EPA-600/R-96-131
November 1996
DEMONSTRATION OF A LIQUID CARBON DIOXIDE PROCESS
FOR CLEANING METAL PARTS
By:
E. A. Hill and K. R. Monroe
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, NC 27709-2194
EPA Cooperative Agreement No. CR818419
EPA Project Officer: Charles H. Darvin
Air Pollution Prevention and Control Division
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Prepared for:
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC 20460

-------
BIBLIOGRAPHIC INFORMATION
PB97-121149
Report Nos: RTI-93U-5171-019
Title: Demonstration of a Liquid Carbon Dioxide Process for Cleaning Metal Parts.
Date: Nov 96
Authors: E. A. Hill and K. R. Monroe.
Performing Organization: Research Triangle Inst., Research Triangle Park, NC.
Performing Organization Report Nos: EPA/600/R-96/131
Sponsoring Organization: *Environmental Protection Agency, Washington, DC. Air
Pol lutionTontroi mv.
Contract Nos: EPA-R-818419
Type of Report and Period Covered: Final rept. Jun 95-Jun 96.
NTIS Field/Group Codes: 68A (Air Pollution & Control), 71Q (Solvents, Cleaners, &
Abrasives)
Price: PC A06/HF A01
Availability: Available from the National Technical Information Service, Springfield,
VA. 22161
Number of Pages: 80p
Keywords: *Degreasing, *Environmental chemical substitutes, ^Solvents, Cleaners,
Surface-cleaning, Chloroform, Ethanes, Material substitution, Alternatives,
Performance evaluation, Air pollution abatement. Demonstration programs. *Liquid
carbon dioxide, Metal parts, Trichloroethane.
Abstract: The report gives results of a demonstration of liquid carbon dioxide (LC02)
as an aiternative to chlorinated solvents for cleaning metal parts. It describes the
LC02 process, the parts tested, the contaminants removed, and results from preliminary
laboratory testing and the on-site demonstration at the Air Logistics Center (ALC) at
Robins Air Force Base (RAFB), Georgia. The process chosen to be replaced was vapor
degreasing in 1,1,1-trichloroethane (TCA), a solvent that is ozone-depleting, a
hazardous air pollutant (HAP), and one of the 17 chemicals on the Environmental
Protection Agency (EPA) 33/50 list of priority pollutants. Carbon dioxide (C02)
degreasing was chosen for demonstration as a potential alternative to TCA because it
is noncombustible, nontoxic, listed on the EPA Significant New Alternatives program
(SNAP) list of approved cleaning alternatives, and not ozone depleting.

-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before compl ||| |||| || |||||| |||| |||j || 11| |||
1. REPORT NO. 2.
EPA- 600 /E- 96-131
PB97-121149
4. TITLE AND SUBTITLE
Demonstration of a Liquid Carbon Dioxide Process
for Cleaning Metal Parts
S. REPORT DATE
November. 1996
6, PERFORMING ORGANIZATION CODE
7, AUTHORtSJ
E. A. Hill and K. R. Monroe
8. PERFORMING ORGANIZATION REPORT NO.
93U- 5171-019
9, PERFORMING ORGANIZATION NAME AND ADDRESS
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709-2194
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
CR 818 419
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air Pollution Prevention and Control Division
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 6/95 - 6/96
14. SPONSORING AGENCY CODE
EPA/600/13
ie. supplementary notes ^pp£jj) project officer is Charles H. Darvin, Mail Drop 61, 919/
541-7633.
ie. abstract The report gives results of a demonstration of liquid carbon dioxide (LCG2)
as an alternative to chlorinated solvents for cleaning metal parts. It describes the
LC02 process, the parts tested, the contaminants removed, and results from pre-
liminary laboratory testing and the on-site demonstration at the Air Logistics Cen-
ter (ALC) at Robins Air Force Base (RAFB), Georgia. The objective of this project
was to find and demonstrate innovative parts cleaning technologies to replace envi-
ronmentally damaging chemicals with more benign processes. The process chosen
to be replaced was vapor degreasing in 1,1,1-trichloroethane (TCA), a solvent that is
ozone-depleting, a hazardous air pollutant (HAP), and one of the 17 chemicals on the
Environmental Protection Agency (EPA) 33/50 list of priority pollutants. Carbon di-
oxide (C02) degreasing was chosen for demonstration as a potential alternative to
TCA because it is noncombustible, nontoxic, listed on the EPA Significant New, Alter-
natives Program (SNAP) list of approved cleaning alternatives, and not ozone deple-
ting. Liquid C02 is distinct from the better known supercritical C02 because it can
be maintained at lower pressures and temperatures than supercritical, so the clean-
ing equipment may be less expensive. Both liquid and supercritical C02 have the ad-
vantage of permeating into tiny holes like a gas and have good solvency.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRiPTORS
bjDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution Ozone
Carbon Dioxide
Solvents
Degreasing
Metal Cleaning
Chloroform
Pollution Prevention
Stationary Sources
Liquid Carbon Dioxide
Supercritical Carbon
Dioxide
1,1,1-Trichloroethane
13B
07B
11K
13 H
07C
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO- OP P^ES
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Fofm 2220-1 (9-73)

-------
FOREWORD
The U, S. Environmental Protection Agency is charged by Congress with pro-
tecting the Nation's land, air, and water resources. Under a mandate of national
environmental laws, the Agency strives to formulate and implement actions lead-
ing to a compatible balance between human activities and the ability of natural
systems to support and nurture life. To meet this mandate, EPA's research
program is providing data and technical support for solving environmental pro-
blems today and building a science knowledge base necessary to manage our eco-
logical resources wisely, understand how pollutants affect our health, and pre-
vent or reduce environmental risks in the future.
The National Risk Management Research Laboratory is the Agency's center for
investigation of technological and management approaches for reducing risks
from threats to human health and the environment. The focus of the Laboratory's
research program is on methods for the prevention and control of pollution to air,
land, water, and subsurface resources; protection of water quality in public water
systems; remediation of contaminated sites and groundwater; and prevention and
control of indoor air pollution. The goal of this research effort is to catalyze
development and implementation of innovative, cost-effective environmental
technologies; develop scientific and engineering information needed by EPA to
support regulatory and policy decisions; and provide technical support and infor-
mation transfer to ensure effective implementation of environmental regulations
and strategies.
This publication has been produced as part of the Laboratory's strategic long-
term research plan. It is published and made available by EPA's Office of Re-
search and Development to assist the user community and to link researchers
with their clients.
E. Timothy Oppelt, Director
National Risk Management Research Laboratory
EPA REVIEW NOTICE
This report has been peer and administratively reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.
PROTECTED UNDER INTERNATIONAL COPYRirwr
ALL RIGHTS RESERVED	COPYRIGHT
f^NAJL TECHNICAL INFORMATION SERVICE
U.S. DEPARTMENT OF COMMERCE
ia

-------
ABSTRACT
This is the Final Report for the demonstration of liquid carbon dioxide (LC02) as an
alternative to chlorinated solvents for cleaning metal parts. The report describes;
•	the LC02 process,
•	the parts tested,
•	the contaminants removed,
•	results from preliminary laboratory testing and the on-site demonstration performed at the
Air Logistics Center (ALC) at Robins Air Force Base (RAFB), Georgia.
The objective of this research project was to find and demonstrate innovative parts
cleaning technologies to replace environmentally damaging chemicals with more benign processes.
The process chosen to be replaced was vapor degreasing in 1,1,1-trichloroethane (TCA). This
solvent needs to be replaced because it is an ozone-depleting compound, a hazardous air pollutant
(HAP), and one of the 17 chemicals on the Environmental Protection Agency (EPA) 33/50 list of
priority pollutants. Carbon dioxide (C02) degreasing was chosen for demonstration as a potential
alternative to TCA because it is not an ozone-depleting material, is noncombustible, has low
toxicity, and is listed on the EPA Significant New Alternatives Program (SNAP) list of approved
cleaning alternatives.
Liquid C02 is distinct from the more well known supercritical C02 because it can be
maintained at lower pressures and temperatures than supercritical, so the cleaning equipment may
be less expensive. Both liquid and supercritical COz have the advantage of permeating into tiny
holes like a gas and have good solvency for many oils, greases and other contaminants. There is
no solvent waste or other media to dispose of because C02 returns to gas phase after cleaning.
The contaminants can be collected for reuse or disposal. In the system demonstrated, COz was
recycled between cleaning cycles, so material costs also were reduced.
The LC02 technology was demonstrated on-site at the Warner Robins ALC at RAFB,
GA. The primary parts tested were fuel system tubing, brass filters, miscellaneous machined
metal parts, and rags from the maintenance processes. Preliminary soaking in a warm mixture of
hydrocarbon oils and surfactants (hot oil process or HOP) was necessary to remove difficult
contaminants from many parts. This high-boiling, nonflammable mixture was then removed by
the LCQ2 process and could be recovered for reuse.
Results of this demonstration show that LC02 performs similarly to TCA vapor
degreasing. It is capable of cleaning debris from a wide variety of substrates, including precision
and nonprecision applications. The HOP/LCOz process removed drawing compounds from
aluminum and titanium tubes used in fuel systems, heavy grease from bolts, hydraulic fluid from
brass filters, and general shop dirt from aluminum, brass, and stainless steel parts. Components
from breathing oxygen systems and aluminum honeycomb core also were cleaned, but additional
testing would be required to validate the LC02 process for these applications.
This report is submitted in partial fulfillment of Cooperative Agreement No. CR818419 by
Research Triangle Institute under sponsorship of the U.S. EPA.
ii

-------
TABLE OF CONTENTS
ABSTRACT							 							 ii
FIGURES					v
TABLES 					vi
ACRONYMS AND ABBREVIATIONS		vii
METRIC UNITS			viii
ACKNOWLEDGMENTS		 ix
1.0 INTRODUCTION							1
1.1	Background 														2
1.2	Project Objectives					2
1.3	Technical Approach			3
2.0 DESCRIPTION OF LIQUID CARBON DIOXIDE CLEANING TECHNOLOGY .... 4
2.1	Background on Carbon Dioxide Chemistry 					 4
2.2	Comparison of Cleaning with LCOz Versus Other C02 Methods					6
2.3	Capabilities/Range of Applications 				7
2.4	Equipment and Materials 								 8
2.5	Process Descriptions							.11
2.6	Advantages						13
2.7	Disadvantages 							14
2.8	Projected Costs 						15
3.0 CHOOSING THE DEMONSTRATION SITE AND PROCESS			 16
4.0 LABORATORY FEASIBILITY TESTING					17
4.1	Development of the NVR Test Procedure	18
4.2	Results of Feasibility Tests							19
5.0 ON-SITE DEMONSTRATION		 . 21
5.1	Introduction 								21
5.2	Range of Parts Tested					21
5.3	F-15 Fuel System Tubes					22
5.4	Steel Bolts				28
5.5	Brass Filters							31
5.6	Large C-130 Aircraft Bearings	.35
5.7	Aluminum Honeycomb Core			36
5.8	Propeller Pitch Lock Regulators					38
iii

-------
5.9	Flap Jack Screw Yokes 		40
5.10	Small Avionic Bearing Assemblies 								 41
5.11	Electronic Test Circuit Board 					42
5.12	Reflectometer	43
5.13	Rags				43
5.14	Summary of On-site Tests 			 45
5.15	C02 Usage During the On-site Demonstration 	45
5.16	Operational Demonstration Results			46
5.17	Recommendations and Observations			46
6.0 QUALITY ASSURANCE		 47
7.0 REFERENCES 						50
APPENDIX A: Case Studies 				....					53
A. 1 Beryllium Copper Bar Stock - BRUSH WELLMAN, INCORPORATED .... 53
A.2 Machined Metal Parts - ACCRATRONICS 	,57
APPENDIX B: Material Safety Data Sheets			59
APPENDIX C. Nonvolatile Residue Procedure for Tubes			68
iv

-------
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
FIGURES
Phase Diagram for Carbon Dioxide 	
Illustration of the LCOz Cleaning System	
The LC02 Cleaning System Demonstrated at RAFB	
The Rear View of the Recycler	
Typical Aluminum Tubes Cleaned During the Demonstration
Graph of NVR Results for Tubes 					
Bolts Contaminated with Heavy Grease 		
Brass Filters, Contaminated (Left), Clean (Right) .........
Graph of NVR Results for Brass Filters			
Bearings Before (Bottom) and After (Top) H0P/LC02 Cleaning
Aluminum Honeycomb Core Test Samples 		
Steel Pitch Lock Regulator Pieces 	
Contaminated Aluminum Flap Jack Screw Yoke Parts	
Avionics Bearing Assembly 			
Rag, Before (Left) and After (Right) HOP/LCOz Cleaning ,,..
Page
.. 5
..9
..9
. 10
. 22
. 24
. 28
. 31
. 33
. 35
. 37
. 39
. 40
. 41
. 44
v

-------
1.
2.
3.
4.
5A.
5B.
5C.
6.
7.
8.
9.
10.
11.
Al.
A2.
A3.
TABLES
Chemical and Physical Properties of Degreasing Solvents .....
Costs for Equipment and Materials	
NVR Results from the Laboratory Feasibility Tests 	
Average Values and Standard Deviations for Tubes 		
NVR Results for Tubes Contaminated with Drawing Compound,
NVR Results for Tubes Cleaned with TCA			
NVR Results for Tubes Cleaned at RAFB with HOP/ LC02 .
NVR Results for Steel Pylon Bolts	
Comparison of P-D-680 and HOP Hydrocarbon Oil. .......
Individual NVR Results for Brass Filters 	
Averages and Standard Deviations for Brass Filter NVR Tests
Results from Adhesion Tests on Aluminum Honeycomb	
Results of the Full Process Blank for the NVR Test	
Comparison of Waste Production for PERC and LC02	
Economic Comparison Summary from NICE3 Application ...
Details of Annual Costs and Savings								56
vi

-------
APPCD
AQMD
ASTM
CAA
CFC
co2
EPA
HAP
J I£p£
HFE
HOP
kW
lb
lco2
MSDS
MEK
NDI
NESHAP
NMP
NVR
PERC
P2
psi
RAFB
RCRA
RTI
SCAQMD
SNAP
TCA
VOC
WR-ALC
ACRONYMS AND ABBREVIATIONS
Air Pollution Prevention and Control Division
Air Quality Management District
American Society for Testing and Materials
Clean Air Act
chlorofluorocarbon
carbon dioxide
Environmental Protection Agency
hazardous air pollutant
hydrochlorofluorocarbon
hydrofluoroether
hot oil process
kilowatt
pound
liquid carbon dioxide
material safety data sheet
methyl ethyl ketone
nondestructive inspection
National Emission Standards for Hazardous Air Pollutants
n-methyl pyrrolidone
nonvolatile residue
perchloroethylene
pollution prevention
pounds per square inch
Robins Air Force Base
Resource Conservation and Recovery Act
Research Triangle Institute
South Coast Air Quality Management District
Significant New Alternatives Program
1,1,1 -trichloroethane
volatile organic compound
Warner Robins Air Logistics Center
vii

