of HF and other gases in the atmosphere. The behavior of HF when mixed with moist air is
very different than that of an ideal gas, and depending upon conditions the HF-air mixture can be
either denser or much less dense than air. This behavior can significantly influence the dispersion of
HF upon release into the atmosphere. HGSYSTEM follows an  earlier version, HFSYSTEM, which was
designed specifically for HF under contract to the Ambient Assessment  Group, a subcommittee of an
ad-hoc Industry Cooperative HF Mitigation Group sponsored by 20 companies from the chemical and
petroleum industries. The system was based upon an existing model in the public domain called
HEGADAS, a dense gas dispersion model.1

       This system of models was developed with the following specifications:

       K      to account for the thermodynamics of released HF with  air and
              moisture and cloud aerosol effects on ploud density;

       >-      to simulate pressurized releases as well as evaporation  from a pool;

       »•      to predict concentrations over varying surface  roughness conditions;

       »-      to predict concentrations for varying averaging periods;

       »-     to consider steady state, finite, and variable duration releases;

       +     to compute crosswind and downwind concentration profiles;

       ».     to handle both dense gas and passive dispersion; and

        ».     to compute release rates from storage vessels.

       The individual models are designed for simulating the  release of HF and to the subsequent
 dispersion of the gas in the atmosphere. The models may calculate time-dependent spillage of HF
 liquid or vapor from a pressurized vessel (HFSPILL); the steady-state or time-dependant evaporation of
 an HF source if the user is sure that a liquid pool would be formed based on the storage temperature
 and pressure (EVAP); the steady-state near-source behavior of HF (HFPLUME, HFFLASH, and
 HFJET); the ground-level dispersion of HF further downwind (HEGADAS-S for steady-state and
 HEGADAS-T for transient);  and the far-field passive dispersion for plumes which do not slump back to
 the ground (PGPLUME). Another model, PLUME, calculates the steady-state near-source behavior of
 hazardous chemicals other than  HF, and is similar to HFPL.UME  except that HF-specific
 thermodynamics have not been included.  These HGSYSTEM models are outlined in the following
 Exhibit XVI-1.
                                                                                 Page XVI-1

-------
                                           EXHIBIT XVI-1
            MODEL
                                 Descriptions of Models in HGSYSTEM
DESCRIPTION OF MODEL
GAS1
RELEASE2
HFSPILL
EVAP
HFFLASH3
HFJET4
HFPLUME
PLUME
HEGADAS-S
HEGADAS-T
PGPLUME
Spills from pressurized vessel
Evaporation and spreading from a
liquid pool
Flashing of HF
Jet flow
Jet flow and near-field dispersion
Jet flow and near-field dispersion
Ground level heavy gas
dispersion
Ground level heavy gas
dispersion
Elevated passive dispersion
HF
Both
HF
HF
HF
Ideal
Both
Both
Both
PR
UNPR
PR
PR
PR
PR
Both
Both
PR
TR |
ST.TR I
ST3 1
ST 1
ST5 1
ST5 1
ST, FD6 I
TR 1
ST, FD I
              Type of released gas: HF or Ideal gas
              Type of release: pressurized (PR) or unpressurized (UNPR); steady-state (ST), finite duration (FD), or
              transient (TR)
              If HFPLUME Is not run, HFFLASH is used to set HF post-flash data required as input to HEGADAS. Although
              HFFLASH is a steady-state program, it can be interfaced with  both HEGADAS-S and HEGADAS-T.
              HFJET is a simplified version of HFPLUME and should not normally be used.
              Although HFPLUME and PLUME are steady-state programs, they do accept a finite-duration input parameter I
              (e.g., from HFSPILL in case of HFPLUME). This parameter is  not used in HFPLUME/PLUME, but passed
              through to PGPLUME or HEGADAS-T.
              HEGADAS-S is a steady-state program.  HEGADAS-T is recommended for modeling finite-duration releases.
        These models are stand-alone programs which can be run separately or can be linked
together to simulate the various components of a release to the atmosphere.  HGSYSTEM has an
interactive program to allow the user to link the output from one model for use as the input into
another model (e.g., the pool evaporation program, EVAP, calculates the data needed for dispersion
calculations in the HEGADAS, heavy gas dispersion program).2

        HGSYSTEM also has several utility programs and batch files to provide a user-friendly
interactive environment, two data validation programs, three post-processor programs for generating
printouts or plotting files, one program for displaying file contents, and one special program which
checks for adequate memory for running the programs.  The basic input files for the models are
stored in a file and the user can select and edit the input files to specify  individual parameters. For
PageXVI-2

-------
example, for HFSPILL the user can input such information as storage temperature and pressure,
ambient atmospheric conditions, the quantity and composition of the liquid or vapor, and the pipe exit
plane conditions. The output will calculate the mass-discharge rate which along with other parameters
can be input into the model EVAP (evaporating pool model) or to the near-field jet/plume model
HFPLUME.

       HGSYSTEM takes into account changes in source strength over time, and whether the release
is coming from a pressurized source, such as a tank or pipe, or a non-pressurized source, such as a
pool or puddle.  It also factors in "surface roughness," i.e., whether the release is taking place in open
country or in an area where buildings or trees would slow down dispersion. Also, HGSYSTEM
predicts average concentration for any specified time period.  HGSYSTEM, developed specifically to
account for the unique properties associated with HF releases, has been expanded to also allow
evaluation of the release and dispersion of other "ideal" gases.

SLAB

       SLAB is a computer model that simulates releases of dense gases and the atmospheric
dispersion which follows. SLAB is capable of modeling both vapor releases and releases of liquid
droplet-vapor mixtures. SLAB was developed by the University of California, Lawrence Livermore
National Laboratory with support from the U.S. Department of Energy.  Recent versions of SLAB were
supported jointly by the United States Air Force Engineering and Services Center, and American
Petroleum Institute.3 The model is not particular to  HF-any dense gas can be modeled with SLAB.

       The SLAB model is designed to meet the following specifications:

       >      to account for the thermodynamics of liquid droplet formation in the emitted
              substance,  evaporation of the emitted substance, and evaporation of water vapor due
              to ambient humidity;

       >      to consider steady state, finite duration, and instantaneous releases;

       >      to predict concentrations for various averaging periods;

       »•      to predict (downwind) concentrations at various heights;

       >      to account for the thermodynamic effect of ground heating when a cloud is cooler
              than the ground;

       »•      to simulate horizontal jet and vertical jet releases;

       >      to predict time-averaged volume concentrations, including contours and maximum
              centerline concentrations; and

       »•      to account for lofting of a cloud if it  becomes lighter-than-air.

       Atmospheric dispersion is calculated by solving the conservation equations of mass,
momentum, energy, and species. The  equations are spatially averaged, and treat the cloud as a
steady state plume, and/or a transient puff.  Combinations of these dispersion modes are used to
model four types of sources: an evaporating pool, a horizontal or vertical jet, or an  instantaneous
release. See Exhibit XVI-2 for details on the source types. A continuous release (i.e., one with a long
source duration) is  treated as a steady state plume.  A finite duration release is described as a steady
state plume until the release stops.  Then the cloud is treated as a transient puff. An instantaneous
release is predicted using the puff dispersion mode.
                                                                                   Page XVI-3

-------
                                          EXHIBIT XVI-2
                                Descriptions of Source Types in SLAB
                 SOURCE TYPE
RELEASE HEIGHT
DISPERSION MODE
Evaporating Pool
Horizontal Jet
Stack or Vertical Jet
Instantaneous or Short Duration
Evaporating Pool
Ground
Elevated2
Elevated2
Ground
Steady State1, Transient Puff |
Steady State3, Transient Puff |
Steady State3, Transient Puff 1
Transient Puff 1
              If steady state is not reached for Evaporating Pool, SLAB switches to Short Duration Evaporating Pool.
              Release height may be set at zero (ground level) if desired.
              The steady state description is assumed for the duration of the active release.
        The input file for a SLAB run consists of 30 possible parameters, and must be created by the
 user.  Necessary inputs include source type, gas properties, spill properties (i.e., source temperature),
 field properties (i.e., concentration averaging time), meteorological parameters (i.e., surface
 roughness), and a numerical substep.  Unlike HGSYSTEM, SLAB does not account for changes in
 source strength over time; the mass source rate must be specified by the user. For example, if the
 rate specified is an initial rate for a pressurized release in which the rate decreases with time, SLAB
 would consequently tend to make conservative predictions.

        The SLAB model provides a detailed output file consisting of problem description, intermediate
 results in the form of instantaneous spatially-averaged cloud properties, and time-averaged volume
 fraction. Although the computer program has no plotting capabilities, enough information is  provided
 to obtain contour plots of zones that encompass a concentration of interest.

