EPA-R2 73-202

April  1973                Environmental Protection Technology Series
Package  Sorption Device
System Study
                                  Office of Research and Monitoring
                                  U.S. Environmental Protection Agency
                                  Washington. D.C. 20460

Package Sorption Device
          System Study

           MSA Research Corporation
    Division of Mine Safety Appliances Company
         Evans City, Pennsylvania 16033
           Contract No.  EHSD 71-2
          Program Element No. 1A2015
       EPA Project Officer:  D. L. Harmon

          Control Systems Laboratory
     National Environmental Research Center
   Research Triangle Park, North Carolina 27711
                Prepared for

            WASHINGTON, D.C.  20460

                 April 1973

This report has been reviewed by the Environmental Protection Agency

and approved for publication.  Approval does not signify that the

contents necessarily reflect the views and policies of the Agency,

nor does mention of trade names or commercial products constitute

endorsement or recommendation for use.

     A Package Sorption Device System Study was undertaken
to achieve the following objectives:   (1)  identify and
characterize the numerous  small  but  objectionable sources
of gaseous pollutants;  (2) develop ranking of sources
according to relative importance;  (3) assess the equipment
and technology available for controlling these sources;
(4) investigate the potential and need for development of
new technology and new sorbents;  and   (5)  develop detailed
research and development recommendations to improve exist-
ing devices or develop new control  methods.  The primary
products of this study are:  (1)  a  handbook on package
sorption device technology;  (2)  an assessment of the emis-
sions to which this technology may  be applicable; and  (3)
the R&D recommendations.  Emission  sources amenable to
control by package sorption devices contribute 15% of the
total organic emissions from all  sources and can be divided
into two broad categories:  (1) solvent user industries
which emit solvents essentially unchanged and  (2) process
industries which generate pollutants  by chemical, biological
or thermal reactions.  Regenerative activated carbon adsor-
bent systems show promise for the control  of pollutants  from
these sources, particularly at concentrations below 700  ppm.
At higher concentrations, catalytic and thermal incinerators
are more economical except where  the  would-be pollutant is a
valuable solvent recoverable by the adsorbent system.

     This study was conducted for the Environmental Protection
Agency under Contract EHS-D-71-2 with Mr. D.L. Harmon serving
as project officer.

     The study was conducted at MSA Research Corporation,
Evans City, Pennsylvania during the period August 1970 to
March 1973 under the general supervision of Dr. J.W. Mausteller.
The program manager was Dr. A,J. Juhola.  Contributions to var-
ious phases of the study were made by Messrs. R.H. Hiltz, S.J.
Rodgers, W.A. Everson, O.W. Milius, R.M. Hunt, B.B. Carr, J.S.
Greer, and T.A. Ciarlariello.


                   Table of Contents


1.  Sources and Emissions                                1.1

       Introduction                                      1-1

       Information Sources                               1-2

       Emission Sources                                  1-6
          Listing of Industries                          1-6
          Categorizing of Industries                     1-6
          Extent of Emission                             1-8

       Types of Pollutants                               1-13
          Listing of Pollutants                          1-13
          Adsorptive Characteristics                     1-13
          Identification With Industries                 1-14

       Conditions of Pollutant Release by Industry       1-20
          Effect on Choice of Control  System             1-20
          Surface Coating Industry                       1-24
          Degreasing Industry                            1-34
          Dry Cleaning Industry                          1-37
          Graphic Arts Industry                          1-39
          Other Solvent User Industries                  1-43
          Rubber Products Industry                       1-44
          Solvent Extraction of Vegetables               1-48
          Solvent and Fuel Evaporation From Storage       1-48
          Meat Rendering Industry                        1-51
          Fish Rendering Industry                        1-60
          Canning Industry                               1-61
          Restaurants                                    1-63
          Meat Smokehouses                               1-64
          Coffee Roasting                                1-64
          Tobacco Curing, Zinc Plating and Tanning       1-65
          Control of Pollutants from Incinerators         1-65
          Baking and Candy Making Industries             1-66
          Other Emission Sources                         1-66

2.  Environmental Effects                                2-1

       Introduction                                      2-1

       Evaluation of Pollutant Concentration             2-2

       Toxic Effects on Humans                           2-5

                       Contents (continued)
            Odor Effects                                      2-9
               Psychological and Physiological Effects        2-9
               Ranking of Pollutants According to Odor
                 Response                                     2-10
               Odor Control                                   2-13

            Toxic Effects of Pollutants on Vegetation         2-14

            Damage to Materials                               2-16

            Smog Forming Potential of Pollutants              2-16
               Nature of Smog                                 2-16
               Distribution of Smog Forming Pollutants        2-17
               Active Smog Precursors                         2-18
               Ranking of Smog Precursors                     2-18

            Effect of Smog on the Environment                 2-20

            Pollutant Regulatory Laws                         2-23
               Federal Standards                              2-23
               Control District Regulations                   2-23

     3.  Research and Development Recommendations             3-1

            Introduction                                      3-1

            Sources and Emissions                             3-1

            Industry Trends                                   3-2

            Carbon Resorb Systems. Present Technology         3-3

            Carbon Resorb Systems, New Technology             3-5

            Adsorption Theories                               3-6

            Monitoring Instruments                            3-7

            Research and Development Programs                 3-7
               Emission and Source Characterization           3-8
               Evaluation of New Carbon Resorb System         3-9
               Generation of Adsorption and Regeneration
                 Data                                         3-11
               Monitoring Instrument Development              3-13

            Summary                                           3-14

                  Contents (continued)
4.  Sorbent Types and Sorption Theories

       Sorbent Types
          Important Sorbent Properties
          Chemically Reactive Adsorbents
          Polar Adsorbents
          Nonpolar Adsorbents
          Molecular Sieves
          Effect of Physical Properties  on
            Adsorption                                   4-10

       Theories of Adsorption                            4-18
          Forces of Adsorption                           4-18
          Adsorptive Capacity                            4-19
          Pore Structure Determination                    4-27
          Adsorption Thermodynamics                      4-31
          Dynamic Adsorption Processes                    4-44
          Adsorption of Vapor Mixtures                    4-58
          Resistance to Airflow                          4-66
          Regeneration of Activated Carbons              4-67

5.  Theories and Types of Catalysts                      5-1

       Background                                        5-1
          Terminology                                    5-1
          Development                                    5-1
          Description                                    5-1

       Theoretical Models                                5-2
          Types                                          5-2
          Geometric                                      5-2
          Electronic                                     5-3
          Boundary Layer                                 5-3
          Crystal Field                                  5-3

       Adsorption                                        5-3
          Description                                    5-3
          Theoretical Principles                         5-3
          Application                                    5-4

       System Design Theory                              5-4
          Approaches                                     5-4
          Reaction Rates                                 5-4
          Thermodynamics                                 5-6

              Contents (continued)

   Types of Catalysts                                5.7
      Noble Metals                                   5.7
      Transition Metals                              5.7

   Methods and Parameters Used to Evaluate
     Catalysts                                       5.7

Package Sorption Systems                             6,-j

   Introduction                                      6.^

   Basic Functions                                   6.2

   Air Purification  Systems,  Present  Technology       6-4
      General  Discussion                             6.4
      Carbon  Bed Designs                             6-6
      Materials of Construction                       6-11
      Operation of Air  Purification Systems           6-13

   Solvent  Recovery  Systems,  Present  Technology       6-18
      General  Discussion                             6-18
      Stationary Bed  Systems                          6-21
      Moving  Bed Systems                             6-28
      Fluidized Bed  Systems                           6-30

   Carbon Resorb Systems, New  Technology              6-32
      General  Discussion                              6-32
      Gas Regeneration with Incineration              6-35
      Gas Regeneration with Secondary
        Adsorption                                    6-38
      Gas Regeneration with Condensation and
        Recycle                                       6-40
      Two-Stage  Stationary Bed System                 6-44
      Two-Stage  Regeneration of Stationary Bed        6-46
      Regeneration by Pressure Reduction              6-48
      Multistage with Bed Displacement                6-49
      "Hypersorption" System                          6-52
      Rotating Axial-Flow Bed with Vapor Recycle     6-54
     Materials of Construction                      6-57
     Auxiliary Equipment                            6-58
     Operation of Carbon Resorb System              6-63

  Polar Adsorbent Systems                           6-71

  Impregnated Adsorbent Systems                     6-72

                  Contents (continued)


       Catalyst Impregnated Carbon Systems               6-72

       Economic Analyses                                 6-73
          General  Discussion                             6-73
          Operating Costs                                6-74
          Pollution Control Costs                        6-78

       Adsorbents  and Adsorbent Systems1  Manufacturers
         and Vendors                                     6-87
          Activated Carbons                              6-87
          Noncarbon Adsorbents                           6-93
          Sorption Systems                               6-94

7.  Catalytic Incineration System                        7-1

       Basic Functions and Limitations                   7-1
          Oxidation and Reduction Functions              7-1
          Advantages                                     7-2
          Disadvantages                                  7-3

       Potential  Applications                            7-4

       System Design Parameters                          7-5
          Type of  Catalyst                               7-6
          Type of  Pollutant                              7-6
          Pollutant Concentration                        7-9
          Operating Temperature                          7-11
          Residence Time                                 7-12

       Operating  Costs                                   7-15
          Equipment                                      7-15
          Depreciation                                   7-15
          Operating Labor                                7-15
          Maintenance                                     7-16
          Electrical  Power                               7-16
          Fuel                                            7-16

       Auxiliary  Equipment                               7-16
          Ducts and Supports                              7-17
          Filters  and  Scrubbers                          7-17
          Blowers                                         7-17
          Preheaters                                      7-17
          Heat Exchangers                                7.18
          Monitors  and Controls                          7-18

                  Contents (continued)
       Care of Incinerator                               7-20
          Start-up                                       7-21
          Maintenance                                    7-21
          Storage                                        7-21

       Catalytic Equipment Manufacturers                 7-21

8.  Pollutant Detection                                  8-1

       Introduction                                      8-1

       Methods of Sampling Gases                         8-2

       Techniques and Instruments for Monitoring         8-4
          Introduction                                   8-4
          Methods of Analysis                            8-6

       Recommended Methods for Specific Classes of
         Impurities                                      8-23
          Hydrocarbons                                   8-23
          Oxygen Containing Organic Compounds            8-23
          Halogenated Compounds                          8-23
          Inorganic Oxides                               8-23
          Odors                                          8-24
          Aromatic Hydrocarbons                          8-24
          Polynuclear Hydrocarbons                       8-24

Appendix A - Glossary                                    A-l

Appendix B - Information Sources on A1r Pollution        B-l

Appendix C - Tables of Data                              C-l

Appendix D - Methods                                     D-l

       Method for Calculating Vm                         D-l
       Lower Explosive Limits of Mixtures                D-2
       Van der Waals1 Constant a                         D-3
       CC14 Test Method        ~"                         D-4
       Activated Carbon Regeneration                     D-7

Appendix E - Sample Calculations                         E-l

       Heat of Adsorption                                E-l
       Free Energy of Adsorption                         E-3

                  Contents (continued)

       Adsorption Zone                                   E-5
       Calculation of Lc at Different Cb/Ci               E-8
       Calculation of Lz and LC, Klotz Equation          E-10
       Calculations for Adsorbed-Vapor Profile as
         Presented in Figure 4-18                        E-12
       Calculations for Adsorbed-Vapor Profile as
         Presented in Figure 4-19                        E-15
       Estimate of Secondary Adsorber Size for Gas
         Regeneration with Secondary Adsorption
         System                                          E-24
       Calculations on Carbon-Resorb with Gas Regen-
         eration, Condensation and Vapor Recycle         E-27
       Calculation of Service Times for Multistage
         with Displacement of Beds                       E-31

Appendix F - Literature Cited                            F-l

Appendix G - Bibliography                                6-1



                     List of Tables

Table No.                                                Page

  1-1        Number of Complaints per Year per
             Million Population                          1-4

  l-II       Small  Stationary Emission Sources
             of Air Pollutants Grouped According
             to Preferred Control Device,  Esti-
             mated  Emission Rates                        1-7

  l-III      Rates  of Gaseous Pollutant Emissions
             from Nine Survey Areas,  Pollutant
             Emission Rate, Ib/da per 1000 People         1-10

  1-IV       Amounts of Organic Pollutants Emitted
             by Small Stationary Sources Based on
             San Francisco Bay Area  (1969) Report         1-11

  1-V        Number of Pollutants Identified with
             Each Industry                               1-15

  1-VI       Most Generally Used Nonhalogenated
             Solvents                                    1-17

  1-VII      Halogenated Solvents Used in Degreasing      1-17

  1-VIII      Solvents Used in Dry Cleaning               1-18

  1-IX       Components of Gasoline  Vapors               1-18

  1-X        Pollutants Appearing at  Highest Con-
             centration from Meat Rendering Oper-
             ations                                      1-19

  1-XI       Malodorous Pollutants Emitted from
             Fish Rendering                              1-19

  1-XII      Pollutant Vapors Emitted from Coffee
             Roasting, Tobacco Curing, Meat Smoke-
             houses, Tanning and Zinc Plating In-
             dustries                                    1-21

  1-XIII      Estimated Air Pollutant  Emission Rates,
             Surface Coating Solvents                    1-27

  1-XIV      Surface-Coating Formulas on an As-
             Purchased Basis (Weiss  1967)                 1-27

                   (Tables continued)

Table No.                                                Page

  1-XV       Gray Epoxy Appliance Primer Form-
             ulation                                     1.28

  1-XVI      Acryl1c-Epoxy Appliance  Enamel
             Formulation                                 1-28

  1-XVII      Vinyl  Lacquer Formulation  for
             Aluminum Siding                             1-29

  1-XVIII    Base for Furniture  Rubbing or  Flat
             Varnish                                     1-29

  1-XIX      Solvent Emissions from Surface  Coatings
             Operations (Bay Area, 1969)                  1-30

  1-XX       Composition of a Common Lacquer Thinner      1-32

  1-XXI      Carbon-Resorb System Size,  Operating
             Conditions,  Regenerating Steam  Cost
             for  Control  of Spray-Booth  Solvent
             Emission                                    1-32

  1-XXII      Amount  of  Solvent Emitted  by Plastics,
             Rubber,  Paper Coating and  Pharmaceutical
             Industries                                  1-43

  1-XXIII    Properties of Solvents Emitted  from
             Natural  and  SBR Rubber                       1-46

  1-XXIV      Properties of Solvent Emitted from the
             Synthetic  Rubber Industry                    1-47

  1-XXV      Emission Factors from Synthetic  Rubber
             Plants  Pounds per Ton of Product,
             (MacGraw,  1970)                              1-49

  1-XXVI      Concentration of Pollutants in  Plant
             Air  at  Various  Operating Areas  or
             Exhaust  Vents                               1-53

  1-XXVII     Pollutant  Vapors Emitted from Meat
             Rendering  Processes                          1-56

  1-XXVIII    Most Abundant Pollutants in Vapor Mix-
             tures from Crackling, Grease and Tallow
             During  Rendering                             1-58

                   (Tables continued)

Table No.                                                Page

  1-XXIX     Pollutant Emissions from F1sh Rendering     1-60

  1-XXX      Pollutant Emission Factors for Three
             Types of Incinerators                       1-66

  2-1        Federal Ambient Air Quality Standards       2-3

  2-II       Relative Amounts of Pollutants Emitted
             from Various Sources (California 1969)      2-4

  2-III      Ranking of Pollutants in Decreasing
             Order of Toxicity to Humans                 2-6

  2-IV       Ranking of Pollutants in Decreasing
             Order of Odor Recognizability               2-11

  2-V        Odor Interaction in Mixtures                2-13

  2-VI       Relative Concentrations of the More
             Reactive Hydrocarbons That Produce
             Biological Response Similar  to That
             Produced by Ethene                          2-15

  2-VI I      Sources of Smog Forming Pollutants of
             High and Low Activities                     2-17

  2-VIII     Pollutants from Small Emission Sources
             that Are Active Smog Precursors             2-19

  2-IX       Pollutant Reactivity Ranking  as Smog
             Precursors                                  2-21

  2-X        Ranking of Organic  Pollutants Producing
             Eye  Irritant and  Vegetation  Damaging
             Smogs  (Data  from  Altshuller,  1966)          2-22

  3-1        Summary of One  Million Dollar Research
             and  Development Plan (A)                    3-15

  3-II       Summary of Three  Million  Dollar Research
             and  Development Plan (B)                    3-16

  4-1        Types  of  Impregnated Adsorbents             4-3

  4-II       Effect of Pollutant Physical  Properties
             on  Adsorptive  Affinity  for Silica  Gel,
             Decreasing  Order  of Affinity Downward      4-5

                   (Tables continued)
Table No.
  4-III      Effect of Pollutant Physical  Properties
             on Adsorptlve Affinity for Activated
             Carbon at 100F Increasing Order of
             Affinity Downward                           4-6

  4-IV       Types of Molecular Sieves                    4-8

  4-V        Effect of Molecular Sieve  Pore Diameter
             on Selective Adsorption of Pollutant
             Vapors                                      4-9

  4-VI       Physical Properties of Adsorbents,
             Surface Areas and Pore Volumes              4-11

  4-VII      Physical Properties of Adsorbents,
             Densities and Mean Pore Diameters           4-11

  4-VIII     Physical Properties of Activated
             Carbons, Densities, Carbon Tetra-
             chloride Activities and Surface Areas       4-17

  4-IX       Adsorption Data and Calculated AH and
             k for Butane on Activated  Carbon Type BP    4-31

  4-X        Free Energy Change AF During Adsorption
             of Butane on Activated Carbon               4-34

  4-XI       Entropy Change AS During Adsorption of
             Butane on Activated Carbon                  4-34

  4-XII      Variation of AF or ci with Temperature
             and Amount u> of Propane Adsorbed on
             Type L Carbon                               4-35

  4-XIII     Lc Calculated from Lz Using Equation
             4-27                                        4-53

  4-XIV      V and Dv for Frequently Encountered
             Pollutants and Solvent Vapors               4-57

  4-XV       Selected Conditions for Adsorption of
             a Ternary Solvent Mixture  Used in
             Spray Coating Operations                    4-61

  4. XVI      Summary of Adsorbed-Vapor  Profile Data
             of Figure 4-18                              4-61

                                                             xvi i
                   (Tables  continued)

Table No.                                                 Page

  4-XVII      Summary of Adsorbed-Vapor  Profile
             Data  of Figure 4-19                          4-64

  4-XVIII     Regeneration of Trichloroethene-
             Activated  Carbon System Based  on
             Figure  4-21  Profiles                         4-75

  4-XIX      Estimated  Operating  Capacities  and
             Steam Requirements  for  Recovery of
             Trichloroethene, 4-Methyl-2-pentanone
             and Propanone  for a  One-Hour Adsorption-
             Regeneration Cycle  Time,  Carbon Equili-
             brated  at  3000 ppm                           4-76

  4-XX       Estimated  Steam Requirements to Regen-
             erate Carbons  Equilibrated with 4-Methyl-
             2-pentanone, Trichloroethene and Propanone
             at 10 ppm  Influent  Concentration and  at
             Varied  Operating Capacities. 212F  Steam
             Used                                         4-77

  6-1        Correlation of Pollutant  Concentration
             with  Odor  Level, Emission Source,  and
             Type  of Adsorbent System  Required            6-3

  6-II       Effect of  Pressure  on Adsorptive Capacity
             of Carbon  for  Propanone,  10 ppm Concen-
             tration                                     6-48

  6-III      Service Time for Pollutant Vapors  at
             Varied Concentrations in  a Two-Stage
             System with Displacement  of Beds             6-49

  6-IV       Auxiliary  Equipment for Sorbent Type  Air
             Pollution  Control Systems                   6-58

  6-V        Effect of  Temperature on  Adsorptive
             Capacity and Equilibrium  Concentration
             for 4-Methyl-2-pentanone  Adsorbed  on
             BPL V-Type Carbon                           6-68

  6-VI       Packaged Vapor Adsorption and  Recovery
             Systems, Approximate Equipment Cost and
             Size                                        6-76

                        (Tables  continued}
     Table  No.                                                Page
       6-VII       Cost  Factors  for A1r Pollution Control
                  by Carbon  Resorb Systems                    6-81

       6-VIII      Cost  Factors  for Air Pollution Control
                  by Catalytic  and Thermal Incinerators       6-82

       6-IX        Effect  of  Pollutant Molar Volume and
                  Concentration on Air Cleaning Cost When
                  No Pollutant  Recovery or Disposal Cost
                  is Considered                               6-83

       6-X         Cost  Factors  for Air Pollution Control
                  by Carbon-Resorb Systems at Two Flow
                  Capacities                                  6-84

       6-XI        Effect  of  Cycle Time on Cost of A1r
                  Cleaning for  3,800 fta/min Carbon-Resorb
                  System, 500 ppm Pollutant Concentration
                  with  Steam Regeneration                     6-85

       6-XII       Estimated  Operating Cost, Continuous
                  Recycle Unit  Including Solvent Recovery     6-88

       6-XIII      Estimated  Operating Cost, Continuous
                  Recycle Unit with Concentration of
                  Pollutant  for Subsequent Recovery of
                  Disposal                                    6-89

       6-XIV       Activated  Carbon Suitable for Air
                  Pollution  Control Applications              6-90

       7-1         Minimum Temperature for Initiating
                  Catalytic  Oxidation and Autolgnition
                  of Organic Pollutant Vapors in Air,
                  Degrees Fahrenheit                          7-8

       7-II        Pollutant  Concentration Required for
                  Self-Sustained Catalytic Combustion
                  at 90%  Conversion Level                     7-10

       7-III       Heat  Input from Pollutant Combustion at
                  Varied  Concentrations                       7-10

       7-IV        Auxiliary  Equipment for Catalytic
                  Incinerators                                7-16


                   (Tables continued)

Table No.                                                Page

  8-1        Classification of Analytical  Methods         8-5

  8-II       Common Analytical Methods Used for
             Air Pollutant Detection                     8-7

  8-III      Partial Listing of Gases  and  Vapors
             Detectable in the ppm Range                 8-12

  C-I        Properties of Pollutant Vapors Emitted
             from Small Stationary Sources               C-l

  C-II       FlammabiHty Characteristics  of Organic
             Vapors in Air                               C-9

  D-I        Numerical Values for the Constants of
             Equation 2                                  D-8

  D-II       Numerical Values for the Constants of
             Equation 3                                  D-8

  D-III      Numerical Values for Constants in
             Equation 12                                 D-ll

  E-I        Data Used for Calculating AF  in Table
             4-12                                        E-3

  E-II       AF Based on  p0 and f0                       E-4


                    List of Ficures
Figure No.                                               Paqe
<^**'""**"^^l'"*-                                               taMMM

   1--1       Degreaser Using Halogenated Hydro-
             carbon Vapors                               1-35

   1-2       Dry Cleaning Machine with Integrated
             Carbon Solvent Recovery System.
             Courtesy of Vic Mfg. Co.                     1-35

   1-3       Typical  Rates of Odor Emissions  and
             of Vapor Emissions from a Batch-Type
             Rendering Cooker Reducing Inedible
             Animal Matter                               1-54

   4-1       Pore Size Distribution Curves for
             Activated Carbons with Cumulative Pore
             Volume Calculated Per Unit Carbon Weight
             I Darco  Carbon, IV Superactivated Coco-
             nut, II  Type BP Pittsburgh Activated
             Carbon,  and III Type SXA Union Carbide
             Corp.                                       4-14

   4-2       Pore Size Distribution Curves for Acti-
             vated Carbons with Cumulative Pore
             Volume Calculated Per Unit Carbon Bulk
             Volume Basis, I Darco Carbon, IV Super-
             activated Coconut, II Type BP Pittsburgh
             Activated Carbon, and III Type SXA Union
             Carbide  Corp.                               4-15

   4-3       Adsorption Isotherms of Hydrocarbon
             Vapors,  Amount Adsorbed w at  Pressure p
             on Type  Columbia L Carbon at  100F, Liquid
             Volume of w at Boiling Temperature.         4-21

   4-4       Adsorption Isotherms of Butane at Three
             Temperatures on Pittsburgh BPL Type Car-
             bon, Amount w Adsorbed at Pressure p.
             Data from Grant (1962).                     4-23

   4-5       Isotherms of Water on Activated  Carbons
             Showing  Hysteresis Between Adsorption
             and Desorption Branches,  I Darco Carbon,
             II Typical of Pittsburgh BP or Union
             Carbide  SXA.                                4-24

   4-6       Nitrogen Isotherm on Activated Carbon at
             -195C Showing Hysteresis Loop Extending
             from p/p0 = 0.5 to 1.0                      4-28

xx 1 i
                      (Figures continued)

    Figure No.                                               Page

       4-7       Isosteres for Butane on Activated
                 Carbon, Type BP                             4.33

       4-8       Variation of Free Energy AF or Polanyi
                 Potential e-j with Amount of Vapor w
                 Adsorbed on Activated Carbon,   w Measured
                 per Gram of Vapor and AF and ei  per Mol
                 of Vapor                                    4-36

       4-9       Variation of Free Energy AF or ei,  with
                 Amount of Vapor to Adsorbed  on  Activated
                 Carbon,  u, AF and CT Measured per  Gram
                 of Vapor                                    4-38