-------
METRIC UNITS
English units have been listed throughout the report because they are the primary units
used by most of the intended readership. The multiplying factors for converting from the English
units to their metric equivalents are given in the table below.
METRIC CONVERSION FACTORS
Symbol
When You Know
the Number of
Multiply By
To Find the
Number of
Symbol
LENGTH
mil
thousands of an inch
0.0254
micrometers
fz m
in
inches
2.54
centimeters
cm
ft
feet
0.305
meters
m
VOLUME
gal
gallons
3.79
liters 1
MASS
lb
pounds
0.454
kilograms j kg
PRESSURE
psi
pounds per
square inch
0.0689
bars
bar
atm
atmospheres
1.01
bars
bar
TEMPERATURE
°F
degrees
Fahrenheit
5/9
(after subtracting
32)
degrees
Centigrade
°C
DENSITY
lb/ft3
pounds per cubic
foot
16.0
kilograms per
cubic meter
kg/m3
viii

-------
ACKNOWLEDGMENTS
The authors wish to thank the personnel from Robins Air Force Base, DEFLEX
Corporation, and Nobles Manufacturing, Incorporated for their cooperation and support.
Without the enthusiastic efforts of all persons involved in this demonstration, it could not have
been successfully completed.
Randy King of the Environmental Management Directorate at Robins Air Force Base
provided extensive on-site coordination during the preparation and duration of the demonstration.
Many people from the Environmental Directorate, F-15 Directorate, C-130 Directorate, and C-
141 Directorate worked diligently to support this demonstration. They were responsible for
facilities preparation and maintenance, providing information on current processes, obtaining test
parts, performing inspections, and many other details.
Barry Carver, Randy Rutledge, and Fred Chen of DEFLEX Corporation ensured the
optimum performance of the equipment while on site. They were supported by Dick Jackson and
Pat Theys. as well as several other DEFLEX personnel during the production and shipping of this
equipment.
Nobles Manufacturing, Incorporated provided a centrifugal spinner for use during the on-
site demonstration.
All phases of this demonstration were supported by Ken Monroe and K. David Carter, Jr.,
of RTI. Ken also assisted at Robins Air Force Base during each of the site visits and
demonstration. David performed the laboratory analyses at RTI.
ix

-------
1.0 INTRODUCTION
This project was initiated by the U.S. Environmental Protection Agency (EPA) to
demonstrate innovative technologies for replacing cleaning processes that rely on
environmentally damaging substances. The specific goal of this work was to demonstrate
environmentally benign cleaning processes to replace 1,1,1 -trichloroethane (TCA) for degreasing
metal parts. The technology chosen for demonstration was degreasing with liquid carbon dioxide
(LCO,). The TCA vapor degreasing process is widely used in many industries for cleaning
metals, so information developed through this demonstration should be valuable to a large
number of civilian and other military operations with metal cleaning applications.
During the preparatory work for this project, nonvolatile residue (NVR) testing was
performed by Research Triangle Institute (RTI) on metal parts cleaned with the LC02 process at
the equipment manufacturer's facility. Based on favorable results of this preliminary testing, full-
scale degreasing equipment was moved on-site at Robins Air Force Base (RAFB), GA. The
equipment was operated on-site for two weeks, during which time samples of each type of part
cleaned were sent back to RTI for NVR testing. Samples of dirty parts and parts cleaned by the
current processes were also collected and tested at RTI. The technology demonstrated during this
project is described in this report, along with the results from the cleanliness tests, the cleaning
and testing procedures, data handling, and corrective actions for the project.
The LC02 cleaning and recycling equipment, the hot oil process (HOP) and formulation,
and related processes, designs, and operations used in this investigation and described in this
report are proprietary to DEFLEX Corporation and are the subject matter of issued patents and
pending patents and applications.1"6
1

-------
1.1	Background
The work in this demonstration project was performed primarily at Warner Robins Air
Logistics Center (WR-ALC) at RAFB, GA. RAFB provides worldwide engineering and
management for the F-15 fighter, €-130 and C-141 transport aircraft, and all United States Air
Force helicopters. The Air Logistics Center carries out maintenance and repair activities on these
aircraft as well as on electronics for airborne avionics, communications, and radar equipment.
They also support the fleet of Air Force vehicles and the Air Force Reserve. Cleaning is often
required at several points during these maintenance processes. Like many other U.S. facilities
involved in repair and machining procedures, RAFB has extensively relied upon chlorinated
solvents for parts degreasing. RAFB has already eliminated or changed many of their processes
that used hazardous materials. One of the few remaining processes that involves a hazardous
solvent is vapor degreasing in the F-15 repair area, where two large covered vapor degreasers
containing 1,1,1 -trichloroethane (TCA) are in operation.
Production of this solvent has been phased out under Title VI, section 604 of the Clean
Air Act Amendments of 1990 and has not been produced in the United States since December
31,1995. It is also one of the 17 targeted chemicals in the U.S. EPA's 33/50 Program for
reduction of hazardous emissions. Thirdly, the National Emission Standards for Hazardous Air
Pollutants (NESHAP) for Halogenated Solvent Cleaning have been promulgated. This NESHAP
regulates the use of six hazardous air pollutants (HAPs) in solvent cleaning machines: TCA,
methylene chloride, perehloroethylene, trichloroethylene, carbon tetrachloride and chloroform. It
contains requirements for specific equipment features and operator behavior, as well as multiple
reporting requirements, and fines for noncompliance. Any one of these requirements would lead
to the elimination of the TCA process, but the combination raises the replacement of this
chemical to a very high priority.
1.2	Project Objectives
The overall objective of this project was to demonstrate the feasibility of innovative
alternatives for cleaning metals without using hazardous solvents. Part of this objective was to
document the performance of the chosen technology, and its current capabilities, possible range
2

-------
of applications, costs, advantages, disadvantages, and future potential. This information,
gathered into a single document, is intended to assist others with similar cleaning applications.
1.3 Technical Approach
The technology chosen for demonstration was LC02 degreasing, which is discussed in
detail in Section 2.0. Preliminary feasibility tests were performed by cleaning samples of metal
parts at the facility where the CO, degreasing equipment was manufactured and testing them in
RTFs Surface Cleaning Laboratory. Once shown to be feasible, the LCO, equipment was moved
to RAFB for an on-site demonstration. Also, to provide as much information as possible on this
technology, case studies from businesses that have tested this process on their parts or are using it
are included as Appendix A.
A team comprised of representatives of the sponsoring agencies was assembled to ensure
successful completion of the pollution prevention technology demonstrations at RAFB. One
member was designated from each of the following four groups to spearhead their organization's
efforts on this project. These people were the primary points of contact between organizations
and were responsible for obtaining the resources and assistance of personnel within their own
organizations. The sponsoring agencies involved in this demonstration and their representatives
were:
1)	U. S. EPA, Air Pollution Prevention and Control Division (APPCD), Research Triangle
Park, North Carolina (Charles Darvin, EPA project officer),
2)	RTI, Research Triangle Park, North Carolina (Elizabeth Hill, RTI project leader),
3)	WR-ALC, RAFB, Georgia (Randy King, Environmental Management Directorate), and
4)	DEFLEX Corporation (DEFLEX), Burbank, California (Barry Carver, vice president).
The team was jointly responsible for maintaining program progress and solving problems
as they arose. The EPA/APPCD initiated this project and provided guidance throughout its
duration. RTI coordinated the demonstration, provided technical expertise in cleaning and
cleanliness testing, and reported to the EPA project officer. WR-ALC at RAFB, GA was the host
3

-------
for the on-site portion of the demonstration. The primary responsibility of RAFB was to support
the demonstration by providing the facilities, parts, current process criteria or specifications, and
site coordination for carrying out the demonstration. The RAFB technical representative
supervised the project at the site. DEFLEX Corporation supported both the preliminary cleaning
trials and the on-site demonstration by making available equipment, chemicals, and personnel
with extensive experience in using the technology.
2.0	DESCRIPTION OF LIQUID CARBON DIOXIDE CLEANING TECHNOLOGY
2.1	Background on Carbon Dioxide Chemistry
Carbon dioxide is a gas at normal room temperature and pressure. It is odorless,
colorless, noncombustible, and low in toxicity. It is not an ozone-depleting material, volatile
organic compound (VOC), or HAP. It is a greenhouse warming gas, but commercial supplies of
C02 are obtained by recycling by-products from other processes. Therefore, its use results in no
net addition of C02 to the atmosphere. Carbon dioxide is used in carbonated beverages, fire
extinguishers, municipal water treatment, and for air enrichment in greenhouses. It has a
threshold limit value of 5000 ppm in air (see MSDS in Appendix B). At concentrations higher
than 10 percent in air, carbon dioxide is an asphyxiant. It is naturally present in ambient air at
about 365 ppm.
As shown in the phase diagram in Figure 1, CO, exists as a gas at standard temperature
and pressure of 32 °F (0 °C), 14.7 psi (1 bar). By increasing pressure and controlling the
temperature, C02 can be changed from gas phase to solid, liquid, or supercritical fluid phases.
Compressing the gas to about 70 bar while maintaining temperature just below the critical
temperature of about 88 °F (31 °C) changes it to liquid phase. When the pressure is reduced, it
changes phase back to a gas. As a liquid or supercritical fluid, C02 has excellent solvent
properties for many oils, greases, and other common machining contaminants. Cleaning with
LC02 is generally performed within the temperature range of 60-80 °F (16-27 °C) and pressure
of 700-1500 psi (48-102 bar). Changing the operating pressures and temperatures within these
ranges allows selective removal and separation of a variety of contaminants.
4

-------
Supercritical
fluid
Liquid
Solid
C 74.8
XI
Critical point
m
k.
3
»
tn

-------
has low surface tension and very low viscosity, which improves the likelihood that the solvent
will wet the surface to be cleaned and penetrate into small crevices and blind holes in the parts.
TABLE 1. CHEMICAL AND PHYSICAL PROPERTIES OF DECREASING SOLVENTS
Solvent
Solubility Parameter
Viscosity at 25 °C
Surface Tension at

(MPa/!)(9)
(centipoise)(1B)
20 °C (dynes/cm)®
lco2
20-22(a)
0.06w
5.0
TCA
17.7
0.79
25.5Cc)
Perchloroethylene
20.3
0.84
31.7
Methylene chloride
20.3
0.41
26.5
Isopropyl alcohol
23.5
2.04
21.7
Acetone
20.0
0.31
23.7
Water
47.8
0.89(b)
72.8
(a)	Calculated from the Giddings Equation on p. 224 of reference 9.
(b)	From reference 8.
(c)	Value at 25 °C from supplier literature.
2.2 Comparison of Cleaning with LC02 Versus Other C02 Methods
There are several forms of cleaning that rely on carbon dioxide, including supercritical
fluid extraction,11,12 pellet blasting,13,14 and snow spraying.15"17 All cleaning methods using C02
share one important characteristic. Carbon dioxide will return to gas phase at normal pressure
and temperatures, so there is no liquid waste stream as in most solvent-based processes.
Contaminants removed by the C02 are dropped as it returns to gas phase. These contaminants
can be collected for reuse or disposed of by proper methods. Even if the contaminants cannot be
reused or recycled, the total volume of disposed material is greatly reduced because it does not
include a cleaning solvent.
LC02 and supercritical CO, cleaning are similar in that both are conducted in closed
vessels under elevated pressure and temperature. Maintaining C02 as a supercritical fluid
requires higher pressures and temperatures, which require larger pumps, compressors, and thicker
6

-------
pressure vessels than LC02. All this makes the system more expensive to build and operate.
Supercritical C02 can extract or dissolve a wider range of contaminants because many are more
soluble in the solvent at higher temperatures and pressures than achievable in the liquid phase.
This can be an advantage or disadvantage, depending on whether the user is trying to remove a
particular contaminant or a wide range of materials. Supercritical fluids cannot be agitated
effectively, so cleaning is more dependent on how well the fluid dissolves the contaminant. They
also are poorly cavitated by ultrasonics, so they are not very effective for particle removal. LCO,
can be stirred, sprayed, or sonicated in a manner more similar to other liquids. Adding
mechanical energy with these process enhancements increases the particle removal capability of
the C02.
LC02 cleaning is quite different from C02 pellet blasting or snow cleaning. Pellets and
snow do not require a pressure vessel for use. They normally are sprayed through a nozzle
towards the surface to be cleaned. Both pellet and snow cleaning are line-of-sight processes that
cannot penetrate into crevices or complex surfaces. Pellet blasting is similar to abrasive blasting
and is used for removal of paints, adhesives, corrosion, etc. Snow cleaning is a much gentler
form of spraying that is used to remove dust and small particles from surfaces without abrading
or scratching them. As a result, snow will not remove paint or thick layers of organic materials.
Neither technique is appropriate for removal of oils or greases, except in special circumstances.
2.3 Capabilities/Range of Applications
LC02 removes most light and medium weight hydrocarbon oils, gross particulate
contamination, drawing compounds, and other machining fluids. The best applications are those
where organic solvent vapor degreasing also would work, such as removing machining oils from
metals or cleaning organic films from inside small tubes. Also, silicone monomers have been
extracted from silicone tubing for the medical industry. Parts that have successfully been cleaned
with liquid C02 include:
•	machined metal parts (valves, brackets, bearing assemblies, complex metal parts),
•	electrical components (spacers, switches, relays, heat sinks),
•	medical devices (tubing), and
7

-------
• other materials (bar stock, rods, tubing, optics, ceramics, natural fibers).
Parts with difficult contaminants may be precleaned by soaking in a mixture of
lightweight hydrocarbon oil and surfactants. The oil/surfactant mixture is then removed with the
LC02. This precleaning step is called the hot oil process (HOP) and is described in the following
section along with the LC02 degreasing process.
LC02 will not remove paint, rust, conversion coatings, or most adhesives. It is also
unlikely to be able to remove small particles at this stage of development. A dedicated system
designed with megasonic or ultrasonic transducers to provide mechanical energy may be able to
dislodge particles. LC02 is not applicable for degreasing parts that will be damaged by pressure,
or when leaching of plasticizers from elastomers is not desired. Examples of inappropriate
applications are cleaning complete bearings where grease removal is not desired, and cleaning
elastomeric O-rings. Also, the cleaning vessel size limits the size of parts that can be cleaned.
2.4 Equipment and Materials
The equipment in this demonstration was primarily supplied by DEFLEX Corporation.
They provided the LC02 DEFLEX Superfuge™ cleaning system, the recycling unit, and the
equipment for the HOP pretreatment step. RTI provided an ultrasonic tank that was also used for
the HOP. A centrifugal spin dryer (Model Turbo 28) was loaned for the demonstration by
Nobles Manufacturing, Incorporated.
An illustration of the LC02 cleaning and recycling equipment is shown in Figure 2, It
consists of two main pieces: the cleaning system that removes the contaminants from the parts
with LCO,, and the COz recycler where the LCO, is allowed to change phase and drop the
contaminants. A typical LC02 parts cleaning system contains a pressure vessel to hold the parts,
a pump, a compressor, and necessary plumbing, valves, and controls. Safety devices prevent
filling the tank without sealing the vessel or opening the vessel while it is under pressure. Other
safety devices release the pressure if it exceeds preset limits.
8