        SLAB can perform multiple,  consecutive runs of a release scenario with different
 meteorological inputs.  SLAB also predicts average concentrations for any specified time period.
 However, if the duration of the release is much smaller than the averaging time, the dose will be
 averaged with concentration values of zero.
Page XVI-4

-------
                                 ENDNOTES

International Conference and Workshop on Modeling and Mitigating the Consequences of
Accidental Releases of Hazardous Materials, May 20-24, 1991, Fairmont Hotel, New Orleans,
Louisiana, American Institute of Chemical Engineers, Center for Chemical Process Safety, New
York, 1991, "The HGSYSTEM Dispersion Modelling Package:  Development and Predictions,"
by Puttock, J.S., K. McFarlane, A. Prothero, P.T. Roberts, F.J.  Rees, H.W.M. Witlox, and
Douglas N. Blewitt. (416.5)

Witlox, H.W.M., K.  McFarlane, F.J. Rees, and J.S. Puttock, Development and Validation of
Atmospheric Dispersion Models for Ideal Gases and Hydrogen Fluoride, Part II: HGSYSTEM
Program User's Manual, Shell Research Limited, Thornton Research Centre, Chester England,
November 1990. (509)

Ermak, Donald L, User's Manual for SLAB: An Atmospheric Dispersion Model for Denser-than-
Air Releases, University of California, Lawrence Livermore Laboratory, June 1990, Document
Number:  UCRL-MA-105607. (139.8)
                                                                           Page XVI-5

-------
                                 This page intentionally left blank.
Page XW-6

-------
                                  APPENDIX XVI!
                 INPUTS FOR HGSYSTEM AND SLAB MODELS
       This appendix presents the inputs that were used in HGSYSTEM and SLAB models for each
scenario examined in Chapter 9.  Exhibit XVII-1a, 1b and 1c show the HGSYSTEM inputs. Exhibit
XVIMa presents inputs for scenarios using the HFSPILL and HFPLUME models (scenarios 2 and 4-7).
Exhibit XVII-b presents inputs for other scenarios that also use HFSPILL and HFPLUME models
(scenarios 8-10 and  13-16). Exhibit XVlMc shows inputs for scenarios that were modeled using
HFSPILL, and HFPLUME or EVAP-HF (scenarios 1, 3, 11 and 12). The first page in each exhibit
shows the inputs for the HGSYSTEM subprograms HFSPILL, HFPLUME and/or EVAP-HF.  The second
page shows the remaining inputs required for the dispersion program HEGADAS, which can be run in
either the transient or steady state mode. Note that most of the scenarios were run in the HEGADAS
transient mode except for scenarios 10, 11, and 16, that were run in the HEGADAS steady state
mode.  For each scenario, HF was input as the gas to be modeled. The gas properties listed in the
input tables are HF properties that are pre-defined by the HGSYSTEM program.  Additionally, "N/A"
indicates that a particular input parameter is not required  by the subprogram.

       The SLAB model can be run for an evaporating pool, a horizontal or vertical jet release, and
an instantaneous release.  Input parameters similar to HGSYSTEM are required to run each SLAB
model. The inputs to SLAB scenarios 1-16 are presented in Exhibit XVII-2.
                                                                             Page XVII-1

-------

-------
                                                      Exhibit XVII-1 a
                                      Input for HGSYSTEM Models: HFSPILL and HFPLUME
Input
Total Storage Capacity (m3)
Liquid Mass (kg)
Effective Release Diam. (m)
Water Mass Fraction (%)
Reservoir Temperature (°C)
Reservoir Pressure (atm)
Ambient Pressure (atm)
Orifice Height (m)
Release Discharge Angle
Last Downwind Displacement (m)
Ambient Temperature (°C)
Wind Speed (rn/s)
Wind Speed Reference Height (m)
Relative Humidity (%)
Surface Roughness (m)
Pasquill Stability Level
Model
Prog.
S
S
S,P
S
S,P
S,P
S,P
p
p
p
P,H
P.H
P,H
P,H
P,H
P,H
Scenario 2:
Derailment
Empty Rail Car
(D Stab.)
79.6
65,000
0.15
0.0
27
1.5
1
3.00
-5°
1000
20
5
10.0
50
0.03
D
Scenario 4:
Hose Failure
Empty
(F Stab.)
24
18,600
0.05
0.0
30
6.5
1
1.00
-5°
1000
20
1.5
10.0
50
0.03
F
Scenario 5:
Hose Failure
Mitigated
(F Stab.)
24
18,600
0.05
0.0
30
6.5
1
1.00
-5°
1000
20
1.5
10.0
50
0.03
F
Scenario 6:
Hose Failure
Empty
(D Stab.)
24
18,600
0.05
0.0
30
6.5
1
1.00
-5°
1000
20
5
10.0
50
0.03
D
Scenario 7:
Hose Failure
Mitigated
(D Stab.)
24
18,600
0.05
0.0
30
6.5
1
1.00
-5°
1000
20
5
10.0
50
0.03
D
Note:  S = HFSPILL  P = HFPLUME  H = HEGADAS

-------
                                                   Exhibit XVIMa (continued)
                                    Inputs for HGSYSTEM Models:  HEGADAS-S and HEGADAS-T
Input
Cloud Data Output Code
Surface Transfer Code
Cloud Meander Averaging Time (sec)
Crosswind Dispersion Coefficient
GAS PROPS.: Molecular Weight (g/mol)
Specific Heat (J/mol-°C)
Fraction water pick-up
Post-Flash Temperature (°C)
Thermodynamic Model
Cloud Output Step Length (m)
Last Downwind Distance (m)
Cone, to Stop Calculations (kg/m3)
Upper, Lower Concentrations (kg/m3)
TSTAR observer (minimum) (sec)
TSTAR observer (maximum) (sec)
Model
Prog.
T,SS
T,SS
T,SS
T.SS
T,SS
T,SS
T,SS
T,SS
T,SS
T,SS
SS
T,SS
T,SS
T
T
Scenario 2:
Derailment
Empty Rail Car
(D Stab.)
0
3
840
2
20.0
29.1
0.0
19.5
2
100
N/A
5E-6
5E-5, 5E-6
1000
7000
Scenario 4:
Hose Failure
Empty
(F Stab.)
0
3
432
2
20.0
29.1
0.0
19.5
2
100
N/A
5E-6
5E-5, 5E-6
5000
16000
Scenario 5:
Hose Failure
Mitigated
(F Stab.)
0
3
60
2
20.0
29.1
0.0
19.5
2
100
N/A
5E-6
5E-5, 5E-6
5000
20000
Scenario 6:
Hose Failure
Empty
(D Stab.)
0
3
432
2
20.0
29.1
0.0
19.5
2
100
N/A
5E-6
5E-5, 5E-6
200
10000
Scenario 7:
Hose Failure
Mitigated
(D Stab.)
0
3
60
2
20.0
29.1
0.0
19.5
2
100
N/A
5E-6
5E-5, 5E-6
900
1800
Note: T = HEGADAS-T (Transient)  SS = HEGADAS-S (Steady State)

-------
                                                     Exhibit XVIMb
                                     Inputs for HGSYSTEM Models: HFSPILL and HFPLUME
Input
Total Storage Capacity (m3)
Liquid Mass (kg)
Effective Release Diameter (m)
Water Mass Fraction (%)
Reservoir Temperature (°C)
Reservoir Pressure (atm)
Ambient Pressure (atm)
Orifice Height (m)
Release Discharge Angle
Last Downwind Displacement (m)
Ambient Temperature (°C)
Wind Speed (m/s)
Wind Speed Reference Height (m)
Relative Humidity (%)
Surface Roughness (m)
Pasquill Stability Level
Model
Prog.
S
S
S,P
S
S,P
S,P
S,P
P
P
P
P,H
P,H
P,H
P,H
P,H
P,H
Scenario 8:
Settler Leak-
Bottom
(F Stab.)
163
26,000
0.05
0.1
40
8.5
1
1.00
-5°
1000
20
1.5
10.0
50
0.03
F
Scenario 9:
Settler Leak-
Bottom
(D Stab.)
163
26,000
0.05
0.1
40
8.5
1
1.00
-5°
1000
20
5
10.0
50
0.03
D
Scenario
10: Settler
Leak-Bottom-
Mitigated
(D Stab.)
163
26,000
0.0158
0.1
40
8.5
1
1.00
0
1000
20
5
10.0
50
0.03
D
Scenario
13: Settler
Leak-Inlet
Pipe
(F Stab.)
24
20,000
0.0177
2.0
40
16
1
1.00
-5°
1000
20
1.5
10.0
50
0.03
F
Scenario
14: Settler
Leak-Inlet
Pipe (D
Stab.)
24
20,000
0.0177
2.0
40
16
1
1.00
-5°
1000
20
5
10.0
50
0.03
D
Scenario 1 5:
Pump Seal
Failure
(F Stab.)
24
18,600
0.01
0.0
40
7.5
1
1.00
-5°
1000
20
1.5
10.0
50
0.03
F
Scenario 16:
Pump Seal
Failure
(D Stab.)
24
18,600
0.01
0.0
40
7.5
1
1.00
-5°
1000
20
5
10.0
50
0.03
D
Note:  S = HFSPILL  P = HFPLUME  H = HEGADAS