       4-10      Generalized Isotherms for Propane,
                 Butane and Pentane                          4-40

       4-11      Generalized Adsorption Curves  for Three
                 Widely Different Types of Activated Carbon   4-41

       4-12      Generalized Adsorption Curves  Typical of
                 Activated Carbons Used in Gas  Masks and
                 Solvent Recovery                            4-42

       4-13      Movement of Vapor Concentration  Distri-
                 bution Curve in Carbon Bed  with  Increased
                 Adsorption Time                             4-45

       4-14      Effluent Concentration Curve of  Butane
                 Vapor  from an Activated Carbon Bed  as
                 Function of Time                            4-47

       4-15      Butane Vapor Distribution in Activated
                 Carbon Bed at Time when Vapor  has Started
                 to Penetrate Bed                            4-49

       4-16      Service Time - Bed Length Curve  for Ad-
                 sorption of Butane on Activated  Carbon       4-51

       4-17      Effect of Gas Velocity, U,  on  Adsorption-
                 Zone Length, LZ for Carbons of  Mean
                 Particle Diameter, Dp, and  Vapor Molar
                 Volume, V, 4 to 10 Mesh * 0.35 cm,  6 to
                 16 Mesh % 0.24 cm and 12 to 20 Mesh *
                 0.12 cm.                                     4-56

Figure No.
                  (Figures  continued)
   4-18      Adsorbed Vapor Profile  in  an  Activated
             Carbon Bed After Steady State is  Estab-
             lished but with no Coadsorption              4-60

   4-19      Adsorbed Vapor Profile  in  an  Activated
             Carbon Bed After Steady State is  Estab-
             lished with Coadsorption                    4-63

   4-20      Adsorbed-Phase Profile  for Trichloro-
             ethene at Service Time  when Vapor Starts
             to Penetrate Bed, BPL V Type  Carbon          4-68

   4-21      Adsorbed-Phase Profiles for Trichloro-
             ethene at Various Stages of Regeneration,
             BPL V Type Carbon                           4-68

   4-22      Amount of Steam Required to Regenerate
             BPL V Type Carbon Equilibrated wi.th  Tri-
             chloroethene at Varied  Concentrations        4-71

   4-23      Amount of Regenerating  Agent Required  to
             Regenerate BPL V Type Carbon  Equilibrated
             with Propanone at Varied Concentrations      4-72

   4-24      Amount of Regenerating  Agent Required  to
             Regenerate GI Type Carbon  Equilibrated
             with 4-Methyl-2-Pentanone  at 10  and  3000
             ppm Concentrations                          4-73

   6-1        Central  Ventilation System Showing Par-
             ticulate Filter and Carbon Bed for Air
             Purification                                6-7

   6-2        Multiple Cell Activated Carbon Filter
             for Air Purification                        6-9

   6-3        A Chemical-Biological-Radiological
             Filter,  Size 12 by 12 by 8 in.,  Operated
             at 48 ft'/min for Small Protective Shel-
             ters. Courtesy MSA Co.                       6-10

   6-4        Pleated Type Carbon Filter for Control
             of Radioactive Pollutants, Size  24 by  24
             by 11.5 in., Operated at 1000 ft3/min
             Flow. Courtesy of MSA Co.                    6-10

                      (Figures continued)

    Figure No.                                               Page

       6-5       Radial  Flow Canister Design for Air
                 Purification Systems.   Courtesy of
                 Connor Engineering Corp.                     6-12

       6-6       Arrangement of Canisters  on a  Slanted
                 Manifold Plate for Medium and  Large Air
                 Purification Systems                        6-12

       6-7       Self-Contained Canister Type Air Puri-
                 fier for Small Occupied Spaces.   Cour-
                 tesy of Connor Engineering Corp.             6-12

       6-8       Pressure Drop Through  Granular Carbon
                 Beds of Varied Mesh  Size  Fractions as
                 Function of Flow  Velocity                   6-15

       6-9       Stationary  Bed Carbon  Resorb System
                 with Auxiliaries  for Vapor Collection
                 and  Solvent Separation  from Steam Con-
                 densate or  Incineration of Pollutant
                 Vapors                                       6-20

       6-10       Cut-Away Diagram  of  Solvent Recovery
                 Adsorber, Showing  Vapor-Airflow  Pattern
                 During  the  Adsorption  and Steam  Flow
                 During  Desorption.   Courtesy of  Vic
                 Manufacturing Company.                       6-22

       6-11       Stationary  Bed Solvent  Recovery  System
                 of 3000 ftVmin Flow Capacity  for  Each
                 Adsorber.   Used for  Water Immiscible
                 Solvents.   Courtesy  of  Vic  Manufacturing
                 Company.                                     6-22

       6-12       Flow Diagram of a  Solvent Recovery
                 System.   Courtesy  of Vulcan-Cincinnati,
                 Inc.                                         6-23

       6-13       Solvent Recovery  System for Recovery of
                 Naphtha  from a Rubber  Products  Plant,
                 12,000  ft3/min Flow  Capacity.   Courtesy
                 of Vulcan-Cincinnati,  Inc.                   6-23

       6-14       Small Cabinet-Enclosed  Solvent  Recovery
                 Unit, 700 ft3/min  Capacity,  Steam  Regen-
                 eration.  Courtesy of  Hoyt  Mfg.  Corp.        6-27

                  (Figures  continued)

Figure No.                                                page

   6-15      Continuous Rotary Bed,  Based  on  Sutcliffe
             Speakman Co.,  LTD Illustrations,  66
             Palmer Ave.,  Bronxville,  N.Y.                6-29

   6-16      Fluidized Bed  Solvent Recovery System        6-31

   6-17      Air Pollution  Control  System  Utilizing
             Carbon-Resorb  with Gas  Regeneration  and
             Incineration  of Desorbed  Pollutants
             (from Mattia,  1970).                         6-37

   6-18      Adsorption-Phase Profile  After Carbon
             Bed Has Been  Regenerated,  When L  is  4
             to 8 in., and  wr is Residual  Capacity.       6-37

   6-19      Pollutant Desorption  Rate  Variation  with
             Regeneration  Time                           6-39

   6-20      Air Pollution  Control  System  Utilizing
             Carbon-Resorb  with Gas  Regeneration  and
             Secondary Adsorber for  Desorbed  Pollutants
             (from Mattia,  1970).                         6-39

   6-21      Air Pollution  Control  System  Utilizing
             Carbon-Resorb  with Gas  Regeneration,
             Condensation,  and Recycle  of  Uncondensed
             Vapor.                                      6-42

   6-22      Adsorbed-Phase Profiles at the End of
             Adsorption and Regeneration for  a Gas
             Regeneration-Condensation-Recycle Process,
             Lt s 21.3 in.                                6-45

   6-23      Two-Stage Stationary  Bed  System              6-47

   6-24      Stationary Bed with Two-Stage  Steam
             Regeneration                                 6-47

   6-25      Multistage Pollutant  Control  System,
             Internal Construction of  a Two-Stage
             System, Courtesy T. Melsheimer  Co.,  Inc.     6-50

   6-26      Two-Stage Pollutant Control System,
             10,000 ft3/min Capacity,  Courtesy T.
             Melsheimer Co., Inc.                         6-50

xxv i
                      (Figures continued)

    Figure No.                                               Page

       6-27      Moving-Bed System for Separation of
                 Low Molecular-Weight Gases (Berg,
                 1947)                                       6-53

       6-28      Rotating Axial-Flow Bed, Plan and
                 Elevation                                   6-55

       6-29      Rotating Axial-Flow Bed, Showing Seal
                 Detail and Mechanism for Raising and
                 Lowering Bed                                6-55

       6-30      Rotating Axial-Flow Bed with Condenser
                 and Recycle for Uncondensed Vapors,
                 3,800 ft3/min Airflow Capacity, 6 ft
                 Bed Diameter                                6-56

       6-31      Comparative Costs of Pollution Control
                 with Carbon Resorb, Catalytic Incinera-
                 tion and Thermal Incineration, 3,800
                 ft3/min Capacity System, Calculations
                 Based on 4-Methyl-2-Pentanone and Pro-
                 panone, 2080 hr/yr Operating Time.          6-79

       6-32      Steam Requirement for Regeneration vs
                 Pollutant Loading and Cycle Time,
                 Carbon-Resorb System on 4-Methyl-2-
                 pentanone, 3,800 ft3/min Flow, 2080
                 hr/yr Operating Time.                       6-86

       7-1       Basic Catalytic Incinerator Components      7-1

       7-2       Effect of Temperature on Catalytic
                 Combustion of 2-Butanone on Pt/alumina,
                 4" Depth Honeycomb Bed, 80,000 Space
                 Velocity and 10% LEL Concentration          7-12

       7-3       Effect of Residence Time on Catalytic
                 Combustion Conversion Efficiency            7-13

       7-4       Basic Catalytic Incinerator with Gas
                 Burner to Supply Preheat                    7-19

       7-5       Catalytic Incinerator with Heat Ex-
                 changer to Supply Preheat                   7-19

       7.6       Catalytic Incinerator with Burner to
                 Supply Preheat and Secondary Exchanger
                 to Recover Heat for Process Use             7-19

                  (Figures continued)

Figure No.                                                Page
   D-l       Static CC14 Test Apparatus                  D-6
   D-2       Logic Diagram for Air-Regeneration          D-12
   D-3       Logic Diagram for Steam-Regeneration        D-14
   D-4       Logic Diagram for Steam-Regeneration,
             Figure D-3 continued                        D-15

xxvii i

                                                            xx 1x

Objectives and Scope of Study

     The packaqe sorption device system study was conducted
for the Environmental Protection Agency,  under Contract Number
EHS D 71-2, by MSA Research Corporation,  Evans City,  Pennsyl-
vania.  It deals with packaged solid sorption devices for the
control of gaseous air pollutant emissions.

     The study had as its primary objectives  the preparation
of a handbook on packaged sorption device technology, an
assessment of the emissions to which the  technology may be
applicable, and the development of recommendations for future
research and development of packaged sorption devices.

     The term "packaged" connotes a size  limitation on the
device and thereby, also, on the size of  the  pollutant emission
source which the device would control.  For  the purpose of this
study, these sources are referred to as small emission sources
to differentiate them from mobile and other  larger stationary

     The term "packaged device" is loosely defined as  (1) a
factory assembled device transportable as a  unit and/or  (2) a
device which can be assembled on site from a  few prefabricated
and readily transportable components.  Such  packaged  devices
are characterized by:

     1.  Smal1 size
     2.  Simple, easy installation
     3.  Availability in essentially off-the-shelf form
     4.  Relatively low maintenance
     5.  Relatively low cost

     In order to match emission volumes from  the small
emission sources with potential packaged  control devices,
the air handling capacities of various sized  activated carbon
solvent recovery systems were used as a measuring guide.
Factory assembled solvent recovery systems that are essentially
off-the-shelf sizes are commercially available up to  10,000
ft3/min airflow capacity.  Such a system  would have the
approximate overall length and width dimensions of 38 and 12
ft and overall height of 16 ft.  Solvent  recovery systems up
to 50,000 ftVmin airflow capacity are available that fit the
prefabricated-component definition.  Emission sources of over
50,000 ftVmin flow rate and custom built systems to handle
these higher flow rates were considered outside the scope of
this study.

          Although  the  major  emphasis  in  this  study was on sorbent
     devices,  some  necessary  but  limited  investigations were also
     conducted  on competing devices  involving  catalytic and thermal

          Small  emission  sources,  amenable  to  control by packaged
     devices,  are exemplified  by  such  sources  as   (1) small solvent
     users  -  dry cleaners, degreasers, and  printing shops and   (2)
     small  processing and  service  industries - meat and fish render-
     ers,  coffee roasters, restaurants and  bakeries.  The general
     characteristics of these  small  sources are:   (1) they are
     numerous,   (2)  the emission  streams  are generally of the order
     of  10,000  ft3/min  or  less, and   (3)  because of their small
     size  their  polluting  effects  are  generally felt only in local

          To  achieve the  primary  objectives the study involved
     effort  in  three areas:

          1.   Identify  the small  emission sources  and determine
        the  nationwide  extent  and  type of gaseous  air pollution
        from  these  sources.

          2.   Evaluate  the existing  technology of  small fixed
        bed  sorbent  and catalytic  incineration package devices
        for  the  control  of these  sources.

          3.  Assess potential new technology  to control pol-
        lution  from  these  sources  using fixed  bed  sorbent
     Report Content

         The  report consists of  two parts.  Part I is oriented
     toward the needs of management personnel in Government and
     industry, responsible  for planning of future action  in pollu-
     tion control.  Part II  is directed more toward the needs of
     design engineers, fabricators, and users of control  devices.
     Part II also contains  information directed to the scientists
     interested in adsorption.

         Part I presents the acquired information on the emission
     sources and recommendations  for future research and  develop-
     ment of packaged sorption devices.  The information  on emission
     sources was collected  from Government publications (federal and
     local), trade journals, monographs, encyclopedias, interviews
     with trade associations and  plant visits.  In this part, the
     operating entities that were  reported to be major small emis-
     sion sources were studied in  detail to determine the extent,

                                                             xxx i
 type, and circumstances of the pollutant emitted from each.
 On  the basis of this knowledge several evaluations were made.
 Industries were ranked according to total amount of pollutants
 emitted by each and again according to potential harm to
 environment of pollutants emitted by each.  Determinations
 were also made regarding type of device (sorbent, catalytic
 incineration, or thermal incineration) which would be the
 most economical one to use to control the pollutant emission
 from each major operating entity.  As primary objective of
 this phase of the study, recommendations are presented for
 future study to fill in the information gaps.

     Part II, also entitled Package Sorption Device Technology
 Handbook, is the most important product of this study.  In
 this part of the report several aspects of pollution control
 technology are considered, with the major emphasis, again, on
 the sorbent type systems.  A theoretical basis is presented
which permits the prediction of behavior of sorbents for the
 variety of pollutants emitted and circumstances of emission
 that would be encountered.  Various sorbent systems are dis-
 cussed with respect to their adaptability to air pollution
 control.   Some systems are commercially available (activated
 carbon solvent recovery systems), some are proposed systems
 that have undergone limited pilot plant tests but have not
as yet been reduced to commercial practice, and some are still
 in the conceptual  stage, appearing sound on theoretical  bases.
Economic analyses  are presented for the operation of carbon
adsorbent systems, catalytic incinerators, and thermal incin-
erators under several operating circumstances.  These results
present a cost level that can be expected for pollution  con-
trol and  also indicate the operating conditions under which
each of the three  competing systems would be the most eco-
nomical  to use.

     The  appendix  contains a bibliography pertinent to air
pollution and its  control by sorption and incineration.

xx xi 1

                                                            xxx iii
                Summary and Conclusions

Study Approach

     To attain  the objectives of the study,  a procedure  of
action was devised to collect available information on:
(1) small emission sources,  (2) environmental  effects  of
the emitted pollutants, and  (3) packaged sorbent and cata-
lytic incineration technology suitable for control  of such
sources.  The collected information was analyzed and used
to:  (1) categorize the emission sources as  to  volume output
and pollutant type,  (2) evaluate the adverse environmental
effects of pollutants from small emission sources and  (3)
evaluate the control  technology in present use  and  further
investigate the reed  and potential for development  of new
sorption system technology.

     Information on emissions and control technology was
sought from numerous  sources:  trade associations,  local and
state pollution control agencies, sorption device users  and
manufacturers,  trade  journals, text books, monographs,  govern-
ment documents, and vendor sales literature.  The methods of
data collection included plant visits to manufacturers  and
users of control systems.  A questionnaire survey of pollution
control agencies was  also used to elicite information to iden-
tify and characterize small sources and their emissions.

Emission Sources Characteristics

     Information Search

     To fulfill the study objectives regarding  emission  sources
in the most satisfactory manner, the study required detailed
information on:  (1)  identity of small emission sources,  (2)
amount of pollutant emitted at each source,   (3) concentration
of pollutants in each emission gas stream, and   (4) types of
pollutants emitted in each gas stream.  During  the  processing
of data from the various information sources, it became
apparent that emission data in the desired detail was not
available.  The questionnaire, which was sent on a  trial basis
to nine large pollution control districts, was  expected  to tap
a rich source of detailed information of the required type.
However, the response to the questionnaire was  incomplete
because of general lack of data of the type requested.  The few
agencies which  do collect detailed emission data are not
sufficiently staffed  or funded to allow compilation of avail-
able data for the numerous requests they receive.  The
agencies in the State of California, in particular, have
collected extensive data but would not provide it  in any form
other than formal reports and publications.  Because of this

     problem, the study was of necessity developed from more
     approximate data as presented in the control  district
     formal  reports,  trade association journals,  and  special
     reports.  Information on types of pollutants  were  extracted
     from process descriptions as  presented in  text books, mono-
     graphs  and encyclopedias.

          Extent of Small  Sources  Emissions

          Based on survey  figures  given  in  formal  reports  from
     nine large air pollution control  districts,  - California
     South Coast Basin, San Francisco  Bay Area, San Diego  Bay
     Area, Delaware Valley, Denver, Hamilton  County in  Ohio, St.
     Louis,  Washington (D.C.), and Jacksonville in Florida, -
     the estimated average rates of various types  of  gaseous
     emissions  during the  years 1968 to  1970  were:

       Sources  and type       Average  rate, Ib/da  per 1000 people

       All sources organic
        vapors  (includes
        motor vehicle fuel
        and  exhaust emis-
        sions)                                 800

       Industrial  processes
        organic vapors                        100

          The combined population  of the  survey areas was  about
     27  million,  or over 13%  of the total  U.S. population. These
     are also the  areas of greater population density;  hence, the
     emission rates given  are characteristic  of areas where pollu-
     tion problems are most acute  and  controls are most needed.

          The 100  Ib/da of industrial  process organic vapors
     virtually  comprise emissions  from small  emission sources.
     This rate, however, is not typical  of  all control  districts;
     it  is based  on such widely varying  rates as 20 Ib/da  for
     Denver  to  156 Ib/da for  San Francisco  Bay Area.  To these
     rate figures  should be added  the  emissions from  an important
     small emission source, the petroleum distillate  marketing
     chain,  which  did not  appear in the  industrial  process emission
     rate figures.  An average distillate emission rate for the
     nine districts could  not be determined,  but it is  expected to
     be  substantial  because the average  for the three California
     districts  was 60 Ib/da per 1000 people.  Because of the
     variations in emission rates  from area to area and lack of
     complete data for all  areas,  it was  not  possible to determine
     the  proportion of emissions from  small  sources relative to

organic vapors from all  sources
portion is 15% with the  balance
with exactness.  The
being primarily from
     Small Emission Sources Identity and Ranking

     Over fifty industries were considered in the study but
for many of them comprehensive information permitting deter-
mination of their contribution to total  organic emissions  from
small sources was lacking.  The best guide in this respect was
the California control districts which were the only ones  to
present any extensive breakdown of industrial processes into
specific industries.  On the basis of their data, 22 industries
were identified as important small emission sources because of
quantity of pollutants emitted or because of intense adverse
effects of their pollutants on the environment.  The emissions
from them are predominantly organic except for the zinc plating
operations which emit only NH3 and HC1.   Rubber products and
domestic incineration also emit S02 and the rendering indus-
tries, H2S.
      Surface coating
      Dry cleaning
      Graphic arts

      Plastic products
      Rubber products
      Adhesives production
      Paper coating
      Vegetable oil extraction

      Organic liquid marketing
       and storage

      Meat rendering
      Fish rendering
      Food serving
      Coffee roasting

      Tobacco curing
      Meat smoke  treatment
      Domestic  incineration
      Zinc plating
      Emission rate,   '
    Ib/da per 1000 people




           Baking                                 0,04
           Candy making                         -

     (1) For some industries (as for instance canninq)  no
         rate figures were given, but are recognized as pollutant
         emission sources of importance.
          The industries comprising the small  emission sources
     were classified into two groups:   (1)  the solvent user
     industries, the first ten in the  list, and  (2)  the process
     industries, the last twelve in the list,  which generate
     pollutants through thermal, chemical,  or  biological pro-
     cesses.  The solvent user industries  clearly emit a larger
     volume of pollutants (solvents)  than  the  process industries,
     except for domestic incineration.   Domestic incineration was
     permitted in the San Francisco Bay Area at the time of the
     survey, but the general trend in  the  country is  to ban all
     domestic incineration,  and dispose of  solid refuse by means
     which do not pollute the atmosphere.

          The surface coating industry, the largest solvent
     emitter, is comprised of many small  operating units.  Of
     these,the industrial spray booths  are  the most numerous and
     are responsible for 45% of the surface coating emissions.
     Emissions from ovens and dryers  used  to cure the coatings
     contribute 17% of the total.  Nonindustrial means such as
     outdoor painting contribute 36%.
     The next largest organic vapor emitting industry is th
industry indicated as organic liquid marketing and storage.
This includes, as the major entity, the petroleum dlstillat
distribution and marketing chain where the largest vapor
losses occur from gasoline marketing.

     The other solvent user industries drop off considerably
in emitted volumes.

     Type of Pollutants

     The chemical species of pollutants emitted from solvent
user industries and process industries are quite different.
Solvent emissions from the degreasing and dry cleaning indus-
tries are predominantly the chlorinated solvents, - trichloro-
ethene, tetrachloroethene, trifluorotrichloroethane - the
three most commonly used ones.  Solvent emissions from other
users are:  hydrocarbons (aromatic,olefinic, naphthenic and
aliphatic), alcohols, ketones, ethers, glycol ethers, esters,

                                                          xxxvi i
and nitroparaffines.  In these industries the nature of the
emission is quite predictable from the solvent blends entering
the process except where high temperature drying or baking
operations are carried out.  In these cases the pollutant mix-
ture may also contain decomposition products of drying oils,
resins, and high boiling solvents.

     Pollutant emissions from the process industries are
predominantly decomposition products of materials entering
the process.  The chemical  species found in the emissions
from various thermal processes are predominantly aldehydes,
organic acids, amines and sulfides.

     Conditions of Pollution Emission

     In addition to pollutant type, the conditions of pollution
emission such as concentration, total gas flow, gas temperature,
humidity, and particulate matter in the gas stream also vary
between emission sources and must be known before the source
can be characterized and the type and size of control system
best suited for its control determined.  Recovery value of
the pollutant is an additional factor that has an important
bearing on the selection of control system.  To obtain the
necessary information on operating conditions and determine
the best suited control system, an in-depth study was done on
each of the 22 industries as listed.

     For pollutants of low recovery value, the favorable
conditions for control by incineration are, high concentration
and/or elevated temperature.  Emissions from coffee roasting
and meat smokehouses occur under these conditions.  For
pollutants of high recovery value, carbon adsorbent systems
are highly economical to operate when the concentration is
high and gas temperature is moderate.  Emissions from most
of the solvent user industries occur under these conditions.

Environmental Effects

     Pollutants produce effects in the environment through
various means that make life unpleasant or difficult.  Most
important is the toxic effect which can produce discomfort,
illness and lower the performance efficiency of humans and
animals.  Pollutants can cause serious damage to vegetation.
Other effects are damage to property through corrosion,
discoloration and soiling,  A nuisance effect which arouses
the most immediate complaint is the bad odor response to
many pollutants.  Some pollutants, although not directly
damaging to the environment, react with nitric oxides with
the aid of the sun's radiation to produce smog which obscures
visibility and produces discomfort and illness since it con-
tains oxidants that are toxic.

xxxvi ii
          The pollutants emitted by the solvent user industries
     are not highly toxic nor malodorous (except for I^S  from
     rubber products), but about 20% of them are smog precursors.
     Their damage to the environment is primarily by this route.
     Smog precursors have been identified with branched  chain
     ketones, substituted aromatics and hydrocarbons containing
     unsaturated bonds.   Aldehydes are also smog precursors  but
     are rarely used as  solvents.   Los Angeles County Air Control
     District recognized this factor as a major cause of  air
     pollution and issued Rule 66  which drastically limits  the
     use of smog active  solvents.

          Compared to solvent emissions, the pollutant emissions
     from process industries  are more toxic, malodorous  and
     corrosive.  The latter emissions also contain a large  portion
     of aldehydes and olefinic compounds and thereby also con-
     tribute to smog formation.   Because of the small quantity of
     pollutants emitted, the  threat to the environment by the
     process industries  is considerably less than that from  the
     solvent user industries.

     Packaged Air Pollution Control  Devices

          Information Sources

          Most of the information  that proved of value in this
     area was obtained from technical  reports in scientific
     society and trade journals, from monographs, sales  literature,
     and discussion with technical  representatives of manufacturers
     of devices, catalysts and sorbents.  Information was sought
     and found on theoretical aspects  of adsorption and  catalytic
     combustion, types of available devices, their potential  appli-
     cations to small  emission sources, their operations  and  main-
     tenance, and their  operating  costs.