-------
.Parts
fi
Clean CO 2
1
Dirty CO2
Parts Cleaner
C02 Recycler
FIGURE 2. ILLUSTRATION OF THE LC02 CLEANING SYSTEM.
The system demonstrated at RAFB is shown in Figure 3. The parts cleaning section of
the system is the smaller, cube-shaped equipment in the foreground of the figure. The recycler is
the larger unit immediately behind. Gas bottles to the right supply C02 to the system as needed.
The vessel that holds the parts to be cleaned is 12 inches (30 cm) in diameter and 12 inches (30
cm) deep. The wire mesh parts basket is shown sitting on the floor in front of the unit.
ra
.X

flii
1* A
> •Mi l
III fill
' M| |M
FIGURE 3. THE LCO, CLEANING SYSTEM DEMONSTRATED AT RAFB.
9

-------
Other, similar units have been manufactured with parts vessels up to 30 inches in
diameter. These vessels open at the top and are called vertical units. Horizontal machines have
been manufactured for degreasing long parts like tubes and bar stock. The longest horizontal
cleaning chamber manufactured so far by DEFLEX was 17 feet long with an internal diameter of
8 inches. This machine is in operation at an industrial facility and is described in Appendix A
(Case Studies).
The dirty C02 from the degreasing machine can be separated into clean C02 and residual
contaminants in the attached recycling unit (Figure 4). The recycler contains a tank to hold the
dirty C02 (the white tank shown on the right side of the picture), a compressor/condenser, and a
75 gallon tank to store the clean, recycled C02 (the larger, black tank on the left side of the
picture). The recycled C02 can be reused, and the contaminants can be collected for reuse or
disposal. The efficiency of C02 recovery depends in part on the ambient temperature around the
equipment. At room temperatures of about 78 °F (26 °C) 90-95 percent recovery of the C02 can
be expected. In a system with a 10 gallon cleaning vessel, about 9 gallons would be recovered
after each cleaning cycle, and 1 gallon would be vented to the atmosphere.
FIGURE 4. THE REAR VIEW OF THE RECYCLER.
10

-------
HOP Equipment
The only equipment necessary for the HOP presoak was a tank large enough to hold the
parts that could be heated and allowed some form of agitation. Air agitation, ultrasonics,
and mechanical stirring were all used during the demonstration and found to be effective.
Materials
The materials used during this demonstration were C02 and the hot oil mixture. The C02
was 99,8 percent purity, supplied by Air Liquide in 40 gallon cylinders with siphon tubes. Six
cylinders at a time were connected to the system by using a manifold of braided stainless steel,
Teflon™-lined hoses. The HOP material (SuperSolv PAS from DEFLEX) was low vapor
pressure, high flash point mixture of a lightweight hydrocarbon oil and a surfactant. The MSDS
for the mixture is in Appendix B.
2,5 Process Descriptions
HOP Pretreatment
For this demonstration, most parts were cleaned with the HOP prior to the LCO,
degreasing. The oil was heated in a tank to about 140 °F (60 °C). The purpose of the HOP
pretreatment step was to solubilize or mechanically remove difficult soils and materials that were
not soluble in the LC02, such as waxes and soaps from greases. This process is the subject of a
patent pending application. After the parts were soaked in the HOP, they were drained or spun in
a centrifugal spinner to remove most of the oil mixture. Then they were degreased in the LC02,
which removed the remaining oil mixture containing the solubilized contaminants and particles.
CO, Degreasing Cvcie
After the HOP, the oily parts were placed in the parts basket in the LC02 cleaning vessel.
The vessel was sealed with a quick closing mechanism and filled with LCO,. In the demonstration
system, the basket rotated inside the vessel, creating turbulent flow and shear forces across the
parts' surfaces. This carried fresh solvent to the part surface, which increased transport of the
dissolved contaminant away from the part. The rotation of the basket and movement of the CO,
11

-------
across the parts also added other forces to increase removal of soluble and insoluble
contaminants, specifically centrifugal and fluid shearing forces. Upon completion of the cleaning
cycle, the C02 was transferred automatically from the cleaning vessel to the recycler, and the parts
were removed clean and dry. Only one chamber volume of LC02 is typically required to degrease
a batch of dirty parts. The oil removed from the parts by the C02 could be filtered and added
back to the HOP tank.
A typical LC02 degreasing cycle takes about 20 to 25 minutes including loading and
unloading parts. The cycle is divided into five steps: filling, cleaning, draining, purging, and
venting. Parameters that can be changed by the operator are the duration of the cleaning step and
the speed and direction of basket rotation. These values are chosen based on the type and amount
of contamination and the complexity and fragility of the part to be cleaned. The process can be
optimized by cleaning representative samples while varying the adjustable parameters. Once the
unit has been programmed, these parameters are controlled by microprocessor. The operator
needs only to load the parts, close the vessel, and push the start button. The average cycle used at
RAFB was as follows;
•	Fill- 3 min. IX02 is pumped into the cleaning vessel. The pumping continues until the
vessel is filled to within an inch of the sealed top. As the vessel fills, basket
rotation starts. The fill time depends on the pressure of the air supply that
drives the C02 pump and is typically 2-3 .5 minutes.
•	Clean-10 min. The basket of parts continues to rotate, half of the cycle time in one
direction, then in the reverse direction for the second half. Typical rotation
speed is 400 rpm. Fragile parts are rotated more slowly, usually around
150 rpm.
•	Drain-3 min. As the basket rotates, the now dirty LC02 and most residual C02 gas is
drained to the recycler. The C02 draining from the vessel cools the parts.
Drain time also depends on the pressure of the air compressor driving the
pump. Typical rotation speeds for this step are 150-250 rpm.
12

-------
•	Purge-3 min. Warm air is pumped through the vessel to warm the parts so moisture in
the air won't condense on them when the vessel is opened. The basket
rotates at about 150-250 rpm and stops at the end of the purge cycle.
•	Vent- 30 sec. The last 5-10 percent of the C02 and the warm air in the cleaning vessel are
vented to the atmosphere. The vessel is ready to be opened.
Each total cleaning cycle requires about 40 minutes from the time the parts are immersed
in the HOP to completion of the LC02 cleaning and removal of parts from the machine. To clean
the most parts in the shortest amount of time, the second batch of parts should be cleaned in the
HOP while the first batch is being degreased in the L€02 and so on.
While parts are being degreased in the cleaning vessel, the dirty CO, from earlier batches
is recycled. The LC02 drained from the cleaning vessel is held in a "dirty" storage tank. A small
amount of heat is added to the dirty LC02 storage tank to vaporize the liquid. This is similar to
vapor distillation of conventional solvents. Oils and particles are left behind in the dirty storage
tank where they can be drawn off, filtered, and returned to the process for reuse or sent to proper
waste disposal. The C02 gas is driven out of the dirty tank and through a condenser that
recondenses the gas into a clean liquid. The clean LC02 is held in the larger tank for subsequent
cleaning cycles.
2.6 Advantages
The main advantages of LC02 cleaning are:
•	C02 is not a VOC or HAP and is not regulated by the EPA. It is listed in the EPA's
Significant New Alternatives Program (SNAP) list of acceptable materials for precision
cleaning and is exempt from permits under Rules 201 and 203 in the California South
Coast Air Quality Management District (SCAQMD).
•	LC02 penetrates into blind holes, crevices, and small tubes where other liquids will not,
•	C02 can easily be recycled, reducing material costs,
13

-------
•	LC02 changes to gas phase when the pressure is released during the recycling process,
leaving the removed contaminants behind as a concentrated waste stream,
•	parts do not have to be dried,
•	LC02 cleaning is carried out near room temperature,
•	systems designed for LC02 can be less expensive than supercritical systems,
•	no water or sewer connections are needed, and
•	C02 does not have to be replaced like spent solvent, unless it becomes contaminated with
a material that cannot be separated during the recycling. The most likely materials that
wouldn't be separated are volatile solvents with low boiling points (see Disadvantages).
2.7 Disadvantages
The main disadvantages are the result of physical laws whereby C02 exists as a liquid only
when held under elevated pressures. This leads to the following requirements.
•	A pressure vessel, compressor, pump, and other equipment are necessary to maintain the
pressures. They must be designed to operate under these conditions in accordance with
mechanical, safety, and electrical standards.
•	The pressure vessel size limits the size of parts that can be cleaned.
•	The process must be run in batch rather than continuous mode.
•	Extra time is necessary to pressurize and depressurize the system, which adds to the
overall cleaning cycle time,
•	Parts must be able to withstand isostatic pressures of 800-900 psi to be cleaned with this
technique. This is not a problem for most metal or plastic parts, but can be a limitation for
some hermetically sealed packages on circuit boards.
The following disadvantages are not related to the pressure requirements:
•	C02 is a greenhouse warming gas (Commercial supplies are obtained by recycling by-
products from other processes, so its use results in no net addition to the atmosphere).
Five to 10 percent of the cleaning vessel volume of C02 is released to the atmosphere each
cycle.
•	The process is dependent on the HOP prctreatment for many applications.
14

-------
•	Volatile solvents removed from parts into the LCO, are difficult to separate from the C02
in the recycling process. Some solvents could be removed with the addition of special
equipment to the recycler. The rest would be mixed with the "clean" CO, and eventually
released to the atmosphere during the venting portion of the cleaning cycle.
•	Plasticizers and other soluble organics may be leached from plastics and elastomers.
•	Some elastomers may swell or warp.
2.8 Projected Costs
Equipment Costs
The equipment cost depends on the size of the cleaning vessel. The size of unit used at
the demonstration costs about $175,000 including the recycler. Larger standard units cost up to
$350,000. A centrifugal spinner that holds a similar sized basket (12 inch diameter by 12 inches
deep) costs from $3000-5000. A tank with a heater for the HOP pretreatment costs less than
$2000. The equipment and materials costs are summarized in Table 2.
Materials Costs
Carbon dioxide costs range from about $0.20/gal for bulk to a high of about $0.85/gal for
COz in cylinders. Each cylinder used in this demonstration cost about $28 and held about 40
gallons of C02, about $0.70/gal. It is estimated that 5-10 percent of the C02 in the degreasing
vessel will be lost in each cycle. The vessel in the demonstration was 10 gallons, so losses would
be expected of 0.5-1 gallons per cycle. At $0.85/gal, this would mean a loss of $0.43-$0.85 per
cycle in C02.
HOP mixture is about $8.9G/gallon in 55 gallon drums It is not very volatile, so the major
losses are the material dragged out on the parts. If the oil removed by the CO, degreasing were
returned to the HOP tank, these losses would be minimal. Depending on the contaminants
removed from the parts, the oil may need to be filtered, or may need to be disposed of
periodically. If the contaminant removed is a hazardous waste, the oil also must be a disposed of
as hazardous waste. Even in this case, the amount of material to be disposed of is greatly reduced
from liquid-based cleaning processes.
15

-------
TABLE 2. COSTS FOR EQUIPMENT AND MATERIALS
Equipment or Material
Cost
Equipment

C02 cleaning unit and recycler
$ 175,000-350,000
Centrifugal spinner
$4,000
HOP tank with heater
$ 2,000
Materials

co2
$ 0.20-0.85/gal
HOP mixture
$ 8,90/gal
3.0 CHOOSING THE DEMONSTRATION SITE AND PROCESS
RAFB volunteered to act as host site for a technology demonstration of LC02 degreasing
and worked with engineers from EPA and RTI to choose appropriate processes needing
alternatives. The main selection criteria were:
•	the current RAFB process must involve cleaning with a hazardous solvent,
•	there must be some difficulty in switching to common alternatives such as aqueous
cleaning,
•	the parts to be cleaned must not be damaged by high pressure,
•	the contaminants must be removable with LC02, and
•	the personnel in the area must be interested in evaluating new technologies.
RAFB personnel from many areas on the Base met with EPA and RTI representatives to
discuss current cleaning problems. The team then visited several facilities to evaluate the most
likely candidates. One of these areas was the plating shop in Building 142. This shop was still
using two TCA vapor degreasers to degrease parts prior to anodizing and plating processes.
Both vapor degreasers had cooling coils, lids, freeboard areas, and hoists for emissions reduction.
16

-------
Even with these precautions, reported usage of TCA was 8900 pounds from January 1, 1995 to
August 29, 1995. This was the highest usage of hazardous solvents in the building for this time
period. The next highest solvent emission was about half that of TCA, No other reportable
solvents were used in quantities greater than 250 pounds for this same time period.
RAFB personnel considered elimination of these TCA degreasers a high priority. Many
parts formerly degreased in TCA could be cleaned with aqueous solutions, and several power
washers were being installed to handle the cleaning of these parts. However, a few parts were not
being adequately cleaned by the aqueous machines.
One of these was the metal tubing used for F-15 fuel systems. Tube stock of various
diameters of aluminum and titanium are cut into sections and bent into the exact shapes needed to
replace parts during aircraft maintenance, To aid the bending process, thick drawing compound is
smeared in the tubes and pushed through. The drawing compound is a combination of tallow,
fatty acids, and alkaline salts. The bending process leaves a heavy residue of drawing compound,
metal chips, shavings, and other shop debris inside and outside the tubes.
The current cleaning process for these tubes starts with vapor degreasing in one of the two
TCA degreasers, followed by acid brightening and aqueous cleaning. Aqueous cleaning in the
power washers alone was not adequately cleaning the tubes. The various diameters and shapes of
the tubes were making it difficult to get enough flow through them for cleaning and drying. This
made them a good candidate for LCO, because of its ability to penetrate through small spaces.
The next step was to conduct preliminary laboratory tests to determine if LC02 was capable of
removing the drawing compound.
4.0 LABORATORY FEASIBILITY TESTING
Before committing to a full scale demonstration on site at RAFB, samples of the tubes and
drawing compound were cleaned at the LC02 equipment manufacturer's facility and tested for
cleanliness at RTI. The cleaning conducted at the manufacturer was carried out using early
prototype equipment that was functional but not as fast or effective as the next generation to be
produced. Some features that were to be added to the next generation of equipment, such as
17

-------
bidirectional spinning of the parts basket during cleaning, were expected to increase the cleaning
capability of the machine. However, the prototype equipment was considered to be good enough
to establish feasibility.
Sample tubes of various lengths and diameters were sent to the LCOz equipment
manufacturer along with a container of the drawing compound. These first parts were used to
develop a process that produced visually clean parts. RTI then brought additional tubes for
cleaning and returned with them to RTI for verification by NVR testing.
4.1 Development of the NVR Test Procedure
To begin the process of choosing an appropriate solvent for the NVR test, the MSDS for
the drawing compound was examined. Based on the chemical nature of the drawing compound
components, seven solvents were chosen for testing: TCA, perehloroethylene (PERC), acetone,
ethyl acetate, methyl ethyl ketone (MEK), n-methyl pyrrolidone (NMP), and hexane.
Small amounts of drawing compound were placed in glass vials, and 20 ml of one of the
solvents were added to each one. The vials were observed for changes in the drawing compound.
After 24 hours, the acetone, ethyl acetate, MEK, and NMP had not dissolved any significant
amount of the drawing compound. The TCA broke the compound into smaller globules, some of
which floated on the surface, and some of which stuck to the glass vial walls, indicating poor
solubility. Hexane partially dissolved the material, but some solids dropped out on the bottom of
the vial. The only solvent to completely dissolve the drawing compound and keep it in
suspension was PERC. The next step was to determine that all the drawing compound could be
removed with PERC using the proposed NVR test procedure.
Drawing compound was added to precleaned and preweighed aluminum weighing dishes.
The dishes were then stripped twice with PERC, and the stripped material was collected in
separate clean dishes and compared to the original weight of drawing compound. The results
were 1) the drawing compound was removed completely (to less than 0.01 mg) from the dishes
with the first solvent strip and 2) the material was recoverable by the NVR test procedure.
18