-------
                                                   Exhibit XVIMb (continued)
                                     Inputs for HGSYSTEM Models: HEGADAS-S and HEGADAS-T
Input
Cloud Data Output Code
Surface Transfer Code
Cloud Meander Averaging Time(sec)
Crosswind Dispersion Coefficient
GAS PROPS.: Mol. Weight (g/mol)
Specific Heat (J/mol-°C)
Fraction water pick-up
Thermodynamic Model
Post-Flash Temperature (°C)
Cloud Output Step Length (m)
Last Downwind Distance (m)
Cone, to Stop Calculations (kg/m3)
Upper, Lower Concentrations (kg/m3)
TSTAR observer (minimum) (sec)
TSTAR observer (maximum) (sec)
Model
Prog.
T.SS
T,SS
T.SS
T,SS
T,SS
T.SS
T,SS
T.SS
T,SS
T,SS
SS
T,SS
T,SS
T
T
Scenario 8:
Settler
Leak-
Bottom
(F Stab.)
0
3
541
2
20.0
29.1
0.0
2
19.5
100
N/A
5E-6
5E-5, 5E-6
5000
160,000
Scenario 9:
Settler
Leak-
Bottom
(D Stab.)
0
3
541
2
20.0
29.1
0.0
2
19.5
100
N/A
5E-6
5E-5, 5E-6
200
10,000
Scenario 1 0:
Settler Leak-
Bottom-
Mitigated
(D Stab.)
0
3
541
2
20.0
29.1
0.0
2
19.5
100
100,000
5E-6
5E-5, 5E-6
N/A
N/A
Scenario 13:
Settler Leak
Inlet Pipe
(F Stab.)
0
3
1200
2
20.0
29.1
0.0
2
19.5
100
N/A
5E-6
5E-5, 5E-6
7000
17,000
Scenario 14:
Settler Leak
Inlet Pipe
(D Stab.)
0
3
1200
2
20.0
29.1
0.0
2
19.5
100
N/A
5E-6
5E-5, 5E-6
500
2500
Scenario
15: Pump
Seal
Failure
(F Stab.)
0
3
1200
2
20.0
29.1
0.0
2
19.5
100
N/A
5E-6
5E-6, 5E-5
3500
14000
Scenario
16: Pump
Seal
Failure
(D Stab.)
0
3
1200
2
20.0
29.1
0.0
2
19.5
100
1.0E+5
5E-6
5E-5, 5E-6
N/A
N/A
Note: T = HEGADAS-T (Transient)   SS = HEGADAS-S (Steady State)

-------
                                                     Exhibit XVIMc
                                Inputs for HGSYSTEM Models: HFSPILL and HFPLUME or EVAP-HF
Input
Total Storage Capacity (m3)
Liquid Mass (kg)
Effective Release Diam. (m)
Water Mass Fraction (%)
Reservoir Temperature (°C)
Reservoir Pressure (atm)
Ambient Pressure (atm)
Orifice Height (m)
Release Discharge Angle
Last Downwind Displacement (m)
Ambient Temperature (°C)
Wind Speed (m/s)
Wind Speed Reference Height (m)
Relative Humidity (%)
Surface Roughness (m)
Pasquill Stability Level
Model Type
Formula
Spilled Liquid Temperature (°C)
Minimum Pool Height (m)
Timestep, Maximum Time (sec)
Min Evap.Flux, Min Evap.Rate(kg/m2-s,kg/s)
Model
Prog.
S
S
S,P
S
S,P
S,P
S,P
p
p
p
P.E.H
P,E,H
P,H
P,H
P,H
P,H
E
E
E
E
E
E
Scenario 1 : Vessel
Rupture-Bulk Storage
(D Stab.)
2023
1,858,000
5
0.0
27
1.5
1
N/A
N/A
N/A
20
5
10.0
60
0.03
D
3
1
20
0.002
60, 36000
0.01,0.1
Scenario 3:
Derailment
70% HF (D Stab.)
79.6
65,000
0.15
30
27
5
1
N/A
N/A
N/A
20
5
10.0
50
0.03
D
3
1
27
0.001
50, 3600
0.01, 0.1
Scenario 1 1 : Vessel
Leak
(D Stab.)
20.2
18,600
0.06
0.0
27.0
1.5
1
10
-5°
1000
20
5
5.0
50
0.03
D
N/A
1
N/A
N/A
N/A
N/A
Scenario 12: Vessel
Leak-Mitigated
(D Stab.)
20.2
18,600
0.06
0.0
27
1.5
1
10
-5°
1000
20
- 5 •;'
5.0
50
0.03
D
N/A
1
N/A
N/A
N/A
N/A
Note:  S = HFSPILL  P = HFPLUME  E = EVAP  H = HEGADAS

-------
                                              Exhibit XVII-1c (continued)
                               Inputs for HGSYSTEM Models:  HEGADAS-S and HEGADAS-T
Input
Cloud Data Output Code
Surface Transfer Code
Ambient Temperature Height (m)
Cloud Meander Averaging Time (sec)
Crosswind Dispersion Coefficient
GAS PROPS.: Molecular Weight (g/mol)
Specific Heat (J/mol-°C)
Fraction water pick-up
Thermodynamic Model
Post-Flash Temperature (°C)
Cloud Output Step Length (m)
Last Downwind Distance (m)
Cone, to Stop Calculations (kg/m3)
Upper, Lower Concentrations (kg/m3)
TSTAR observer (minimum) (sec)
TSTAR observer (maximum) (sec)
Model
Prog.
T.SS
T,SS
T,SS
T,SS
T,SS
T,SS
T,SS
T,SS
T,SS
T,SS
T,SS
SS
T,SS
T,SS
T
T
Scenario 1 : Vessel
Rupture-Bulk Storage
(D Stab.)
0
3
N/A
600
2
20.01
29.0
0.0
2
N/A
200
N/A
5E-6
5E-5, 5E-6
5000
25000
Scenario 3:
Derailment 70% HF
(D Stab.)
0
3
N/A
840
2
20.01
29.0
0.0
2
N/A
50
N/A
5E-6
5E-5, 5E-6
400
900
Scenario 11:
Vessel Leak
(D Stab.)
0
3
2.20
1200
2
20.0
29.1
0.0
2
19.5
100
101,000
5E-6
5E-5, 5E-6
N/A
N/A
Scenario 12: Vessel
Leak-Mitigated
(D Stab.)
0
3
2.20
180
2
20.0
29.1
0.0
2
19.5
100
N/A
5E-6
5E-5, 5E-6
100
4000
Note: T = HEGADAS-T (Transient)  SS = HEGADAS-S (Steady State)

-------
                                             Exhibit XVII-2
                                         Inputs for SLAB Models
Input
Spill Source Type (see Note below)
Numerical Substep
Molecular Weight (kg/mol)
Vapor Heat Capacity (constpr.) (j/kg-K)
Boiling Point Temperature (°C)
Liquid Mass Fraction
Heat of Vaporization Q/kg)
Liquid Heat Capacity (j/kg-K)
Liquid Source Density (kg/m3)
Saturation Pressure Constant, spb
Saturation Pressure Constant, sbc
Temperature of Source Gas (°C)
Mass Source Rate (kg/s)
Source Area (m )
Continuous Source Duration (sec)
Instantaneous Source Mass (kg)
Source Height (m)
Concentration Averaging Time (sec)
Maximum Downwind Distance (m)
Cone. Measurement Height, zp(i), i=1,4 (m)
Surface Roughness Height (m)
Ambient Measurement Height (m)
Ambient Wind Speed (m/s)
Ambient Temperature (°C)
Relative Humidity (%)
Atmospheric Stability Class Value
Scenario 1 :
Vessel
Rupture-Bulk
Storage
(D Stab.)
1
1
0.020006
1450
292.67
0.0
373200
2528
957
-1
0
292.67
720
235900
1200
0
0
1800
35000
0
0.03
10.0
5.0
293.0
50
4
Scenario 2:
Derailment
Empty Rail
Car
(D Stab.)
2
1
0.020006
1450
292.67
0.94
373200
2528
957
3404.51
15.06
292.67
77.4
0.0177
841
0
3
1800
40000
0
0.03
10.0
5.0
292.8
50
4
Scenario 3:
Derailment
70% HF
(D Stab.)
1
1
0.020006
1450
292.67
0.0
373200
2528
957
-1
0
292.67
2.5
2827
1200
0
0
1800
35000
0
0.03
10.0
5.0
292.8
50
4
Note:  1 = Evaporating Pool   2 = Horizontal Jet Release  3 = Vertical Jet Release  4 = Instantaneous Release

-------
                                                Exhibit XVII-2
                                           Inputs for SLAB Models
                                                 (continued)
Input
Spill Source Type (see Note below)
Numerical Substep
Molecular Weight (kg/mol)
Vapor Heat Capacity (const, pr.) (j/kg-K)
Boiling Point Temperature (°K)
Liquid Mass Fraction
Heat of Vaporization (j/kg)
Liquid Heat Capacity (j/kg-K)
Liquid Source Density (kg/m3)
Saturation Pressure Constant, spb
Saturation Pressure Constant, sbc
Temperature of Source Gas (°K)
Mass Source Rate (kg/s)
Source Area (m2)
Continuous Source Duration (sec)
Instantaneous Source Mass (kg)
Source Height (m)
Concentration Averaging Time (sec)
Maximum Downwind Distance (m)
Cone. Measurement Height, zp(i), i=1,4 (m)
Surface Roughness Height (m)
Ambient Measurement Height (m)
Ambient Wind Speed (m/s)
Ambient Temperature (°K)
Relative Humidity (%)
Atmospheric Stability Class Value
Scenario
4: Hose
Failure
Empty
(F Stab.)
2
1
0.020006
1450
292.67
0.918
373200
2528
957
3404.51
15.06
292.67
43.3
0.002
432
0
1
1800
37000
0
0.03
10.0
1.5
292.8
50
6
Scenario
5: Hose
Failure
Mitigated
(F Stab.)
2
1
0.020006
1450
292.67
0.918
373200
2528
957
3404.51
15.06
292.67
43.3
0.002
60
0
1
1800
30000
0
0.03
10.0
1.5
292.8
50
6
Scenario
6: Hose
Failure
Empty
(D Stab.)
2
1
0.020006
1450
292.67
0.918
373200
2528
957
3404.51
15.06
292.67
43.3
0.002
432
0
1
1800
30000
0
0.03
10.0
5.0
292.8
50
4
Scenario
7: Hose
Failure
Mitigated
(D Stab.)
2
1
0.020006
1450
292.67
0.918
373200
2528
957
3404.51
15.06
292.67
43.3
0.002
60
0
1
1800
35000
0
0.03
10.0
5.0
292.8
50
4
Note: 1 = Evaporating Pool   2 = Horizontal Jet Release   3 = Vertical Jet Release   4 = Instantaneous Release