          Characteristics of  Pollution Control  Devices

          Sorbents  - The scope of  this study included all  types of
     sorbent systems that would  be  adaptable for air pollution
     control.  There are three types  of sorbents in use for  various
     purposes;  namely, chemisorbents,  polar adsorbents, and  non-
     polar adsorbents.  Chemisorbents  and polar adsorbents are not
     generally applicable to  control  of gaseous pollutants under
     the conditions of their  release  from small emission  sources.
     Chemisorbents  would have very  limited application; they  are
     highly  selective  since sorption  is determined by highly
     reactive functional  groups  1n  the pollutant molecule.   Polar
     adsorbents,  of which silica gel,  activated alumina,  and  mole-
     cular sieves are  the best known  examples,  are effective  for
     dehumidifying  organic gases.   Since pollutant-air mixtures

                                                            xxx ix
 invariably  contain moisture, polar adsorbents, because of
 their  affinity for moisture, would not be effective for
 pollution control for any practical length of time.  The
 nonpolar activated carbons preferentially adsorb organic
 vapors, the type of vapors emitted from small emission
 sources.  Further study thereby narrowed almost exclusively
 to sorption control devices using activated carbons.

     Activated Carbon Adsorbent Systems - Activated carbon
 adsorbent systems, as presently manufactured for industrial
 and commercial uses, are oriented toward "air purification"
 and "solvent recovery".  In either system the solvent or
 pollutant and air mixture is passed into the granular acti-
 vated  carbon bed wherein the adsorbable gas or vapor is
 retained by the carbon and the purified air passes through.
 When the carbon has reached its capacity to retain vapor,
 the gas mixture flow is stopped or diverted to a system
 containing  fresh carbon.

     Air Purification Systems - Air purification is carried
 out in a closed system, i.e., the carbon purified air is
 recirculated into the occupied space.   Make-up air from
 outdoors is added to the recirculating airstream to maintain
 C02 and moisture concentrations at acceptable levels.

     The carbon beds in air purification systems are thin;
 depths of 0.5 to 3.0 in. are in frequent use.  Bed areas are
 sized  to control airflow velocity in the 20 to 60 ft/min
 range with most operated at about 40 ft/min.  Coarse mesh
 carbons of 4 to 6, 4 to 8 and 4 to 10  U.S.  sieve number are
 used.  The thin bed, the relatively low airflow velocity,
 and coarse mesh all contribute to low  airflow resistances
 which are generally below 0.25 in. of  H0 pressure drop.
 The service times attained vary from 6 mo,  when the odor
 concentration is described as heavy, to much longer periods
 of up to 2 yr, when odor concentration is described as light.
 These are nonregeneratiye systems, i.e., the carbon bed cannot
 be readily regenerated in place.   At the end of the service
 time, the spent carbon beds are removed from the cells and
 the carbon is reactivated with high temperature steam at the
 carbon manufacturer's plant.   In  an atmosphere where the
 pollutant concentration is higher than 1.0  ppm, the frequency
 of carbon bed replacement would greatly increase and the
 operating costs would rise to prohibitive levels.  When the
 concentration is 10 ppm or higher a regenerative type system
 becomes a definite economic necessity.

     Although these systems have  been  intended for internal
 use,  they are well suited for control  of emission sources
where pollutant concentrations are low but need to be con-
 trolled because of highly malodorous or toxic emissions.

     Solvent Recovery Systems - In solvent recovery systems
the spent carbon is regenerated within the system with low
temperature (212F) steam.  The solvent is separated from
the steam by condensation and decantation or distillation.
Profitable solvent recovery is possible when the solvent is
relatively high priced and vapor concentrations are high
during the adsorption phase.  Profit-loss breakeven concen-
tration for most solvents is near 700 ppm, hence, the interest
in the past has been on solvent vapor concentrations well
above 700 ppm level.

     Three types of solvent recovery systems are in use as
differentiated by the manner in which the carbon bed is
maintained or handled during the adsorption and regeneration
phases of the operating cycle.  These are:

     Fixed bed
     Moving bed
     Fluidized bed
     In the fixed
 tained in a verti
 downward through
 flow velocity at
 36 in.  A coarse
 sieve number, is
 practical level.
 from 6 to 18 in.
 carbon decreases
 temperatures are
 bed system the carbon bed is usually con-
cal or horizontal  tank with airflow usually
the bed.  Bed areas are sized to control air-
near 100 ft/min.  Bed depths vary from 12 to
mesh activated carbon, usually 4 to 6 U.S.
used to maintain airflow resistance at a
 The pressure drop through the bed can vary
of H20.  Because the adsorptive capacity of
with increase in temperature, high inlet air
avoided, temperatures below 130F are recom-
     Other components of the solvent recovery system may
include duct work, a particulate filter, inlet air cooler,
blower, condenser for the solvent-steam mixture, decanter,
distillation column, steam generator, and control instruments.
The cyclic operations of the recovery system are very often
automated with the various valves and other moving parts
motor operated.

     The fixed bed system is the least complicated of the
systems available, has the least moving parts and is the most
trouble free.  It is also the largest in size per unit air
handling capacity.

     In the moving bed system, the carbon may be confined in
a retainer and the retainer moved alternately between the
part of the system where adsorption occurs and the part where
regeneration occurs.  In one system that employs this concept,

the carbon is retained in the annular space between  two  per-
forated cylinders.   As the cylinder or "rotary bed"  is  rotated,
divided sections of the bed pass  through  an area  where  the
vapor-laden air is  directed radially inward through  the  bed
vupui-iuuuu u r i  19 uircibcu i a u i a i i y iiiwaiu in i u u^u  tue
to carry out the adsorption phase.   Simultaneously at a
different area  of the rotary bed, previously spent sections
of the bed are  regenerated by low pressure steam directed
outward through the bed.   On each rotation, all  sections of
the rotary bed  undergo an adsorption-regeneration cycle.

     The advantages of the rotary bed system are:  (1) the
system is compact and  (2) it utilizes shallower beds than
the fixed bed system and  thereby, for a given airflow capacity,
is operated at  a lower pressure drop.  The disadvantages are
associated with maintenance of the larger number of  moving
parts and seals that are  necessary to separate the adsorption
and regeneration sections of the system.

     In fluidized bed systems, a high upward airflow velocity
(240 ft/min) is maintained through the carbon bed which ex-
pands the bed and produces continuous movement or agitation
of the carbon particles.   In the fluidized bed adsorber, the
carbon is caused to flow  in a fairly thin layer (4.0 in. deep)
across successively lower horizontal trays that are  stacked
vertically.  The solvent  laden air flows upward through the
trays making contact with the carbon in a countercurrent
manner.  The spent carbon, withdrawn from the bottom tray, is
elevated to an  overhead tank in which the carbon is  moved
downward countercurrent to the flow of low pressure steam to
desorb the solvent and regenerate the carbon for recycle
through the system.  The  countercurrent contacting of the
fluids and carbon in both phases of the adsorption-regener-
ation cycle offers a more efficient utilization of the carbon
and regenerating steam.  Large volumes of air can be stripped
of vapors with  a relatively small recovery system and consump-
tion of steam is greatly reduced.  An important disadvantage
is the loss of  carbon by attrition.  This also produces carbon
dust in the exhaust stream which should be filtered out to
avoid air pollution.

     The results of this study have shown that modifications
to the present-type solvent recovery systems and changes  in
their operating procedures offer some of the more economical
approaches to controlling pollutant emissions, specifically
at concentrations below 700 ppm, down to the 1.0 ppm  level
where the air purification system becomes more economical to
operate.  To indicate the broader application for the  (regen-
erative) solvent recovery type systems  they  are  designated
as carbon resorb systems  in this report.

xli i
          Carbon  Resorb  Systems  -  A controlling  factor  in  the
     operation  of carbon resorb  systems  at low  concentrations  is
     the  small  operating capacity  and  difficulty of  regeneration.
     At decreased vapor  concentrations the available operating
     capacity becomes  smaller but  the  amount  of  regenerating agent
     required becomes  larger.   If  steam  is used, the amount  re-
     quired  is  so large  that it  is  usually impractical  to  separate
     the  vapor  from the  steam.   By  using  air  or  inert gas  at
     elevated temperatures  for  regeneration the  desorbed-vapor
     disposal problem  is alleviated.   The desorbed vapor can now
     be incinerated, or  readsorbed  onto  a secondary  carbon bed
     and  recovered by  conventional  steam  regeneration or recovered
     by condensation.   In each  case the  auxiliary device for the
     disposal operation  is  small and operable at low cost  because
     of the  high  vapor concentration in  the regenerating gas

          Incinerators or Afterburners -  To place the effectiveness
     and  economics of  carbon resorb systems for  pollution  control
     in proper  perspective,  a  limited  evaluation of  the competing
     systems, - catalytic and  thermal  incinerators,  - was  necessary.

          Catalytic and  thermal  (direct-flame)  incinerators  are
     effectively  used  in many  applications to control pollutant
     emissions  where other  means would be less  economical.   The
     use,  however, becomes  of  questionable value when the  pollutants
     contain  halogenated, sulfonated,  or  nitrated hydrocarbons since
     these functional  groups are converted to inorganic acids  which
     may  be  more  harmful than  the  pollutant in  its original  form.

          The efficiency of  each system  increases as the operating
     temperature  and gas residence  time  are increased.  In practice
     90%  to  98% efficiencies have  become  acceptable  performances
     because  of various  factors  which  limit the  maximum temperature
     and  residence time.  Tor  thermal  incineration the  generally
     used  furnace temperature  range is 1200"  to  1500F  and for
     catalytic  incineration  it  is  600 to 900F. The residence
     time  for both is  generally  between  0.05  and 0.5 sec.

          The factors  that  are  important  in determining the
     merits  of  the incinerators  relative  to carbon resorb  systems
     for  pollution control  are  the  cost  of heating the  air to  the
     incineration temperature  and  the  catalyst  life. At concen-
     trations below 100  ppm, the combustion of  the pollutants
     contribute very little  to  heat Input; the  major portion of
     the  heating  load  is borne  by  an auxiliary  heater.  As the
     combustible  pollutant  concentration  Increases the  supple-
     mentary  heating requirements  decrease.   At  a specific con-
     centration for each pollutant  the combustion becomes  self-

                                                           x 11 i 1
sustaining.  The pollutant emission streams often contain
trace amounts of inorganic acids on particulates which can
deactivate the catalyst, making necessary costly catalyst

     Economic Analyses

     Operating costs were estimated for carbon resorb,
catalytic incineration and thermal incineration systems at
pollutant vapor concentrations ranging from 10 to 7,000 ppm.
The results of these analyses are presented graphically in
Figure 6-31 which shows the cost of cleaning air in $/hr per
1000 ft3/min of airflow.  At 10, 500 and 4,000 ppm pollutant
concentration levels, the cleaning costs for the three types
of control systems for an influent airstream at 77F are:
                                $/hr per 1000 ftVmln
  	System	     10 ppm   500 ppm   4,000 ppm

  Carbon resorb               0.65      0.85       0.90
  Catalytic incineration      0.95      0.85       0.60
  Thermal incineration        1.27      1.10   0.70 to 0.95
     As the vapor concentration rises above 500 ppm, more
frequent regenerations are required and the operating cost
of the carbon resorb systems rises.  Unless the pollutant
vapor has recovery value, the carbon resorb systems are not
as economical to operate as the incinerators at these higher
concentrations.  At concentrations above 4,000 ppm air
cleaning costs with carbon resorb systems increase to the
$0.90/hr per 1000 ft3/min level while the costs with incin-
erators drop to $0.60/hr per 1000 ft3/min level and possibly

     At elevated airstream temperatures and low vapor concen-
trations the incinerators become increasingly more economical
to operate and can approach the operating cost level of the
carbon resorb systems.

Research and Development Recommendation

     The results of the review and evaluation of existing,
commercially available package devices for control of
emissions from small sources show that for the areas of air
purification (pollutant concentrations below about 1 ppm)
and solvent recovery (recoverable solvent concentrations

     above  700  ppm)  existing  adsorbent  system  technology  is
     generally  adequate  and  the  control  devices  are  available.
     Significant  short  range  benefits in  emission  reduction
     could  be obtained  by  accelerated implementation  of  the
     existing adsorbent  devices  technology,  particularly  to  the
     solvent user industries.  The  problem areas are  in  the
     process industries  where  research  and development is  re-
     quired to  demonstrate the effectiveness of  carbon resorb

         The research  and development  plans recommend work  in
     four areas:   (1) emission and  source characterization,
     (2) evaluation  of  new adsorbent systems,   (3) generation
     of  fundamental  adsorption data, and  (4)  monitoring  instru-
     ment development.

         The objective  of the emission  and  source character-
     ization is to provide detailed chemical and engineering data
     necessary  for the  selection  and design  of optimally  effective
     control systems.   Evaluation of new  adsorbent systems would
     involve laboratory  feasibility studies  on a potentially
     promising  system and  a  final plant  scale  demonstration  in one
     of  the industries  experiencing control  problems.  Adsorption-
     regeneration data  useful  to  the design  engineer  of  adsorbent
     systems is lacking; a program  to generate this  data  would
     greatly expedite design work.  To  attain  effective  performance
     from many  of the new  proposed  systems,  monitoring and control
     instrumentation need  to be  developed.

                      Part I

Sources and Emissions, Environmental  Effects,  and
     Research and Development Recommendations

              1.  Sources and Emissions


     The systematic Investigation of small  pollutant sources
and their emissions required the formulation of certain pre-
liminary selection or screeninq criteria.   The general  quide-
lines provided at the outset of the proqram simply required
investigation of all potential sources that could be amenable
to control by packaged or off-the-shelf  devices.   The first
and most important selection criterion was  thus one of  size
or total emission volume.  Analysis of the  sizes  of typical
sorbent or catalytic oxidation devices placed an  upper  limit
on total gas volume readily handled as be1ng!of the order of
10,000 scfm.  In order to include air purification systems,
where large volumes of air are handled for  recirculation to
working or inhabited areas, some consideration has been
given to treatment of air volumes as large  as 50,000 scfm.
Above 50,000 scfm, sources are considered to be completely
outside of the scope of the study.

     As the study progressed and specific pollutant emissions
were Identified and related to the sources, additional  cri-
teria were developed and applied to the  candidate sources.
These criteria were based on the physical and chemical  char-
acteristics of the pollutants and on the operating parameter
ranges for sorbent or catalytic oxidation systems.  These
criteria may be briefly summarized as follows:

     Carbon Sorption Systems

          (1)  Molar volume  (Vm) of pollutant  1n the
              range from 80 to 190 cm3/mole.

          (2)  Pollutant  concentration in the  range
              from  about 10 ppm to near 25  percent
              lower explosive  limit (LEL).

          (3)  Exhaust stream  either free of  partlcu-
              lates or easily  filtered.

          (4)  Exhaust stream  moisture content con-
              trollable  to  below  saturation,  pre-
              ferably below 80*  RH.

          (5)  Exhaust stream  temperature controllable
              to  130F or  below.

          Catalytic Oxidation Systems

              (1)  Pollutant readily combustible.

              (2)  Pollutant concentration  between
                   300 ppn and 25%  LEL  if gas  temper-
                   ature is near ambient,

              (3)  Pollutant concentration  of  less
                   than 300 ppm if  exhaust  stream can
                   be heated from 300P  to 500F.

              (4)  Exhaust stream either free  of
                   particulates or  readily  filtered.

              (5)  Exhaust stream free  of catalytic

          Thus,  applying these criteria to  a wide variety  of
     potential  emission sources resulted in the selection  and
     classification of small emission  sources  suitable  for more
     detailed study.

     Information Sources

          The major information sources utilized in  the study
     included the trade magazines,  technical literature, reports
     of related  government contract studies, and data obtained
     by letter  and interview contact with trade associations and
     pollution  control agencies. To aid 1n the preliminary
     literature  search, computerized key work  abstract  printouts
     were obtained from APTIC.

          The trade associations and pollution control  agencies
     contacted  and reports of related  government contract  studies
     are listed  in Appendix B.

          At the outset of the program  it was  felt that a  signif-
     icant body  of information pertinent to the study resided In
     the files  of regional, state and  municipal pollution  control
     agencies.   Many of these agencies  have been involved  in
     pollutant  source inventory studies and control  activities
     for years  and thus should be a rich source of information.

          A questionnaire was designed  to elicit Information from
     these agencies relevant to the Identification and  character-
     ization of  small sources and their emissions.   The question-
     naire, shown in Appendix B, consisted  of  two questions with
     a  form for  listing the desired Information.  The first

question asked for a breakdown of the number of citizen  com-
plaints received per year by the aqency with a check  list of
what were felt to be some of the more common sources  with a
cross listing of the type and nature of the complaint (i.e.,
odor type; effect on health, materials and vegetation,  etc).

     The second question asked for a listing of specific
data on the sizes, rates and types of emissions and concen-
trations of contaminant species for each of the identified
small source categories.

     Preliminary contact was made with a number of control
agencies to determine whether they had the type of informa-
tion desired.  Only nine agencies so indicated and thus  were
sent copies of the questionnaire.  These nine were:

	Area	  Population  (1969-1970)
San Francisco Bay Area, California              4,100,000
New York, New York                              7,800,000
Houston, Texas                                  1,200,000
Jacksonville, Florida                             510,000
Washington, D. C.                                 800,000

State of Illinois, Springfield, Illinois       11,000,000
Albuquerque, New Mexico                           240,000
Pittsburgh, Pennsylvania                          510,000
Vancouver, Washington                              40,000

     Of these nine agencies which were sent questionnaires,
six replies were received.  Of these six, three agencies
indicated that they could not supply the desired Information
for various reasons.  These were San Francisco Bay Area,
Houston and the State of Illinois.

     The Washington, D. C. control office stated that they
did not have sufficient information available to complete
the questionnaire, but they did provide a record of complaints
for 1970 grouped under the general headings of heating plants,
Incinerators, vehicles, open burning and miscellaneous.  Lead-
ing the list of complaints, which totaled 1304, was the Incin-
erator source category with 283.

     Jacksonville and Albuquerque agencies completed the
first question but were not able to supply relevant infor-
mation for the second.  The results of Question 1 are summar-
ized in Table 1-1 for the two areas.  To permit combining
of the numbers of the two areas, they are reported on the
per million population basis.

      Table  1-1   Number of Complaints per Year per Million
      Dry cleaning

      Paint shop

      Print shop

      Gasoline transfer


      Food stores

      Small chem. mfq.


      Woodwaste burners

      Sewaqe plants

      Pulp mills

      Large chem. mfq.


ft   E  O
3   3
O   0>  +>
%-   r-  C
3   O  0
*-   U  >

T-   >,
f.   V
3   01
CO   Q
               ca a n a en an
            i  n  11  I
I   ii   ii   irnnni  i  n   i
iii   IT~II  ^rgcni  irnni   i
i   ii   n   n  ii  iraoii   i rni   11i

     No reply was received from the Pittsburgh-Alleqheny
County A1r Pollution Control Office.  However, on a prior
visit to their office, data were obtained on complaints
they had investiqated.  These were under headings such as
smoke, open fires, coal deliveries, dust-fly ash, dumps,
fumes-odors, and incinerators.  The incinerator complaints
for the first ten months of 1970 totaled 179.

     The compilation of public complaints gives some indi-
cation of the types and frequency of noxious emissions,
however, such data may be misleading.  For example, a single
source of annoyance may be the cause of all  of the complaints
within a municipality relating to that source category.  The
manner in which the complaint data are collected and compiled
do not allow analysis of the number and distribution of
Individual sources which are the cause of complaint.

     Certain source categories were identified by the
survey which had not been previously considered.  Signif-
icant among these are some large sources which, however,
may have small unit operations amenable to control by pack-
aged devices, such as wood waste burners, paper mill opera-
tions and sewage disposal plants.  Large chemical complexes
are also often sufficiently fragmented as to be treated as
a group of small sources.

     After having contacted a number of local pollution
control agencies, the original conception of the ready
availability of relevant data from these sources was found
to be incorrect.  The most general response in the initial
contacts, reinforced by the questionnaire returns, was a
general lack of the type of data required.  Additionally,
those few agencies which have made any effort in this
direction are not sufficiently staffed or funded to allow
compilation of available data for the numerous requests
which they receive.  The agencies in the State of California
in particular have collected extensive Information but can-
not provide data in any form other than their scheduled
reports and publications.

     Most of the activities of pollution control agencies
have been directed toward definition and control of the
major emission sources, with little attention thus far to
those small sources which constitute only a small portion
of the total pollutant emission load.  Some relevant Infor-
mation was obtained from published regional inventories,
however, because of the definition of certain small sources
which when lumped together constitute a so-called major
source.  Examples of such small sources are the numerous
Individual solvent users (graphic arts, dry cleaning and

      surface  coating establishments) which are generally lumped
      together as major sources 1n emission Inventories.  The de-
      tailed treatment of  this type of data 1s presented In a
      later section  of this  chapter.
      Emission  Sources
           Listing of  Industries

           Table  l-II  lists 22  Industries which, 1n part or 1n their
      entirety, are  1ndentif1ed as small, stationary, emission
      sources of  vapor  phase air pollutants.  These Industries were
      Investigated to  varying degrees of depth to determine the
      chemical nature  of  the pollutants, the amounts and concen-
      trations of the  emissions, and the operating conditions or
      circumstances  of  the emissions.  This 11st 1s by no means
      complete, but  Includes those Industries most frequently men-
      tioned  1n surveys conducted 1n various air pollution control
      districts.  The  variety of Industries represented gives a
      good  Indication  of  types  of pollution problems which will be
      encountered for  a considerable period of time 1n the future.
      Other existing and  potential sources of emission are listed
      on  page  1-67.

           The 11st  of  Industries 1n Table l-II 1s divided Into
      two groups  on  the basis of pollutant control methods.  Those
      1n  the  first group  have processes amenable to control with
      activated carbon  adsorption systems, although 1n some cases
      operated 1n conjunction with absorption and/or catalytic or
      thermal Incineration systems.

           Those  1n  the second  group are not amenable or are eco-
      nomically borderline 1n regard to pollutant control by the
      carbon  adsorption systems.   Catalytic  and thermal  Inciner-
      ation systems  are more economical 1f the pollutant concen-
      tration 1s  high,  the emitted gas temperature 1s high or the
      pollutant mixtures  contain low molecular weight pollutants
      not controllable  by adsorption.  In no case 1s the emitted
      pollutant of sufficient value to be recovered.

           Categorizing of Industries

           The first nine Industries 1n the first group are solvent
      users wherein  the solvent passes through the operations essen-
      tlally  unchanged  and Is ultimately released by one or more  of

Table l-II
Small  Stationary Emission Sources  of Air
Pollutants Grouped According to Preferred^)
Control Device. Estimated Emission Rates
              Emission rate,
              Ib/da per 1000
              people exposed
        -Preferred Control With Sorbent Devices-
Surface coating
Degreasi ng
Dry cleaning
Graphic arts

Rubber products
Paper coatings
Vegetable oils

Organic storage evap,

Meat rendering
Fish rendering

Candy making






-Preferred Control With Catalytic or Thermal Incineration-
Coffee roasting
Tobacco curing
Meat smokehouses
Domestic incinerators
Zinc plating
  (1)  Based on economic
  (2)  Rate based on the
       petroleum fuel in
             three California
        survey areas for

     several means.  Because of the solvent volatilities, the
     major means of release is by vaporization while some may
     be retained by the product produced or retained in sludge
     or solid wastes produced by the operation.  Unless some
     form of control is practiced, the evaporated portion becomes
     an atmospheric pollutant.  Except 1n the case of industry
     start up or where incineration 1s used for pollution control,
     the amount of solvent sold per year to each Industry is a
     measure of its pollutant emission rate.  Some of the solvent
     users of combustible solvents, such as the surface coating
     and graphic arts, use Incineration; hence, the solvent sales
     to them Is an over estimated measure of their pollutant
     emission rates.

          Industries listed with identification numbers 51 and
     larger are chemical or thermal processing Industries.  Pol-
     lutant vapors are emitted from them as an undesirable by-
     product.  In these cases direct measurements are required to
     determine the pollutant emission rates which, in many cases,
     have been difficult.

          The Identification numbers associated with the Indus-
     tries are used for reference purposes to Identify the pollu-
     tant with the industry.  Since the 11st does not include all
     the present (and also future) potential emission sources,
     the numbering system has been designed to allow future In-
     sertions into the two broad groups.  Those from 1 to 50
     Inclusive are the solvent users and those numbered 51 and
     above are chemical,  thermal, and biological process Indus-

          Extent of Emission

          In order to:  (1) determine the extent of pollutant
     emissions from small sources relative to the extent from all
     sources and  (2) determine the amount of pollutants emitted
     by small sources for quantitative ranking, a review of pub-
     lished area emission inventories was conducted.  Emission
     inventories were obtained of eight metropolitan areas, made
     by local or federal pollution control agencies.  These in-
     ventories have reported, with varying degrees of precision,
     the total pollutant emissions from all sources as well as
     breakdowns by type and source category.  The most complete
     and accurate inventory available is that published by the
     San Francisco Bay Area APCD.  This Inventory gives a rather
     detailed breakdown by major pollutant type and Individual
     source category, Including a number of categories Identified
     as being of pertinence to this study.

     Other inventories available were much less explicit in
their reported emission estimates and, in general, are based
on rapid emission estimating methods rather than actual

     Table l-III presents the estimated qaseous pollutant
emission rates reduced to pounds per day per thousand persons
(Ib/da per 1000 people) for each of the reported area surveys
Column 2 gives the rates for all sources, stationary and
mobile, and all types; i.e., oxides of sulfur and nitroqen,
carbon monoxide and the rates for organic vapors from ail
sources, namely: petroleum refineries, chemical plants, in-
cinerators, metallurgical plants, organic solvent users,
power plants, agricultural burning and motor vehicles.
Column 4 gives the emission rates of solvents and industrial
organic vapors - those vapors that are of primary concern to
control by the package sorbent systems.  The extent of the
latter, relative to pollution from all sources, is given as
percentage in Column 5.