-------
4.2 Results of Feasibility Tests
Once the NVR procedure had been verified, the tubes from the feasibility tests were
examined both visually and by NVR testing. These included tubes cleaned at the LC02 equipment
manufacturer, tubes cleaned at RAFB by the TCA process, and dirty tubes contaminated with the
drawing compound.
Inspection of the tubes immediately alter cleaning at the equipment manufacturer showed
that the LC02-cleaned tubes were visually cleaner than tubes cleaned with TCA. No visible
drawing compound remained on any tubes, but the TCA-cleaned tubes were covered inside and
out with a fine yellow dust. There was no dust visible on the LC02-cleaned tubes. Samples of
each set of tubes were sealed in nylon bags and returned to RTI for NVR testing.
The NVR results listed in Table 3. The maximum acceptable level of for NVR on
Category 1 and II oxygen equipment at RAFB is 0.003 g/ft2. None of the processes, including the
current TCA process, achieved this level. The LCOz process removed the drawing compound as
effectively as the TCA process from some of the tubes but not all of them. Tubes with smaller
diameters or more bends were not cleaned as well as larger diameter, straight tubes.
Since some of the tubes from the LC02 process were cleaned as well as the current TCA
process, even while using this prototypical equipment, it was decided that the results were good
enough to warrant further testing and demonstration at RAFB once the improved LCOz
equipment was available.
W

-------
TABLE 3. NVR RESULTS FROM THE LABORATORY FEASIBILITY TESTS
Part Descriptions
Nonvolatile
Residue (g/ft2)*
Tube Diameter
(o.d., inches)
Tube Length
(inches)
Comments
Dirty
1.612
1.25
8.22
Straight

2.513
1.25
7.31
Straight

2.808
1 *<25
6.69
Straight
TCA cleaned
0.008
1.25
10.72
Straight

0.005
1.25
10.44
Straight

0.008
1.25
9.94
Straight

0.005
1.25
8.34
Straight

0.008
1.25
6.72
Straight
HOP/LCOz cleaned
0.006
1.25
10.44
Straight

0.007
1.25
6.50
Straight

0.097
0.75
11.62
Right angle

0.007
0.75
11.00
Straight

0.022
0.75
9.12
45° bend, bits
of metal in pan

0.006
0.75
9.00
45° bend

0.022
0.50
8.31
Straight

0.029
0.50
4.69
Straight

0.024
0.50
4.56
Straight

0.029
0.375
12.62
Many bends

0.181
0.375
12.25
4 bends

0.021
0.375
11.19
Straight

0.010
0.375
11.00
Straight
*NVR is reported in grams/square foot because it is the unit used in the parts specifications.
20

-------
5.0
ON-SITE DEMONSTRATION
5.1	Introduction
The demonstration at RAFB was set up in an air-conditioned hangar in Building 137, Bay
4. The air conditioning was important because the efficiency of the condenser on the LC02
equipment depended on the temperature of the surrounding air. The demonstration unit would
not condense the used C02 at ambient temperatures higher than 80°F. Units with larger
condensers can be manufactured for use in areas with higher ambient temperatures.
The LC02 equipment, HOP tank, and spin dryer were arranged in the hangar in a U-shape
so that, if desired, parts easily could be soaked in the HOP mixture and spun to remove excess
material before C02 degreasing. A large table was provided to hold samples of dirty and cleaned
parts so demonstration visitors could view various applications for the technology.
5.2	Range of Parts Tested
Other parts besides tubes had been proposed for the demonstration, and even more were
suggested as feasibility tests and planning for the on-site demonstration proceeded. It was
decided that anyone with a promising application could contact the primary RAFB team member,
who would then coordinate the C02 cleaning and user inspection of the parts. Cleaning experts
from the equipment manufacturer and RTI agreed to remain on site for the entire duration of the
demonstration to ensure the best possible operation of the system and develop processes for the
parts as they were brought in. As word of the demonstration spread, many other parts were
brought to Building 137. These parts were cleaned if they were appropriate for the LC02
process, but the primary focus of the demonstration was metal tubes.
Prior to cleaning, parts were inspected and photographed, and information was collected
on the contaminants, current cleaning process, inspection criteria, and usage of the part. After
cleaning, the parts were rephotographed and inspected by at least three people, including the user
when possible. When enough parts were available, samples of dirty parts, LC02-cleaned parts,
and parts cleaned by the current process were sealed in nylon bags and returned to RTI for NVR
testing. Parts cleaned during the demonstration included:
21

-------
•	Aluminum and titanium tubing pieces for F-15 fuel and breathing-oxygen systems,
•	Brass filters (from the C-130 propeller assembly) contaminated with black grime and
used hydraulic fluid,
•	Small pieces of aluminum honeycomb core contaminated with light oils,
•	Propeller assembly parts for C-l 30 pitch lock regulators (mechanical gears and machined
parts) contaminated with oil and black particulate wear products,
•	Small precision bearing assemblies for avionics,
•	An electronic test circuit board,
•	A reflectometer for avionics, and
•	Rags contaminated with shop dirt, drawing compound, grease, and solvents.
5.3 F-15 Fuel System Tubes
These were the original parts chosen for the demonstration and the ones tested in the
laboratory feasibility tests. Most of the tubes were aluminum and titanium, with outside
diameters from 3/8 to 1-1/2 inch (Figure 5). In the tubing shop, long sections of tube stock are
cut and bent into parts for repairing fuel systems and breathing oxygen systems for the
F-15 aircraft. Drawing compound containing fatty acids, tallow, and potassium salts is applied to
the inside of the tubes before bending. After bending, the tubes are heavily contaminated with
drawing compound, metal chips, and shop dirt.
FIGURE 5. TYPICAL ALUMINUM TUBES CLEANED DURING THE DEMONSTRATION.
22

-------
The current cleaning process after tube bending is to:
•	vapor degrease in TCA,
•	brighten by dipping in a mixture of dilute hydrofluoric and nitric acids in water
(MIL-C-38334), and
•	rinse in hot water.
The tubes are then anodized immediately and air dried.
Visual Inspection
All parts were visually inspected by at least three people unless the parts were cleaned
especially for RTI's NVR test. In that case, the parts were transferred directly from the LC02 unit
to nylon bags and sealed.
Seven batches of tubes were cleaned during the on-site demonstration, all heavily
contaminated with drawing compound. Before LC02 degreasing, they were agitated for less than
five minutes in the HOP tank at temperatures from 115-150 °F (46-66 °C). Warmer oil
temperatures dissolved the drawing compound more quickly, but lower temperatures also were
effective. All visually inspected tubes appeared to be completely free of drawing compound after
the HOP/LCO, process. It removed large metal chips and debris but did not remove all the fine
particles. In all cases, a haze of particles or small piles like dried puddles could be seen inside
the tubes. In some, these particles could only be seen with a small flashlight pointing through the
tube towards the viewer.
In one test, the HOP pretreatment step was deliberately skipped. The tubes cleaned only
with LC02 had easily visible residues of white, soap-like material. The LC02 used without the
HOP removed metal chips and the oily portion of the grease, but not the soap-like portion.
One aluminum tube, about 18 in. (7.2 cm) long and 2 in. (0.80 cm) diameter with several
bends was cleaned with the H0P/LC02 process for the supervisor of the tubing shop. After
inspecting the part visually with the small flashlight, he stated that it was not as clean as the
current process because he could see a haze of particles left on the inside of the tube. He later
added that sometimes the parts must be sent through the current process twice to meet cleanliness
requirements.
23

-------
Five aluminum tubes cleaned by the H0P/LC02 process were sealed in a nylon bag and
taken back to the plating shop to be anodized. The parts were inspected by shop personnel who
verified that the tubes were sufficiently clean for the anodizing process. After anodizing the
interior and exterior of the tubes, they were reexamined by the shop personnel. The five tubes
also passed visual inspection after anodizing. Four of them were kept for further tests by RAFB
personnel. No results have been reported from these additional tests.
NVR Test
Several sets of tubes were sent to RTI for NVR testing. This included parts cleaned by
the current TCA process, parts contaminated with drawing compound and cutting debris, and
tubes cleaned in several different batches in the H0P/LC02 process. They were tested with the
same method as those in the laboratory feasibility studies. The results of the NVR tests are
shown graphically in Figure 6, and averages and standard deviations are listed in Table 4. Full
data for all the tubes cleaned during the demonstration listed in Tables 5 A-C.
0.018
0.016
J °014'
° 0.012-
©
I 0.01-

o 0.004
z
0.002
0
FIGURE 6. GRAPH OF NVR RESULTS FOR TUBES.

iDirty Tubes = 5.462 g/sq.ft. j








-








-





-




TCA LC02 Batch 10 Batch 36 Batch 37
Process Type
24

-------
Figure 6 shows averages and standard deviations for four sets of tubes tested for NVR:
one cleaned with TCA and three separate sets cleaned with the IIOP/LC02 process. The average
NVR results for all three LC02 batches were as low as for the TCA tubes and the LC02 standard
deviations were slightly lower. The NVR levels measure mainly the drawing compound, other
organics, and large metal chips. Small particles are more easily seen than weighed. They do not
weigh enough to be significant in this test unless present in large amounts.
It is interesting to compare the average NVR results for the TCA-cleaned tubes tested
during the lab feasibility tests to the TCA-cleaned tubes tested during the on-site demonstration
(Table 4). The average NVRs for the two sets of tubes were similar (0.007 g/sq. ft. during the
feasibility tests vs. 0.009 g/sq. ft. during the on-site demonstration). This shows that the TCA
degreasing process remained fairly consistent during this period.
TABLE 4. AVERAGE VALUES AND STANDARD DEVIATIONS FOR TUBES

Average NVR (g/ ft.2)
Standard Deviation (g/ ft.2)
Dirty Tubes
5.462
2.471
TCA Cleaned
0.009
0.009
C02, Batch 10
0.005
0.005
C02, Batch 36
0.008
0.004
C02, Batch 37
0.003
0.002
Also, as expected, the HOP/LCO , equipment demonstrated on-site performed better than
the prototype equipment used in the laboratory feasibility tests. The NVR results show the on-
site cleaned tubes to be much cleaner than those cleaned in the prototype LC02 equipment, even
though the parts were more heavily contaminated than in the feasibility tests. The standard
deviations were also much lower during the on-site tests.
Tables 5A-C contain NVR data for all tubes tested during the on-site demonstration and
the size and shapes of those tubes. There is no correlation between NVR and diameter or length
or type of bends in the tubes. Tubes with the smallest diameters and most bends have both the
highest and lowest NVR values out of the samples tested.
25

-------
TABLE 5A. NVR RESULTS FOR TUBES CONTAMINATED WITH DRAWING COMPOUND
Part
Nonvolatile Residue
Diameter,
Length
Tube Shape
Description
fs/ft2)
o.d. (in.)
(in.)

Dirty
4.848
0.625
14.56
Straight
i«
7.523
0.625
11.19
Straight
a
5.917
0.625
10.38
Straight
66
7.964
0.625
10.25
Straight
u
5.669
0.625
10.00
Straight
it
4.553
0.625
9.75
Straight
U
9.382
0.625
9.19
Straight
a
5.224
0.500
8.50
Straight
a
2.183
0.375
4.81
Straight
66
1.360
0,313
3.81
Straight
TABLE 5B. NVR RESULTS FOR TUBES CLEANED WITH TCA
Part
Nonvolatile Residue
Diameter,
Length
Tube Shape
Description
(g/ft2)
o.d. (in.)
(in.)

TCA Cleaned
0.003
1.00
18.50
2 bends, one over 90°
ct
0.001
1.00
13.25
90° bend
u
0.003
1.00
12.75
90° bend
66
0.005
1.00
12.50
"S" shaped
U
0.006
1.00
12.50
2 bends, both 90°
66
0.014
0.75
17.00
2 bends, one 90°
66
0.003
0.75
14.25
2 bends, both 90°
66
0.019
0.75
12.75
2 small bends
66
0.028
0.75
10.50
90° bend
26

-------
TABLE 5C. NVR RESULTS FOR TUBES CLEANED AT RAFB WITH HOP/LC02
Part
Nonvolatile Residue
Diameter,
Length
Shape
Description
(g/ft2)
o.d. (in.)
(in.)

Batch 10
<0.001
0.625
10.62
Straight
a
0.006
0.625
9.69
Straight
a
0.011
0.625
7.44
Straight
a
0.002
0.625
7.38
Straight
<<
0.005
0.50
9.75
Straight
U
0.003
0.50
9.19
Straight
U
0.012
0.25
13.75
Straight
ti
<0.001
0.25
14.12
2 bends, like twisted "Z"
Batch 36
0.003
2.00
10.25
1 bend

0.005
1.00
17.25
2 bends, 2 axes, 1 over 100°
u
0.006
0.625
18.75
2 bends, 1 about 270°

0.006
0.625
17.00
2 bends

0.014
0.375
12.50
4 bends
U
0.012
0.375
10.25
3 bends, 1 about 100°
Batch 37
0.002
2.00
9.62
120° bend
«<
0.001
0.75
10.50
90° bend
u
0.002
0.625
16.50
Bent in 2 axes, 1 over 90°

0.001
0.50
15.00
3 bends, 2 axes
a
0.006
0.50
14.50
"S"shaped
a
<0.001
0.375
14.25
2 bends, 1 about 100°
a
0.006
0.25
12.00
2 bends, 1 over 100°
27

-------
Conclusions
The combination of the NVR data, the visual inspections, and the anodizing results
indicate that the H0P/LC02 process has good feasibility to be developed for this part. It is
capable of removing drawing compound, large particles, and shop debris from the tubes. Better
small particle removal is needed in the H0P/LC02 process.
The HOP/LC02-cleaned tubes all have small particles remaining inside after cleaning but
pass visual inspection after anodizing. The anodizing process appears to be removing the small
particles. If these parts were going to be used for breathing-oxygen equipment, they would be
anodized only on the outside. If used for fuel lines, they would be anodized inside and out.
More detailed tests for organic and particle removal would need to be performed to determine if
the parts would be acceptable for breathing oxygen equipment and fuel system use after going
through the rest of the cleaning and anodizing process.18
5.4 Steel Bolts
Steel pylon bolts are removed from aircraft during repair operations, cleaned, inspected,
and reused. When removed, they have heavy deposits of grease and embedded dirt (Figure 7).
FIGURE 7. BOLTS CONTAMINATED WITH HEAVY GREASE.
28

-------
The current cleaning process for pylon bolts is to:
•	soak in P-D-680 (Type III), a mixture of petroleum distillates (MSDS in Appendix B),
•	scrub with brushes to remove all visible grease,
•	vapor degrease with TCA, and
•	abrasive blast with walnut shells to remove rust and carbon deposits.
Then they were inspected visually and sent for nondestructive inspection (NDI).
The H0P/LC02 demonstration process was started by soaking the bolts in the HOP
mixture for 30-80 minutes at 150 °F (66 °C) with stirring or ultrasonic agitation, but no hand
scrubbing. Most of the heavy grease was removed by the soaking and stirring action. The grease
packed into the comers inside the cap heads of the bolts would not come out with stirring alone.
Once most of the grease was removed with the HOP, the bolts were cleaned in the LC02. This
completely removed all visible traces and odor of oil, but some grease remained in the bottom of
the cap heads. The rust and paint were not removed, but some paint was flaking off. The next
step in the normal cleaning process for these parts would be walnut shell blasting to remove the
rust and paint. The bolts were returned after I .CO, degrcasing for inspection by RAFB
personnel. According to later reports, they were cleaned as well as those from the P-D-680 and
TCA portions of the current process.
NVR Test
Samples of the dirty bolts and those cleaned with the H0P/LC02 process were sealed in
nylon bags and returned to RTI for NVR testing. Bolts that had been degreased in the current
process but had not been all the way through the walnut shell blasting process were not available
because they were in short supply and were needed on the aircraft. Therefore, comparisons could
not be made between NVR results from the current TCA and the H0P/LC02 processes, but only
general observations about how much grease remained on the parts.
The PERC used in all NVR tests so far would dissolve small amounts of grease and
loosen large clumps so that it would fall off the bolts, but not dissolve them completely. Two of
the most heavily contaminated dirty bolts were soaked in TCA after being soaked in PERC to see
if the grease was more soluble in that solvent, but the grease was not very soluble in TCA either.
29