-------
                                              Exhibit XVII-2
                                         Inputs for SLAB Models
                                               (continued)
Input
Spill Source Type (see Note below)
Numerical Substep
Molecular Weight (g/mol)
Vapor Heat Capacity (const, pr.) (j/kg-K)
Boiling Point Temperature (°C)
Liquid Mass Fraction
Heat of Vaporization (j/kg)
Liquid Heat Capacity (j/kg-K)
Liquid Source Density (kg/m3)
Saturation Pressure Constant, spb
Saturation Pressure Constant, sbc
Temperature of Source Gas (°C)
Mass Source Rate (kg/s)
Source Area (m2)
Continuous Source Duration (sec)
Instantaneous Source Mass (kg)
Source Height (m)
Concentration Averaging Time (sec)
Maximum Downwind Distance (m)
Cone. Measurement Height, zp(i), i=1,4jn)
Surface Roughness Height (m)
Ambient Measurement Height (m)
Ambient Wind Speed (m/s)
Ambient Temperature (°C)
Relative Humidity (%)
Atmospheric Stability Class Value
Scenario
8: Settler
Leak-
Bottom
(F Stab.)
2
1
0.020006
1450
292.67
0.84
373200
2528
957
3404.51
15.06
292.67
49.5
0.002
541
0
1
1800
40000
0
0.03
10.0
1.5
292.8
50
6
Scenario
9: Settler
Leak-
Bottom
(D Stab.)
2
1
0.020006
1450
292.67
0.84
373200
2528
957
3404.51
15.06
292.67
49.5
0.002
541
0
1
1800
40000
0
0.03
10.0
5.0
292.8
50
4
Scenario
10: Settler
Leak-Bottom
Mitigated
(D Stab.)
2
1
0.020006
1450
292.67
0.84
373200
2528
957
3404.51
15.06
292.67
4.95
0.002
541
0
1
1800
40000
0
0.03
10.0
5.0
292.8
50
4
Scenario
1 1 : Vessel
Leak
(D Stab.)
2
1
0.020006
1450
292.67
0.94
373200
2528
957
3404.51
15.06
292.67
12.40
0.0028
1200
0
10
1800
40000
0
0.03
10.0
5.0
292.8
50
4
Scenario
12: Vessel
Leak
Mitigated
(D Stab.)
2
1
0.020006
1450
292.67
0.94
373200
2528
957
3404.51
15.06
292.67
12.40
0.0028
180
0
10
1800
40000
0
0.03
10.0
5.0
292,8
50
4
Note:  1 = Evaporating Pool   2 = Horizontal Jet Release  3 = Vertical Jet Release   4 = Instantaneous Release

-------
                                               Exhibit XVII-2
                                          Inputs for SLAB Models
                                                (continued)
Input
Spill Source Type (see Note below)
Numerical Substep
Molecular Weight (kg/mol)
Vapor Heat Capacity (const.pr.) (j/kg-K)
Boiling Point Temperature (°K)
Liquid Mass Fraction
Heat of Vaporization Q'/kg)
Liquid Heat Capacity G/kg-K)
Liquid Source Density (kg/m3)
Saturation Pressure Constant, spb
Saturation Pressure Constant, sbc
Temperature of Source Gas (°K)
Mass Source Rate (kg/s)
Source Area (m2)
Continuous Source Duration (sec)
Instantaneous Source Mass (kg)
Source Height (m)
Concentration Averaging Time (sec)
Maximum Downwind Distance (m)
Cone. Measurement Height, zp(i), i=1,4 (m)
Surface Roughness Height (m)
Ambient Measurement Height (m)
Ambient Wind Speed (m/s)
Ambient Temperature (°K)
Relative Humidity (%)
Atmospheric Stability Class Value
Scenario
13: Settler
Leak-Inlet
Pipe
(F Stab.)
2
1
0.020006
1450
292.67
0.84
373200
2528
957
3404.51
15.06
292.67
9.56
0.0025
1200
0
1
1800
40000
0
0.03
10.0
1.5
292.8
50
6
Scenario
14: Settler
Leak-Inlet
Pipe
(D Stab.)
2
1
0.020006
1450
292.67
0.84
373200
2528
957
3404.51
15.06
292.67
9.56
0.0025
1200
0
1
1800
40000
0
0.03
10.0
5.0
292.8
50
4
Scenario
15: Pump
Seal
Failure
(F Stab.)
2
1
0.020006
1450
292.67
0.84
373200
2528
957
3404.51
15.06
292.67
1.8
8E-5
1200
0
1
1800
30000
0
0.03
10.0
1.5
292.8
50
6
Scenario
16: Pump
Seal
Failure
(D Stab.)
2
1
0.020006
1450
292.67
0.84
373200
2528
957
3404.51
15.06
292.67
1.8
8E-5
1200
0
1
1800
30000
0
0.03
10.0
5.0
292.8
50
4
Note: 1 = Evaporating Pool   2 = Horizontal Jet Release   3 = Vertical Jet Release   4 = Instantaneous Release

-------
                                     BIBLIOGRAPHY
       In gathering data and information for this report, an extensive library of references and source
material was compiled.  Although many of these sources were not cited directly in the report, they are
referenced in this bibliography for anyone seeking additional information on various topics.  In an
attempt to group the sources by similar topics, they are presented by chapter in the HF report. Within
these chapter headings, the references are arranged in alphabetical order.
                                                                                    Page B-1

-------
Chapter 1

ICF Inc., Preliminary Report: Background Information on Hydrogen Fluoride, Fairfax, VA,
       July, 24, 1991. (380)

National Petroleum Refiners Association, "EPA Launches HF Study," Washington Bulletin, Washington,
       D.C., October 25, 1991. (139.505)
Page B-2

-------
Chapter 2

Allied-Signal, Recommended Medical Treatment for Hydrofluoric Acid Exposure, The HF Products
       Group, Morristown, NJ, 1991. (422.5)

Beal, John W., "Wind-Blown Acid Mist Blamed for Damaged Cars, Airplanes," Kansas City Star,
       March 1, 1976. (53)

Caldwell, Jean, "Fumes in Springfield Roust Thousands Again," The Boston Globe, June 19, p 1. (77)

Canadian Centre for Occupational Health and Safety (CCOHS), Hydrogen Fluoride, Hamilton, Ontario,
       April 30, 1985. (265)

Citizens Clearinghouse for Hazardous Wastes, Inc., Environmental Health Monthly, Volume 4, Number
       5, Falls Church, VA, February 1992. (139.515)

Comite Technique Europeen Du Fluor, Medical  Sub Group: A) Advice for Hydrofluoric Acid
       Intoxication, Document Number: STS 89/70,  B) Recommandations en Cas d'Accident a I'Acide
       Fluorhydrique, Document Number: STS  89/69. (130.15 A,B)

Du Pont Chemicals Literature Search on Hydrogen Fluoride, November 10, 1989. (138.5)

Dayal, Hari K, et al, "A Community-Based Epidemiologic Study of Health Sequelae of Exposure to
       Hydrofluoric Acid," Annals of Epidemiology, Volume 2, Number 3, May 1992, pp 213-230.
       (134.8)

Hydrogen Fluoride Task Force of the South Coast Air Quality Management District, Health Effects Due
       to Hydrogen Fluoride Inhalation: A Literature Review, March 1989. (148)

Kaiser, G.D., Science Applications International  Corporation, ancl Adam Hasen, Hydrogen Fluoride
       Storage and Use, (comments on various materials that were distributed by the South Coast Air
       Quality Management District while it was preparing Rule 1410),  March 1991. (130.2)

Energy Safety Council, "Lethal Chemical Threat to Millions of Americans Exposed by New Organization
       Focusing on Energy Hazards," Energy Safety News, January,  1992. (293.2)

Lifelines: OCAW Health and Safety News, "OCAW Debates Hydrogen Fluoride Safety," May-June 1989,
       p 15.  (367)

McFarland, Mack, Investigations of the Environmental Acceptability of Fluorocarbon Alternatives to
       Chlorofluorocarbons," Proceedings of the National Academy of Sciences, Volume 89, pp 807-
       811, February 1992. (328.649)

Millar, Fred, Press Release: Tens of Thousands At Peril From Hydrogen  Fluoride at Texas City
       Refineries; New Research Data not Shared with Populations At Risk, Environmental Policy
       Institute, November 19, 1987. (469)

Moran, Kevin, "Hospital Staff Rallied to Code Yellow," Houston Chronicle, November 1987.(342.5)