     The combined population of the survey areas is about
27 million or over 13% of the total U.S. population.  These
are also the areas of greater population density; hence, the
emission rates given are most characteristic of areas where
pollution problems are most acute and controls are needed

     Air pollution from all sources and all types is of the
order of 5,000 Ib/da per 1,000 people.  About 3,500 Ib, or
70%,  of it is carbon monoxide, emitted mostly by motor vehi-
cles. Organic vapors from all sources are emitted at a rate
of the order of 800 Ib/da per 1,000 people.  In this case
also, the emissions from motor vehicles constitute about 550
Ib, or 70%, of the total organic emissions.  Of the balance,
10% to 15% are emitted by the solvent users and the rest by
the process industries including incineration and evaporation
of organic liquids from storage or during transfer.  The
volume of emissions from the process industries is quite
small except where domestic incinerators are in use, as in
the San Francisco Bay Area (1969).  The emission rate of
small sources is of the order of 2% of the total emissions
or, as stated above, about 10% to 15% of all organic vapors.

     Of the various pollutant inventory surveys conducted
throughout the country, the three California survey areas
(first three in Table l-III) identify the sources to a
greater extent than the others.  The emission rates as
given in Table l-II are from a 1969 San Francisco Bay Area
survey, except for the organic storage evaporation.  The
Bay Area data are shown in greater detail 1n Table  1-IV.

Table l-III  Rates of Gaseous Pollutant Emissions From Nine Survey Areas. Pollutant
             Emission Rate.  Ib/da per 1000 People

Survey area
South Coast Basin
San Francisco Bay Area
San Diego Bay Area
Delaware Valley
Denver, Colorado
Hamilton County, Ohio
St. Louis Study Area
Washington, D.C. Region
Jacksonville, Florida
All types,


Cal. (1968)
Cal. (1968)
Cal. (1968)
Wohlers (1965)
Dobler (1971)
Smith (1970)
McGraw (1968)
Maryland (1970)
Sheehey (1963)

    1)   Surface  coating  and  dry  cleaning  solvents  only
    2)   Dry  cleaning  solvents  only
_. cleaning solvents
Percentage of Industrial
process organlcs relative to all  types

Table 1-IV
             Amounts  of Organic Pollutants  Emitted  by  Small Stationary Sources Based
             on San Francisco Bay  Area  (1969)  Report
Surface coating
   Spray booths
   Ovens and dryers
   Flowcoaters, dip
     tanks & washers
   Brushing, rolling,
     nonindustri al

Dry cleaning
   Halogenated solvent 1
   Stoddard solvent   J
Graphic arts, printing
Plastics, resin, putty mfg
Surface coatinq, rubber-
  products, pharmaceutical

Other solvent users, pos-
  sibly adhesives, rubber
  products, paper coatings,
  extraction of vegetable

Total for solvent users
                             Number of
                          operating units






       Lb  per  day
       per 1,000

      Lb per day
    per operating





Table l-iv    (continued)
                              Number of
                               Lb per day
                               per 1,000
 Tobacco  curing
 Candy manufacture
 Smoke houses

 Breweries & distilleries

 Domestic Incinerators
 Conical  Incinerators
 Sulflte  pulp mills

 Auto-refueling stations
   Underground tank
   Auto-tank filling
 Pipe coating
 Asphalt blowing
   20,000  24.3
   50,000  37.3


                              Lb per day
                            per operating
Animal rendering
F1sh rendering
Feed, fish oil & protein
Coffee roasting
operating units
Tons /day
v   m

Of all of the small emission sources, the surface coating
industry emission rate is the larqest.  Fuel evaporation,
primarily that of gasoline, is the second larqest rate.
In both cases the emissions come from many small  operations;
in the latter case it is the automobile filling stations.
The surface coating industry is constituted of various sized
operations, some of the major ones being the lacquer spray
booths in small local shops, and varying in size  to larger
ones in the machinery manufacturing plants.

     These and other significant small sources are discussed
in detail in later sections of this chapter.

Types of Pollutants

     Listing of Pollutants

     The number of pollutants emitted from small  stationary
sources is expected to exceed 200.  Table C-I, in Appendix
C, lists 167 compounds that have been identified  with the
22 industries listed in Table l-II.  The chemical formula,
the molar volume, Vm, boiling point and lower explosive
limit, LEL, are important properties to be considered in
developing the subject of sorbent technology.  In the last
column of the table are the industry Identification numbers.

     Adsorptive Characteristics

     The pollutants in Table C-I are listed in decreasing
order of Vm, the molar volume at normal boiling point, cal-
culated according to the method of Le Bas (1915), Appendix
D.  This order of listing gives a general ranking of the
pollutants in decreasing order of adsorbabilities.  The
ranking applies  more consistently to adsorbabilities on ac-
tivated carbons than to the polar adsorbents: silica and
alumina gels.  Adsorption on the polar adsorbents is highly
selective, being influenced by the polarity or polarizabi11ty
of the molecule as induced by negative groups such as the
halogens, hydroxy and double bonds.  Small hydroxy compounds
such as water, inorganic acids and salts are strongly adsorb-
ed, preferentially to organic compounds of much larger molar
volume.  On activated carbons the compounds of larger Vm
tend to displace those of smaller Vm, with certainty of dis-
placement increasing as difference in Vm increases.  The dis-
placement is not always complete; some coexistance can occur.
On polar adsorbents, the same preferential adsorption occurs
in mixtures of nonpolar organic compounds.  Theoretical  as-
pects of adsorption are discussed in greater detail 1n Chapter

          Because of the nonselectlv1ty of activated carbons,
     except that due to differences 1n Vm, and the low level of
     Interference of water with organic vapor adsorption, activated
     carbons are used almost exclusively 1n pollutant vapor control
     when moisture 1s present 1n the exhaust gas stream.

          On activated carbons, those pollutants with small Vm,
     listed in the lower end of Table C-I, become increasingly
     difficult to control with adsorption systems.  Those with
     Vm  larger than 80 cm3/mol offer no real problem, at least,
     in  the adsorption phase of the operating cycle, while those
     below are variable.  Propane, propene, ammonia, methanol
     and methanolc add require excessively large carbon beds or
     systems operation with very short adsorption-desorptlon
     cycle times to effect any pollutant control.  Propanone,
     ethanolc add, ethanol and dlchloromethane are more amenable
     to  control with practical sized carbon beds.

          With regenerative carbon systems (also designated as
     carbon resorb systems), the ease or difficulty of pollutant
     desorptlon 1s also an Important factor to be considered.
     Those 1n the upper end of Table C-I, with large Vm, are In-
     creasingly difficult to desorb with commonly used regenerat-
     ing agents and at the relatively low temperatures commonly
     used.  Regeneration with 212F steam does not always effect
     complete desorptlon of compounds with Vm larger than 190 cm1/
     mol, while those with smaller Vm are readily desorbed unless
     polymerization occurs.  The accumulation of heavy or large
     Vm  compounds will gradually deactivate the carbon over
     successive adsorption-desorptlon cycles, so that ultimately
     1t will be necessary to discard the carbon or have 1t re-

          The Vm range 1n which the carbon resorb type system 1s
     most effective is from 80 to 190 cm3/mol.

          Identification With Industries

          The number of pollutants that have been Identified with
     each of the 22 Industries Investigated are given 1n Table 1-
     V.  Not all possible pollutants from each of the sources have
     been identified, but the closest approach to complete identl.
     flcatlon exists 1n the surface coating, degreaslng, dry
     cleaning and graphic arts Industries which, as of 1969, were
     the major small stationary emission sources 1n regard to
     total emission quantity.  The type of emissions from the
     other solvent users and also the meat rendering Industry are
     well documented.  For the other emission sources the Import-
     ant and predominant pollutants have been reported, I.e.,
     those that will have a bearing on the type of control tech-
     nology required.

Table 1-V  Number of Pollutants  Identified  With  Each
                                 Number  per  Industry
 1.  Surface coating
 2.  Decreasing
 3.  Dry cleaning
 4.  Graphic arts
 5.  Plastics

 6.  Rubber products
 7.  Adheslves
 8.  Paper coating
 9.  Vegetable oils ext.
10.  Organic storage

51.  Meat rendering
52.  Fish rendering
53.  Canning
54.  Restaurants

55.  Coffee roasting
57.  Tobacco curing
58.  Baking
59.  Candy making
60.  Tanning

61.  Meat smokehouses
63.  Domestic Incineration
65.  Zinc plating
Number in Vm
over 190 80 to 1
9 50
11 62
3 11
1 28
2 9
2 28


90 under 80



           The  breakdown  of  the  totals  into  Vm  ranges presents
      the most  important  criterion  for  determining  the  feasibility
      of using  a  carbon system for  pollution  control.   Most of  the
      emitted solvents and the pollutants  from  the  meat-rendering
      plants are  in  the 80 to 190 cms/mol  yni  ranne; hence, they
      are amenable to  control when  adsorption and regeneration  are
      the controlling  factors.   Some of  the  solvents from surface-
      coatinq and qraphic-arts operations  have  Vm larqer than 190
      cm3/mol and may  not desorb completely  durinq  reqeneration,
      while some  have  Vm  smaller than 80 cm3/mol and may not be
      sufficiently adsorbed  to make a carbon  adsorption system
      economically favorable when compared to competing systems.
      Pollutants  from  the thermal processinq  industries, such as
      the coffee  roasting, tobacco  curing  and smokehouses,  tend
      to be small molecules with Vm smaller  than 80 crnVmol.  A
      fair  portion of  those emitted from the  meat rendering In-
      dustries  also  falls into this Vm  range.   Further  information
      on exact  molecular  species and the emission source operating
      conditions  1s  required before a final  decision can be made
      on type and also size of the  control system.

           The  classes of compounds used and/or emitted by the
      various industries  follow  a pattern.   The solvent users,
      except for  the decreasing  and dry  cleaning industries, use
      the hydrocarbons (aromatic, naphthenic  and aliphatic), alco-
      hols, ketones, ethers, glycol ethers,  esters  and  nltropar-
      afflns.   Classes of compounds appearing as pollutants from
      the thermal process industries are predominantly  aldehydes,
      organic acids, amines  and  sulfides.  Chlorinated  solvents
      are predominantly used 1n  the cleaning  industries.

           The  most  generally used  nonhalogenated solvents are
      listed 1n Table 1- VI.  In  the surface  coating and graphic-
      arts  Industries  the solvents  are  used  1n  blends of two to
      twelve to obtain the desired  viscosity, solvency  power,
      drying rate and  compatibility with the  drying oils and
      resins.   Most  frequently,  the blends consist  of two to four
      solvents.  The two  most frequently appearing  solvents 1n
      the blends  are toluene and xylene.   All of these  solvents
      are recoverable  with carbon solvent  recovery  systems.

           In  the degreasing industry  five halogenated  hydro-
      carbons  are used as listed In Table  1-VII.  In 1967, about
      9055 of  the  tonnage  of  solvent sold for this  purpose 1n Los
      Angeles  County was  trlchloroethene and most  of  the  remaining
      10*.  tetrachloroethene (Walsh,  1967).   Most  greases dissolve
      readily  at  188F 1n trlchloroethene, hence  Us  wide use.
      Tetrachloroethene  1s used  when  a  higher temperature  1s
      required.  For cleaning of electrical  wire  colls, motor
      components  or  other temperature-sensitive materials,  the

lower bolHnq solvents are used.   Since the application  of
Rule 66, the use of tr1chloroethene has been restricted  1n
Los Angeles County because of Its smog forming properties
and 1,1,l-tr1chloroethane 1s now  the accepted solvent.
Table 1-VI  Most Generally Used Nonhalooenated Solvents

     Solvent                          Vmp cmVmol

     Heptane                              163
     4Methyl-2-pentanone                 141
     Hexane                               140
     Xylene                               140
     Toluene                              118
     Cyclohexanone                        118

     Ethyl acetate                        106
     Butanone                              96
     Benzene                               95
     Propanone                             74
     Ethanol                               61
Table 1-VII  Halogenated Solvents Used 1n Degreaslng

      Solvent                   Vm. cmVrool  temperature. F

      Trlchloroethene                98            188
      Tetrachloroethene             116            250
      1,1,1-THchloroethane         106            165
        1,2,2-Trlfluoroethane       120            118
      Dlchloromethane                65            104

     In the dry cleaning Industry two types of solvents are
in use:  the hydrocarbons and halogenated hydrocarbons.
Those most frequently used are   listed In  Table 1-VIII.
The trend of solvent usage is away from the Stoddard and
140F solvents to the halogenated hydrocarbons, particularly
the trichloro-trlfluoroethane.

     Evaporation of organic vapors from storage tanks and
during transfer occurs 1n many Industrial operations where
volatile reactants or solvents are used.  The largest emis-
sion rates occur 1n motor fuel storage and transfer.

     The emission rate of 60 Ib/da per 1,000 neople, given 1n
     Table l-II, is an average fiqure for motor fuels in the three
     California survey areas during 1968 (California, 1969).  The
     list in Table  1-IX  gives the mol fractions of the compounds
     in a typical gasoline vapor mixture.
Table 1-VIII Solvents Used in Dry Cleaning

Solvent                         VITI, cm 3/mol   Boiling range,
          Stoddard  -
     140F solvent  -i
                 45.7% paraffin
                 42.2% naphtha
                 _12.1% aromatic

                 46.5% paraffin
                 41.9% naphtha
                 11.6% aromatic


        185 ave

        185 ave


            349 to 410

            358 to 396
     Table l-ix   Components of Gasoline Vapors

                                               ,.    Boiling
         Organic compound    Vm, cm'/mol  Mol _%*  '   point, F
         2-Methyl  butane


Ca 84

       (1)  Balance air, 58.1  mo1%
     Of the gasoline fractions  the propane will  not be recover-
     able  but it Is  present at  a low concentration; hence, there
     will  not be a significant  economic loss.

     In meat rendering operations, the pollutant vapors
emitted are decomposition products of tallow or fats and
proteins.  About 43 compounds have been identified as listed
in Table C-I, Industry No. 51, (IITRI, 1971).  The aldehydes
are the most frequent and most abundant class of compounds
emitted, with organic adds, dlols, amines, sulfides making
up the balance.  Some of the compounds appearing at the
highest concentrations, 10* to 50% of the total pollutant-
vapor mixtures, are as listed 1n Table 1-X.

Table 1-X  Pollutants Appearing at Highest Concentration
           From Meat Rendering Operations

     Pollutant vapor         Vm, cmVmol     point, F

     Hexanal                     141            263
     3-Methyl butanal            118            198
     Butanoic add               108            327
     2-Methyl propenal            89            155
     Propanolc add               86            286

All of these vapors give bad odor responses, particularly
the butanoic add.  Some of the amines and sulfides which
are more offensive are emitted at lower concentrations.
One of these is dipropyl sulflde.

     Five very malodorous decomposition products have been
Identified with the fish rendering Industry; these are
listed 1n Table1-XI with their odor detection thresholds
(Conrad, 1971).
Table 1-XI  Malodorous Pollutants Emitted From Fish

                                          Odor detection
     Pollutant vapor        Vm, cm3/mol    threshold, ppm

     Dimethyl dlsulflde         103          0.0076
     Trimethyl amlne             92          0.0002
     Ethanethiol                  70          0.001
     Propenal                    67          0.210
     Carbon disulfide            66          0.210

 Pollutant  vapor  concentrations  1n  exhaust  air  are  low but,
 because  of the very  low  odor  detection  threshold concen-
 trations,  the pollutant  emissions  create an odor problem.

      Fish  and meat cannlnq  and  also meat preparation 1n
 restaurants  emit some  of the  decomposition products of the
 renderlnq  operations but of a considerably less malodorous

      The other processing industries -  coffee  roasting,
 tobacco  curing,  tanning, meat smokehouses  and  zinc plating  -
 emit  sufficient  quantities  of low-molecular-weiqht pollutant
 vapors to  make overall control  with sorptlon systems Im-
 practical.   The  reported pollutants for each Industry are
 listed in  Table  1-XII.   Those for  coffee roasting and meat
 smokehouses  were reported as  classes of compounds, but
 with  each  class  consisting of some of the  lowest molecular.
 weight members.

 Conditions of Pollutant  Release by Industry

      In the  preceding  text, the general pattern of types
 and amounts  of pollutants emitted by the various industries
 were examined.  In order  to select the most economical
 pollutant  control system, it is necessary to have an under-
 standing of  the  operating conditions under which each com-
 peting system is most effective and also to have detailed
 information  on the operations  of the emission source that
 affect or determine the  type of pollutants  and conditions
  n^K   r?lease- ,Based 0" feasibility and economic studies
  number of generalizations have been worked  up which estab-
      ?6 eftect ofPllutant release conditions on  choice of
       fonn!TS'   T5?SC ^neral1"t1ons  are discussed immeS.
duHnn f!l  ?1?q' uA1so'  the more 1mPrtant pollutant pro-
ducing industries have  been investigated to varying depths

      !ro1^thH.ClrCUm;tanCeS f the P'^t
      are also discussed  under this section.
          Effect on Choice of Control  System

          Thermal  Incinerators - Thermal  Incineration  would  be
          1.   The  pollutant has  no  recovery  value.

          2.   The  pollutant is combustible and yields  no
              new  pollutants  on  combustion.

Table 1-XII Pollutant Vapors  Emitted  From  Coffee  Roasting.
          Tobacco Curing, Meat  Smokehouses,  Tanning  ana
          Zinc  Plating  Industries
    Pollutant vapor
    Organic adds
    Ethanoic acid
        cm Vmol
                       Coffee  roasting*
                       -Tobacco  curing-
140 for hexane
105 for phenol
 55 methanethlol
 42 methanolc acid
 30 methanal
    Organic acid
                      Meat smokehouses'
                        Zinc plating-
 74 for propanone
105 for phenol
 42 methanolc
 30 methanal

           3.   The  concentration  is  between  1,000  ppm  and
               25%  of  lower  explosive  Unit.  At  this  con-
               centration  range,  the gas  mixture  can be  at
               the  100F level.

           4.   The  concentration  Is  below 700  ppm  but  gas-
               mixture temperature Is  relatively  high, 1n
               the  500 to 700F  range,

           5.   The  gas mixture contains  vapors which would
               deactivate  combustion catalysts,

           6.   The  qas mixture contains  considerable amounts
               of vapors of  Vm larger than 190 cm'/mol,  which
               cannot  be effectively desorbed  from carbons,  or
               that contain  considerable  amounts  of vapors  of
               Vm smaller  than 80 cmj/mol which are very
               poorly  adsorbed by carbons.

           7.   The  gas mixture contains  completely combustible
               (to  C02 and HgO)  particulates.   Noncombustlble
               partlculates  would have to be scrubbed  or dry-
               filtered out.

           8.   The  moisture  content  1s very  high.

           Catalytic Incinerators -  Catalytic  Incineration  would
      be used  1n preference  to the competing systems when:

           1.   The  pollutant has  no  recovery value.

           2.   The  pollutant 1s  combustible  and yields no  new
               pollutant on  combustion.

           3.   The  concentration  1s  between  300 ppm and  2555  of
               the  lower explosive Hm1t.  At  this concentration
               range,  the  gas  mixture can be at the 100F  level.

           4.   The  concentration  1s  below 300  ppm  but  the  gas-
               mixture temperature 1s  elevated -  1n the  300  to
               500F range.

           5.   The  gas mixture does  not  contain vapors which
               would deactivate  combustion catalysts.

           6.   The  gas mixture contains  considerable amounts  of
               vapors  of Vm  larger than  190  cmj/mol, which  cannot
               be effectively  desorbed from  carbons, or  contains
               considerable  amounts  of vapors  of  Vm smaller  than
               80 cm'/mol, which  are very poorly  adsorbed by  car-

     7.   The gas  mixture contains  no  appreciable  amount  of
         particulates,  or contains participates which  can
         be readily dry-filtered out.

     8.   The gas  mixture 1s  relatively  dry.

     Absorption Systems - Absorption  systems  using  water
spray would be used in  preference  to  the  other competing
systems  when:

     1.   The pollutant  has no recovery  value.

     2,   The pollutant  is soluble  in  water.

     3.   The pollutant  is poorly adsorbed on  carbons;
         i.e., Vm less  than  80 cm3/mol.

     4.   The pollutant  is biologically  degradable.

     5.   The concentration is below 700 ppm  if pollutant
         is combustible.

     6.   The gas mixture contains  particulates that are
         difficult to dry-filter  out.

     Carbon Air Purification Systems* - Carbon air  purification
systems  would be used 1n preference to  the other  competing
systems  when:

     1.   The pollutant  has no recovery  value.

     2.   The pollutant  Vm is larger than 80  cm'/mol.

     3.   The concentration is of the order of 1  ppm or

     4.   The gas mixture has no appreciable  amount of
         particulates or has particulates that are
         readily filtered.

     5.   The gas mixture temperature is 130F or lower.

     6.   Relative humidity of gas mixture is below
         saturation, preferably below 80%.

     Carbon Resorb Systems* - With the carbon resorb  systems,
the  limitations on their use depends to some extent on  whether
the  pollutant has recovery value.  When the pollutant has no
recovery value, the resorb system would be used  1n  preference

*See Chapter 6, pages 6-4 and -32 for  discussion of these two

      to  competing  systems  when:

           1.   The  pollutant Vm  range  1s  80  to  190  cm'/rnol.

           2.   The  gas  mixture has  no  appreciable  amount  of
               particulates or has  partlculates  that  are
               readily  filtered.

           3.   The  qas  mixture temperature  1s  130F or  lower.

           4.   The  relative humidity of the  qas  mixture 1s
               below saturation,  preferably  below  80%.

           5.   The  pollutant concentration  1s  1n the  range  10
               to 700 ppm when pollutant 1s  combustible and
               yields no new pollutants on  combustion.

           6.   The  pollutant concentration  1s  10 ppm  or higher
               for  noncombustlble pollutants or up to 25% of
               lower explosive limit for pollutants that yield
               new  pollutants on combustion.

           When the pollutant has recovery  value,  Items 1 to 4 are
      the same as above.  Restriction on concentrations are  differ-
      ent than Items 5  and 6,  For recoverable pollutants,  the
      concentration 1s  10 ppm and higher for the noncombustlble
      and up to 25% of  lower explosive limit for the  combustible.

           Surface  Coating Industry

           The surface  coating Industry can be divided  Into  two
      separate entitles:  those  who manufacture the material and
      those who use the material.  The manufacturing  end of  the
      business does not appear to be a principle concern to  this
      study.  Most  companies are medium to  large size with  oper-
      ations using  closed cycle  systems so  that release of  material
      to  the air 1s extremely small.  The exceptions  to this gen-
      eralization are the emissions from solvent storage and odors
      from varnish  cookers,  Of  principal concern  are emissions
      from the various  users of  the surface coatings.

           Types of Surface Coatings - There are five main  types
      of  surface coatings: namely, paint, enamel,  varnish,  lacquer
      and shellac.   Some of the  differences  are due more to  modifi-
      cations  In formulation to  meet use requirements rather than
      to  basic differences.

           Paint Is a highly plgmented drying  oil, thinned  with
      a low solvency power solvent, referred to as a  thinner.   The

piqment can be a metallic oxide or a high molecular weight
organic compound and is suspended in the vehicle.   The dry-
inq oil is an unsaturated polymer which on exposure to air
oxidizes and polymerizes further to form the resinous pro-
tective film.  It is only partially soluble in the thinner.
Mostly natural drying oils are used but the shift  is toward
the synthetic type.  The formulation is directed to give a
coating that protects the substrate from the weather and
beautifies it, but tends to be low on abrasion resistance.
Painted surfaces also tend to have a dull finish.

     Enamels are basically the same as paints except that
synthetic drying oils are used to a larger extent.  The
formulation is directed to give a coating of higher abrasion
or wear resistance.  Enamel coatings are generally glossy,
smoother, hence easier to wash, but are harder or  more
brittle than paint coatings.  Resistance to weathering tends
to be lower.

     Varnishes are colorless, or tinted with stain or small
amounts of pigment, to maintain a high degree of transparency.
The substrate beauty is enhanced by the coating.  A synthetic
drying oil is used, usually an aromatic modified tung oil,
and thinned with a low solvency thinner both with  higher
solvency than those used in enamels.  An almost complete
solution of the drying oil occurs.  The emphasis is on wear
resistance and also weather and light resistance.   The attempt
here is to combine all the good qualities of paints and enam-
els and still have a tough resilient and transparent coating.

     The above three coating materials after being applied
to the surface, dry and cure by evaporation of the thinner
and by an oxidation-polymerization of the drying oil to form
the resinous film.  In contrast to the above, lacquer and
shellac vehicle consists of a resin dissolve in a  high sol-
vency power solvent.  On application, the drying occurs by
evaporation of the solvent and deposition of the resin-
pigment film.  The early lacquers consisted of a cellulose
nitrate resin while the present ones are made with acrylic
resins.  Shellac resin is a natural occurring product syn-
thesized in the skin glands of certain insects of  India.
Lacquer and shellac coatings have the disadvantage that the
resin film will redissolve.