-------
Ultrasonic agitation was added to the NVR test procedure for these parts to help break down the
grease in the cap heads of the bolts. The NVR values in Table 6 include the grease removed by
both PERC and TCA. Results are reported as milligrams stripped from each bolt.
TABLE 6. NVR RESULTS FOR STEEL PYLON BOLTS
Part Description - Bolts
Nonvolatile Residue (rag)
Comments
Dirty
653.40

£4
1234.15
Stripped in PERC and TCA
a
898.11

a
800.72

a
354.02

C02 Cleaned
<0.01


0.82


68.78
Stripped in PERC and TCA
u
276.12
Stripped in PERC and TCA
u
5.10

Conclusions
NVR tests showed varying levels of cleanliness achieved, which was due mostly to grease
remaining inside the cap head. The HOP process removed grease slowly with stirring or
ultrasonic agitation. Scrubbing or spraying with hot oil would increase the cleaning speed and
effectiveness. Grease left after HOP cleaning was not removed by LC02. The grease was also
not very soluble in TCA or PERC, especially without the addition of ultrasonics.
The HOP portion of this process is similar to the P-D-680, and the LCOz is similar in
function to the TCA step. Table 7 shows a comparison of the P-D 680 to the hydrocarbon oil
portion of the HOP mixture. Both compounds are petroleum distillate mixtures with high boiling
points and high flash points. It is likely that the P-D-680 could be used prior to LC02 degreasing
instead of HOP oil or be replaced with the HOP oil. Either way, the TCA could be eliminated.
30

-------
Other processes that might be effective in cleaning this part are steam stripping, baking in
an oven or abrasive tumbling. Since this part undergoes an abrasive process anyway, it might be
worth while investigating tumbling in walnut shells, corn cobs or other mild media to remove the
grease. This also might be a case where supercritical C02 dissolves the grease better than LC02.
TABLE 7. COMPARISON OF P-D-680 AND HOP HYDROCARBON OIL
Property from MSDS Sheets
P-D-680
HOP Oil
Chemical Family
Petroleum Distillate
Petroleum Oil
CAS#
64742-46-7
64742-53-6
Boiling Point (°F)
430-530
400
Flash Point (°F)
200-210
300
Specific Gravity
0.82-0.83
0.88
5.5 Brass Filters
These parts are flat disks of two layers of brass wire mesh, about 1.5 inch (3.8 cm)
diameter, with a 0.5 inch (1.3 cm) hole in the center (Figure 8). They are used in the propeller
assembly to filter hydraulic fluid. The contamination to be removed is hydraulic fluid, aircraft
grease, and thick carbon deposits.
FIGURE 8. BRASS FILTERS, CONTAMINATED (LEFT), CLEAN (RIGHT).
31

-------
The current cleaning process for brass filters is to;
•	soak them in P-D-68Q (Type HI) overnight,
•	hand scrub in P-D-680,
•	agitate with ultrasonics in water-based general purpose detergent,
•	rinse in water, and
•	dry in an oven.
The cleanliness criteria for these filters is that they are clean enough when light can be
seen through all the open areas of the mesh. Even though this part is not cleaned with TCA, the
owners were interested in a new process because the current one takes too long and does not
clean well. Black residue is visible on the surface of the filter after the current cleaning process.
For the demonstration, the first filters were cleaned only with LC02, which removed
hydraulic fluid but not carbon build-up. Then several parts were presoaked overnight in the HOP
mixture at about 115 "F (46 °C) with 30 minutes ultrasonic agitation but no hand scrubbing. It
was later determined that the parts currently were being cleaned by overnight soaking and hand
scrubbing with a brush. The tests were repeated with a 0-5 minute presoak and the addition of
hand scrubbing in the HOP prior to the LC02 degreasing. This greatly improved removal of the
crusty black carbon buildup and produced parts that passed visual inspection.
Visual Inspection
Filters cleaned without hand scrubbing were still covered with crusty black residue left
behind when the oil and grease were removed, and light could not be seen through the mesh.
Once the demonstration process was changed to include hand scrubbing in HOP mixture (with a
0-5 minute presoak), the filters came out visually clean and shiny, and light could be seen
through all open portions. The owners of the parts were very impressed. Parts cleaned by the
current process were brought to the demonstration site for comparison. The filters cleaned by the
current P-D-680 process had black residue blotches on the surfaces; the H0P/LC02 cleaned parts
did not. To determine how difficult it would be to clean these parts on a larger scale, the RAFB
team representative cleaned a set of 30-50 filters. He reported that the process was not difficult.
Total hand scrubbing time was about 25 minutes, the LCOz process took 20 minutes.
32

-------
Additional batches of filters were cleaned with the H0P/LC02 process and returned to
RTI for NVR testing. Of those parts cleaned for testing at RTI, seven filters were blown out with
shop air between the HOP and LC02 processes to see if this would dislodge any of the
contamination between the mesh screens. Seven other filters were not blown out Samples of
dirty parts and parts cleaned by the current process were sent to RTI for comparison. All parts
were returned to the RAFB team representative after testing.
"V ¦»-» IT* 
c
o
45-
40
30
25
20
15
10



Dirty filters — 289.02 mg


















-



1	1		 ' !
;			 J
Current C02 /Air Blown C02/No Air
Process Type
C02/No Air
FIGURE 9. GRAPH OF NVR RESULTS FOR BRASS FILTERS.
33

-------
TABLE 8. INDIVIDUAL NVR RESULTS FOR BRASS FILTERS (mg/filter)
Dirty
Current
C02 Cleaned,
C02 Cleaned,
C02 Cleaned, Not Blown
Filters
Process
Blown Out
Not Blown Out
Out (Another Batch)
503.34
21.58
12.38
17.58
36.80
54.96
12.25
14.59
29.84
13.60
487.77
14.22
14.08
18.25
6.13
196.79
14.95
19.86
39.37
40.94
174.73
22.66
17.44
18.01
30.77
559.35

18.40


266.64

20.88


527.28




63.95




55.36




Note: Each column represents a different set o:
statistic.
items cleaned to de
ine the applicable parameter
TABLE 9. AVERAGES AND STANDARD DEVIATIONS
FOR BRASS FILTER NVR TESTS (mg/filter)

Average NVR
Standard Deviation
Dirty filters
289.02
210.13
Cleaned by current process
17.13
4.68
C02 cleaned, blown out
16.80
3.18
C02 cleaned, not blown out
24.61
9.73
C02 cleaned, not blown out (another batch)
25.65
15.09
Results
The HOP was most important for removing the black crusty carbon material, and the
LC02 removed the oily portion, hydraulic fluid, and HOP mixture residue. The H0P/LC02
seems to be viable for cleaning this part if used with an air blow-out. This air blow-out might be
34

-------
replaced by fixturing the parts in the HOP for cleaning with turbulent liquid spray under the
surface or ultrasonic agitation. Using the H0P/LC02 process would eliminate the overnight
soaking in P-D-680, shortening the cleaning cycle. In addition, replacing the water and detergent
cleaning and oven drying with LC02 would eliminate the water waste stream and problems from
incomplete drying.
Another option would be to investigate cleaning these parts in hydraulic fluid alone or
prior to LC02. Sometimes the fluid that carries the contaminant to the part can also remove it.
Used hydraulic fluid might be clean enough to use as the cleaning fluid before being disposed of,
eliminating waste from the P-D-680 process.
5.6 Large C-130 Aircraft Bearings
These parts are removed from C-130 aircraft, cleaned, checked, and repacked with grease
for reuse. The grease on these bearings is very heavy (Figure 10).
FIGURE 10. BEARINGS BEFORE (BOTTOM) AND AFTER (TOP) H0P/LC02 CLEANING.
35

-------
The current cleaning process for these bearings is:
•	30 minute soak in P-D-680 (Type III),
•	hand scrub in P-D-680,
•	rinse in P-D-680, and
•	air dry.
During the demonstration, these parts were soaked in the hot HOP oil for about 40
minutes. Some of the grease was removed in clumps, but it was not very soluble in the HOP oil.
The bearings were then sonicated in the hot HOP for 35 minutes, but this did not remove all the
grease. Then they were hand scrubbed with warm HOP oil before being degreased in LCO,.
This produced better results and removed the visible grease.
Visual Inspection
The grease was not very soluble in HOP mixture. After LC02, all of the heavy grease
was gone, but a thin film of soap-like material was left under the roller bearings. The assembler
said it looked clean enough to be used, but was not as clean as parts cleaned with P-D-680. This
was a curious result since both P-D-680 and HOP oil are petroleum hydrocarbon mixtures with
similar chemical properties. It may have more to do with the scrubbing technique than the
chemistry. Regardless, neither HOP nor LC02was capable of removing all components of the
grease. This is not a good application for H0P/LC02 cleaning.
5.7 Aluminum Honeycomb Core
Three pieces of aluminum honeycomb core were brought for cleaning, each roughly 3 in.
x 6 in. x 1 in, thick (7.6 x 15.2 x 2.5 cm). One piece had black felt tip marks on the edge (Figure
11). The main concerns were whether the rotation of the parts in the machine would damage the
parts and whether the process would clean the honeycomb well enough for bonding aluminum
skins to it.
The current cleaning process prior to bonding is vapor degreasing with TCA.
36

-------
FIGURE 11. ALUMINUM HONEYCOMB CORE TEST SAMPLES.
Visual Inspection
There were no visible signs of contamination on the parts either before or after cleaning.
The black ink was not removed. There was no visual sign of damage to the honeycomb.
WR-ALC personnel supplied eighteen additional samples of aluminum honeycomb core
for further testing. The eighteen pieces were visually inspected and wrapped in clean, wax-free
Kraft paper. Six pieces were selected at random and cleaned by soaking in HOP oil at 150 °F
(66 °C) for 20 minutes, followed by LC02 degreasing. Six other pieces were cleaned by the
current TCA vapor degreasing process, and six were not cleaned at all. Anodized pieces of
sheetmetal were bonded to both sides of each piece, and the sandwiched panels were subjected
to peel testing for adhesion by Military Specification MIL-A-25463B, "Adhesive, Film Form,
Metallic Structural Sandwich Construction," Section 4.6.1 on Sandwich Peel Strength. This
specification is similar to ASTM D-1781. The results of the peel test are shown below. The
minimum passing value for the adhesive used in this test is 28 inch-pounds/inch. All of the
samples were well above this limit. RAFB personnel reported that the values were all good and
37

-------
the differences were not significant. The most interesting result actually was that the uncleaned
parts did so well. Uncleaned parts are not normally tested, so it was not known whether their
results were normally in this range.
TAUT t? 1A DCCTTT TO	A TMJRCTA\T TI7CTC
lArSLJi J.U. KcoULio rKUM AL^ra.cM
-------
During the demonstration, the parts were soaked in HOP oil with air agitation for five
minutes, followed by LC02 degreasing.
i
FIGURE 12. STEEL PITCH LOCK REGULATOR PIECES.
Visual Inspection
The process removed the hydraulic fluid and grease, but a black carbon residue remained
on some pieces that could be moved with cotton-tipped swab. There were also a few particles on
the inside of the smooth bore that were removed easily with C02 snow. Carbon dioxide snow
blasting was not officially part of this demonstration. However, C02 snow equipment had been
provided by DEFLEX in case it was needed. The snow gun was supplied with C02 from the
same tanks as the LCOa equipment. The snow was effective at removing the remaining particles
from the inner bore of this part.
One of the gears with black residue was recleaned by first soaking it in HOP oil for 2-3
hours, scrubbing with a brush, then degreasing again in LC02. Some black debris still remained
in the crevices that was moveable with a cotton-tipped swab. The owners of this part said it did
not meet their cleanliness criteria of no moveable contamination, but they currently used Scotch
Bright® in a subsequent abrasive step to reach the necessary cleanliness level.
39

-------
5.9 Flap Jack Screw Yokes
The yokes are tools used to repair F-15 aircraft. They are assemblies of machined steel
parts and contain a sealed bearing (Figure 13). The parts brought to the demonstration area were
heavily crusted with dirt, grime, and heavy oil. Under the contamination, the parts were partially
covered with two other layers, a chromate conversion coating and silver paint.
FIGURE 13. CONTAMINATED ALUMINUM FLAP JACK SCREW YOKE PARTS.
The current cleaning process for flapjack screw yokes is hand scrubbing with Marsol
over a drip pan (Marsol MSDS is in Appendix B). They are clean enough when there is no
moveable contamination.
During the demonstration, half of the parts were soaked in room temperature HOP oil
overnight, and half were not soaked but were hand-scrubbed in HOP oil. Both sets of parts were
then cleaned in LC02.
Visual Inspection
The oil and grease were removed from all parts. The parts that were not hand-scrubbed
had areas of black residue in a few places that could be moved with a cotton-tipped swab. Soap
40

-------
from the grease in the bearings was visible on the outside of the seals. The owner of the parts
said both the scrubbed parts and the parts that were not prescrubbed were clean enough for use.
FIGURE 14. AVIONICS BEARING ASSEMBLY.
5.10 Small Avionic Bearing Assemblies
This part is a stainless steel bearing assembly [approximately 4 inches (10 cm) in
diameter], and is typical of the mechanical parts cleaned in the avionics shop (Figure 14). Most
assemblies are disassembled prior to cleaning. Common contaminants are lubricants, oils, dirt,
and fine sand.
Two years ago, these parts were routinely cleaned in a CFC-113 vapor degreaser, but this
process has been replaced with superheated steam cleaning. A cleaner/lubricant is sprayed onto
the part to be cleaned, followed by steam cleaning with a small bench-top steam unit.
The bearing assemblies were disassembled before being cleaned. The first bearing was
cleaned in the HOP at 150 °F (66 °C) for 30 minutes, followed by LC02 degreasing. The second
was cleaned only in LC02, with no HOP preclean.
41

-------
Visual Inspection
The first bearing had no signs of oil or other contamination after the H0P/LC02 cleaning.
It also was inspected by the engineers in charge of the avionic bearing assembly. Both
concluded that it was "immaculate," and "the cleanest they had ever seen." It was visually
determined to be cleaner than was obtained with the CFC-113 degreaser.
A second bearing was brought to the demonstration site that had already teen cleaned
once using a hydrofluorinated ether (HFE) solvent in a vapor degreaser. After the HOP/LCOa
process, RAFB personnel inspecting the part said it was cleaner than when it came out of the
HFE process. This second part had small gearing slots on the outer surface. The bottoms of
some of these grooves were black, but the material did not move. The avionics personnel
believed it to be corrosion products that were not removable. The small screws for this bearing
assembly were tied in a cloth rag and cleaned in the same batch. They came out free of oil, but
still had some black material on them. The black material did not move with a cotton-tipped
swab.
5.11 Electronic Test Circuit Board
This was an avionics test board designed to measure the effect of processes on circuitry
operations. The board had both through hole and surface mount components on it. There is no
regular current cleaning process for this part. The goal of this test was to see if the components
would withstand the S00 to 900 psi pressure of the LC02 process. The board was cleaned with
the same procedure as the most rugged parts: 30 minutes in 150 °F (66 °C) HOP oil followed by
an LC02 process of 3 minutes fill, 10 minutes clean at 400 rpm, 3.5 minutes drain, and 4
minutes purge at 200 rpm.
Visual Inspection
By visual inspection, the board was clean, and the HOP/LC02 process did not appear to
damage the components or affect the inks used to stamp them. It was taken back to avionics for
electronic component testing and was determined to be functioning properly.
42