Public Inquiry Into the Proposed Northwick Village Development, Applications byAldersgate
       Developments Ltd to Castle Point District Council, Proof of Evidence Submitted by the Health
       and Safety Executive, October 1990.  (374)
                                                                                   Page B-3

-------
Sax, N. Irving and Richard J. Lewis, Sr., Dangerous Properties of Industrial Materials, Volume 3, Van
       Nostrand Reinhold, New York, pp 1906-1907. (427)

Stein, George, "Assembly Targets South Bay in Study of Chemical Risks," Los Angeles Times March
       17, 1989. (463.2)

Trevino, MA, G.H. Herrmann, and William L Sprout, Treatment of Severe Hydrofluoric Acid
       Exposures," Journal of Occupational Medicine, Volume 25, Number 12, December 1983,
       p 861-863. (485.5)

Turner, R.M., and S. Fairhurst, Assessment of the Toxicity of Major Hazard Substances, England.
       (489.6)

U.S. Department of Health and Human Services, NIOSH Pocket Guide to Chemical Hazards, Centers
       for Disease Control, National Institute for Occupational Safety and Health, Washington, D.C.,
       June 1990. (362)

U.S. Environmental Protection Agency, Integrated Risk Information System (IRIS), Fluorine,
       August 1,  1989. (283)

U.S. Environmental Protection Agency, Summary Review of Health Effects Associated with Hydrogen
       Fluoride and Related Compounds, Health Issue Assessment, Document Number: EPA/600/8-
       89/002F, December 1988. (467.5)

U.S. Nuclear Regulatory Commission, Assessment of the Public Health Impact from the Accidental
       Release of UF6 at the Sequoyah Fuels Corporation Facility at Gore, Oklahoma (Appendices),
       Washington, D.C., March 1986.  (492)

Valentine, Rudolph, Acute Inhalation Toxicity of Hydrogen Fluoride in Rats, Du Pont, Newark, DE,
       August 14, 1990, Document Number: HLR 365-90. (13)

Weiss, G., ed., Hazardous Chemicals Data Book, Noyes Data Corporation, Park Ridge, N.J., 1985,
       pp 568,  573. (505)

Wing, Jessie, S., Lee M. Sanderson, Jean D. Brender, Dennis M. Perrotta, Richard A. Beauchamp,
       "Acute Health Effects in a Community After a Release of Hydrofluoric Acid," Archives of
       Environmental Health, Volume 46, Number 3, May/June 1991. (507)
Page B-4

-------
Chapter 3

Allied-Signal, Inc., North American HF Industry Overview, The HF-Products Group, Morristown, NJ.
        (363)

Aviation Week and Space Technology, "Alpha Missile Defense Laser Is Fired for the First Time,"
        April 17, 1989, pp 23-24.  (10.42)

Bernhardt, Steven H., HF Value to Society, Fluorocarbons Manufacture, Du Pont Chemicals,
        Novembers, 1991. (55)

Chemical and Engineering News, "Production by the U.S. Chemical Industry, Little Change in Chemical
        Output Last Year," June 24, 1991, pp 30-81. (372.7)

Chemical and Engineering News, "Gasoline Alkylation Process to be Commercialized,"
        February 10, 1992, p 24.  (141)

Chemical Engineering Progress, "CFC Phase-Out to Benefit Hydrofluoric Acid Makers," June 1990, p 8.
        (80)

Chemical Marketing Reporter, "CFC Substitute Race; ICI Announces Plants," November 28, 1988, p 3.
        (90)

Chemical Marketing Reporter, July 30, 1990, p 37. (105)

Chemical Marketing Reporter, "Hydrofluoric Acid Demand Will Increase to 480,000 Tons From 375,000
        Tons in 1972," February 5, 1973,  p 91.  (195)

Chemical Marketing Reporter, "Hydrogen Fluoride Looks in Good Shape for '88; a Maybe in '89,"
        March 7, 1988, p 7. (230)

Cunningham, Larry D., Columbium (Niobium) and Tantalum, U.S. Department of the Interior, Bureau of
        Mines, Annual Report, 1990. (294.3b)

Environmental Policy Institute, Districts in Which Refining Facilities  Utilizing Hydrofluoric Acid are
        Located and U.S. Refineries Using HF, Washington, D.C. (135.5)

General Chemical Corporation, Bay Point Plant, Brochure, Suisun Bay, PA, April, 1992. (141.56)

J.J. Kelley & Associates, Hazardous Materials Guide, Shipping, Materials Handling and Transportation,
        Inc., Neenan, Wl, 1982. (147.7)

Lawler, Gloria M., ed., Chemical Origins and Markets, Fifth  Edition, Chemical Information Services,
        Menlo Park, CA, 1977. (292)

Morse, David E., Minerals Yearbook - Fluorspar, U.S. Department of the Interior, Bureau of Mines,
        Washington, D.C., 1988.  (345)

National Fire Protection Association, NFPA 49:  Hazardous  Chemicals Data, 1984.  (361)

National Library of Medicine, Hazardous Substances Data Bank, (On-Line Database), "Hydrofluoric
        Acid," July 15, 1992. (361.3)
                                                                                     Page B-5

-------
 •Preparation of Hydrogen Fluoride," U.S. Patent No. 3,102,787.

 South Coast Air Quality Management District, A) Hydrofluoric Acid Inventory in the District,
        B) Businesses that Use Hydrofluoric Acid or Hydrogen Fluoride, C) HF Inspection Survey,
        D) Phone numbers of Semiconductor Producers/Distributors, E) Semiconductor Fabricators in
        SCAQMD, F) Mailing List for Proposed Rule 1164, G) Mailing List for Proposed Rule 1410.
        (443 A-G)

 South Coast Air Quality Management District Board, Recommendations on the Potential Hazards of
        Hydrogen Fluoride Transportation, Storage, and Use, El Monte, CA, April 4, 1990. (442.5)

 "Survey of Operating Refineries in the U.S. (State Capacities as of January, 1, 1988)," Oil and Gas
        Journal, March 21, 1988. (468)

 Thrash, Lou Ann, "Worldwide Refining," Oil and Gas Journal, December 31, 1990, p 84. (473)

 Will, Ray, R. Willhalm, and S. Mori, SRI International, Chemical Economics Handbook Product Review:
        Fluorine Compounds, Preliminary Draft,  Menlo Park, CA, May 20, 1992.  (102)
Page 8-6

-------
Chapter 4

Ale, B.J.M., The Implementation of an External Safety Policy in the Netherlands, Ministry of Housing,
       Physical Planning, and Environment, the Netherlands. (10.15)

Amendola, A. and S. Contini, Community Documentation Center on Industrial Risk, National
       Approaches to the Safety Report a Comparison, Joint Research Centre, Commission of the
       European Communities, January 1991. (10.45)

California State Assembly, "Hydrogen Fluoride Chemical Plants," Chapter 6.3,  Division 20 of the Health
       and Safety Code, Sacramento, CA, February 12, 1990. (10.7)

The Chemical Engineer, "HF Guidelines Change," June 25, 1992, p 7. (76.3)

(UNEP) IRPTC PC database, Data on International Regulations on HF, Document Number: 17.01 Legal,
       1965-1990, pp 1-23. (133.5)

Contra Costa County Health Services Department, Hazardous Materials Release Response and
       Inventory Law, Business Plan 3/92, General Chemical Corporation, Pittsburg, CA. (141.50)

Gladstone, Mark, "Bill to Restrict Refinery Chemical Shelved," Los Angeles Times, April 4, 1990, p  B8.
       (142)

Gladstone, Mark, "Tucker's Bill on Hazardous Acid Dies in Committee," Los Angeles Times,
       Aprils, 1990. (143)

Hatch, George, "Foes in Acid  Issue Trade Jabs, File Arguments," Los Angeles Times,
       December 29, 1989, p B3. (147.3)

"Hazardous Materials Regulations," 49 CFR 171-180. (156.5)

Martin, Hugo, "Torrance Council Defends Its Methods in Battle with Mobil Refinery," Los Angeles
       Times, April  14, 1989, p 8. (327)

McGuire, S.A., A Regulatory Analysis on Emergency Preparedness of Fuel Cycle and Other Radioactive
       Material Licensees, Draft Report for Comment, U.S. Nuclear Regulatory Commission,
       June 1985. (329)

Mitchell, Jeff, "Tucker Bid to Ban HF Acid Killed in State Legislature," Easy Reader, April 1991.  (333)

"Mobil Settles in California on Hydrofluoric Acid Use," Wall Street Journal, October 1990. (337)

New Jersey Department of Environmental Protection and Energy, Cleanup Standards for Contaminated
       Sites, Site Remediation Program, Document Number (Proposed New Rule): N.J.A.C. 7:26D,
       June 2, 1991. (570)

Promoting Safer Highway Routing of Ultrahazardous Cargoes: DOT Oversight, Hearings Before a
       Subcommittee on Government Operations House of Representatives, One Hundredth
       Congress, October 14 and 19, 1987. (373)

Public Law 101-549, Clean Air Act Amendments of 1990, November 15, 1990, p 104 STAT.2560. (375)
                                                                                   Page B-7

-------
 Rabin, Jeffrey L, "Drive to Ban Mobil Acid Use Launched," Los Angeles Times, April 6, 1989  p 8
        (421.1)