     Methods of Application - There are five methods of
application which are of possible Interest to this study
namely, spraying, roller, dipping, flow coating, and tumbling.
rollpr*nni?^^HSedKm!t!]?ds are nand brushing and also hand
roller application, but these are of limited Interest since

      they are  used  in out door construction or maintenance jobs
      where  the job  site changes frequently, hence means of vapor
      collection are  Impractical.  The other methods are used 1n
      applying  Industrial finishes, where continuous operation at
      a  stationary site makes use of emission control devices

          Each of the five types of coating materials can be
      applied by each of the six methods of application when
      appropriate modifications are made to the formulations,
      although  some  coatings are more readily applied by some
      methods than others.  As an example, lacquer 1s best applied
      by  spraying.

          Each method 1s best suited for certain types of articles
      to  be  coated,  as for Instance dip coating for pipes, flow
      coating for automobile frames and bed springs, roller coating
      for flat  sheets or panels, tumbling for small rounded objects
      such as buttons and spray coating for large uneven surfaces
      such as automobile bodies.  For each application the coating
      viscosity and  drying rate must be carefully controlled by
      the solvent or  thinner formulation.  For Instance, 1n the
      dip and flow coating applications the coating should become
      fixed  to  a uniform thickness and not continue to drain toward
      the lowest point on the article after being coated.

          An essential part of the applications 1s the drying-
      curing operation.  The drying-curing operation may be with
      air at ambient  temperature, or 1n ovens with forced heated
      air or radiant  heat.  A1r drying occurs at temperatures below
      100F, hot air  drying at 100 to 200F and baking at 200 to

          Surface Coating Solvents - Little or no volatlles are
      expected  to be  released from the pigments, drying oils or
      resins during  applications and during drying unless recom-
      mended drying  temperatures are exceeded.  Baking produces
      some odorous decomposition products.  The vapor release 1s
      essentially the thlnners and solvents used.  Table 1-XIII
      contains  a listing of some of the thlnners and solvents
      which  have been Identified with the surface coating Industry.
      The thlnners are aliphatic hydrocarbons, turpentine, mineral*
      spirits VM&P naphtha.  They are mainly 1n paints, enamels
      and varnishes.  The solvents are the aromatlcs, alcohols,
      ketones, ethers and esters.  They are used to varying degree*
      1n enamels and  varnishes but mainly 1n lacquer and shellac.
      Table  1-XIVshows this breakdown of solvents used In surface
      coatings  (data  selected from Table 112 of Weiss, 1967).  in
      enamels, varnishes and shellacs, the volatile portion 1s
      about  50% and  1n lacquers, about 77% by weight.  In brush-on
      paints, the volatile content may be lower than 50*.

Table 1-XIII  Estimated Air Pollutant Emission Rates.  Surface
              Coating Solvents
     Emission rate,
lb per day per 1000
o, p & m
                               P  18
     n-Pentyl acetate
     2-Ethoxyethyl acetate
     Ethyl acetate
     2-Methoxy ethanol
Table  1-XIV   Surface-Coating  Formulas on an As-Purchased
              Basis  (Weiss  1967)

                      Composition  of  coating. %	
  Type    Nonvolatile      Hydrocarbons                     &
coating    portion     Aliphatic  Aromatic Alcohol  Ketone Ether
      Many of the coatings  are cured or baked at the higher
 temperatures (200  to 450F)  1n which case the resins and
 other components can release  small  amounts of high molecular*
 weight vapors.

           To  show  the  general  make-up of  coating formulations
      with  respect  to solvent  content, four have been selected
      from  data  given by  Parker (1965) and are presented 1n Tables
      1-XV, -XVI*. -XVII and  -XVIII.  The first three - a primer,
      an  enamel  and a lacquer  - are  Industrial finishes for appli-
      ances, machinery, metal  furniture, automobiles, etc.  The
      varnish  formulation  1s for wood furniture, a manufacturing
      or  repair-shop operation  where solvent vapor control 1s

      Table 1-XV Gray  Epoxy Appliance Primer Formulation

      Material	     Nonvolatile. %    Volatile. %

      Rutlle,  titanium  dioxide           8.0
      Lamp  black                        0.1
      Blsphenol  A-ep1chlorhydr1n
       epoxy  resin                    24.7
      Urea  formaldehyde resin           18.7

                         Totals        51.5             48.5

           The above primer  Is  cured by baking 20 minutes  at 385*F%

      Table 1-XVI   Acryl1c-Epoxy Appliance Enamel Formulation

      Material	   Nonvolatile. %   Volatile.  %

      Rutlle,  titanium  oxide              26.9
      Acrylold AT-52 (Rohm & Haas
       50% acrylic resin                 19.8
        12.5 cellulose  acetate             4.9
       37.5%  xylene                                      14.8
      Epoxy resin
       50% blsphenol  A type resin        13.2
       50% 2-ethoxyethyl  acetate                          13.2
      Raybo 3  ant1s1lk1ng agent            0.1

      2-Ethoxyethyl acetate                	
                         Totals           64.9

           The above enamel  1s cured by baking for  30 minutes at

Table 1-XVII  Vinyl  Lacquer Formulation for Aluminum  Siding
Rut He, titanium dioxide
Geon 427, vinyl  resin
Geon 433, vinyl  resin
Admex, epoxy polymer
Thermolite, tin  stabilizer
4-Me thy1-2-pentanone
Solvesso 100
Nonvolatile. %   Volatile.  %

Solvesso 100 1s 94% aromatic with probable boiling range from
317 to 350F.
Table 1-XVIII  Base for Furniture Rubbing or Flat Varnish

Material                      Nonvolatile, %    Volatile. %
Modified phenolic resin
Z 2 dehydrated castor oil
Liquid olticia oil
Mineral spirits

     The nonvolatile* of the varnish base are cooked at 510
and 460F before being diluted to 50% solids with mineral
spirits.  Mineral spirits generally designate the predomi-
nantly paraffinic and naphthenlc solvents of CIQ to C]2 range
with boiling points from 300 to 400F.  To produce a rubbing
varnish from the base, 1t is blended with VMP naphtha, aro-
matic naphtha, lead naphthanate, cobalt naphthanate and an
antiskinning agent, yielding a final product that 1s 57%
volatHes.  VMP naphtha solvents (varnish makers and painters)
are Cs to Clo aliphatics, and the aromatic naphthas are 1n
the same high boiling ranges as the mineral spirits.

     The primer and enamel may be applied to appliances by
dipping or flow-coating.  To reduce the enamel to spray vis-
cosity 1t is diluted further with xylene and 2-ethoxyethyl

     Pollutant Emission from Surface Coating Industry  - The
combined Information of Tables l-II, -IV and  -XIV  allow

      quantitative  estimating  on  the  various  types of compounds
      being  emitted.   In  Table 1-IV,  the  brushing, rolling, and
      nonlndustrial  spraying methods  appear to apply to outdoor
      applications  and  the  emissions  are  therefore not controllable.
      This  reduces  the  emission rate,  of  concern  to this study,
      from  96  to  61  Ib  per  da  per  1000  population.  Of the 61 lb,
      43  lb  are emitted from spray booths which use mostly lacquers
      and some enamels.  The 43 Ib/da  per 1000 population 1s then
      a  rough  estimate  of the  emission  rate of the aromatic, alco-
      hol,  etc type  solvents.   Of  these solvents, the aromatlcs
      constitute  about  40%  of  the  total as based  on the percentage
      figures  given  1n  Table 1-XIV for  lacquers and enamels and
      also  on  a number  of industrial  finish formulations described
      1n  Principles  of  Surface Coating  Technology (Parker, 1965).
      On  the basis  of  these considerations the emission rates as
      given  in Table 1-XIII were  estimated.

           Solvent  Emission Control -  As  of 1969, the surface-
      coating  industry  1n the  California  survey areas emitted the
      largest  quantity  of solvents Into the atmosphere.  The rates
      of  emissions  from the various surface coating operations as
      reported by the  1969  San Francisco  Bay  Area survey are given
      1n  Table 1-XIX.
      Table  1-XIX
Solvent Emissions from Surface Coatings
Operations (Bay Area, 1969)
       Type  of

      Rol Ung
     Number of


1000 people


Lb/da per



About one half of total emission volume comes from many spray
booths, with an average rate of 70 Ib/da per operating unit.
These are various shops such as auto body repair shops and
production Hne spray booths for Industrial  machinery and
home appliances.  Part of the drying oven emissions are from
the lacquer coated products and part from enamel coated pro-
ducts as received from the flow coating, dipping and roller
coating operations, the 35 Ib/da emissions from'brushing and
nonindustrlal spraying are not controllable  by adsorption,
leaving 61 Ib/da of the total 96 Ib/da as possibly control-

     The survey of the literature and of the control equip-
ment manufacturers has not brought out evidence of any wide-
spread use of control devices in the surface coating Industry.
Carbon sorbent systems do not appear to be in use for solvent
recovery purposes.  Under the prevailing operating procedures,
solvent recovery is not profitable.  Where bad odors are
emitted, as from baking ovens, a limited use of "catalytic and
thermal incinerators 1s reported.  The reasons for lack of
use of carbon solvent recovery systems vary  with type of
surface coating operation.

     Spray booths are operated with face velocities of 100 ft/
min and over as required by varied state laws.  Spray booths
of sizes to fall within the scope of small emission sources
are commercially available.  Booths of front opening area
from 13 to 100 ft2 and operated at 100 to 180 ft/m1n air
face velocity are described in sales literature (Paasche,
1971 and Spray. 1971).  Because of over spray of the articles
being coated, particulates consisting of the pigments, drying
oils and resins are released Into the exhaust 'airflow.  To
avoid air pollution by a lacquer or enamel mist, either dry
filters or water sprays are used to separate out the mist.

     For effective use of a carbon resorb system, the spray
coating particulates and water mist should be completely
removed from the exhaust air.  This 1s not always accomplished.
The 100 ft/min and over air velocities tend  to keep the sol-
vent concentration low, below the approximate 700 ppm economic
break even point.  The use of mixed solvents reduces the car-
bon adsorptive to some extent; I.e., a single organic compound
at a given concentration requires less carbon than a mixture
of several with the same combined concentration.  Recovery of
a solvent mixture from the steam condensate, after regener-
ation, 1s more Involved and hence more costly than a single-
compound solvent.

           To  determine  the  nature  of  air  pollution  control  from  a
      spray booth  using  a  carbon  resorb  system,  estimates  v/ere  made
      of  unit  size,  total  airflow,  and cost  of  regenerating  steam,
      when  total  solvent emission rate 1s  70 Ib  per  8  hr da  and
      concentrations  are 10,  100, 1,000  and  4,800  ppm.  The  calcu-
      lations  were made  for  an  acrylic lacquer  thinner  of  compo-
      sition given 1n Table  1-XX.

      Table 1-XX   Composition of  a  Common__L.a_cgjje_r_ T_h1_n_n_e_r

                                                Cone of each  at
                                                4,800  ppm  total
      Organic  compound         Weight,  %        ppm        vol,%

      2-Ethoxyethyl  acetate         17              480      10.0
      Toluene                       43            1,740      36.3
      Propanone                     40            2,570      53.7

      The 4,800 ppm  1s 25% LEL  for  the mixture  calculated  with  an
      equation based  on  Le Chateller's modification  of  the mixture
      law (Le  Chatelier, 1891).   The equation  1s  given  1n  Appendix

           Because of pressure  drop and  other  considerations,  the
      linear velocity through carbon resorb  system beds 1s varied
      from  about  20  to 110 ft/m1n.  The  minimum  face velocity  for
      spray booths is 100  ft/mln; hence  the  area  of  the carbon  bed
      must  be  sized  equal  to, or  larger  than,  the  spray booth  front
      opening  for  a  given  total airflow.   The  results  of the  calcu*
      lations  are  given  1n Table  1-XXI when  the  airflow is kept at
      the 110  ft/nin  maximum which  then  gives  the  minimum  carbon
      bed area.   The  110 ft/m1n velocity for the  spray  booth  gives
      the maximum  front  opening area.  E1oht one-hour  cycles  are
      assumed  to  recover the 70 Ib/da  solvent  emitted.

      Table 1-XXI  Carbon-Resorb  System  Size,  Operating Condi-

      (1)  See
tlons. Regenerating Steam Cost
for Control
of Spray-Booth Solvent Emission

AP Cost of 212
A, ft2 W, Ib 1n.H20 $/l
1.3 97 25.6
6.4 194 10.8
63.6 800 4.5
636 3,000- 3.1
f for definitions of terms
Ib solvent

F steam

     The results of Table 1-XXI  show that at 4,800  ppm the
steam cost 1s low enough (3
          Degreasing Industry

          Description of Industries - Degreasing is a necessary
     step In the preparation of metal fabricated articles for
     surface coating, electroplating or other finishing operations
     The cleaning operation is done in a packaged unit, termed the*
     degreaser, which utilizes an organic solvent as cleaning
     medium, Table 1-VII.  Depending on the complexity of the
     unit, the soiled article may be contacted by immersion in
     the hot solvent, by a warm rinse, by liquid spray and conden-
     sate.  A high degree of cleanliness of the articles is attain-
     ed by suspending the (cool)  article in the solvent vapor.  By
     condensation of the vapor the grease, oil or other soil  is
     dissolved and  washed down into the liquid phase (Kearney.

          The basic vapor degreaser consists  of a vapor generating
     sump containing the solvent  and heater,  a vapor zone above
     the sump and freeboard.   A narrow water-cooled jacket around
     the upper part of  the degreaser sets the maximum height  of the
     vapor zone.   Soiled articles are lowered into the vapor  zone
     through the  open top and freeboard and withdrawn by the  same
     route when cleaned.  Vapor condensate cleaning stops when the
     article temperature rises  to the vapor temperature.  The
     spray step aids  in  the  cleaning of irregularly shaped articles
     or articles  otherwise difficult to clean and is done while the
     article is submerged in  the  vapor zone.   The essential feature*
     of the  degreaser are Illustrated 1n Figure 1-1.               *

          Size  of Decreasing  Industry - San Francisco Bay Area
     (1969)  reported, (Table  1-IV), 600 degreasers  In operation
     with average emission rate of 143 Ib of  solvent per day  per
     operating  unit.  The amount  emitted per  day averaged 21  Ib
     per 1000 people. The latter average would apply more to
     Industrialized  areas and not give a good Indication of nat-
     ional average.

          Table 8 of  ASTM STP 310 Report lists typical  appli-
     cations  of vapor degreasing.   The production rates  give  an
     indication of  size  of degreasing units used 1n Industry.
     For degreasing  by vapor-spray-vapor system, the smallest
     rate reported  is 600 Ib/hr of steel  spark plugs to  remove
     machining  oil,  and  the  largest is 40,000 Ib/hr of steel
     automatic  transmission  parts to remove machining oil, chips
     and shop dirt.   For degreasing by the warm Hquld-vapor
     system,  the  smallest rate  reported  is 500 Ib/hr of  magnesium
     aircraft castings to remove  polyester resins,  and the largest
     is 12,500  Ib/hr  of  aluminum  tubing  to remove drawing lubri-
     cants.   For  degreaslng  by  the boiling liquid, warm liquid and
     vapor system,  the smallest is 50 Ib/hr of gold  and tin

                   Frtt boird
                   Vapor lone
                             'l| I i-t-l I'll
                                 Basket containing
                                 liquid solvent
Vent duct
to carbon
Water cooling
jacktt crouna
 - Su*ip
Figure  1-1  Degreaser  Using Halogenated  Hydrocarbon  Vapors
                                            AMBIENT AIR
                                             yhig carbon bed*)
Figure  1-2  Dry  Cleaning  Machine With  Integrated  Carbon
              Solvent Recovery System.   Courtesy  of V1c
              Manufacturing  Company

      plated  transistors  to remove silicone and light oil, and
      the  largest  is  5,000 Ib/hr of steel stampings to remove chips
      and  light  oil,  Most of  the materials handling is by conveyor
      and  some by  hoists.  In  the small operations, such as the
      transistor cleaning, the  articles are lowered into the de-
      greaser manually  in wire  mesh baskets.

           Air Pollution  from  Degreasers - Solvents escape from
      degreasers uncer  several  circumstances.  Degreasers must of
      necessity  be  operated with the  top open completely, or
      partially  oper, or  opened at intervals.  Most of the vapors
      escaped ty this route.   Water-cooled condensers or jackets
      operated at  68F  can hold the vapor concentration at the
      vapor-zone-freeboard boundry to  approximately the following
      concentrations  for  three  of the  most commonly used solvents.

           Solvent                        Volume, %    ppm

           Tetrachloroethene                   1.9      19,000
           Trichloroethene                     7.9      79,000
           1,1,1  Trichloroethane              13       130,000

      In  a  very  still  freeboard atmosphere,  a  concentration grad-
      ient  can form with  very  slow  diffusion  of vapors out of the
      c'egreaser.  Vapo  movement  through  the  freeboard is greatly
      accelerated by  air  currents passing over an open or partial 1**
      open  degreaser,  by  opening  and  closing  of the lid  if one 1s
      used,  and  by the movement of  the  articles in and out of the
      degreaser,  Movements necessary during  spray cleaning create
      turbulences whicn accelerate  vapor  escape.  Degreaser man-
      ufacturers  estimate 0.05 Ib/hr  per  ft2  of open  tank area as
      the vapor  escape rate.   Another potential means  of vapor es.
      cape  is by  way  of liquid solvent  adhering to the articles
      being  transferred out.   This  happens when the articles have
      not reached the  vapor temperature.  Also, pockets  and
      crevices in the  articles can  carry  out  solvent.  By careful
      operating  procedures vapor  losses can be minimized.

           For a  degreaser cleaning 40,000 Ib/hr of transmission
      parts, the  solvent distillation rate 1s 600 gal/hr and
      solvent make-up rate Is  2.4 gal/hr  (Kearney/1960).  Assumln
      trichloroethene, the total make-uo  or potential pollutant   '
      emission rate 1s 233 Ib  per 8 hr  day.  The mean calculated
      emission rate from  San  Francisco data is 143 Ib per de
      greaser per day.  Possible reasons for the difference can
      be that the average degreaser of  San Francisco Bay Area
     would be smaller than the 40,000  Ib/hr unit and some of th.
     San Francisco degreasers have solvent recovery systems

     If 1t is assumed that a trichloroethene  vapor is
drawn off the (larqest)  40,000 Ib/hr degreaser at a con-
centration of 1,000 ppm, the airflow would be 1,450 ft3/
min.  The calculation is:

1,000 ppm x IP'6 x 131 Ib mol wt   _ 2.4 gal/hr x 12.7 Ib/gal
         391 ft3 mol vol           "        60 min/hr

         F  1,450 ft'/min.

The solvent can be profitably recovered under these conditions
by several of the package solvent recovery systems on  the
market.  Much of the solvent  recovery system advertising is
directed toward recovery of trichloroethene.

     Emission Control - To  control air pollution by way of a
carbon solvent  recovery system, slotted ducts are  attached
to the top rim  of the degreaser as shown  in  Figure 1-1.  The
airflow  through the  slots can be optimized tot   (1) minimize
or avoid creating turbulence  in the  freeboard, (2) prevent
or minimize  escape of solvents past  the slotted  ducts, and
(3) maintain solvent  concentration in  airstreams  above 700
ppm, the approximate  economic breakeven point.

     This  has been  accomplished in a considerable  number of
installations.  A case  in point is a recovery operation re-
ported by  Berg  (1968) in which 7,300 Ib/wk (183  Ib/hr) of
trichloroethene were  recovered with  a  carbon  solvent-recovery
system at  3,000 ftVrnin  airflow.  The  calculated  concentration
was 3,000 ppm.   The  recovery cost was 0.34c/lb of  solvent.  It
included electricity  for  controls and  exhausters,  steam, cool-
ing water,  and  labor  to  clean filter bags  ahead  of the carbon
beds each  day.  A  comparison  between operating cost and sol-
vent selling price  (11.54/lb) indicates a  profitable  opera-

     The  2.4 gal/hr  (29  Ib/hr) solvent make-un rate reported
above  for  one of  the  larger degreasers  can be readily con-
trolled  by  solvent  recovery systems  of  less  than  3,000 ft3/
min  air-handling  capacity.   The use  of  carbon resorb  systems
in  the degreasing  industry  represents  one  of the  more straight-
forward  and  trouble-free  applications.  For  all  practical  pur-
poses, the  logical  solution to pollution  control  in the  in-
dustry is  the installation  of more  recovery  systems.

     Dry Cleaning  Industry

     Description  of  Operation - The  procedure  consists of
charging a weighed  amount of  clothing  into the dry cleaning

     unit, washing by agitation in a solvent bath and rinsing
     with clean solvent.  On completion of the washing step, the
     solvent is drained off and the liquid solvent in the cloth-
     Ing is removed by spinning in a centrifuge, termed the ex-
     tractor.  In the third and final step, the clothing is dried
     by means of heating  air circulated through a tumbler.  The
     solvent from the washer is reclaimed for reuse by filtration
     and distillation.  Table  1-VIII lists the most used solvents.

          Size of Dry Cleaning Industry - The dry cleaning  indus-
      try  of  the  United  States  is  comprised of some 35,000 individ-
      ual  operations  ranginq  from  the small - grossing less  than
      $40,000  per year -  to  large  complex operations grossing over
      $300,000  per year.   At  present  approximately 7S% of  the
      plants  utilize  tetrachloroethene.  These tend to be  the
      smaller  operations  and  account  for only 55% of the volume  of
      clothes  cleaned  (National Drycleaners Institute  interview).
      In  Los  Angeles  County  (1956), because of fire laws,  the newer
      smaller  commercial  and  coin  operated dry cleaning establish-
      ments,  which  are generally near residential and  shopping
      areas,  use  the  chlorinated solvents.  The  larger dry cleaning
      establishments  in  industrial  and  commercial areas use  the
      petroleum  solvents  (Lunche,  1957).  About  60% (1957) use
      chlorinated  hydrocarbons.

           The  Dry  Cleaners'  Association estimates  that an average
      per day  load  of  clothing  for small operations, which are  the
      majority  of  installations, would  be approximately 400  Ib.
      They further  feel  that  there are  few operations  doing  in
      excess  of  800  Ib per day.  Based  on tetrachloroethene  vapor-
      recovery  figures for Los  Angeles  County,  the  estimated daily
      clothing  load  for  the  average size  tetrachloroethene plant
      is  570  Ib.

           Based  on  the  35,000  dry cleaners  figure, the number  of
      dry cleaners  per  1000  people in  the United States is 0.17.
      For the  San  Francisco  Bay Area  the  estimate  for  1969  1s 0.26.

           Air  Pollution From Dry Cleaning -  The San  Francisco  Bay
      Area survey  (1969)  on  1,050 dry  cleaning  units  gave  estimates
      on  the  amount  of emitted  chlorinated solvents at  4.7  Ib/da
      per 1,000  people.   The  combined  amount  emitted  per  operating
      unit was  estimated at  42  Ib/da.

           Solvent  can escape into the  atmosphere  under several
      circumstances.   In the  older dry  cleaning  plants,  the  washer-
      extractor  and  tumbler  (or dryer)  are two  separate operating
      components.   When  clothes are transferred  from  one  to  the
      other,  solvent  vapor is  released  into  the  atmosphere.   Vapor
      released  during  the  tumbler step  is passed through  a water-
      cooled  condenser which  removes  part of  the vapors  1n  the  air-

     A survey of 60 (of 600) tetrachloroethene-type dry
cleaners in Los Angeles County, 1962, showed that per 1,000
Ib of clothinq cleaning, 10 qal of tetrachloroethene are
required when no carbon adsorbent recovery unit is employed.
When an adsorbent unit is employed, the solvent loss is 3.5
qal or a recovery of 6.5 qal (88 Ib).  A 1971 estimate, by
the National Dry Cleaners Institute, for tetrachloroethene
recovery for coin operated machines 1s much smaller.  For a
2,000-load cycle of approximately 17,500 Ib of clothing, the
recovered quantity was 25 gal or approximately 1.5 gal per
1 ,000 Ib of clothing.

     Emission Control - When a carbon solvent recovery system
is used, it is installed at the discharge of the ventilating
system to collect the vapor laden air from the tumbler and
vents placed near the doors of the washer-extractor, the
tumbler, and the floor near the tumbler.  The floor vent
collects solvent released from the clothing when removed from
the tumbler.

     In the construction of new dry cleaning machinery the
trend is toward a completely contained machine which includes
an activated carbon solvent recovery system.  A flow diagram
of this type of machine is shown in Figure  1-2.  In a 14-min
cycle time, the coin operated machine of this make cleans 8
Ib of clothing; and a commercial unit, 25 Ib.  One gallon of
1 ,1 ,2-trichloro-l ,2,2-tri f luoroethane is used per pound of
clothinq in the wash phase of the overall cycle.  Very little
solvent is lost to the atmosphere.

     The effective use of carbon resorb type solvent recovery
systems has been demonstrated in the dry cleaning plants
where halogenated solvents are used.  Because of the high
cost of tetrachloroethene, the value of its recovery makes
possible the amortization of the adsorption unit within 1 to
2 yr; hence a financial incentive exists for the prevention
rL  r Pllution with the chlorinated solvents.  Approximately
50% of these plants have activated carbon recovery units
Plants employing the petroleum type solvents, however, make
ho^n!!mof JJ V?Pr recovery. at least in the United States,
because of the lower cost of these solvents.