-------
5.12	Reflectometer
This part also was brought to the demonstration site by Avionics. It was a metal tube
with rectangular cross-section, joined by welding to another similar piece with curved ends. The
internal structure was not described in detail, but there was some type of complex structure
inside. This piece is not currently cleaned or reused when removed from a piece needing repair.
It was visually clean when brought to the site.
This part was dipped in 150 °F (66 °C) HOP oil followed by LC02 degreasing.
Visual Inspection
When removed from the LC02 vessel, oil was dripping from holes in the part; It was
recleaned in LC02, and appeared to be clean and dry. However, more oil leaked from inside the
part after it sat overnight. This is the only part cleaned that showed any sign of oil reappearance
after LC02. It is possible that the internal cavities contained some feature that held the oil, but
without further understanding of the internal structure, it is not possible to determine why the oil
was not removed. Regardless, this part is not appropriate for cleaning in the H0P/LC02 process.
5.13	Rags
We did not begin this demonstration with any intention of cleaning rags. However,
during discussions of cleaning and waste disposal problems at the start of the on-site portion of
the demonstration, it was clear that one of the largest hazardous waste streams on the base is
shop rags. They are currently used once, sealed in steel drums, and shipped to a hazardous waste
disposal company. The cost is determined by weight, including the weight of the steel drums.
The contaminants are usually grease, oils, solvents, or other shop debris. We agreed to try
degreasing some rags with the understanding that even if successful, this process probably would
need to be optimized in order to be economically feasible.
The recipe was changed slightly for the rags. The spin speed during drain and purge
cycles was slowed down to prevent the rags from packing tightly against the walls of the basket.
43

-------
If they packed too tightly, C02 would be left behind as clumps of snow and leave oil spots when
changing back to gas phase.
Visual Inspection
A variety of dirty cloth and blue paper rags were cleaned in several batches. Blue paper
towels used to clean up HOP oil drips around the demonstration site were cleaned and reused
throughout the week. Cloth rags contaminated with drawing compound, oils, cured green
adhesive, and other unidentified shop dirt were brought to the demonstration site for testing.
The results were similar to those for metal parts (Figure 15). When the HOP precleaning
step was used, drawing compounds were completely removed, leaving no oily feel or smell. If
not, the soap-like portion of the drawing compound was left behind. The H0P/LC02 process
removed grease and oils, but not black carbon grime. The process did not have any effect on the
cured green adhesive compound. There was no fraying or other visible damage, even to the
paper towels.
FIGURE 15. RAG, BEFORE (LEFT) AND AFTER (RIGHT) HOP/LC02 CLEANING.
One batch was filled too full to clean properly. It held eight thick white towels and nine
rags, filling the basket about half full. After the LC02 cycle, there was still some oil odor and an
oily feel to the rags. The load was split into two and recleaned with LC02 This removed the oil,
44

-------
leaving no oily smell. A later batch was also filled too full, with seven large cloth rags. Small
clumps of C02 ice were left on the rags that left oily spots behind as they evaporated. These rags
were split into two batches and recleaned. No oily feel or smell remained on either batch.
This system was not built to clean rags, but did well enough to establish that it is
technically feasible. Questions remain as to whether the equipment could be manufactured or
modified to hold and clean enough rags to support the needs of the users and whether it would be
economically feasible. Other work on removing oils from fabric LCOz looks promising,19 This
is clearly an area of interest worth pursuing further.
5.14	Summary of On-Site Tests
The LC02 is similar in performance to TCA vapor degreasing. It can dissolve oils and
most greases, but is not effective for fine particle removal. The HOP is necessary to remove fine
particles and assists greatly with removal of other contaminants. The HOP oil mixture is easily
removed by LC02. When the HOP precleaning step was used, the drawing compound was
removed well, leaving no oily feel or smell. If not, the soap-like material was left behind. The
H0P/LC02 did not remove black carbon grime. The process did not replace the need for hand-
scrubbing but did show excellent degreasing capability. Further testing and process optimization
would be needed to implement the process effectively for specific parts.
The aluminum tubes were cleaned to as low an NVR level as with the current TCA
process. Brass filters could be cleaned equally as well as the current process if an air blow-out
was used and nearly as clean even without it.
5.15	C02 Usage During the On-site Demonstration
The cleaning vessel holds 10 gallons of LC02, and the recycler holds 75 gallons. The
system automatically pulls LC02 from separate supply bottles into the recycler to replace C02
lost in each cleaning cycle. Approximately 6 gallons is added to the recycler during each refill.
The CO , recovery rate with the recycler primarily depends on two variables. The volume
of the basket filled by parts affects how much C02 is added to the vessel in each cycle. More
parts means less C02 per cycle. The other variable is the temperature of the ambient air. The
cooler the ambient air, the more efficiently the condenser on the recycler works, so more of the
45

-------
C02 is recovered. During this demonstration, the baskets rarely had more than half a dozen parts
in them, and the temperature was maintained at approximately 80°F (27 °C).
A total of 37 cleaning batches were run during the 2-week demonstration. The system
refilled after running 8 batches. At 6 gallons usage for 8 batches, C02 losses were estimated at
0.75 gallon per batch, a 92.5 percent recovery rate for the 10-gallon system. The system refilled
again after running a total of 19 batches, or 11 batches since the first refill (6 gal/11 batches =
0.55 gal/batch, a 94.5 percent recovery rate). The third refill was after 30 batches total (11
batches since the second refill for a recovery rate of 94.5 percent). The fourth system refill was
after the 39th batch (nine batches on 6 gallons, or 0.67 gallons per batch, 93.4 percent recovery.)
The average recovery rate during the demonstration was 93.8 percent. Total COz lost was about
24 gallons,
5.16	Operational Demonstration Results
The HOP and LC02 equipment operated well throughout the demonstration, with no
downtime caused by equipment failure. Routine maintenance included opening a valve on the
recycler about once per day to check for and drain any collected contaminants and oil separated
from the LC02 by the recycler. The inside of the vessel was occasionally wiped out with a rag
dampened with an alcohol/water mixture to remove oil residue (or lint from cleaning rags).
The demonstration unit needs a larger refrigeration unit. Running 3-4 loads in quick
succession in the machine would heat the LC02 above its optimum operating temperature. It was
necessary to allow the LC02 to cool so that it would clean adequately. This problem could be
solved with the addition of a larger condenser.
5.17	Recommendations and Observations
The HOP demonstrated at RAFB was in separate equipment from the LCOa step. It may
be possible to perform the HOP in the same equipment as the LC02. The vessel would first be
filled with the HOP mixture. After draining the vessel and spinning off the majority of the
remaining mixture, the LCO, would fill the vessel as it does now. This is similar to a wash and
rinse cycle in conventional aqueous cleaning equipment.
46

-------
For precision cleaning applications, nondestructive testing technology can be directly
integrated with the LC02 system to determine when the parts are clean, These technologies
include measurements with quartz crystal microbalances (QCMs) and chromatography on LCCK
from the cleaning vessel.
6.0 QUALITY ASSURANCE
The critical measurements for this project related to the following project objectives:
•	measure cleanliness of the parts
-	qualitatively by visual inspection
-	quantitatively by NVR, Quantitative NVR depends on stripping effectiveness,
weight measurements, and surface area measurements,
•	evaluate operational suitability of the LC02 process in a production environment
-	throughput (time per item)
-	reliability (number of events; percentage downtime)
The two quality assurance and quality control objectives were to maintain integrity of the
nonvolatile residue measurement process throughout the demonstration and to document the
suitability of the technology for operation in a production environment.
Visual Inspection
Visual inspection for residue, discoloration or visible damage was carried out in room
light without magnification. A small flashlight was used on some parts to check for
contamination on interior surfaces that were too shadowed to inspect with normal room light.
The criteria for visual inspection were determined by the RAFB personnel supplying the parts.
The most common criteria was no visible sign of oil, grease, or loose debris. Often, a wipe with
cotton-tipped swab to check for moveable contamination was also included in the visual
inspection. Visual inspection was performed by at least three people immediately after removing
them from the LC02 equipment, except for parts which were immediately sealed in bags for
NVR testing at RTI. These parts were inspected after bagging, but contaminants were not easily
47

-------
seen through the translucent bag material. The bagged parts were visually inspected by the
person performing the NVR extraction once they were unbagged at RTI. Results of visual
inspections are included in the individual sections for each part.
Quantitative Assessment of NVR
Integrity of the NVR measurement was maintained by several procedures. Full process
blanks were performed in triplicate to measure recovery of the drawing compound by the
method. Five percent of the parts were stripped twice for residual NVR analyses to ensure that
there was no significant amount of extractable material remaining after one extraction. Solvent
blanks were performed on each new bottle of solvent used and with each set of parts stripped.
The microbalance calibration was checked to the nearest 0.01 mg with a 100 mg calibration
weight during every set of weighings, at least once per day. The zero was checked before every
sample. Multiple samples were stripped for each part. They were stripped separately to establish
the variability of the contamination or cleanliness levels. Averages and standard deviation were
calculated for all samples.
The drawing compound was a combination of fatty acids, waxes, and salts. Some of the
organic materials were volatile, A simple test was performed to estimate how much of the
drawing compound would evaporate off parts or could be volatilized during the NVR testing
procedure. A small amount of fresh drawing compound was added to each of three clean, dry
aluminum weighing pans, and they were immediately weighed. Ten milliliters of PERC were
added to each pan to dissolve and distribute the drawing compound in the pans. The
contaminated pans were then warmed slightly (not boiled) on a hot plate to drive off the PERC
and allowed to cool in a desiccator. The pans were reweighed once cool, and the losses of
drawing compound were calculated. This loss averaged 23,3 percent and is an estimate of the
maximum losses expected from volatility during the NVR testing procedure.
A full process blank was performed in triplicate to ensure that the NVR process was
capable of removing and recovering the drawing compound from tubes. In this test, small,
preweighed tubes were contaminated with drawing compound and reweighed. Then they were
stripped with PERC by the same procedure as the other parts tested by the NVR test. The
stripped tubes were dried in a desiccator and reweighed for comparison to the drawing compound
48

-------
removed and recovered in the weighing pans as NVR. The losses shown in Table 11 show that
an average of 9.53 percent of the drawing compound was lost during the NVR test, less than
would be expected due to the volatility of the drawing compound.
TABLE 11. RESULTS OF THE FULL PROCESS BLANK FOR THE NVR TEST
Drawing Compound
Added (nig)
Drawing Compound
Recovered (mg)
Amount Lost (mg)
Amount Lost (%)
123.15
113.78
9.37
7.6
142.38
129.28
13.10
9.2
237.23
209.28
27.95
11.8
One modification was made to the NVR test procedure during the analysis on the pylon
bolts. The solvent planned for extracting the grease from the bolts (PERC) was not completely
effective on bolts with heavy contamination inside the cap heads. These parts were extracted a
second time with TCA, and the weight of the additional material removed was added to that
removed by the first extraction. The NVR test also was modified for brass filters because simple
soaking and swirling in PERC would not remove the black carbon residue from the brass mesh.
These parts were sonicated for 5 minutes in PERC to improve the removal, All other NVR
procedures were maintained.
Tubing Dimensions
The length of the tubes were determined using a meter stick. Curved or irregular tubes
were measured several times and averaged. Tubing outside diameter and tubing thickness were
measured with calipers reading to at least 0.001 inch (0.03 mm). The tubes were standard
diameters. The sizes of the outer diameters were often stamped on the outside, but were
confirmed with calipers regardless. The total surface area of each tube was calculated from the
measurements of length, diameter, and tubing thickness. The formula used is in Appendix C.
49

-------
Cycle Times
The times for the HOP process were measured with a wrist watch. They were estimated to
the nearest minute when the parts were soaked during the working day. Soak times for parts
soaked overnight were not measured and are stated in the text as being soaked overnight. In
general, the length of the soak times in the HOP were not critical measurements.
The cycle times for the LC02 cleaning process were controlled and measured by the
equipment. After the times were confirmed with a wrist watch, they were used as displayed on
the readout of the machine. Average cycle time for the LC02 portion of the cleaning process was
20 minutes per batch. Throughput measured as time per item depended on the size of the item
and how many could be placed in the cleaning vessel. The time for the HOP process was
dependent on the type of contamination and cleanliness level desired.
Reliability
The 2-week demonstration was too short to establish the reliability of the equipment as
percentage of downtime. However, during the 2-week demonstration, the equipment
experienced no downtime due to equipment failure.
7.0 REFERENCES
1.	U.S. Patent 5,368,171, "Dense Fluid Microwave Centrifuge," reissue application pending.
2.	U.S. Patent 5,344,493, "Cleaning Process Using Microwave Energy and Centrifugation in
Combination with Dense Fluids," reissue application pending,
3.	Patent Pending, "Dense Fluid Centrifugal Separation Process for Substrate Treatment."
4.	Patent Pending, "Liquid Carbon Dioxide Dry-Cleaning Apparatus and Method."
5.	Patent Pending, "Dense Fluid Spray Cleaning Process and Apparatus."
6.	Patent Pending, "Process and Apparatus for Recycling and Reusing Dense Fluids."
7.	Bok, Edward, K. Dieter, and K.S. Schumacher, "Supercritical Fluids for Single Wafer
Cleaning," Solid State Technology. June 1992, pp.117-120.
8.	Handbook of Chemistry and Physics. 56th Edition, CRC Press, Cleveland, OH, 1975.
50

-------
9.	Barton A.F.M., Handbook of Solubility Parameters and Other Cohesion Parameters. CRC
Press, Boca Raton, FL, 1983, pp. 153-158.
10.	High Purity Solvent Guide. Burdick and Jackson, 1984.
11.	McHardy, J., T.B. Stanford, L.R. Benjamin, T.E. Whiting, and S.C. Chao, "Progress in
Supercritical C02 Cleaning," SAMPE Journal. Vol. 29, No. 5, September/October 1993,
pp,20-27.
12.	Barton, Jerome C., The Los Alamos Super Scrub™ Supercritical Carbon Dioxide System
Utilities and Consumables Study. Report No. LA-12786, issued June 1994.
13.	Moore, Timothy D., "C02 Pellet Blasting Technology as Employed at the Pearl Harbor
Naval Shipyard," Presented at the Naval Industrial Engineering Symposium, Norfolk,
VA, June 19-20, 1991.
14.	Schmitz, Wayne N., "CO, Pellet Blasting...Problems at First, But Worth a Second Look,"
Presented at the Second Annual International Workshop on Solvent Substitution,
Phoenix, AZ, December 10-13, 1991.
15.	Sherman, Robert, D. Hirt, and R. Vane, "Surface Cleaning with the Carbon Dioxide
Snow Jet." J. Vac. Sci. Technol. A 12(4), July/August, 1994, pp, 1876-1881.
16.	Hill, Elizabeth A., "Carbon Dioxide Snow Examination and Experimentation," Precision
Cleaning Magazine. February, 1994, pp. 36-39.
17.	Vito, Richard R., "Cleaning Large Optics with C02 Snow," SPIE Vol. 1236 Advanced
Technology Optical Telescopes TV, 1990, pp. 952-971.
18.	MEL-STD-1359B, "Military Standard Cleaning Methods and Procedures for Breathing
Oxygen Equipment."
19.	Smith, H., Mike, R.B. Olsen, C.L.J. Adkins, and E.M. Russick, "Oil, Grease, and Solvent
Removal from Solid Waste Using Supercritical Carbon Dioxide," Presented at the
American Institute of Chemical Engineers (AIChE) Boston Summer National Meeting,
Boston, MA, July 30-August 2,1995.
51