 Rabin, Jeffrey L, "Councilman's Petition Drive to Target Acid at Oil Refinery," Los Angeles Times,
        December 15, 1986, p 12. (420.7)

 Rabin, Jeffrey L, "Mobil Lobbies More as Pressure Builds in Torrance," Los Angeles Times,
        March 24, 1989, p 8. (421.2)

 Radcliffe, Jim, "AQMD Staff May Urge Ban of Deadly Chemical," The Daily Breeze, Torrance, CA,
        January 12, 1990. (421.9)

 Radclrffe, Jim, "Report on Chemical Draws Torrance Ire," The Daily Breeze, Torrance, CA,
        January 10, 1990. (421.95)

 Robak, Warren, "Chemical Ban Gets AQMD Nod," The Daily Breeze, Torrance, CA, April 7, 1990.
        (424.4)

 Robak, Warren, "House Panel Accepts Toxics Amendment," The Daily Breeze, Torrance, CA, p A3.
        (424.55)

 Robak, Warren, "Proposal Extends Refinery Acid Use," The Daily Breeze, Torrance, CA,
        February 27,1991.  (424.6)

 Robak, Warren, "Study Tells Plant How to Cut Risk," The Daily Breeze, Torrance, CA,
        January 12, 1990., p. A1. (424.65)

 South Coast Management District, Hydrogen Fluoride Storage and Use, Rule 1410, CA, April 5 1991
        (250)

 Stein, George, "600 Neighbors of Oil Refinery in Torrance File Suit Against Mobil," Los Angeles Times,
        February 14,1990, p B3. (463.1)

 Stein, George, "AQMD Votes to Ban Hydrofluoric Acid," Los Angeles Times, April 7, 1990. (463.15)

 Sternberg, Ken, "Use of Hydrofluoric Acid Comes Under the Gun in California," Chemical Week,
       October 31, 1990. (465)

 Sweeney, James  P., "Panel  Kills Chemical Bill, Hydrogen Fluoride Was Targeted," The Daily Breeze,
       Torrance, CA, p A3. (468.5)

 Ultramar Refining, Regulatory Use of Risk-Based Standards and Techniques,  Wilmington, CA,
       May 1991. Prepared by Science Applications International Corporation. (421)
Page B-8

-------
Chapter 5

American Institute of Chemical Engineers, Spring National Meeting, Houston, TX, April 11, 1991,
       "Alkylation's Role in Reformulated Gasoline," by Kenneth R. Masters, Stratco, Inc. (10.4)

"Annual Refining Survey," Oil & Gas Journal, March 18, 1991, pp 85-105. (30)

Encyclopedia of Science and Technology, Sixth Edition, Volume 7, McGraw Hill Book Company, New
       York, 1987, pp 580-583. (328.9)

Hammershalmb, H.U. and B.R. Shah, "Trends in HF Alkylation," Hydrocarbon Processing, June 1985,
       p 73-76. (147.25)

"HF Alkylation," Hydrocarbon Processing, September 1974, p 207. (149)

"HF Alkylation," Hydrocarbon Processing, September 1984, p 124. (149.1)

"HF Alkylation draws Safety and Process Questions," Oil and Gas Journal, June 24, 1991, p 48-54.
       (149.3)

Hydrocarbon Processing,  "Alkylation," September 1986, p 101. (10.3)

High-Octane Gasoline Requires High-Quality Alkylate, Panel Discussion April 30,  1990, pp 62-64. (158)

Lee, Patrick, "Are  Refineries Unsafe?," Los Angeles Times. (293)

Masters, Kenneth R., Alkylation's Contribution to Low RVP, Clear Pool Octane, Stratco, Inc.,  Leawood,
       KS, 1990. (328.6)

Mark, Herman F., Donald F. Othmer, Charles G. Overberger, and Glenn T, Seaborg, eds., Kirk-Othmer
       Encyclopedia of Chemical Technology, Third Edition, Volume 2, John Wiley and Sons, New
       York, 1978. (284.5)

Myers, Philip, Krishna Mudan, and Henry Hachmuth, The Risks of HF Sulfuric Acid Alkylation,  Journal
        article, Phillips Petroleum and Technica, 1990. (357)

"Refiners Discuss HF Alkylation Process and  Issues," Oil and Gas Journal, April 6, 1992. (421.97)

 Rogers, W.R. and Muller, K., "Hydrofluoric Acid Manufacture," Chemical Engineering Progress,
        Volume 59, Number 5, May 1963. (425)

Vista Chemical Company, Chloroparaffin vs. HF Processes for LAB, Houston, TX, April  1992. (497.7)

 Vora, Bipin V. and Charles P. Luebke, "Process Hydrogenates Unwanted Diolefins and Acetylenes," Oil
        and Gas Journal, December 5, 1988, p 40. (499)
                                                                                     Page B-9

-------
  Chapter 6

  American Petroleum Institute, "Safe Operation of Hydrofluoric Acid Alkylation Units," API
         Recommended Practice 751, First Edition, June 1992, Washington, D.C. 20005. (10.6)

  American Petroleum Institute, Process Hazards Management, Washington, D.C. (19.97)

  American Petroleum Institute, "Management of Process Hazards," API Recommended Practice 750,
         First Edition, January 1990, Washington, D.C. (19.99)

  BP CIRRUS, An Integrated Consequence Modelling Package, User Guide, Corporate Safety Services,
         BP International Limited, January 1991. (51.5)

  BP Alliance Refinery Safe Practices, Procedure No. 23, "Handling of Hydrofluoric Acid (HF),"
         February 13, 1986, and Procedure No. 11, "Minimum Safe Dress," December 17, 1991. (73)

  BP, The Use and Value of Quantified Risk Assessment Within the BP Group, Corporate Safety Services
         Document (GSC Report) Number: 04-90-0005,  March 1990. (497.5)

  BP International, Ltd., "The Use and Value of Quantified Risk Assessment Within the BP Group,"
         Corporate Safety Services Guidance Note, London, September 1991. (497.6)

  BP International, Ltd., "Concept Safety Evaluations," Corporate Safety Services Guidance Note
        London, September 1991. (130.3)

  BP International, Ltd., Consequence Modelling in Design, Emergency Planning, and Quantified Risk
        Assessment, Corporate Safety Services, London, September 1991. (130.5)

  Comrte Technique Europeen Du Fluor (European Fluorine Technical Committee), Working Group
        Storage, Transport, Safety, Twenty-three Assorted Recommended Practices for Handling and
        Transporting HF Safely, CEFIC, CTEF-250, Brussels. (130.16A-X)

 The Development of Acceptability Criteria for the BP Group,: A Corporate Services Safety Discussion
        Document, July 1990. (135.3)

 Du Pont, Off-Site Consequence Analysis, Antioch Works, Contra Costa County, CA
        September 30, 1991.  (11)

 Environmental Policy Institute, Statement by Dr. Fred Millar Before the California Assembly Committee
        on Safety in the  Workplace, Carson, CA, October 19, 1988. (463)

 Hatch, George, and George Stein, "Ban of Toxic Chemical Rejected in Draft Report; Safely: Despite
        Significant Public Health Risks, Draft Report To AQMD Calls for Improved Safeguards at Plants
        Rather Than Banning Use of Hydrogen Fluoride," Los Angeles Times, December 29 1989
        p.B3. (147.5)

 Kansas City Star, "Refinery  Replaces Pipe Fittings To Guard Against Acid Leaks," March 3, 1976. (422)

 Mobil Oil Corporation Torrance Refinery, Hydrofluoric Acid Alkylation Unit Risk Management and
       Prevention Program Report, Torrance, CA, April 1991. (187)
Page B-10

-------
Official Journal of the European Communities, "Council Directive of 24 June 1982 on the Major-
       Accident Hazards of Certain Industrial Activities (82/501 /EEC)," Volume 25,  Number L230,
       August 5, 1982, pp 1-18. (530a)

Powerine Oil Company, Hydrogen Fluoride Alkylation Unit, Santa Fe Springs Refinery, Risk Management
       and Prevention Program, Santa Fe Springs, CA, May 11, 1990. Prepared by Radian
       Corporation. (424.31)

Rabin, Jeffrey L, "Mobil Sees Little Risk in Acid Use at Refinery," Los Angeles Times, March 17, 1989,
       p8.  (421.3)

Rabin, Jeffrey L., "Torrance Rejects Consultants' Report on Mobil as 'Flawed'," Los  Angeles Times,
       January 19,  1989, p 8.  (421.5)

Risk Management Plans, HazOp, Safety Evaluations, etc., for various anonymous
       HF Refining/Alkylation facilities. (424.34D-J)

Robak, Warren, 'AQMD Says Acid Storage Requires New Procedures," The Daily Breeze, Torrance,
      ' CA,  December 28, 1989, p A1. (424.35)

Robak, Warren, "Expert Tells of Chemical Plant Hazard," The Daily Breeze, Torrance, CA,
      ' March 22, 1988, p A3.  (424.5)

The 6th International Symposium, Loss Prevention and Safety Promotion in the Process Industries,
       Oslo, June 1989, "The  Use of Cost Benefit Analysis in Evaluating the Acceptability of Industrial
        Risks:  An Illustrative Case Study," Fleishman, A.B., and M.S. Hogh. (139.95)