     Graphic Arts Industry

     Size of Industry - This multifaceted  industry is
pertinent as an emission source because of solvents or nil

by the -uch

          The  printing  activity  is  concentrated in urban areas,
      although  distributed  throughout  the  country.  About one-half
      of  all  commercial  printing  is  done in  four states:  New York,
      Illinois, Pennsylvania  and  California  (Gadomski,  1970).

          Printing  Processes  - The  individual  printing  processes
      which  comprise the graphic  arts  industry  include  letterpress,
      lithographic,  gravure and screen process.

           In the  letterpress  process, the printing area of  the
      printing  plate or  roller is raised above  the nonprinting
      area.   Ink applied to the raised area  is  subseguently  trans-
      ferred to the  paper or  other substrate to be printed.   Books,
      newspapers and magazines are printed by this process  (Bruno,

           In lithography,  the image and nonimage  areas  are  on  the
      same plane.   The image  area is grease  receptive  and water
      repellent, while the  nonimage  area is  the reverse. The ink
      applied to the image  area is transferred  to  the  substrate  to
      be  printed.   General  commercial  literature,  books, catalogs,
      greeting  cards, letterhead  business  forms,  checks, art repro-
      ductions  and labels are  printed  by this process.

           In the  gravure process, the printing area  is  recessed;
      it  1s  comprised of wells etched  to different depths.   The
      ink is transferred to the roller surface  and excess removed
      with a blade.   Longrun  magazines, newspaper  supplements,
      preprints for  newspapers and plastic laminates  are printed
      by  this process.

           In the  screen process, a  stencil  is  applied  to a  silk
      or  stainless-steel screen to which a thickened  paste  ink  1s
      added,  The  ink is transferred to the  printing  surface by
      rolling with a squeegee. This process finds application  in
      printing  displays, posters  and signs.

          When the  image is  transferred directly  from  the  image
      carrier to the paper, it is known as direct  printing.   When
      the image is transferred by an intermediate  roller or  plate
      it  is  known  as offset printing.   All screen  process and most
      letterpress  and gravure  processes are  direct printing.  Most
      lithographic printing is offset.

          When the  plates  or  rollers  used in the  letterpress pro-
      cess are  of  rubber, the  process  is known  as  flexography.
      This process is used  widely in multicolor printing on  a wide
      variety of substrates.

     Printing Inks - The three important properties of inks
are drying, printablHty and color.  Of greatest concern to
air pollution is the dryinq properties of the ink which can
occur by one or more of seven different physical or chemical
mechanisms; i.e.. absorption, solvent evaporation, resin pre-
cipitation, oxidation, polymerization, cold setting and gela-
tion.  The choice of ink is determined by the printing pro-
cess and the substrate to be printed, and the ink composition,
in turn, determines the air pollution potential of the print-
ing process.

     Inks used in letterpress and lithographic processes are
referred to as oil inks or paste inks.  Most of these inks
dry by means other than solvent evaporation, and those few
that depend to a minor extent on solvent evaporation use sol-
vents in 250 to 350F boiling range.  Liquid body is imparted
to the ink to some extent by liquid reactants such as linseed
oil.  The solvents used are mineral oil, high boiling hydro-
carbons 1n the 400to 530F boiling range and triethylene

     Letterpress and lithographic newsprint inks contain
aliphatic hydrocarbon solvents of 425 to 530F boiling
range.  Initial  or immediate drying after ink application
occurs by absorption into the paper, with drying completed
In heated ovens.  Inks used on folding cartons and Kraft-
liner containers dry by oxidation and'some inks, used 1n
the latter, dry by precipitation.  Metal-container inks dry
by curing for 10 min at 300 to 400F.  Polymerization,
oxidation, and crossllnking occur.

     Inks used in flexographlc and  gravure processes are
referred to as solvent inks and differ from those used in
letterpress and  lithography in that they are of very low
viscosity and almost always dry by  evaporation of highly
volatile solvents.  Those identified with the graphic arts
industry in Table C-I are primarily those used'by the flex-
ography and gravure processes.  However, because of the use
of rubber plates and rollers, certain ones of these solvents
cannot be used in the flexographlc  process.  Toluene and
xylene cannot be used at all; ketones can be used with butyl
rubber, while hydrocarbons  can be used with buna rubber.
Alcohols, glycols and low concentrations (25%) of esters are
acceptable.  The boiling range of the solvents extends from
147 for methanol to 343F  for 2-butoxy ethanol.

     Air Pollution From Graphic Arts - Without pollution
control, volatile solvents  are emitted into the atmosphere
from the flexographlc and gravure printing operations during
the ink application and also during subsequent drying.  As

      listed in Table C-I, 83 organic solvents have been Identified
      with these two printing processes.  The emission rate, essen-
      tially of these solvents, 1s estimated at 5 Ib/da per 1000
      people based on the San Francisco survey.  Pollution from
      the letterpress and lithographic printing operations will
      differ considerably.  Very little vapor emission will occur
      during the Ink applications but, where high temperature oven
      drying is carried out, high boiling solvents will be emitted
      accompanied by some decomposition products of the drying oils
      and resins.  The high boiling solvents emitted have Vm values
      considerably 1n excess of 190 cm3/nol and would make carbon
      resorb or regenerative systems impractical for solvent re-
      covery.  The decomposition products add to the regenerative

          Emission Control - The extent of emission control 1n
      the various printing plant operations has been investigated
      by Gadomskl (1970) through visits to 57 plants distributed
      in 10 states, namely:  California, Illinois, Wisconsin,
      Georgia, Tennessee, Florida, New York, New Jersey, Pennsyl-
      vania and Ohio.  Of the 57 plants, 40 plants use no solvent
      emission control devices.  Of the remaining 17 plants 5
      different types of devices or methods are used to control
      emissions, vOne plant, Involved in metal decorating opera-
      tions, uses a deodorizer unit wherein a deodorizing vapor
      is fed into the exhaust alrstream.  Another plant, also  in
      metal decorating operations, uses a spray chamber.  Seven
      plants Involved in metal decorating and lithographic opera-
      tions use direct flame afterburners, with and without heat
      recovery.  Six plants, also in metal decorating and litho-
      graphic operations, use catalytic combustion incinerators,
      with and without heat recovery.  Of the 57 plants, 8 plants
      are Involved with gravure operations.  Two of them use
      activated carbon solvent recovery systems, while the other
      six use no control devices.  The two using solvent recovery
      systems may be classified as large plants, although not  all
      of the nonusers can be classified as small.  The airflow
      per press unit is 1n the range 3,000 to 11,000 ft3/m1n,  and
      air temperatures are about 100F, a favorable temperature
      for adsorption.  The airflows from the lithographic and
      metal decorating operations are over 300F and generally  1n
      the 400 to 600F range, temperatures favorable for catalytic
      and thermal incineration pollution control.

          Duhl (1970) has analyzed the factors affecting economic
      recovery of oravure solvents with carbon solvent recovery
      systems.  For an example calculation, he used an 80 press
      installation with total exhaust  airstream of 120,000 ft3/m1n
      at 100F with approximate solvent load of 4,400  Ib/hr.   The
      solvent mixture being utilized was principally the acetate
      esters Including, at times, the ethyl, Isopropyl, butyl  and

amyl and, at times, some quantities of ethanol  and propanone
along with aliphatic hydrocarbons, toluene and  xylene.   With
these solvents the concentration was of the order of 3,000
ppm.  The system consisted of six Vulcan-Cincinnati  Adsorbers
on-stream preceded by a glass fiber filter to remove paper
and dust particles, and a fin tube cooling unit to control
air temperature to approximately 95F.  Economic analysis
showed a recovery of $1,600,000/yr of solvent and payout
time of 1.5 yr for the installation.

     One of the large cost items in solvent recovery in
gravure and surface coatings industries, where  water soluble
mixed solvents are used is the separation of the solvents
from the steam condensate and from each other.   For large
installations where continuous and larger, more efficient,
separation systems can be used the cost decreases consider-
ably.  Large Installations can show a profit on solvent re-
covery.  This may not be the case with snail installations.

     Other Solvent User Industries

     The combined solvent emission from the remaining solvent
user Industries is of the order of 32 Ib/da per 1,000 people
based on the San Francisco Bay Area (1969) survey report.
Table 1-XXII gives a partial breakdown of the industries and
estimated emission rates for three groups of industries.
Table 1-XXII  Amount of Solvent Emitted by Plastics. Rubber.
              Paper Coating and Pharmaceutical Industries

Industry                             Lb/da per 1,000 people

Plastics, resin, putty mfq.                    14
Surface coating, rubber products,
   pharmaceutical mfg.                          6
Other solvent users, possibly
   adhesives, paper coatings,
   extraction of vegetable oils
   plastic products                            Jl_2_

                          Total                32
     The solvents used in these industries have been identi-
fied, as listed in Table C-I, and are the same as used in
surface coatings, and flexographlc and gravure printing pro-
cesses.  The problems of pollutant control are also quite

          A number of carbon solvent recovery systems are 1n
     successful operation in fiber or synthetic filament pro-
     duction (plastics), rubber coatings of gloves,  pharmaceuti-
     cals, adhesives and paper coatings (Vulcan-Cincinnati,  Inc).

          Rubber Products Industry

          General - Certain raw and synthetic rubber products
     processing steps emit uncontrolled organic solvents and
     participates into the atmosphere.  In addition, pollutants
     are emitted in the production of certain types  of synthetic
     rubber such as butadiene-styrene (SBR).  While  most of  the
     rubber products industry consists of large operations such
     as tire manufacture, small specialty Items are  also produced.
     Most rubber products are processed in the same  manner except
     for changes in compounding recipes and vulcanizing operations,

          Product Processing Steps - With a few minor exceptions
     the manufacture of rubber products involves:  compounding,
     mixing, forming and vulcanization.  Compounding is the  pro-
     cess of selecting and weighing the various ingredients  re-
     quired by the formulation and Involves the selection of
     rubber, rubber chemicals, reinforcing pigments  and processing
     aids.  Mixing is the blending of the compounding mixture to
     attain dispersion of pigments and production of consecutive
     uniform batches.  In forming process, rubber is extruded
     into the desired shape or calendered (applying  sheeted  rubber
     to a fabric).  Rubber cementing is also a forming process in
     which rubber stock solvents are mixed to form an adhesive.
     Rubber building 1s the combining of stocks of different
     composition.  Vulcanization (curing) 1s the last major  pro-
     cessing step.  The curing process Involves application  of
     heat to convert the green dough Into an elastic state.

     Emission Sources from Rubber Products  Processing  -  Most
pollutants are emitted from rubber mills  used  in the mixing
process and in applications of rubber cements  in the forming
step.  Particulates (dusts and fumes, oil mists) and odors
are emitted.  Mixing emissions vary depending  on the size  of
the mill, size of batch and composition of  the mix.  Solvent
vapors are uncontrolled.  Large quantities  of  uncontrolled
organic solvent vapors are also given off  in the form  of
rubber cements (D'Imperio, 1967).

     Rubber Cements - Rubber is more or less soluble in  most
aliphatic and aromatic solvents.  The cements  are made by
mixing stock rubber and solvent together  in churns to  form a
solution which will flow at 10% concentration  (Rogers, 1968).
Cements find application in production of  products such  as
hospital sheeting, balloon and life raft  fabrics, nipples,
gloves and in part of the tire building process.

     Emission Types and Properties of Rubber Solvents  -  The
physical properties of organic solvent vapors  emitted  from
products processing are given in Tables 1-XXIII and  -XXIV.
The processing solvent used varies according to the  base
polymer.  Solvents such as benzene, toluene and xylene are
used with natural and SBR rubber.  Various  ketones,  etc.,
used with nitrile and neoprene synthetics.   The proportion
of solvent used is determined by the solvent power of  the
stock, consistency desired and character  of the stock.

     Synthetic Rubber Processing - The process for producing
most synthetic rubber consists of mixing  monomers of buta-
dienestyrene (SBR).  Soaps and mercaptans  are  added  and  the
mixture polymerized.  Oil and carbon black  are then  added
and the mixture coagulated and precipitated from a latex
emulsion.  Rubber particles are dried and  baled (MacGraw,

Table 1-XXIII  Properties of Solvents Emitted  from Natural  and  SBR  Rubber
Carbon tetra-
Diethyl ether
formul a
caHs to ceH-|4
Mol wt
Boi 1 inq
ranqe, F

Table 1-XXIV Properties of Solvent Emitted from the Synthetic Rubber Industry


1 ,2-Dichloro-

          Boiling     LEL,
Mol wt   range, F    pom
          172-178    20,000
CH3COCH2CH(CH3)2      6.68      100      237-243     14,000

          Emission Sources from Synthetic Rubber - Organic com-
     pounds are emitted from reactor and blow-down tanks; odors
     are emitted from the drying process.

          Emission Types and Properties from Synthetic Rubber -
     Emission factors are presented on Table 1-XXV as given by
     MacGraw (1970).  In the SBR process, gases are recycled, but
     some odors are emitted.

          Pollutant Emission Control - Rubber mills have exhaust
     hoods to remove dusts, fumes and mists from the rolls.  The
     most common control device is a bag house to remove partic-
     ulates.  Scrubbers are sometimes used to control oil mists.
     There does not appear to be odor control devices.  In the
     SBR process, gases are recycled back through the process.

          It is not known if any kind of air pollution equipment
     is used in tire retreading shops since pollutant data and
     emission rates are difficult to obtain in this area.  It Is
     likely that the same organic solvents are used in retreading
     as are used 1n other parts of the rubber industry.

          Solvent Extraction of Vegetables

          Solvent extraction 1s an efficient way to recover oil
     from seeds, such as soybean and cotton, and from lemons.
     Hexane or heptane is used for the seed extraction where the
     process may be continuous in large plants or batch-wise in
     the smaller ones (Cowan, 1969).  A closed system 1s employed,
     keeping vapor emissions very small.  A carbon resorb would
     be amenable to the recovery of any escaping solvent vapors.

          2-Propanol is used in the extraction of lemon pulp in
     the plant of Exchange Lemon Products, Corona, California
     (Barnebey, 1965).  2-Propanol vapors escaped from the plant
     operation in exhaust air at 2,000 ftvmin, 8300 ppm and
     120F.  The solvent 1s being recovered for reuse with an
     activated carbon resorb system.  Because of particulate
     matter and small amount of HC1, a liguid scrubber 1s used
     to clean the air-solvent mixture before entry into the carbon

          Solvent and Fuel Evaporation From Storane

          Another potential emission source 1s the organic vapor
     loss from above-ground fixed-roof storage tanks, which are
     subject to day-night temperature fluctuations.  Almost every

Table 1-XXV Emission Factors from Synthetic Rubber Plants
            Pounds per Ton of Product, (MacGraw, 1970)
          Compound                    Emissions* '

            Butadiene                     40
            Methyl propene                15
            Butyne                         3
            Pentadiene                     1

            Dimethyl heptane               1
            Pentane                        2

          Ethanenitrile                    1

            Propenoic nitrile             17
            Propenal                       3

(1)  Butadiene emission is not continuous and is greatest
     after a batch of partially polymerized latex enters
     the blowdown tank.

      chemical  manufacturing  operation  has storage tanks for
      solvents  or  reactants,  while  the  petroleum Industry nay be
      the  largest user of  storage  facilities.  For gasoline storage,
      29,000  tanks  are  reported 1n  use  1n the country, v/hich range
      in size from  42,000  to  21,000,000 gallon capacity.  Gasoline
      vapor loss from  this  source 1n  the United States is estimated
      at 7.5  Ib/da  per  1,000  people.  Gasoline evaporative losses
      of all  types  approach 25 Ib/da  per 1,000 people, which in-
      clude losses  from floating  roof tanks and during transfer.
      In the  San Francisco  Bay Area (1969), the estimated indus-
      trial solvent evaporation rate was 5.1 Ib/da per 1,000

          Some calculations  were made which indicate that carbon
      adsorbent devices are feasible for this application and that
      further study is warranted.   For a 3,360,000 gal tank, the
      estimated evaporation is 71,400 gal/yr, API Bulletin 2518.
      The  concentration of  gasoline vapors in the vent gas during
      warm-up is reported  to  vary from 30% to 5058 by volume.  If
      the  assumptions given below are made, the intake flow during
      cooling is about 20  ft'/min and out flow about 40 ftVmin.

          Mean ullage                  50% of capacity

          Cool-down and
            warm-up time               8 to 10 hr

          Temperature fluctuation
            span                       25F

      Because of the high vapor concentration, cooling of the car-
      bon  bed will   be necessary.  This can be done by pumping a
      coolant liouid through  colls  in the carbon bed.  During the
      night, when air is being drawn in, the bed could be desorbed
      by applying heat by means of  the same liquid and colls.  Per
      adsorption phase of  the cvcle the gasoline weight pick-up
      will  be about 1,200  Ib, requiring from 3,500 to 5,500 Ib of
      carbon.  The  amount of  gasoline vented and hence the required
      carbon bed size will vary with the ullage and temperature
      change span.   The main  cost items will be the adsorber and
      carbon, refrigeration and heating.  The amount of gasoline
      recovered will be 71,400 gal/yr which at 13^/gal will place
      the  recovered gasoline  value  at $9,000/yr.  The value of
      the  recovered gasoline  reduces the overall cost of the
      pollutlon control .

          Gasoline is low molecular weight, containing considerable
     amounts of butane and propane, hence the high or near satura-
     tion  vapor pressures and the  need for refrigeration In adsorp^

tion; see Table 1-IX for compositions.   Industrial  solvent
and chemical  reactions that are  of higher molecular weight
will be more  amenable to carbon  adsorption recovery systems.

     Gasoline, other fuel, and solvent  vapor losses can
occur during  transfer of the liquids  between storage tanks
and motor vehicle tanks due to displacement of  vapors from
the tank being filled.  One method to prevent this  type
vapor loss is to Install a vapor line to carry  the  vapors
from the tank being filled to the tank  being emptied.  When
this method Is not feasible, the vapors can be  recovered  by
a carbon resorb type system.  Incineration of the  escaping
vapors has also been proposed.

     Meat Rendering Industry

     Description of Industry - Animal matter (offal) not
suitable for  food for either humans or  pets is  converted  to
salable products by rendering.  Rendering plants  are the
major outlets for disposal of waste from slaughter  houses,
butcher shops, poultry dressers  and other processors.  They
also dispose  of whole animals that have died naturally or
accidentally.  In this respect,  the rendering also  operates
as a scavenger or solid waste disposal  Industry.

     The process is thermal and  yields  as main  products  a
protelnous meal, suitable as animal feed, and tallow.  Bones
are also ground into the process Input feed; hence  the pro-
telnous meal  1s enriched with calcium and phosphate.

     Based on statistics given by Easton (1969) the meat
and poultry industry has been growing at a rate of 5* per
year since 1963 in dollar volume of business.  Since the
meat rendering operations are scavengers for the overall
meat Industry, they of necessity must also expand.

     Most rendering plants are old, but some modern plants
have been constructed 1n recent years.   The older plants
operate batch processes while the modern ones have converted
to continuous operations such as the Anderson Carver Green-
field (Barr,  1966) or the Dupps  Duke systems.

     The estimated number of meat rendering plants  is
between 800 and 900.  Average figures for the older plants
Indicate about 4 cookers per plant, 6 batches of offal pro-
cessed per day and daily processing of about 100,000 Ib,
Norfolk Tallow Company, Chesapeake, Virginia reports a
20,000 Ib/hr processing rate for the continuous-type plant.

          Rendering Processes - The rendering process 1s comprized
     of 3 basic steps:

          1.  Grlnelng

          2.  Cooking

          3.  Separation of tallow from protein-bone meal.

          The grinding breaks down the offal to a homogeneous
     hash which 1s then charged Into the cooker.  In the older
     plants the cooking 1s a batch process wherein a steam Jack-
     eted cooker 1s used.  Steam temperatures at 300F are needed
     to digest the bones, hooves, hides and hair, 1n addition to
     the meats and fats.  Process times range from 1 to 4 hr.
     The moisture content of the charge Is Initially 40% to 50%
     which, during the heating phase,  1s reduced considerably,
     the steam being vented through a  condenser.

          The exhaust steam contains  very odorous gases  which are
     1n part dissolved 1n the condensate and in part escape through
     the condenser 1n gas phase.

          The hot digested  mixture or  slurry from the cooker Is
     then discharged over a screen.  Most of the melted  tallow
     passes  through, while  the solids  or cracklings, consisting
     of the  digested bones  and protelnous constituents,  are re-
     tained.   The cracklings  are pressed to remove most  of the
     adhering tallow.   At this point  the moisture content of the
     cracklings  1s  below 8% allowing  preservation without refrig-
     eration.   The  tallow Is  filtered  and then  dehydrated by
     centrlfugatlon, settling or air  blowing.

          In  the  Anderson Carver Greenfield system,  the  Incoming
     raw  material  Is  finely ground and  mixed with recycled  liquid
     fat.   In  the  second  step,  the fluldlzed ground  slurry  1s
     pumped  through  a  steam heated evaporator operating  at  190
     to 210F  and  27  In.  of Hg vacuum.   The released water  vapor
     and  noncondensable  gases  are  expelled through a water  cooled
     contact  condenser.   The  dewatered  slurry 1s continuously
     centrlfuged  to  remove  the  major portion of  the  tallow  from
     the  cracklings  and  the cracklings  are then  run  through  con-
     tinuous  presses  or  expellers  to reduce the  tallow  content.

          Some materials  (blood  and feathers) that do not contain
     tallow  are  digested  and  dewatered  1n  batch  dry  rendering

     Pollutant Emission - Odorous pollutants escape from
several locations of the older batch operated plants.   Mea-
surements were made by IIT Research Institutes (1971)  to
establish a sem1quant1tat1ve assay on the magnitude of
various plant type vapor samples.  The estimates  are reported
1n micrograms (ug) per liter of plant air in Table 1-XXVI.
In the third column the concentrations have been  converted  to
volume ppm, assuming a mean molecular weight of 90.
Table 1-XXVI  Concentration of Pollutants  in Plant Air At
              Various Operating Areas or Exhaust Vents

                           	Concentration ranee,
Area or vent                  ug/1	        ppm	

Raw pile area               1.9 to  3.8      0.5 to 1.0
Cooker condenser           10.8 to  11.2     2.9 to 3.1
Cooker dump area            3.1 to  5.8      0.8 to 1.6
Press area                  4.5 to  8.5      1.2 to 2.3
Press ventilator           27.4 to  50.4     7.4 to 13.7
Grease drying operation     7.7 to  9.6      2.1 to 2.6
     Mills, et.al. 1967, reported estimates of odorant
emissions from rendering operations in the Los Anneles area.
From a survey of beef cattle operations (swine, sheep, poul-
try, etc. were not considered in these estimates) and
assumptions of average weights of cattle and inedible refuse
(mainly blood and offal), Mills calculated that plants
handling refuse from 28,000 cattle would emit about 3x10"
odor units per day from offal and about 5x10" odor units per
day from blood.  (An odor unit is the number of dilutions
required to dilute the odor Intensity of a unit volume to
Its recognition threshold concentration.)

     For a batch rendering process, Mills (1967) reports
total emissions and odorant emissions as a function of time
during processing.  Peak odor emissions take place during
the first hour of operation, as shown In Figure 1-3.

     The continuous rendering plant is to a considerable
extent a closed system.  In a well maintained plant, pollu-
tant vapors could escape into the atmosphere only at a few
locations; i.e., at the raw material receiving bins, evap-
orator condenser, and at the crackling or meal presses.
The latter two emission points can be readily connected  to
pollutant control systems.  Odors from the receiving bins
are variable depending on the type of material being pro-

                                 Odor emission  rate
                               1             2

                            Time for one batch, hr
      Flaure 1-3  Typical Rates of Odor Emissions and of Vapor
        M         Emissions from a Batch-Type Rendering Cooker
                  Reducing Inedible Animal Matter

cessed, age of the material  and ambient temperature.

     Identity of Pollutants  Emitted -  Table  1-XXVII  lists
43 pollutant vapors that have been found 1n  air  samples  from
various rendering operations (Straus,  1964;  IITRI,  1971).
An analysis of study results conducted by IITRI  (1971)  give
Indications that the major part of pollutants  are concen-
trated in about nine compounds.  The study used  the  tech-
nique of dual column gas chromatography and  mass spectro-
graphy.  Thirty chemical compounds were identified  in syn-
thetic laboratory vapor samples prepared from  prepressed
cracklings, finished crackles and press grease solid  samples
from the Chicago rendering plants.  Those compounds,  as
identified by the Kovats Index, that were in excess  of 10%
of the total pollutant mixture are listed in groups  1n
Table 1-XXVIII (Kovats, 1964).  Of the nine  compounds pre-
dominating 1n the vapor mixtures, five were  identified,
namely, 2-methyl propenal, 3-methyl butanal, n-hexanal,
propanoic and butanoic acids.  Host of the other compounds
were present 1n fractions of a percent.  IITRI did  not
report any amine or sulflde compounds  although they qual-
itatively detected the latter 1n all samples.

     Emission Control - The control of pollutant emissions
from rendering plants poses a difficult problem; one that
has not been solved although attempts have been made.
Factors contributing to the difficulty are:

    ' 1.  Low pollutant concentration, hence negligible
         fuel value and thereby a disadvantage in
         catalytic and thermal incinerations.