-------
APPENDICES
52

-------
APPENDIX A: CASE STUDIES
The following case studies are from manufacturing firms that are currently using LC02 in
their processes or have tested their parts with the LC02 degreasing process. Their case studies
are presented here to provide practical information for the potential user.
A.l Beryllium Copper Bar Stock - BRUSH WELLMAN INCORPORATED
Brush Wellman Incorporated produces beryllium-copper alloy bar stock at its Elmore,
Ohio plant for domestic and foreign markets. It is primarily produced in 12 foot ( 3.7 m) nominal
lengths of round, hexagonal, or rectangular bar stock, or heavy wall tube. Different sizes are
made with a bench drawing process. Heavy Draw #50® (Algonquin Chemical Co.), a chlorinated
hydrocarbon grease containing "tackifiers," lubricates the bar stock during this process. Heat
generated by the drawing process creates organic surface contamination from the drawing
compound that is difficult to remove. The stock must be cleaned prior to straightening and
further processing.
The conventional cleaning method was to place bundles of the stock in a
perchloroethylene (PERC)-based solvent vapor degreaser. Health, safety, and environmental
concerns made replacement of PERC a critical issue. As recently as March 1994, Brush Wellman
was releasing 100 tons of PERC to the atmosphere annually. Jack Currie was placed in charge of
investigating alternative cleaning technologies. Traditional alternatives such as aqueous cleaning
were evaluated, but none appeared to clean as well as the PERC. Aqueous and semi-aqueous
cleaning, even if they had performed well, were undesirable because of additional treatment
necessary for the wastewater. Treatment requirements are more stringent than in some other
metalworking operations because the metals being processed are regulated and monitored as air
and/or water pollutants.
Continued research into alternative technologies disclosed LC02-based degreasing as a
possible alternative ("Dense Fluids Clean Parts Without Use of CFCs," Advanced Materials and
Processes. January 1993). Following initial feasibility tests, Brush Wellman funded construction
of a prototype chamber to test the LC02 process on bundled bar stock. Brush Wellman
appropriated $32,000 in October 1993 for the prototype, which was completed early in 1994.
53

-------
The prototype was a horizontal chamber, 17 feet (5.2 m) long and 8 inches (20 cm) inner
diameter, with mechanical agitation of the LC02. Tests in this unit proved the capability of LC02
cleaning of bundled bar stock. Cleanliness was judged by appearance and feel during controlled
tests. During this initial phase, the HOP pretreatment was used prior to LC02 degreasing.
Additional testing is now underway to determine how machining chips and scrap material returned
to Brush Wellman will be cleaned.
Waste reductions were estimated for the LCG2 process and are shown in Table Al.
Cleaning cycle times will be established when the LC02 system is installed at Brush Wellman.
Cycle times will be set to match cleanliness levels of PERC vapor degreasing. Use of the HOP
pretreatment step may be included in the final cleaning process.
TABLE Al. COMPARISON OF WASTE PRODUCTION FOR PERC AND LC02
Waste Generated
(tons/year/unit)
Waste Category**
PERC Process
LC02 Process
Still Bottoms
LHH
49.5
0
PERC Air Emissions
LHT
100
0
Soil Removed
LHN
0*
1.25
TOTAL WASTE

149.5 tons
1.25 tons
C02 Emissions (cu. ft.)



C02 Emissions From
Process (cu. ft.)
GXN
0
3.7 x 104
C02 Emissions From
Energy Use (cu. ft.)
GXN
3.8 x 107
2x 105
~included in still bottoms
** Waste Category:
LHH = liquid, hydrocarbon/organic, regulated under hazardous
LHT = liquid, hydrocarbon/organic, regulated under toxic
LHN = liquid, hydrocarbon/organic, regulated under nonhazardous
GXN = gaseous, not hydrocarbon/organic, regulated under nonhazardous
54

-------
In March 1994, Mr. Currie submitted an application for a DOE/EPA grant to the National
Industrial Competitiveness through Energy, Environment, and Economics (NICE3) Program.
The application was accepted, and a grant for $400,000 was awarded. The State of Ohio also
awarded $25,000 to this effort. The purpose of the grant was to demonstrate LC02 cleaning in an
industrial application and was contingent on Brush Wellman's providing a commitment equal to
or greater than the federal grant. In July 1994, Brush Wellman appropriated $920,000 to fund the
replacement cleaning process. In the NICE3 application, estimated annual savings of $282,000
were defined and used as part of the justification for funding the LC02 cleaning process. Table
A2 and the accompanying detail (Table A3) include the estimated costs for both processes.
TABLE A2. ECONOMIC COMPARISON SUMMARY (FROM NICE3 APPLICATION), $
Item
Costs for
Current Process
Costs for
Proposed Process
Savings
Costs:



Capital
(from supplier quotation)
n/a
880,000
n/a
Installation
(from engineering estimate)
n/a
350,000
n/a
Total cost

1,230,000
(1,230,000)
Savings per year:



Solvent
70,000
9,600
60,400
Steam
146,100
0
146,100
Electricity
24,300
1,400
22,900
Water
2,800
500
2,300
Disposal
38,100
300
37,800
Maintenance
15,400
2,500
12,900
Total per year
296,700
14,300
282,400
55

-------
TABLE A3. DETAILS OF ANNUAL COSTS AND SAVINGS, $
Item
Annual Usage
Annual
Costs
Annual
Savings:
Solvent
Current; 200,000 Ib/yr @ $.35/lb
Proposed: 20,000 gal/yr @ $.33/gal = $6,600
C02 tank rental/service = $3,000
70,000
9,600
60,400
Steam
Current: 24,352,000 lb/yr @ $6.00/1000 lb
Proposed: Zero consumption
146,100
0
146,100
Electricity
Current: 38.5 kw x 8760 hr/yr x $.072/kwhr
Proposed: 7.5 kw x 2500 hr/yr x $.072/kwhr
24,300
1,400
22,900
Water
Current: 90 gal/hr x 8760 hr/yr x $3.50/1000 gal
Proposed: 60 gal/hr x 2500 hr/yr x $3.50/1000 gal
2,800
500
2,300
Disposal
Current, still bottoms: 99,000 lb/yr x $.385/lb
Proposed, non-contaminated oils: 3,000 lb/yr x $. 10/lb
38,100
300
37,800
Maintenance
Current: 1991/1992 actual average
Proposed: estimated
15,400
2,500
12,900
The process life expectancy is 20 years. Based on receiving the federal grant cofunding
requested ($425,000) and total Brush Wellman funds of $805,000, the return on investment will
be 32 percent. This equates to a payback in 3.6 years.
The LCO, stock cleaning system is scheduled to be installed at Brush Wellman in the first
quarter of 1996. As part of the NICE3 grant, Brush Wellman will allow other companies to
evaluate the cleaning process. Mr. Currie said he would "confidently recommend consideration
of LC02 as an alternative cleaning technology."
56

-------
A.2 Machined Metal Parts - ACCRATRONICS
Accratronics, a business in southern California, machines conventional metals to
specifications provided by their customers. They primarily manufacture valve bodies, cartridge
bodies, explosive bolts, and explosive valve bodies for ordinance applications. These are small
parts; most would fit in a 10 inch (25 cm) cube. The metals in these parts include 300 series
stainless steels, 15-5, 13-8, and 17-4 precipitation hardened steels, alloy steels, and aluminum
During machining, they are contaminated with organic or synthetic machining fluids, coolants,
some heavier lubricants, and metal chips. The next process is shipping the parts to another site
for passivation and/or plating. Since the parts are cleaned again just prior to passivation or
plating, gross cleaning before shipping is adequate. Cleanliness testing is visual inspection under
room light. If the parts look clean and feel clean, they are adequate for the next process.
The company also manufactures hermetic glass to metal seals for cartridges used in fire
extinguishers. These extinguishers release halons and other fire retardants in commercial and
military aircraft and other vehicles (tanks, buses, etc.). Contained within the detonator is a
ground and lapped ceramic substrate that is potted on one side. The grinding and lapping
processes generate particulate contamination that must be removed from the ceramic substrate
before assembly. The contamination typically contains particles of epoxy, ceramic, and metals.
The current cleaning process for all of the parts is vapor degreasing and immersion with
ultrasonics in TCA. Cycle time for this process is 5-10 minutes. Replacing TCA cleaning with an
aqueous process was considered a poor alternative because the facility has no aqueous process
emissions now and does not want any in the future.
Samples of machined metal parts and hermetic seals were sent to the manufacturer of the
LCQ2 equipment for cleaning tests. The HOP pretreatment was not used on any of these parts.
There was a significant improvement reported in removing machining fluids and chips from blind
holes in the machined parts over that obtained using the TCA vapor degreaser. The LCOz was
also reported to be far better than TCA degreasing for cleaning the components and glass sealed
assemblies. After cleaning, no particles were visible on the parts under black light inspection.
This was verified with white light inspection at 10X magnification.
57

-------
The LC02 process has several advantages over TCA vapor degreasing in these
applications. In addition to being an environmental problem, the elevated temperature of the TCA
in the vapor degreaser softened the potting material used to position the ceramic, jeopardizing
reliability. The lower operating temperature of the LC02 process reduces reliability concerns.
Another significant advantage of the LC02 process is that neither the Air Quality Management
District (AQMD) nor the water quality control board of southern California requires the company
to obtain permits because C02 is not a volatile organic compound, hazardous air pollutant, or
water pollutant.
Accratronics has ordered a vertical LC02 machine with a 12 inch (30 cm) cleaning vessel
and a recycler. The machine will be installed in 1996.
58

-------
APPENDIX B: MATERIAL SAFETY DATA SHEETS
Carbon Dioxide
HOP Mixture (SuperSolv PAS)
P-D-680, Type IE
Marsol
59

-------
/L
LIQUID AIR CORPORATION
IHOUSTRIAL OASCR DIVISION
Material Safety Data Sheet
LiQUtO AIR COKTOHATION
i«oiurm«i. ***m okauok
One California P l«ia. Suit* 350
2121 N, California Blvd.
Walnut Cna*k, CalHornla 34S94
issue date octobch i, ims
amo neviwote cowan ate safety dspt.
productH"«e Carbon Dioxide,
Refrigerated Liyuid 	
TfiUSWIOne (415) #77-8500
EMCnCOICY BCSPONSC INfORMATION ON PAUfc 2
TKA0S NAMfc and SYNONYMS
See last page.
CHEMICAL NAME AMD SYNONYMS
Dioxide
CdrHori
Carbonir. Anhydride	
FOMMULA
MOLECULAR WEU2MT
CO 7
44.01
CAS NUMBER
124-36-9
CHEMICAL FAMIL.Y
Carbonate
HEALTH HAZARD DATA (SEE NOTE ON LAST PAGE)
TiMfi wttoHTiD AveoAae «xi»asu*£ uMit
5,000 Molar I'I'M.
15,000 Molar PPM to 30,000 Molar PPM (ACC)I
Its STEL is proposed to he changed from
, i984-as).
SYMPTOMS Qf iXFMUSE
Nervous system control of respiration is dependent on the C0z level breathed in air. By
reducing the oxygen level in air, €02 can Cause suffocation. Symptoms of overexposure
include headache, dizziness, shortness of breath, muscular weakness, drowsiness and
ringing in the ears. High concentrations produce a faint acid taste and can cause
paralysis of the breathing control centers of the nervous system; Zt by volume in the
atmosphere will cause a SOX increase in hreathing rate; 3%, a 100% rate increase; >416
produces labored breathing and is jlangerous Tor .even a few {continued on last page)
TomaxoaiCAL mw^cttiis	~
Car&on dioxide Is the most powerful cerebral vasodilator known. Inhaling large concentrations causes rapid circula-
tory insufficiency leading to coma and death. Chronic, harmful effects are not known from repeated inhalation of
low (3-5 molar %) concentrations.
Rat. inhalation LCLo 657.190 ppm for 15 minutes.
Bat (10 days preg.). Inhalation TCLo $0,000 ppm, 24 hours teratogenic effects.
Human, Inhalation TCLo 2,000 ppm pulmonary effects.
FroatDlie effects are a change in the color of the skin to gray or white possibly followed by blistering.
Listed as Carcinogen	National Toxicology Yes I I	LA.R.C. Yes I 1	OSHA Yes 1 ]
or Potential Carcinogen	Program	No 1*5	Monographs No 0	No i*i
aecoMM€noeo nmr *n trcatmgnt PR0'mPT~HEd1cAL ATTENTFOtTlS MANDATORY IN ALL CASFS OF OVER-
EXPOSURE TO CARBGN DIOXIDE. RESCUE PERSONNEL SHOULD BE EQUIPPED WITH SELF-CONTAINED
BREATHING APPARATUS.
Inhalation: Conscious persons should he assisted to an uncontaminated area and inhale
fresh air. Quick removal from the contaminated area Is most important. Unconscious
persons should be moved to an uncontaminated area, given mouth-tu -iiiuu Lh resuscitation
and ^lippl f»mon t a1 r»*yg«rt. Ac c ut-o that wi>mife.c<* material *toc* nu L uLo Li us. I tll« a I fWtiy Uy
use of positional drainage. Medical assistance should he sought immediately.
frostbite: Flush affected areas with lukewarm water. 00 NOT USE HOT WATER. A physi-
cian should see the patient promptly if the cryogenic "burn" has resulted in blistering
of the dermal surface or deep tissue freezing.
JudMnwsnts IS in true-iwUbiSty n< wformatinn decern tor purchase'J purprccs a/e iiecwsaftij	lesfWAiiWitv tliercfoie. iWiuugh aisonlBta cm lu» Mm latum inlhe ptctMiatiiw nt sucli
mluiimiion. Liquid Air Coiyominn extends nu warrjniies. mains, na ieincvcflM(iiln&. and mssiiiimf, huiexpansibility 4* lollw accuracy ut wiiU&iMy olxudi infWflUtMn to ippjcaliwMU (HifcnjSW's
inWkted putpgses or consequences of its use. Sihm | fju«J An C.aipun!m has tin roflltol ovm tli* i.vo ol tins piudiict. it jisumus no UjOilily tar	m |Ust nl product rimltinq tram jiic,p»r («
improim) uie or app!ic»tionof IN jirrxSucl Om Sheets may ks chmgmf tiaifl tirni in time Uc suic In consult ttnUMSt tjiuon
¦	60