Stein, George and Hugo Martin, "Safety Report Criticizes Mobil Oil Co. Refinery," Los Angeles Times,
        January 1, 1989. (464.5)

 U.S. Environmental Protection  Agency, Technical Guidance for Hazards Analysis, Emergency Planning
        for Extremely Hazardous Substances, December 1987. (489.95)

 U.S. Environmental Protection Agency, Chemical Safety Audit Profile, British Petroleum Company,
        Ferndale, WA, March 19-23, 1990. (116.2)

 U.S. Environmental Protection Agency, Chemical Safety Audit Profile, Phillips 66 Company Woods
        Cross Petroleum, West Bountiful, UT, May 2-4, 1989. (116.4)

 U.S. Environmental Protection Agency, Chemical Safety Audit Profile, Ultramar, Incorporated,
        Wilmington, CA, April 17,  1990. (116.6)

 Ultramar, Alky Unit and HF Acid General Safety, Chemical Handling, October 15, 1991. (10.2)

 Ultramar Inc., Quantitative Risk Assessment of the HF Alkylation Unit, Executive Summary, Wilmington,
        CA, November 1991. (417)

 Ultramar Refining, HF Alkylation Unit Risk Management and Prevention Program and Related Studies
        and Activities, Presentation to the U.S. Environmental Protection Agency, December 16, 1991.
        Prepared by Science Applications International Corporation. (149.5)
                                                                                    PageB-11

-------
  Chapter 7


  Blewitt, Doug, N., "Design of Water Spray Mitigation Systems for Amoco HF Alkylation Units." (56)

  Blewitt, D.N., J.F. Yohn, R.P. Koopman, T.C. Brown, W.J. Hague, Effectiveness of Water Sprays on
         Mitigating Anhydrous Hydrofluoric Acid Releases, International Conference on Cloud Modeling.



  Du Pont Chemicals, EPA - Du Pont Meeting, HF Clean Air Act Study. Includes agenda, notes on Du
         Font's HF Philosophy, notes on Process Hazards Management System, Advanced Modeling of
         Accidental HF Release Scenarios, Equipment failure at Quimica Fluor - Matamoros Plant HF
         Safety. (137)


  Fthenakis, Vasilis M., The Feasibility of Controlling Unconfined Releases of Toxic Gases By Liquid
         Spraying," Chemical Engineering Comm., Volume 83, 1989, pp 173-189. (139.64)

  General Chemical Corporation, Bay Point Works - Pittsburg, California, Hazardous Materials Emerqencv
        Action Plan, April 16, 1992. (141.5A)


  General Chemical Corporation, Bay Point Off-Site Consequence Analysis, Engineering and Research
         Inc., pp 8-44. (141.5C)


  Oil and Gas Journal, "Studies Cover HF Spills and Mitigation," October 17,  1988, pp 58-62. (467)

  Prugh, Richard W., The Controlled Dispersion of Liquid Spill and Vapor Emission Incidents by Water
        Spray, Effectiveness  of Water Sprays, Hazard Reduction Engineering, Glen Mills PA
        June 12, 1987. (131)


 Robak, Warren, "Study: Water Best Weapon in Deadly Chemical Leak," The Daily Breeze, Torrance
        CA, August 21, 1989, p A1. (424.7)


 Schatz, K.W. and R.P. Koopman, "Water Spray Mitigation of Hydrofluoric Acid Releases," Journal of
        Loss Prevention in the Process Industries, Volume 3, April 1990, pp 222-243.  (436)

 South Coast Air Quality Management District, On-Line Leak Detectors LinkAQMD to Hydrogen
        Fluoride Sites, Diamond Bar, CA, February 4, 1992. (442)

 Stein, George, "Refineries Lack Means to Handle Corrosive Acid Leaks, Report Says" Los Anaeles
        Times, May 6, 1989.  (464)


 Ultramar, Multi-Agency Training of Response Personnel. Prepared by Kusz, Michael R. and James R
        Jeter, Union Pacific Resources Co., Wilmington, CA. (290.5)

 Wade, Robert C., HF Mitigation Strategies atAmoco's Texas City Alkylation Unit, Amoco Oil Companv
        Chicago, IL (499.5)


 Ward, John, SARA Title III Facility Emergency Response Handbook, BP Oil Company  Markus Hook
        Refinery, Marcus Hook, PA. (499.7)
Page B-12

-------
Chapter 8

"Acid Leak Under Lens at Phillips," The Kansan, June 7, 1978. (03)

Applebome, Peter, "Acid Leak in Texas Leaves a Residue of Questions," New York Times,
       December 20, 1987, p 30. (35)

ARIP Database Event Summary, Hydrogen Fluoride, May 22,1991. (40)

Bell, I.E., "Evacuees Stoic Over Bad News But Still Longing For Showers," Houston Chronicle. (54)

"Blast in Cincinnati Kills at Least 1, Injures Dozens," The Washington Post. (55.5)

Cassidy, Keith, "Hazard Control and Major Hazards," CIMAH Safety Reports, Institution of Chemical
       Engineers, Loughborough University of Technology, January 22-26, 1990. (79)

Community and Worker Right-To-Know-News, "Two Killed, F:ive Injured by HF in accident at Texas
       Refinery," Volume 5, Number 10, Washington, D.C., March 22, 1991. (489.8)

Coulter, Bill, "Disaster Prone Industry Brings Pain and Gain to Texas City," Houston Chronicle,
       November,  1987. (133.1)

The Daily Breeze, "Explosions, Fires Part of Mobil History," July 16, 1988. (139.53)

Daunt, Tina, "Officials Probe Chemical Leak That Injured 7 at Refinery," Los Angeles Times,
       February 11,  1990, p J1. (134)

Daunt, Tina, "Plant Acid Spill Safety Plan Delayed," Los Angeles Times, February 18, 1990. (134.5)

Dawson, Bill and T.E. Bell, "Some Texas City Evacuees Unable to Return Home," Houston Chronicle,
       November 1, 1987, p 10. (134.7)

Du Pont Chemicals, Du Font's La Pone,  Texas Plant's Disaster Scenarios, Air Permit Submitted to
       TACB-1991, Wilmington, DE,  1991. (139)

Energy Safety Council,  Energy Safety News, HAZMAT Conference '92 Update, "Bhopal in America? Will
       E.P.A Act to Prevent It?," Lincolnshire, IL, 60069. (139.05)

Environmental Protection Institute, Press Briefing Statement on Marathon Refinery Accident, Texas City,
       TX, November 16, 1987. (385)

Farrar, Foss, and Bob Friskel, "Blast at Phillips Leaves Smoke Pall," Kansas City Kansan,
       November 8, 1976. (139.55)

Fineman, Mark, "New Generation  Hit; Bhopal - Death Won't Leave City," Los Angeles Times,
       March 13, 1989, Part 1, p 1. (139.6)

HF Release Response Decision Tree, Phillips 66 Refinery.  (135.7)

Kansas City Times, "Oil  Blaze Probe to Start," March 29, 1973. (368)

Lanterman, Ken, "Federal Expert Says Texas City Was Close to Massive Tragedy, Thousands Could
       Have Been Killed," Houston Post, November 22, 1987. (291)
                                                                                   Page B-13

-------
 Marsh & McLennan, One Hundred Largest Losses, a Thirty-Year Review of Property Damage Losses In
        the Hydrocarbon-Chemical Industries, Eighth Edition, May 1985. (328.55)

 Mathews, Jay, "25,000 Evacuated After Gas Leak," The Washington Post, September 4, 1988, p A3.
        (328.63)

 Rice, Gary, and Joseph H. McCarty Jr., "Vapor Cloud Affects 38 at Plant," The Kansas City Times,
        June 6, 1978, p 14A. (424.05)

 Shearer, Cody, "Another Bhopal Accident Could Happen Here," Las Vegas Sun,  December 20, 1987.
        (437)

 Stewart, Don, and Sandi McDaniel, "Three U.S. Spills Show Acid Danger," The Tulsa Tribune,
        April 28, 1988. (466.5)

 Stewart, Don, and Sandi McDaniel, "Danger Adrift, Chaos Reigned After Sun Co. Spill," The Tulsa
        Tribune, April 28, 1988, p 1A. (466)

 Stein,  George, Torrance School Pupils, Faculty Felled by Fumes," Los Angeles Times, March 9, 1989
        p 1. (464.3)

 Toxic Chemical Release Inventory (TRI) 1989, Hydrogen Fluoride, November 5, 1991. (Also available on
        disk) (480)

 Toxic Chemical Release Inventory, (TRI) 1989, "Hydrogen Fluoride Releases in Ibs/year." (489.5)

 U.S. Department of Transportation, Hazardous Materials Information System (HMIS), January 1, 1980 -
        December 31, 1990. (155)

 U.S. Department of Transportation, Hazardous Materials Safety,  Hazardous Materials Information
       System, "Hydrogen Fluoride Incidents Since 1971 by Date," November 25, 1991. (157)

 U.S. Environmental Protection Agency, Hydrofluoric Acid Events in Acute Hazards Events Database,
       June 5, 1991. (200)

 U.S. Environmental Protection Agency, Acute Hazards Events Database, Listing of chemical release
       and number of evacuees, Document Name:  Mainw. (201)

 U.S. Environmental Protection Agency, ARIP Data on Hydrogen  Fluoride, Data on Releases from
       10/31/86 to 3/15/91, January 29,  1992. (42)