     2.  Extremely low detection  levels of many of
         the malodorous pollutants, thereby requiring
         large decreases  1n concentrations, a task
         difficult to accomplish  by mere dilution
         with additional  air.

     3.  Release  of pollutants at  low concentration
         levels  from many locations thereby requiring
         ducting  or hooding from  many locations  or
         use of  plant wide  ventilation.

     4.  High moisture  content and  particulates  in
         the exhaust air.

     Operating  costs have been estimated  for  the incineration
and  carbon  adsorbent systems  and  are  reported in detail in
Chapter  6.   The  results  of  the estimates  are  summarized

Table 1-XXVII  Pollutant  Vapors Emitted  from Meat  Rendering  Processes
Oifurfuryl ether
Oipropyl sulfide
Mol wt
ooint, F
 3-Hethyl  butanolc

 Butanoic  acid

 2-Methyl  tetra
 3-Methyl  butanal






. . .



Table  1-XXVII  (continued)
 Propanoic acid
 Diethyl amine
 2-Methyl propanal

 2-Methyl propenal

 Propene nitrile

 Hydrogen sulfide

Acetic acid



Hoi Wt





point, F

Table 1-XXVIII Most Abundant Pollutants in Vapor Mixtures  from  Crackling. Grease

and Tallow Durinq Rendering
Kovats Index Probable chemical Percentaqe
of pollutant Identity of poll. in vapor
885 "
 2-methyl propenal 29
923 3-methyl butanal 36
994 1_ 11
998 J 53
1077 15
1087 _
 n-hexanal 40
1142 12
1415 '
1425 _
Description of
pollutant odor
aldehyde, musty
aldehyde, musty
aldehyde, musty
aldehyde, musty
aldehyde, musty
minty, unpleasant

Table 1-XXVIII(continued)
Kovats index
of pollutant

    1603  J
Probable chemical
Identity of poll.

propanoic acid
butanoic acid
 in vapor


Description of
pollutant odor

unpleasant, soap
unpleasant, soap

     graphically  1n  Figure 6-31. Results taken from Figure 6-31
     show  that  to operate the three systems at the 10 ppm concen-
     tration  level,  as found 1n rendering operations, the costs to
     clean  the  exhaust air would be approximately as follows:
        Type control system

        Thermal Incineration
        Catalytic Incineration
        Carbon adsorbent
Cost to clean
$/hr per 1000
f low
     Although the carbon adsorbent system can be operated at the
     lowest cost, there 1s some uncertainty that 1t can effect
     complete removal of all pollutants.  As reported 1n Table
     1-V, 28 of the 43 pollutants have Vm between 80 and 190 cm3/
     mol but Included 1n these 28 are the pollutants appearing in
     the exhaust alrstream at the highest concentration; see
     Table 1-X for Identity of these pollutants.  No real problem
     1s foreseen 1n the adsorption and desorptlon of the predom1
     nant pollutant vapors.  13 of the 43 pollutants, however,
     have Vm less than 80 cm3/mol.  Retention of some of them may
     be poor so an early penetration at low concentration may
     occur.   Thermal  and catalytic Incinerators also operate at
     less than 100% efficiencies.  The relative effectiveness of
     carbon  adsorbent and incineration systems can only be deter*
     mined under actual operating conditions.

          Fish Rendering Industry

          The  main pollutants from the fish rendering operations
     as reported by Conrad (1971) are as qlven in Table 1-XXIX.
     The pollutants are listed downward 1n decreasing order of
     concentration and the concentration range 1s 1  to 12 ppb.
     Table 1-XXIX  Pollutant Emissions  From Fish  Rendering

          Pol lutant	          Odor threshold,  ppb
          Dimethyl  dlsulflde
          Carbon dlsulflde

The odor threshold (Stern, 1968) concentrations of tri-
methylamine and ethanethiol are of such low value that even
on considerable dilution in the atmosphere after leaving
the plant operating area they are still detectable.

     Escape of pollutants occurs in several steps of the
process, but one of the larger emission sources is the fish
meal drier,  Mills (1963) has reported exhaust airflow rates
and odor concentrations for various fish meal input rates.
To indicate the pollutant emission rate level, the exhaust
gas volume for a drier of 10 ton/hr meal input rate is
18,500 ftVmin and concentration is 1,500 odor units.  If
dimethyl dlsulfide is assumed to be the predominant odor
the pollutant concentration is of the order of 10 ppm, but
if the predominant odor 1s that of trimethylamine, the
pollutant concentration is of the order of 0.3 ppm.

     The problems of pollution control in the fish rendering
plants are expected to parallel those of the meat rendering
plants.  According to Conrad (1971) the public complaints
are directed mainly toward particulate emissions rather than
odors from the plant.  Also odors from unloading ships are
subject to more complaints than from the fish processing

     Canning Industry

     Vegetable, meat and fish canning Industries have some
odor problems, with the fish canning emitting the most offen-
sive variety while the vegetable canning, the least.

     Vegetable Canning - The basic source of odor comes from
the blanching operation which is done as soon as possible
after the vegetables have been removed from the field.  This
is done to facilitate preparation procedures such as removal
of skin and to prevent deterioration due to the time lag
between picking and canning.  Although odors are emitted
which could be judged objectionable by normal standards,
canning plants are usually located close to the source of
the vegetables in rural areas.  The majority of the people
1n the area are employed 1n the industry and thus odor is
accepted as a necessary factor.

     Fish Canning - These canneries are located near harbors,
and fish rendering plants are located near the canneries  to
process the by-products.  Much of the odorous air pollutants
attributed to canneries come from by-product processes    Only
choice fish parts are canned for human consumption,  the re-
mainder being converted Into by-products such as fish oil
and protein animal feed supplements.

          Canning preserves highly perishable fish foodstuffs.
      Hsh are canned by two methods.  In the older "wet-fish"
      method trimmed fish are cooked directly 1n the can.  The
      newer method, "pre-cooked process", 1s characterized by cook
      1ng the whole, eviscerated fish, and by hand sorting of
      choice parts before canning.

          In wet-fish canning, open cans containing the fish are
      conveyed through hot exhaust boxes where live steam cooks
      the fish.  At the discharge end, cans are mechanically up-
      ended to decant stick-water (condensed steam, juices, oils),
      which is collected and retained for by-product processing.
      Cans of drained fish are filled with oil or sauce, sealed
      and pressure cooked.

          In the precooked process (used for larger fish), whole,
      eviscerated fish are put 1n baskets and then live-steam
      cooked.  After cooking, the fish are cooled and placed on
      a  conveyor belt where the choice portions are picked by hand.
      packed and sealed Into cans, and then sterilized by pressure
      cooking.  The scraps and stick-water are collected and trans*
      ferred to the rendering processes.

          A1r pollution from fish canning 1s due to the malodorous
      organic compounds carried out with the steam vapors vented
      from cookers and 1s also due to nonsoluble and noncondensable
      vapors released from the stick-water.  To abate the odor
      release several control operations are required, the first
      of which is the condensing of the steam.  This operation
      will remove some of the water soluble and condensable vapors.
      The remaining vapors, including varying amounts of air, may
      then be burned by catalytic or thermal incineration.  If
      particulates are present, a water spray and demlster are
      required prior to the catalytic incinerator.  If air en-
      tralnment 1s large so that pollutant concentration 1s low,
      a  carbon resorb system following the water spray and demlster
      may be more economical.

          The odor vapors have not been identified but  their
      Identification will, in this case, not be a factor in deter.
      mining the effective use of a carbon resorb system.  Most of
      the compounds 1n the lower part of Table C-I, up to 80 cm3/
      mol Vm, would be removed by the water spray because of their
      water solubility.  Five compounds are not water soluble -
      namely:  methanethiol, propyne, dichloromethane, carbon
      disulfide and ethanethlol - but of these, carbon disulflde
      and ethanethlol and all those that are above 80 cm3/mol,
      Vm, are adsorbed strongly enough to be removed.

     Meat Canning - The processes Involved 1n meat canning
Involve:  (1) cooking of the meat In a sealed kettle,  (2)
transfer of meat from the kettle to cans,  (3) conveyance
of the cans through an exhaust box and  (4) sealing of the

     Odor release occurs when the kettle 1s opened and
during conveyance through the exhaust box.  In both cases
the odors are released in the presence of considerable
steam.  As in fish canning, condensation of the steam is
the first step in pollution control, followed by Inciner-
ation or adsorption of the Insoluble and noncondensable
gases.  The factor determining whether incineration or
adsorption is used is the pollutant concentration.

     Based on the San Francisco Bay Area survey (Bay Area
1969), the estimated pollution rate from restaurants is
1.6 Ib/da per 1,000 people or, for 15,000 restaurants, at
a rate of 0.4 Ib/da per average restaurant.

     Kitchen odor emissions come from the grill and frver,
vegetable cookers and steam tables.  In small  operations,
these are all vented through a common duct.  In large oper-
ations a separate ventilation system is directed to'frying
and grilling, and emissions from vegetable and steam tables
are controlled by the main ventilation system.

     The primary source of odors is from the frying and
grilling operation, but the identities of the  pollutants
have not been determined.  Qualitatively the pollutant 1n
the restaurant exhaust can be described as oily matter of
a wide range of particle sizes, accompanied by various food
volatiles and decomposition products.  The predominant odor
character is oily.  Acrolein, a dehydration product of gly-
cerol, is present to some extent in effluent from fat fry-
ing, but it 1s not the predominant odor.

     A carbon resorb system can be an effective means for
removing the pollutant vapors from the exhaust to the out-
doors since carbon air purification systems are used in  the
dining areas of the restaurants to remove kitchen odors.
The carbon resorb system would have to be well protected
from the oily partlculates by means of dry filters or wet

          Meat Smokehouses

          Control  of  pollutant emissions from smokehouses appear
      to  be better  controlled by  thermal Incineration than by
      sorbent systems  due  to two  reasons:  (1) the emission con-
      tains poorly  adsorbed methanal, and  (2) the pollutant con-
      centration  Is generally in  the 1100 to 1500 ppm range where
      the sorbent system 1s not economically competitive with the
      catalytic and thermal incineration systems.  The use of ther-
      mal incinerators  has not, however, been entirely successful
      in  controlling smokehouse pollutant emissions.  Pollutants
      identified  in the emissions are as follows:

          Pollutant                      Concentration, ppm

          Methanal                            20 to 40
          Higher aldehydes                   140 to 180
          Formic acid                         90 to 175
          Methanoic and ethanoic add        460 to 500
          Phenols                             20 to 30
          Ketones                            190 to 200
          Resins                                1000

      According to  the  U.S. Industrial Outlook, 1970, smoked products
      are being prepared chemically, hence, smokehouses may be elim-
      inated 1n the near future.

          Coffee Roasting

          Coffee roasting is in  the doubtful category.  The
      effluents are emitted at temperatures in the 500 to 900F
      temperature range and contain both partlculate matter and
      vapors, the vapors being the lower molecular weight alde-
      hydes, lower  molecular weight organic adds, hydrocarbons
      and mercaptan at  total concentration near 500 ppm.  Under
      these conditions  the estimated control cost with a carbon
      resorb system would  be over $0.80/hr per 1000 ft3/min of
      gas flow.  The operation would Involve cooling the gas
      stream, partlculate  filtration and burning of the desorbed
      pollutants.   Catalytic Incineration, at 500F gas tempera-
      ture, would cost  about $0.80/hr per fts/rnin and about
      $0.41 at 900F gas temperature.  Particulates may have to
      be  filtered from  the gas stream at the elevated temperature.
      Thermal incineration would  cost about $0.97/hr per 1000
      ftVmin at  500F  gas temperature and less at 900*F.  The
      latter would  have an advantage in that the partlculates
      would also  be burned to a large extent If not completely.

     Tobacco Curing,  Zinc  Plating  and  Tanning

     Tobacco curing,  zinc  plating  and  tanning  do  not  appear
controllable with carbon  adsorbent systems  operated at  normal
temperatures, because of  the  low molecular  weight pollutants.
The pollutants emitted by  these  industries  are listed below:
     Tobacco curing

     Ammoni a
     Acetic add
                         Zinc plating


The ammonia is not controllable with unimpregnated activated
carbon systems.  A carbon impregnated with phosphoric acid
will react with the ammonia, but regeneration is expected to
be too expensive.  In a dry airstream such as might be the
case with zinc plating, a molecular sieve sorbent system
would be feasible and desorption then done by heating and
returning the ammonia back into the operations.  The methanal
and methanol are also poorly adsorbed by activated carbons,
hence, competing systems or methods have been considered.

     Some catalysts decompose ammonia to N2 and H2 a possible
answer to zinc plating pollution problem, unless traces of
ammonium chloride or zinc chloride poison the catalyst.  A
two catalyst bed would be required for tobacco curing and
tanning; the second being a combustion catalyst to burn
methanal and the other organics.

     Control of Pollutants from Incinerators

     Most of the refuse  in San Francisco was burned  in
residential incinerators  (Bay Area,  1969).   For 1969,  the
number of reported residential incinerators  was 1,080,000
                   tons  of refuse  per day.   Also  in  use  were
                   commercial  Incinerators which  burned  15
                   day and 744 multiple  chamber  commercial
                   burned 398  tons  of refuse per  day.   There
was a  drastic  difference  in the amount of  pollutant" emitted
by  each  type of  incinerator.   The  reported emission  factors,
as  given  in Table  1-XXX  indicate  incomplete  combustion in
the residential  incinerators,  resulting  in excessive emissions
of  organics,  particulates and  carbon monoxide.   The  use of
which burned 2,214
305 single chamber
tons of refuse per
incinerators which

     multiple  chamber  commercial  Incinerators for the burning of
     all  refuse  appears  to  be  one  of  the better solutions to the
     Incineration  pollution  problem.  With the latter the organic,
     partlculate and CO  emissions  are reduced to very lov/ levels.
      Table  1-XXX   Pollutant  Emission  Factors for Three Types
                   of  Incinerators


Emission factor.  Ib  pollutant/ton  refuse
 Organic    Part.     MOx    SOX      CO
0.1    1.0    600
1.0    1.0
2.0    1.0
           A  carbon  adsorbent  system,  to  clean up the effluent from
      the  residential  Incinerators would  not  be practical, because
      the  adsorbent  system  would  have  to  be operated at,   (1) very
      adverse conditions,   (2)  1t would not remove CO, and   (3) the
      desorbed pollutants would have  to be disposed of.
           Baking  and  Candy  Making  Industries

           The  candy manufacturing  and  baking  1
      concentrations of  odors  which most  people
      objectionable.   In candy manufacture,  the
      comes from roasting the  cocoa beans.   The
      the  odors  are  not  reported  1n literature.
      emission  from  the  baking industry is etha
      emissions  from these two sources  is not  a
      trol  of emissions  from the  other  sources
      However,  the use of incinerators  for thei
      have  been  reported.

           Other Emission Sources

           The  industries listed  in Table l-II account  for a high
      nercentage of  the  pollutant volume  emitted  from small sta-
      tionary emission sources and  also those  that  are  of major
                     ndustrles  emit low
                      do not find  highly
                      main source  of odor-
                      chemical  nature of
                       The main identifier
                     nol.   Control  of odo
                     s urgent as the con-
                     discussed  previously
                     r successful  control%

concern because of malodorous emissions.   Other existing
and potential sources solvent and malodor emissions  of lesser
concern are recognized and are listed as  follows:
     Hospital exhausts
     Veterinary laboratories
     Cork products
     Sewage vents
     Perfume manufacture

     Soap manufacture
     Pesticide manufacturing
     Cosmetic manufacturing
     Fur processing
     Textile finishing

     Settling ponds
     Fertilizer manufacture
Sulfite paper mil Is
Pipe coating
Asphalt air-blowing
Plywood manufacture

Electroplating exhaust
Azo-dye manufacturing
Waste water plants

              2.   Environmental  Effects
     The primary concern over the uncontrolled emission of
air pollutants is their adverse effects on the environment.
These effects range from toxicity to humans, animals and
vegetation to materials damage, soiling and loss of visi-
bility.  Of paramount importance are the toxic effects on
humans, resulting in mortality, morbidity, physiological
and pyschological discomfort and impairment of performance
efficiency.  Certain pollutants have toxic effects on vegeta-
tion, leadinq to significant economic loss due to retarded
qrov/th, reduced yield, and, in severe instances, complete
crop destruction,  Other effects are damaqe to property,
as by corrosion of metals, deterioration and soiling of
architectural and other painted surfaces and discoloring
and soiling of  clothing and other personal  effects.

     Many  of  the adverse effects of air pollutants  can be
assessed  in terms of real  economic  factors, such as  health
costs  and  dollar losses due  to  crop and materials damaqe.
Other  effects can only  be  evaluated in  subjective terms,  but
are still  significant  in terms  of esthetic  enjoyment of  life.
Nontoxic,  but odorous  emissions  arouse  perhaps  the  most
immediate  and vigorous  public  response  to  pollutant emissions.
Reduction  of  visibility is  a  pronounced esthetic effect  of
particulate  and aerosol emissions and  atmospheric interactions
such as  photochemical  smog formation.

     Much  has been  published  concerning the adverse effects
of air  pollutants  and  this study  has  not  attempted  to  make
a  definitive  review  of the subject.   For  good  general  reviews
and  related  bibliography,  reference may be  made  to  the air
quality criteria published by DHEW  and EPA  for each of the
major  classes of air pollutants (reference  criteria documents).
This study oresents  only  a brief review of  effects  associated
with those pollutant classes  identified as  beino  emitted by
small  sources and  controllable by packaged  devices.

     The  magnitude  of  the  damage to the environment from ex-
posure to air pollutants  is  dependent on  two  factors:   the
potential  of  the pollutants  to produce the  adverse  effects
mentioned above and  the concentration of  the  pollutants 1n
the  atmosphere.  The evaluation of  these  two  factors, as re-
lated  to emissions  from small  sources, is the purpose of this

      Evaluation  of  Pollutant Concentration

           The  pollutant concentration  that  can  develop  1n  any
      area  will  depend  1n part on  the  pollutant  emission  rates  at
      the  various sources and 1n  part  on  the air mass movements
      and  other meteorological conditions.   As indicated  by  the
      numbers  given  1n  Table  l-III,  the average  emission  rate of
      all  types and  from all  sources is approximately 5,000  Ib/da
      per  1,000 people, with  a 130%  variation for the nine  densely
      populated areas referred to.   In  these, and all areas,  the
      pollutant concentrations will  vary  with the prevailing
      meteorological conditions.   For  Los Anqeles County  and  San
      Francisco Bay  Area, the meteorological conditions  are  unfav-
      orable for the dilution of  pollutants  in the inhabited  areas.
      In some  of the other survey  areas the  meteorological  condi-
      tions are more favorable, but  notso favorable that  no  effort
      1s required to reduce the overall emission rate.

           Pollutant dispersion models  have  been developed  which
      allow the estimation of pollutant concentrations  from  point,
      line  and  area  emission  sources (for example, Martin and
      Tikvart,  1968).  Considerable  Improvement  1s yet  required,
      however,  to allow realistic  modeling of urban, Industrialized
      areas, particularly where atmospheric  Interaction  and  smog
      formation 1s prevalent.

           The  current  best estimates  of  pollutant concentration
      levels required to be protective  of human  health  are  those
      given 1n  the Federal primary and  secondary ambient  air
      quality  standards.  Table 2-1  gives the standards  for  six
      pollutants  as  established by the  Environmental Protection
      Agency (EPA, April 30,  1971).   The  primary standard is  the
      pollutant concentration level  considered to be protective
      of human  health,  with a safety margin. The secondary
      standard  is more  restrictive 1n  regard to  SOx and  partlc-
      ulate concentrations.  It 1s protective also of other  as-
      pects of  human welfare, I.e.,  protection against  known  or
      anticipated effects on  soil, water, vegetation, animals and
      manufactured materials.

           When ambient air tests  show  higher concentrations  than
      allowable  by the  Federal standards, action 1s required  to
      curb  the  pollutant emission  rates.   The Identity  of the
      pollutant very often also Identifies the specific  sources
      from  which  it  was emitted.   In other cases, the pollutants
      cannot be  Identified with a  specific source, but  rather to
      a  group of  sources.

Table 2-1  Federal Ambient Air Quality Standards

                         Primary standard  Secondary standard
 Pollutant	   v/m3     ppm       v/mj      ppm  ""

 Carbon  monoxide    .
   max  8-hr conc.(')      10,000    9       10,000      9
   max  1-hr cone.         40,000    35       40,000      35

 Sulfur  oxides
   annual  arlth mean.         80     0.03         60      0.02
   max  24-hr  conc.(l)       365     0.14        260      0.1
   max  3-hr conc.O)        	     	      1300      0.5

 Nitrogen dioxide
   annual  arlth mean        100     0.05        100      0.05

 Photochemical  oxidants
   max  1-hr  conc.U)        160     0.08        160      0.08

   max  3-hr  cone.v' /
     6  to 9  AM             160     0.24        160      0.24

 Particulates                                     /  .
   annual geometric mean      75      	         60\z)    	
   max 24-hr  conc.vU        260      ----        150       	

 (1) This concentration  1s  not to be  exceeded more  than once
     a year.

 (2) This number 1s to be used as a guide 1n assessing Im-
     plementation of plans  to achieve the 24-hr standard.

           Table 2-II. based on the survey results  of the  three
      California air pollution control districts (Table  l-III).
      chows the relative portions of the main classes of pollut-
      an?s which J Jht appear In any control district.  In these
      survey result?, 84* of the pollutants are emitted  from motor
      vehicles   The amount of pollutants from this source can  be
      expected to be proportional to the population density. The
      pollutants emitted from the stationary sources (15X  of total)
      will also be Influenced to a considerable degree by  the In-
      dustrialization of the area.
      Table 2-II  Relative Amounts of Pollutants Emitted From
                  Various Sources (California 1969)
           Organic vapors
                                       Motor vehicles


                                       Motor vehicles


                                       Motor vehicles


                                       Motor vehicles
                                       Small emission
                                       Other stationary
           The total organic vapor emission 1s 18% of which 12% 1S
        om motor vehicles.  These are combustion products, hence
        11 contain high proportions of aldehydes and organic acids
        e 2.5* other-stationary-source emissions are primarily    *
           J	 J __.  f^ A A ^ 4fe ^ A 4h  ft* 
     It is evident that the orqanic pollutants  from motor
vehicles will tend to be widely distributed in  proportion to
population density.  Pollutants emitted from small  sources,
althouqh smaller percentage-wise, can create higher concen-
trations in local areas and produce toxic effects,  damage to
property and plants, and nuisance effects because of bad odors
On diluting into the atmosphere these pollutants can combine
with the organic pollutants (and NOx)
to raise the concentration level over
contributing to smog formation.
           from  the  other sources
           wider areas,  thereby
Toxic Effects on Humans

     The potential of a pollutant to produ
on humans can be evaluated by two systems.
the threshold limit value (TLV) and is the
concentration of a substance to which 1t i
nearly all workers can be repeatedly expos
without adverse effects.  The other is the
grade number, and consists of five numbers
the severity of symptoms felt when exposed
grade numbers are defined as follows:
               ce a toxic effect
                 One of these is
                maximum airborne
               s believed that
               ed day after day
                vapor Irritant-
                which Indicate
                to a vapor.  The
     Grade 0  Chemicals that are nonvolatile, or the
              vapors which are nonirritatlng to the
              eyes and throat.

     Grade 1  Chemicals that cause a slight smarting
              of the eyes or respiratory system if
              present in high concentrations.  The
              effect 1s temporary.

     Grade 2  Chemical vapors that cause moderate
              irritation, such that personnel will
              find high concentrations unpleasant.
              The effect 1s temporary.

     Grade 3  Moderately irritating volatile chemicals
              such that personnel will not usually
              tolerate moderate or high vapor concen-

     Grade 4  Severe eye or throat irritants, vapors
              which are capable of causing eye or
              lung Injury, and which cannot be toler-
              ated even at low concentrations.
Table 2-III lists the
                                            red by

      Table  2-III
Ranking of Pollutants in Decreasing  Order  of
Toxicity to Humans

Hydrogen chloride
Methanoic acid
Sulfur dioxide
Nitrogen dioxide
Tetrachlorome thane
Ethyl ami ne
Ethanoic acid
Hydrogen sulfide
Carbon disulfide
4-Methyl -3-penten-2-one
1-Ni tropropane
2-N1 tropropane
4-Methyl -2-pentanol
Ethanenl trile
Hexyl acetate
1 ,2-Di chlorobenzene
1 Ceiling limit, not to
2 Includes skin adsorptl
3 (NFPA) 1969 ratings
TLV, irritant
ppm grade No.
0.1 3
3 /,% 2
5 2
5 (1) 3
5 3
5 3
5 ,_. 3
5 (2) 3
10 MI 3(3)
10 (2) 3
10 Ml 2
10 (1) 2
10 , . 3
20 (2) 2
20 4

25 2
25 l(3j
25 2(3)
25 2
25 3
25 2
?5 (2) 2
4 W b . *
25 1(3)
25 1
25 (2) z(3)
40 1
50 2
50 MI
50 (2)

50 1
50 1
be exceeded


4 .
51 ,52


Table 2-III (continued)

Pol lutant
2-Furyl methanol
1 ,2-Di chloroethane
2-Methyl pentyl acetate
Pentyl acetate
^ Pinene
2-Ethoxyethyl acetate
4 -Me thy 1-2- pen ta none
Ethyl benzene
1 -Pentanol
Ni troethane
Nitrome thane
Mineral spirits
Butyl acetate
1-Methyl-propyl acetate
Propyl acetate
Tol uene
Tetrahydrof uran
Methyl acetate
Stoddard solvent
Isopropyl acetate
1 ,1 ,1-Tri chloroethane
Ethyl acetate

grade No.
1 (3)

/ ** \



1 ,4,5,6,7,8,51


      Table  2-III   (continued)

                                TLV,   Irritant
      Pollutant _ _   PP*    grade N<>-  - Industry - .