-------
Page 2
44ZAMOOUS Kl)CTU«e»"«* UfMtH UOUIOS, Sai.loi. O« 0*S€S
. jrms carbonic acid in the presence of water. See REACT1V11Y DATA Section.
PHYSICAL DATA
-1 09 . 3f J[ - 76 . "Cj
¦™u«a ramt
Sublfmation point
tmron f*f»C3SU»e
C 70"F (21 . I°C) = B14.7 lJuid_i'j0c'4 kl'a)
wtuwurv in water (36R"r (20 °C) ISunsert
Cye f f i c j ent _,0704		
Lf*e49AHCC AMD QOOft
70°F
_P_.ii3Z kg/m3)
(-55.57^7
UOUIO nfNSHT »r SOILING POWfT (*i«lc.
REACTIVITY DATA
.fAWurr
U>>MriwW«
conditions to 'void	i % stable under ordinary conditions of. use
and storage. It does not polymerize. It does cause violent
mCOM'aTMIUTY (MaWrlab to •told)
peroxiae with aluminum or_ma<{nes 1 urn
IAZARDOUB DECOMPOSITION PROOUCT3
rwrhrtn mnm>_* 11
««zaroous pot,riteftiz*TioM
*»T Occur
DO 1 Vdieri ZiH I. tnn Ilf .< r ' r-y 1 1 tiriUyrio	othyl 4o«i»1no. It decom
poses to CO and Oz when heated above (continued on last page)
An f?X[il 0
-------
SPECIAL PROTECTION INFORMATION	P*9** 3
nearinATOH* #POTection m*~e«p if**) positive pressure air line with mask or sel f-contairied
urcauiinx aoucirai.ua anouiu uc available iuf cmt:'_"ycf',^y_H5£_:
W«MtlUkTMS»l
iee local exhaust.
loc«uexhaust Jo prevent accumulation
above the TWA.
ugcmmcAt
• FECIAL
OTHER
moTecn*€ amoves
Loose fitting, insulated,
"i?t MOTICTIOM
¦lo re ty yuyij lea ur 'jlaaac:
OTHG* fHOT^CTIVt EOUIPMCNT
Safety shoes
SPECIAL PRECAUTIONS*
arecut. lamjjno inform*fio» Carbon t) 1Qx 1 de
DOT Shipping Name: Refrigerated Liquid	HOT Hazard Class: Nonflammable ^as
DOT Shioo.ina Label: Nonflammable, ass.	I JU	,,m.2UZ_ .. 				 . 	
*p*c»Ai.hamouho*eco«*»**oATio*s 5ec ntite on last page regarding Spill or Leak Procedures.
Also see CGA Pamphlet G-6, Carbon Dioxide and G-6.1, Standard for l.ow Pressure Carbon
Dioxide Systems at consumer sites. Provide general and local exhaust ventilation to
meet TLV requirements. Provide approved supplied-air or self-contained respirators for
use in non-routine or emergency situations with exposure above the TLV. A full face-
piece is required for concentrations - 1(J%. Provide standby person(s) with rescue equip-
ment where work is required at	CQz in air.
Workers should use gloves and may require additional protective clothing (apron, face
shield, etc. which are resistant to low temperatures} to prevent freeze burns and
frostbite if more than momentary contact with CO? at low temperature is possible.
Fo< adouionat handling recommendations consult t'Alr LX|uid«'$ Focyt;leif>edi^ fa Gal 0< Compressed Gaa Assocklion Piwnfflte* P-1.
SPECIAL STORAGE RECOMMNDATfONS
See note on last page regarding Spill or Leak Procedures, Also see CGA Pamphlqt G-6,
Carbon Dioxide and G-6.1, Standard for lo w I'res sure Carbon Dioxide Systems at Consumer
Si tcs.
Do not store cylinders in sub-surface or closed areas. Carbon dioxide is heavier than
air and leaking gas could accumulate tn law areas and cause suffocation.
For additional storage rccomm«nd»iion» conswti (."Air Uquirte's FncyclopaeJia fle Uaz or Gumpreued Gat Association Pam^ftlet P-1.
WtCUL PACKaOIMQ KCOMU6MDATIONS (F0R 6ASE0US CARBON DIOXIDE)
Dry carbon dioxide can be handled with most common structural materials. Moist carbon
dioxide is corrosive by its formation of carbonic acid. For these applications, 316,
309 and 310 stainless steels may he used as well as Hastel1oy A®, B & C and MortelG,
Ferrous nickel alloys are slightly corroded.
At normal temperatures, carbon dioxide is compatible with most plastics and elastomers.
Also see CCA Pamphlet G-6.3 Carhon Dioxide Cylinder filling and Handling Procedures for
Beverage Plants.
OTHER RCCOtHrtCNOATtOMa OH PRECAUTIOUS
Crifnprfa* ^#*<1 ywv i*y1fnclor< choiilrt not bo r <£> t i 1 1 qd	accept by c|U.*l1Mcd producers o F
compressed gases. Shipment of a compressed gas	cylinder which has not been filled
by the owner or with his (written) consent is a	violation of Federal Law (49CFS).
See NOTE on last page.
62

be familiar with (haw reoulalxsm.

-------
P«fl« 4
LIQUID AIR CORPORATION
inwtm •
*OOITIOK*t. DAT*
TRADE NAME AND SYNONYMS: (Continued)
Carbon 0 io* i de , Re f ri gerat.ed Liquid; flu 1 |e Cnrhon Dioxide.
HFAI.TH HAZARD DATA: (Continued)
NOTE: Except where specified, the health hazard da u and mo* I of the other duM
in this material safety data sheet are for qnseous carbon dioxide.
SYMPTOMS OP EXPOSURE; (Continued)
minutes of exposure; >12X causes rapid unconsciousness, a few hours exposure at 25%
results in death.
SUMMARY ; Inhalation; Low concerttr«i tions (3-b molar T.) cause Increased respiration
and headache. Eight to 15 molar t concentra Lions cause headache, nausea and vbmi ting
which may lead to unconsciousness if not moved to open air or given oxyyen.
Higher concentrations cause rupid circulatory insufficiency leading to Coma and death.
When refrigerated 11 quid carbon dioxide 1s valorized through an orifice, it can form
3ui iu put MLiKi o i <-ar uiiri u hijm ue C snow or ury ice" powaer/. continuous aeplta 1
contact with this cold snow could result in frostbite or cryogenic (freeze) "burns."
contact wicn tne liquid or so no can produce i rostoite ano freeze fiurns.
REACTIVITY DATA; (Continued)
CONDITIONS TO AVOID (Continued)
1700°C. This weakly acidic material will react with alkaline materials to fortn
carbonates and bicarbonate*.
Incompatibility (materials to avoid) (continued)
f^lich alkali ntptal<.	1 um. aluminum, titanium, nr Ti^ronlnmJ . Thai r-	Hoc ,
and materials like diethyl magnesium, moist ccsiuin oxide, or lithium acetylide. with
ammonia can ignite in a CO? atmosphere. Dry ice can lorm shock sensitive mixtures
with sodium, potassium, or sodium-potassium alloy.
NOTE ON 0TS1LR RCCDMMFNDATi ON:i UK PRLTAIITI UM% (Continued)
Carbon dioxide, refrigerated liquid is delivered t.o 
-------
^7
DEFLEX/
corporation
Material Safety Data Sheet
Trade Secret Formulation
FORMl 'I.A THUS NAME
I M .I LEX C< JKPURATION
l-WMH-CTNAMK
SUPERSOLV PAS
ISSItEDAlK
JU1 Y 7, l
OIL ODOR
FLAMMABLE LIMIT
not determined
EXTINGUISHING MEDIA
FOAM. IJKY CIIEM1CAI, WATER SPRAY OR FOG, C<>2
special hkk h<;ii ung rwKWHiKts
WEAR SEI.F-CONTAINED HRKA I1 UNO APPARAlXIS, AVOID URIAH JIMCi SM( >K.F. OR VAPOR. USE WATER SPRAY TO COOL CONTAINERS
exposed ro ihgh heat or opicn fijvmes. *
I'imsivu. HRK AND EXPLOSION HAZARDS
ix) No r wi .i .i). a rr. w inch irk or burn empty a intainers
64

-------
PACK 2
HEALTH HAZARD DATA
EXPOSURE !.IMIT(S)
NOl" ESTAIM .ISIIED
MEDIA CWNDITIONS A BV OVKHF-XlXI.St'RE INGESTION
IIKRMATIIUS
MAY CAUSE CKAMI'S AND DIARRHEA
1MIAI.A I ION
EYK< X1NTA< T
MAY CAUSE 1RKII AllON
N(IT NORMA! I,Y EXPECTED
SKIN < t»N I AC T
MAYCAUSE EYE AND SKIN IRRITATION lJIN S< K JRCES. CONTAIN S1MI LAND IHIiVKNT I K< JM ENIKRINU WA I'EK WAYS AND SEWERS SOAK UP SPILL Willi AN
(>11. AI1S( IRIS A NTCOR. WASH SOU EDCLOniING lIEKIRIv RE-USE.
STABtLI IV
S IAISI.E
CONDITIONS TO AVOID
PROLONGED EXCESSIVE HEAT MAY CAUSE PRODUCT IO DECOMPOSE
INCOMPATIBILITY (MATKRIAIS TO AVOID)
STKONU OXIDIZING AOENT		 __ 			
HAZARDOUS DECOMPOSITION ANI) BYPRODUCTS
CO. Ct 12. SMC IKE AND R JMES IN IT IE CASE, (IE INC( 1MPI ,E t E C< IMIW IS HON
HAZARDOUS POLYMERIZATION
WIIJ. NOT OCCUR
CONDITIONS TO AVOID
NONE KNOWN
PRECAUTIONS FOR SAFE HANDLING AND USE
REACTIVITY DATA
65

-------
REPRESENTATIVE MSDS SUMMARY
CHEMICAL OR PRODUCT NAME: P-D-680, Type HI
COMPONENTS (CAS #): Mixed Aliphatic Hydrocarbons (CAS # 64742-46-7)
STANDARDS AND REGULATIONS: EPA TSCA CHEMICAL INVENTORY, Title TIT hazard
classifications: acute and chronic.
BOILING POINT: 221-277 °C
SPECIFIC GRAVITY: 0.83 at 25 °C
VAPOR PRESSURE: 0.4 mm Hg at 38 °C
EVAPORATION RATE (n-butyl acetate=l): N/K
FLASH POINT: 200-210 °F PMCC
LOWER EXPLOSIVE LIMIT: 1.0%
UPPER EXPLOSIVE LIMIT: 6.0%
COMBUSTIBLE: Yes
FLAMMABLE: 1
SOLUBILITY IN H20: Insoluble
APPEARANCE: Clear, colorless liquid
ODOR: Hydrocarbon
REACTIVITY DATA: Stable. Incompatible with: strong oxidizers, acids, bases.
WASTE DISPOSAL: Prevent liquid from entering sewers, waterways, or low areas. Dispose of
according to applicable regulations.
HEALTH EFFECTS: Mixed aliphatic hydrocarbons: slight skin and mild eye irritant. Extreme
exposure by aspiration into lungs may cause pneumonia. Overexposure may cause weakness,
headache, nausea, confusion, blurred vision, drowsiness and other nervous system effects. Greater
overexposure may cause dizziness, slurred speech, flushed face, unconsciousness, or
convulsions. May cause eye, lung, or skin irritation after prolonged or repeated overexposure.
66

-------
REPRESENTATIVE MSDS SUMMARY
CHEMICAL OR PRODUCT NAME: MARSOL (MIL-C-38736B)
COMPONENTS (CAS #): Toluene (CAS # 108-88-3), ethyl acetate (CAS # 141-78-6), methyl
ethyl ketone (CAS # 78-93-3), light aliphatic hydrocarbons (CAS # 64742-89-8), isopropyl
alcohol (CAS # 67-63-0), and xylenes (CAS # 1330-20-7).
STANDARDS AND REGULATIONS: EPA TSCA CHEMICAL INVENTORY, Title III hazard
classifications: acute and chronic.
BOILING POINT: 143 °C
SPECIFIC GRAVITY: 0.838 at 25 °C
VAPOR PRESSURE: 42.7 mm Hg at 21 °C
EVAPORATION RATE (n-butyl acetate=l): N/K
FLASH POINT: -2.2 °C TCC
LOWER EXPLOSIVE LIMIT: 1.7 %
UPPER EXPLOSIVE LIMIT: N/K
COMBUSTIBLE: Yes
FLAMMABLE: 3
SOLUBILITY IN H20; N/K
APPEARANCE: Clear, colorless liquid
ODOR: Ketone
REACTIVITY DATA: Stable. Incompatible with oxidizers, heat, sparks, electric equipment,
open flame.
WASTE DISPOSAL: Prevent liquid from entering sewers, waterways, or low areas. Recycle or
dispose of according to applicable regulations.
HEALTH EFFECTS: Irritating to skin, defatting, dermal absorption through skin increases
exposure. Inhal.: anesthetic. Irritant to respiratory tract. Central nervous system depression.
Acute overexposure can cause damage to kidneys, blood, nerves, liver, and lungs. Ingestion
harmful or fatal. Abdominal irritation, nausea, vomiting, diarrhea. Persons with asthma, chronic
respiratory problems, severe heart, skin, liver, or kidney problems should avoid use.
67

-------
APPENDIX C; NONVOLATILE RESIDUE PROCEDURE FOR TUBES
1.	Place one tube in a clean glass graduated cylinder and fill with perchloroethylene (PERC)
until the tube is completely covered.
2.	Gently swirl the solvent in the graduated cylinder to dislodge all visible contamination.
Remove the tube from the cylinder with stainless steel tongs. Rinse the tube with clean
PERC, combining the rinsate with the contents of the graduated cylinder.
3.	Pour the PERC into a clean, dry Erlenmeyer flask. Rinse the cylinder with clean PERC
and pour into the same flask.
4.	Weigh an aluminum weighing dish to the nearest 0.01 mg using the Mettler AT20 balance.
5.	Slowly transfer the PERC to the pre-weighed aluminum weighing dish.
6.	Gently heat the aluminum weighing dish on a hot plate in a chemical hood. As the liquid
evaporates, add more PERC from the Erlenmeyer flask until the entire volume has been
transferred. Rinse the Erlenmeyer into the weighing dish with fresh PERC. Gently heat
the weighing dish until the volume has been reduced to about 10 ml.
7.	Allow the last 10 mis of PERC to dry in the hood without application of heat.
8.	When the liquid has evaporated, transfer the weighing dish to a large vacuum desiccator
charged with desiccant. Leave the dish in the desiccator overnight.
9.	Remove the weighing dish from the desiccator and reweigh to the nearest 0.01 mg using
the Mettler AT20.
10.	Calculate the difference between the two weighings. Record as NVR.
11.	Measure the length, wall thickness, and inside and outside diameters of the tube.
Calculate the total surface area (inside + outside + ends) in square feet.
Area = 2rcL(D-T) + 2rcT(DT)
Where, Area - total area of tube including inner and outer surfaces and two ends
D - Outside diameter of the tube
T - Tubing wall thickness
L - Length of the tube
68

-------
12.	Calculate and record the NVR as g/square foot.
13.	Calculate an average NVR for each set of conditions (not cleaned; current process; C02
process).
The fundamental equation for reporting NVR is:
NVR(wt./area) = (residue wt.) / (surface area)
Standard Deviation of a set of replicated measurements is calculated by:
Std.Dev. --,
N
mrx)
N-1
2
Where: Xj = Individual value
x = Sample mean
N = Number of data points
69

-------