 U.S. Environmental Protection Agency, U.S. Department of Transportation, U.S. Coast Guard, National
       Response Center, HF Accident Data from ERNS Data Base 1987-1991, November 10, 1991
       (139.52)

 U.S. Nuclear Regulatory Commission, Release of UF6 from a Ruptured Model 48Y Cylinder at Sequoyah
       Fuels Corporation Facility: Lessons-Learned Report, Washington, D.C., June 1986, Document
       Number: NUREG-1198. (493)

 U.S. Nuclear Regulatory Commission, Release of UF6 From a Ruptured Model 48Y Cylinder at
       Sequoyah Fuels Corporation Facility: Lessons-Learned Report, "NRC Staff Responses to the
       Recommendations Made by the Lessons-Learned Group (Supplement Number 1),"
       Washington, D.C., August 1986. (494)
Page B-14

-------
Vobejda, Gloria, "Refinery Fumes Hospitalize Four," The Kansan, June 6, 1978. (498)

Ware, J.W., Lake Charles LAB Plant:  HF Release Scenarios, Vista Chemical Company, May 4, 1992.
       (497.9)

Wilson, David S., "20,000 Flee Fumes in East Los Angeles and 3 Suburbs," The New York Times,
       September 4, 1988, p 26. (506)
                                                                                 Page B-15

-------
 Chapter 9

 Blewitt, D.N., J.F. Yohn, R.P. Koopman, and T.C. Brown, Conduct of Anhydrous Hydrofluoric Acid Spill
        Experiments, International Conference on Vapor Cloud Modeling. (60)

 Brookhaven National Laboratory, Biomedical and Environmental Assessment Group, Development of a
        Three-Dimensional Computer Code for Modeling Water Spraying of Unconfined Releases of
        Hydrogen Fluoride and Other Gases, Upton, NY, November 1, 1991. Proposal to the U.S.
        Environmental Protection Agency, Office of Chemical Emergency Preparedness and
        Response. (135.4)

 Chemical Week, "Which Catalyst for Alkylation?," February 24, 1988, p 11. (500)

 Du Pont Chemicals, EPA - Du Pont Meeting, HF Clean Air Act Study.  Includes agenda, notes on Du
        Font's HF Philosophy, notes on Process Hazards Management System, Advanced Modeling of
        Accidental HF Release Scenarios, Equipment failure at Quimica Fluor - Matamoros Plant, HF
        Safety.  (137)

 Fthenakis, V.M. and K.W. Schatz, Numerical Simulations of Turbulent Flow Fields Caused by Spraying
        of Water on Large Releases of Hydrogen Fluoride, Biomedical and Environmental Assessment
        Division, Department of Applied Science, Brookhaven National Laboratory, Associated
        Universities, Inc., May  1991. (139.66)

 Fthenakis, V.M. and V. Zakkay, "A Theoretical Study of Absorption of Toxic Gases by Spraying,"
        Journal of Loss Prevention in the Process Industries, Volume 3, April 1990, p 197-205. (139.68)

 Hague, William  J., and William Pepe, "Flow Chamber Simulations of Aerosol formation and Liquid Jet
        Breakup For Pressurized Releases of Hydrogen Fluoride," Plant/Operations Progress,  Volume
        9, Number 2, April 1990, pp 125-130. (147.2)

 Industry Cooperative HF Mitigation/Assessment Program Ambient Impact Assessment Subcommittee,
        Development and Validation of Atmospheric Dispersion Models for Ideal Gases and Hydrogen
        Fluoride, Pan I: Technical Reference Manual, Shell Research Limited, Thornton Research
        Centre,  Chester, England, November 1990, Document Number TNER.90.015. Prepared by
        McFarlane, K., A. Prothero, J.S. Puttock, P.T. Roberts, and H.W. M. Witlox. (328.65)

 Industry Cooperative HF Mitigation/Assessment Program Ambient Impact Assessment Subcommittee,
        HGSYSTEM: Dispersion Models for Ideal Gases and Hydrogen Fluoride Tutorial and Quick-
        Reference Guide, Shell Research Limited, Thornton Research Centre, Chester, England,
        May 1991, Document Number TNER.91.007. Prepared by  Witlox, H.W.M. (508)

 International Conference and Workshop on Modeling and Mitigating the Consequences of Accidental
        Releases of Hazardous Materials, May 20-24, 1991,  Fairmont Hotel, New Orleans, Louisiana,
        American Institute of Chemical Engineers, Center for Chemical Process Safety, New York,
        1991, "Numerical Simulation of the Mitigation of HF Cloud  Concentrations by Means of Vapor
        Barriers and Water Spray Curtains," by Meroney, Robert N. (330)

 Koopman, Ronald P., Donald L Ermak and Stevens T. Chan,  "A Review of Recent Field Tests and
        Mathematical Modelling of Atmospheric Dispersion of Large Spills of Denser-Than-Air  Gases,"
       Atmospheric Environment, Volume 23, Number 4, pp 731-745, 1989. (289.2)

 Lehr, W.J., 2nd Review of ALOHA 5.0 (update), Modeling and  Simulation Studies, National  Oceanic and
       Atmospheric Administration, Hazardous Material Response Branch, April 30, 1990. (293.1)
Page 6-76

-------
National Oceanic and Atmospheric Administration, CAMEO™ 3.0 Computer-Aided Management of
       Emergency Operations, Hazardous Materials Response Branch, Seattle, September 1991.
       (361.5)

The Netherlands Organisation of Applied Scientific Research, "Methods for the Determination of
       Possible Damage," CPR 16E, First Edition 1992. (362.5)

International Conference and Workshop on Modeling and Mitigating the Consequences of Accidental
       Releases of Hazardous Materials, May 20-24, 1991, Fairmont Hotel, New Orleans, Louisiana,
       American Institute of Chemical Engineers, Center for Chemical Process Safety, New York,
       1991,  "The HGSYSTEM Dispersion Modelling Package: Development and Predictions," by
       Puttock, J.S., K. McFarlane, A. Prothero, P.T. Roberts,  F.J. Rees, H.W.M. Witlox, and Douglas
       N. Blewitt. (416.5)

Hanna, Steven R., Joseph C.  Chang and David G. Strimaitis, Review of Dense Gas Modeling with
       Emphasis on Application  in Planning for Accidental Releases and Emergency Responses, Final
       Report, Contract No. C91319, South Coast Air Quality Management District,
       February 26, 1992. (441)

South Coast Air Quality Management District, Modeling of Accidental Hydrogen Fluoride and Sulfuric
       Acid Releases and Comparison of Air Quality Impacts,  January 24, 1990. (342)

Spicer, T.O., and J. Havens, Modeling HF and NH3 Spill Test Data Using DEGADIS,  Department of
       Chemical Engineering, University of Arkansas,  Fayetteville, Arkansas, August 1988.  (444)

Stratco, Inc., Sulfuric Acid Alkylation - A Viable Alternative to Hydrofluoric Acid Alkylation,
       March 26, 1992. (467.3)

Tilton, James  N. and Carl W. Farley, "Predicting Liquid Jet Breakup and Aerosol Formation During the
       Accidental Release of Pressurized Hydrogen Fluoride," Plant/Operations Progress, Volume 9,
       Number 2, April 1990, p 120-124. (477)

Turner, R.M. and S. Fairhurst, Toxicology of Substances in Relation to Major Hazards, Hydrogen
       Fluoride, London,  1990. (489.7)

U.S. Coast Guard, Vulnerability Model, A Simulation System for Assessing Damage Resulting From
       Marine Spills, Washington, D.C., June 1975,  Document Number AD-A015245. Prepared by
       Eisenberg, Norman A., Cornelius J. Lynch, and Roger J. Breeding, Enviro Control,  Inc.
       (139.513)

Woodward, John, ed., International Conference on Vapor Cloud Modeling, November 2-4, 1987, Boston
       Marriott Cambridge, Cambridge, MA, American Institute of Chemical Engineers, Center for
       Chemical Process Safety, New York, 1987, "Specialized Techniques for Modeling the Unique
       Phenomena Exhibited in  HF Releases," by Chikhliwala, E.D. and W.J. Hague. (120)

Woodward, John, ed., International Conference on Vapor Cloud Modeling, November 2-4, 1987, Boston
       Marriott Cambridge, Cambridge, MA, American Institute of Chemical Engineers, Center for
       Chemical Process Safety, New York, 1987, "The Mixing of Anhydrous Hydrogen Fluoride with
       Moist  Air," by Clough, P.N., D.R. Grist, and C.J. Wheatley. (130)

Woodward, John, ed., International Conference on Vapor Cloud Modeling, November 2-4, 1987, Boston
       Marriott Cambridge, Cambridge, MA, American Institute of Chemical Engineers, Center for
                                                                                  Page B-17

-------
        Chemical Process Safety, New York, 1987, "FEM3 Modeling of Ammonia and Hydrofluoric Acid
        Dispersion," by Chan, Stevens T., Howard C. Rodean, and Doug Blewitt. (100)
Page B-18

-------
Chapter 10

American Industrial Hygiene Association, Concepts and Procedures for the Development of Emergency
       Response Planning Guidelines (ERPGs), ERPG Committee, December 1989. (10.47)

General

American Paint Coat Journal, January 2, 1989, p 54.
                                                                               Page B-19

-------

-------