      Diethyl  ether             400       2      1,6,7
      2-Prooanol                 40       ]      1,4,7,9,51
      Hexane                    500       0      1,4,6,7,9,10,51
      Octane                    500       1      1,4,6.10.51
      XeplSne                    500       1      1,4,5,6,7.9

      Dllsopropyl  ether         500       2(3)   1,4
      Dlchloromethane            500       2^'   2,5,7
      Naphthas                ^500       1      2
        trlfluoroethane        1000       -      z,3
      Pentane                   1000       1      4,6,10,51

      2-Methyl-butane           1000       -      JO
      1,3-Butadlene            1000       2      4,6
      Propanone                 1000              1.4,
                               1000       1       lu.b/
                               1000       1      1,4,5,7.51,52,58
      Propene                  fOOO        ](3)    57
      Propyne                  400        ]       4
     The TLV values  are from MCA  (1970)  and ACGIH  (1970) and the
     vapor  Irritant  grade  number  from Committee on Hazardous
     Materials  (1970).  The last  column  1n the table Identifies
     the pollutant with the Industry from which 1t was emitted;
     see Table  l-II  for Industry  Identification.

          The most toxic pollutants, up  to 20 ppm  TLV, are emitted
     predominantly from the processing Industries, namely:  the
     meat rendering,  fish  rendering, coffee roasting, tobacco
     curing, tanning, meat smokehouses,  zinc plating and also
     domestic Incinerators.  Two  solvent user Industries, the
     rubber products  and adheslves,  also emit pollutants in this
     TLV range.   Several other  pollutants of lower toxldty (tr1
     ethylamlne,  benzene,  1,2-d1chlorobenzene, ammonia and 1-
     pentanol)  are also emitted from the process Industry.  The
     pollutants  of lesser  toxldty,  I.e., with TLV larger than
     20 ppm, are  predominantly  emitted from the solvent user
     industries;  namely, the surface coating, degreaslng, dry
     cleaning,  graphic  arts, etc.

     The most toxic pollutants are emitted by industries
which, on the per 1,000-population basis for the control  dis-
trict, have the lowest emission rates.   However, the local
areas under adverse meteorological conditions can experience
toxic effects.

     A comparison of the TLV values with the Federal standard
of 0.24 ppm for hydrocarbons indicates  a very large safety
factor for the Federal standard in regard to toxicity.  The
hydrocarbon standard is based solely on the relationship of
nonmethane hydrocarbon concentration to the resulting photo-
chemical oxidant levels.  Correlation of the 6 to 9 AM non-
methane hydrocarbon levels with the peak oxidant levels show
an approximate three to one relationship.  Thus, to maintain
a standard of 0.08 ppm oxidant, the nonmethane hydrocarbon
concentration should not exceed 0.24 ppm.

Odor Effects

     Psychological and Physiological Effects

     Odors may affect humans in different ways depending not
only upon the characteristics of  the odor, but also on the
individual and his environment.   Odors, per se, are not the
cause of organic diseases or toxic effects.  Some highly
toxic gases,  such  as hydrogen sulfide,  are associated with
highly offensive  odors, but the dangerous properties do not
come from the odor Itself.  In this'case, the odor  serves as
a warning of  the presence of a harmful  gas.  Conversely, some
toxic gases may be odorless or have pleasant odors.

     Odors can, however, aggravate attacks of asthma and
other allergic conditions which cause  loss of appetite,
lowered water consumption, breathing difficulties,  nausea,
vomiting and  insomia (Sullivan, 1969).  Asthmatic attacks can
be triggered  by odors from hay, mustard,  animal skins, fer-
mentations, paint, sulfur and cooking  (Horesh,  1966).  Mal-
odorous gases are  harmless to the  average Individual, but
these gases can set off an allergic attack in sensitized
Sfol;c 5nrfK  (  66) re!1ewed the importance  of the role
of odors 1n the management of allergic  diseases and reported
that odorous  irritants are significant  factors  in asthma
hay  fever, eczema, migraine, rhinitis,  allergic crouDtra-
cheitis, dermatitis and gastrointestinal   al erSJ   AddUlon
al substances incriminated as aggravating a?"ergy-  Add^1on.
Include odors from turpentine, dimethyl su

           Ranking  of  Pollutants  According  to  Odor  Response

           Odors  may be ranked on the  basis  of detection  and
      threshold concentration and recognition  threshold  concen-
      tration.   The latter method is  used  by odor  investigators
      more frequently  and values  for  more  odor compounds  are
      readily available.  The more odorous  pollutants  emitted
      from small  sources are listed in Table 2-IV  according  to
      increasing recognition threshold concentration,  or decreas-
      ing odor response.  The values  for this  table were obtained
      from Copley (1970) and Leonardos (1969).

           The ranking of odorous emissions by recognition thres-
      hold concentration provides only a partial  measure of  the
      objectionable nature or perceived effect on  human  subjects.
      The odor quality has significant bearing on  the  subjective
      response or odor perception.  Psychological  differences  in
      exposed humans tend to obscure  this  differentiation, how-
      ever, since what is objectionable to one individual may
      elicit a different response from another.

           Other properties of odorous emissions  have  been used
      for ranking or differentiation.  Moncrlef (1968) attempted
      to differentiate odors by subject acceptability.  Certain
      compounds were generally unacceptable:  mercaptans, sulfides
      and amines.   Individual differences  tended  to limit the
      ranking of other compounds.

           The human response to odor sensations  may be quantita-
      tively expressed by the Weber-Fechner law which  states that

                              I - k In c

      where I 1s the Intensity of the odorant, k  is a  constant and
      c  1s the odorant concentration.  The usual  method of utilizing
      this expression  is by human panel testing of dilution ratios.
      The constant  1n  the Weber-Fechner expression is  a measure of
      the pervasiveness  of the odorant (Sullivan,  1969).  Extremely
      pervasive odors  are those whose odor Intensity remains un-
      changed or shows only slight change (small  k values) on di-
      lution.  The  intensity  of  an odor 1s generally expressed In
      'odor units'  which gives the number of  dilutions required to
      reduce the odor  to its  recognition threshold concentration

           Additional  factors which limit the characterization of
      odors by human response are olfactory fatigue and odorant
      mixtures.  Olfactory fatigue 1s a well  recognized phenomenon
      by which odor perception Is lessened or in some instances
      completely inhibited by prolonged exposure.

Table  2-IV
Odor Recoqni zabili ty

Pol lutant
Hydrogen sulflde
Butanolc acid
Ni trobenzene
Dimethyl sulfide
Dipropyl sulfide
Me thane thiol
Methyl ami ne
Carbon disulfide
Pentyl acetate
Dimethyl amine
Sulfur dioxide
Butyl acetate
1 -Butanol
Ethanoic acid
Nitrogen dioxide
Isobutyl acetate
Gasol ine
Propeneni trile
Propyl acetate
Mineral spirits

q u a 1 1 ty
shoe polish

rotten egg
new bread
burnt, sweet



-  -

~- -


-- -
1 to 4
1 to 20
10 to 25
3 to 60

51 ,52
51 ,52
51 ,55

 _ M

4,51 ,55,57,61
1 4
1 , "



     Table  2-IV   (continued)

      Ethyl  acetate

      Isopropyl acetate
      Tri chloroethene

      Methyl acetate
      1,1,1-Trichloroe thane

150 to



          Odor intensities, as measured by human panel response,
     vary widely when mixed odors are used.  Rosen, et.al.  (1962)
     investigated the odor perception of mixed odorants  and found
     that intensities of mixed systems may exhibit a variety  of
     interactions.   Intensities may be additive, counteractive,
     synerqistic, or independent.  Data on odor interactions  of
     butanole p-cresol  and pyridine are shown in Table 2-V.

          The most malodorous pollutants are emitted from  the
     process industries, primarily the meat and fish rendering.
     Many of the odorous pollutants at the top of the  list  are
     also at the top of the list  in Table  2-III for their  toxicity,
     although odor is not the cause of toxicity.  Except for
     ammonia, methanol, propanone, and methanol, the less mal-
     odorous pollutants are emitted from solvent user  industries.
     Under  adverse meteorological conditions, local areas near
     rendering, coffee  roasting,  tobacco curing, tanning and  meat
     smokehouse operations can experience  periods of malodors.

          A comparison  of the recognition  threshold concentrations
     in  the upper part  of Table 2-IV with  the Federal  standard
     of  0,24 ppm for hydrocarbons indicates that standard  is  not
     restrictive enough, assuming the term hydrocarbon refers to
     all organic pollutants.  A fair number of the pollutants have
     recognition threshold concentrations  considerably below  0.24
     ppm.  As mentioned above, the Federal standard is based  solely

on health  effects  and does  not take odor into consideration.
The A1r Quality Criteria  for Hydrocarbons (NAPCA Publication
AP-64,  March  1970)  concludes that,  hydrocarbons, per se,  do
not exhibit adverse health  effects  at the usually occurring
ambient air concentration levels.
Table  2-V   Odor Interaction in Mixtures
           Fraction  of Odor Threshold Concentration  at
           Which  Odor was  Perceived
rn  ftal


                additive threshold concentrations
     Odor  Control

     Concentration  Reduction Methods  - Odor control  is a
term which nay be  used  to describe any process  which makes
olfactory  experiences  more acceptable to people.   This can
be done  by reducing the odor intensity or concentration, or
by odor  modification.   Concentration  reduction  by sorption
devices  is the main theme of this  handbook (Chapters 4 and
6) with  additional  attention, in decreasing order of impor-
tance,   also  given  to  catalytic and thermal Incineration

      (Chapters  5  and  7).   Concentration  reduction by dilution
      with  Increased air  volumes  1s  practiced  by many Industries,
      but 1t  1s  not a  satisfactory  solution  to  the problem.
      dor Mod_;_f_ 1 ca11 on - It 1s possible to nullify  the  per-
,.,	1  response to a malodorous pollutant by the admixture
of another odorous vapor Into the atmosphere.  This  1s  com-
monly done by spraying aerosols to control, or mask,  mal-
odors coming from cooking, from densely occupied  living
quarters, bathrooms, locker rooms, etc.  Odor modification
1s also  being tried on a larger scale 1n 1~J--'---   --  '
metal decorating operations; see Chapter 1
Control  for Graphic Arts Industry.
                                               Industry, as 1n
                                               " under Emission

           There  1s  no  generally accepted physiological mechanism
      to  explain  odor modification.   It has been shown that nasal
      Inspiration of malodors with masking odors does not prevent
      the person  from smelling  the malodor upon being transferred
      to  a  new  environment;  hence, the mechanism cannot be attri-
      buted to  olfactory anesthesia  (AIHA, 1960).

           The  method has some  apparent advantages 1n Industrial
      applications associated primarily with low Initial equip-
      ment  costs.  It does not  counteract toxic, corrosive and
      smog-forming potential of the  pollutants  and 1s therefore
      limited 1n  Its application.

      Toxic Effects  of  Pollutants on  Vegetation

           The  only  (nonsmog) organic pollutant capable of causing
      plant damage at low concentrations (less  than 1 ppm) 1s
      ethene.   Other organic pollutants must be present at much
      higher concentrations  to  produce equivalent plant damage.
      Concentrations of the  order of  magnitude  required are highly
      Improbable  from the small emission sources.

           Exposure  to  0.1 ppm  of ethene produces eplnasty (down-
      ward  bending of leaves and other plant parts).  Abeles and
      Gahagan (1968) found that 0.1  ppm of ethene produced partial
      abscission  (shedding of plant  parts).  This plant response
      was also  studied  with  other hydrocarbons.  These findings
      are combined with those of other Investigators (Crocker,
      1935;  Zimmerman,  1935; Heck and P1res, 1962 and Burg and
      Berg,  1967), and  are presented  In Table 2-VI.  These are the
      most  active  hydrocarbons  and,  as the results show, they must
      be  present  1n  concentrations of two or more magnitudes higher-
      than  ethene  to produce an equivalent effect.

Table 2-VI  Relative Concentrations  of the  More  Reactive
            Hydrocarbons That Produce  Biological  Response
            Similar to That Produced by Ethene
                                        Inhibition of growth
Hydrocarbon     Abscission   Eplnasty   Pea stem     Tobacco

Ethene                 1            1            1          1
Propene               60          500          100        100
Propyne            1,250          500        2,800        100
1-Butene         100,000+    500,000      270,000      2,000
1,3-Butad1ene    100,000+        -    5,000,000        ---
     The 1norqan1c addle pollutants S02t H2S and N02 have
the capacity to damage vegetation.  Concentrations higher
than 0.1 ppm of the sulfur compounds can cause Injury, while
slightly higher concentrations of NOg are required.  Sulfur
dioxide 1s emitted from the rubber products and sulflte paper
Industries and domestic Incinerators.  Hydrogen sulflde 1s
emitted from the rendering operations and N02 from commercial
Incinerators.  The contribution from the small emission
sources compared to the total emission of these pollutants
from other sources (power plants, motor vehicles, etc) 1s
very small.  Local effects may occur under unfavorable
meteorological conditions.

     Synerglstlc effects of mixed air pollutants  can often
result  1n Increased plant damage over that observed  for the
single  pollutants.  Examples  of such synerglsm are given  by
H1ndaw1  (1971).

     Smog products, primarily ozone  and peroxyacetyl nitrate
(PAN) are principal contributors to  economic  crop loss 1n
such affected  areas as Southern California.

     Excellent descriptions  of the  types  and  manifestations
of vegetative  response to air polluants are  given 1n the
Handbook  of  Effects Assessment (Lacasse and  Moroz,  1969).

Damage to Materials

     Observable corrosive action occurs on unprotected
ferrous metals with S02 at concentrations less than 0.1  ppm.
Nonferrous metals corrode at hiqher concentrations.  In  both
cases the corrosion rate increases with increased concentra-
tion, temperature and humidity.  The presence of particulates
is also a corrosion rate Increasing factor.

     Paper and leather are oxidized to sulfates and some
textiles are weakened.  Sulfur dioxide and hydronen sulfide
discolor painted surfaces.  Dyes are faded by nitrogen di-

     Clothinq can be soiled by the pervasive odor emitted
from rendering plants.  Otherwise, organic pollutants are
not considered as particularly damaging to materials.

     The contribution from the small emission sources com-
pared to the emission of these pollutants from other sources
(power plants, motor vehicles, etc) is very small.  Their
effects will be felt locally, near the emission sources
under adverse meteorological conditions.

Smog Forming Potential of Pollutants

     Nature of Smoo

     A haze or mist, which appears to be a combination of
smoke and fog, 1s produced photochemlcally when certain  or-
ganic vapors in the atmosphere are exposed to solar radiation
in the presence of nitrogen oxides.  The products of the
Photochemical reactions  have been identified as oxldant
(ozone),  nitrogen dioxide, peroxyacetyl  nitrate (PAN), peroxv
     ytra   (PBzN>  metha"a'. Particulates. aerosol,
     Hoic                       .             .        ,  n
     H2S04 mist.  These substances produce eye irritation, plant
     damage, reduced visibility, material damage, Injury to healfh
     and a characteristic odor.  The H2S04 mis? is evidence thJt *
     S02 is also a contributor to smog formation.

          These effects are not produced by the primary contam<
     nants, but appear after the photochemical oJidaJ'on h
     place.  All pollutants are not equally active and also
     duce different smog effects.  For example, the eye  rr
     smog is not the same as the plant toxicant  as evidlnrn
     the fact that the half-life of the two IffectI arl d?55
     Therefore, it Is important to rate po?lStants accord nj
     their  role as smog precursors as  well  as  their direct 1r+
     or nuisance effects as discussed  in the  previous  text.      nt

     Distribution of Smog Forming Pollutants

     Smog producing pollutants are ever-present where the
population densities are high because the main source is
motor vehicles.  Tab-le 2-VII designates the type, source
and rate of the smog active pollutant emissions as calculated
from the California (1969) survey results.   The estimated
emission rates of active organic and NOX pollutants from
motor vehicles are 360 and 200 Ib/da per 1,000 people, re-
spectively, and by far the largest emissions from any source.
Heavy smog can always form where the meteorological condi-
tions are unfavorable for the dispersion of pollutants.
Where appreciable air mass movements occur, the smog thins
out over a larger region.  Ultimate disposal of the smog
can be by a slow absorption into vegetation and terrain  or
by the rapid scrubbing action of rain.

Table 2-VII  Sources of Smog Forming Pollutants of High
             and Low Activities

                          Emission rate,
Type and              Ib/da per 1,000 people     Percentage
source of              HighU)       Low         of organic
pollutants             activity    activity      pollutants

Organic from             360                         46
  motor vehicles                     160             20

Organic from
  solvent users           35                          4
  and fuel                           110             14

Organic from              20                          3
  other stationary                   100             13

Organic totals           415         370             100

NOX from
  motor vehicles         200         ---             ---

NOx from
  stationary              95         ---             ...

(1) High  activity  refers  to  all  olefins, substituted
    aromatics  and  aldehydes.

          Th(*  rate of smog active pollutant emission from the
     solvent user industries is low compared to that from motor
     vehicles, being only 35 lb/da per 1,000 people.  The effects
     will also be felt regionally rather than locally because:
     (1) a time  lapse occurs during which the organic pollutant
     mixes in with NOx from other sources,  (2) the photochemical
     reactions are slow, allowing the air mass movements to carry
     the mixtures over a large region before the reactions are
     completed,  and  (3) emissions produced during nondaylight
     hours are dispersed and then react during the daylight hours.

          A typical polluted atmosphere may have NOx and total
     organic pollutant concentrations of the order of 0.5 ppm
     for each.   Los Angeles (in 1970) defined a day of photo-
     chemical smog as one in which the average oxidant level is
     0,10 ppm for 1 hour.

          Active Smog Precursors

          The smog forming potential exhibited by organic pol-
     lutants has been related to the pollutant chemical structure
     by the investigators 1n the Los Angeles County pollution con-
     trol districts.  Those classified as highly active are all
     of the olefins, substituted aromatics, aldehydes and ketones
     having a branched hydrocarbon structure.  This method of
     classification is the basis for the separation of high ac-
     tivity and  low activity groups in Table 2-VII.  The percentage
     of high activity pollutants emitted from motor vehicles (only)
     1s 69%, while from the solvent users and other stationary
     sources the percentages are 24 and 17, respectively.

          Of the 167 pollutants listed in Table C-I (Appendix C),
     41 can be classified as being highly active smog precursors.
     These are listed in Table 2-VIII with a separation made for
     those emitted from the solvent user industries and those
     from the process industries.  The emissions from the process
     industries  are small, and hence do not make a significant
     contribution to smog formation.  Those from the solvent
     user industries, and including fuel evaporation, contribute
     about 8% of all the active organic pollutants.  Of the 20
     solvents listed, toluene, xylene and trichloroethene are
     emitted at  the highest rates; hence, these three solvents
     are the main contributors to smog from the small emission
     sources,  There is a sizable number of lower activity smoq
     precursors which make a lesser contribution.

          Ranking of Smog Precursors

          There  are several reactivity classification systems
     derived from different sets of data and experimental condi-
     tions.  Of  these, a reactivity scale based primarily on

product yields and biological  effects was  proposed by Alt-
shuller (1966) and Levy (1970).   Altshuller took into con-
sideration the parameters:   ozone or oxidant yield, peroxy-
acetyl nitrate (PAN) yield, methanal yield, aerosol forma-
tion, eye irritation and plant damage.  Levy included the
parameters:  nitrogen dioxide  concentration, oxidant concen-
tration, eye irritation, and methanal concentration.  These
were combined into a single system which rated pollutants
from 0 to 3 in increasing order of activity.  This system
was further modified by MSA Research Corporation by expand-
ing the numerical span to 0 to 10.
Table 2-VIII
Pollutants From Small  Emission Sources That
Are Active Smog Precursors
Pollutants from
user industries
fuel marketing
Butyl benzene

Isopropyl benzene
Propyl benzene

Ethyl benzene

                   Process industries
                   Difurfuryl ether

                   Propenoic  nitrile


           The  expanded  rating  system was used  to rank pollutants
      from  small  emission  sources;  the  ranking  is presented in
      Table 2-IX.   Only  a  limited number of  the pollutants could
      be  ranked because  of lack of  complete  data (data from Levy,
      1970).  The most active pollutants are predominantly the
      ketones having  a branched hydrocarbon  structure, olefins,
      substituted aromatics, and aldehydes,  in  agreement with the
      classification  system used by  the California investigators.
      Three of  the  most  abundantly  emitted solvents, xylene, tri-
      chloroethene  and toluene, are  in  the upper part of the rank-
      ing list.   In the  degreasing  industry, 1,1,1-trichloroethane
      with  a relative reactivity of  one is promoted as a substitute
      for trichloroethene  with  a relative reactivity of 6.  Benzene
      is  quite  nonreactive; substitutions producing xylene and tol
      uene  greatly  increase the reactivity.

      Effect of Smog  on  the Environment

           The  effect of smog on the environment is observed as
      reduction in  visibility,  eye  irritation and damage to vege-
      tation.   Eye  irritation and vegetation damage are two cri-
      teria for determining the pollutant reactivity as smog pre-
      cursors used  in Table 2-IX ranking.  Different pollutants do
      not produce the same type of smog; some smogs cause more eye
      irritation while others cause  more plant  damage.  Table 2-X
      lists  the classes  of organic pollutants in order of decreas-
      ing activity  to produce eye irritant or vegetation damaging

           The  most active precursors producing eye irritant and
      vegetation damaging  smogs  are  xylene,  4-methyl-3-penten-2-
      one,  1-pentene, 1-butene  and propene emitted mostly from the
      solvent user  industries and fuel  storage  evaporation.  01-
      olefins such  as 2,6-dimethyl-2,5-heptadien-4-one, butadiene,
      1,3-pentadiene  and 2,4-dimethyl-l,3-pentadienes are only
      highly reactive in producing eye  irritant smog.  Most of the
      latter are  also emitted from solvent user industries.  The
      largest contributor, also volume-wise, is xylene.

           Epidemiological studies  conducted primarily in Los
      Angeles and San Francisco have attempted  to ascertain the
      existence of  correlations of  respiratory  illnesses, eye
      irritation, and other health effects with oxidant level.
      These  have  been reviewed  in NAPCA Publication No. AP-63
      (March 1970).   The general conclusions are that the thres-
      hold  for  observable  health effects is  about 0.15 ppm oxi-
      dant.  The  highest correlation of effect  with oxidant con-
      centration  is found  for eye irritation.   Excess mortality,
      hospital  admissions, and  respiratory cancer in the general
      public have not shown significant correlation with oxidant

Table 2-IX  Pollutant Reactivity Ranking As  Smog Precursors

Pollutant vapor	       0 to 10 scale   Industry	

4-Methyl-3-penten-2-one         8         1,4
1-Butene                        7         10
Propene                         7         57
Xylene                          7         1,4,5,6,7,8
1-Pentene                       7         10
4-Methyl-2-pentanone            6
Trichloroethene                 6
1 ,3-Butadiene                   6
Toluene                         6
  pentanone                     6         4,7

2,6-Dimethyl-4-heptanone        6         1,4,58
2-Ethoxy-ethanol                6         4,5,7
  4-one                         6         4,7
1-Methyl-l ,4-epoxybutane        6         4,51
Isopropyl benzene               6         1,10

Ethyl benzene                   6         1
Ethene                          5         
Butyl benzene                   4         1
Cyclohexane                     4         1.4,7
Cyclohexanone                   3         14,6,7

Butanone                        3         1,4,5,6,7,8,51
Methanol                        3         1,4,7,55,57
2-Propanol                      3         1,4,7,9,51
Butanol                         3         1,4,7
Diethylamine                    3         51

Triethylamine                   3         51,52
Propanone                       2         1,4,5,7,51,61
Ethyl  acetate                    2         1,4,7,8
Butyl  acetate                    2         1,4,7
Ethanol                          2          1,4,6,10 ,51

 Benzene                          1          1.4,6,7.51
 Nltropropane                     1          1.4,7
 1.11 ,-Trlchloroethane           1          2,4
 Pentane                          0          4,6,10,51
 Propane                          0          10,57

     Table  2-IX   (continued)

     Pollutant vapor	 0  to  10 scale    Industry

     Propyne                          0          4
     2-Ethoxy-ethyl acetate           0          1.4,5,8
     Ethanedlol                       0          4,51
     Phenyl acetate                   0          1,4,7
        trlfluoroethane                0          2,3
      Table  2-X   Ranking  of Organic Pollutants Producing Eye
                 Irritant and Vegetation Damaging Smogs
                 (Data  from Altshuller. 1966)

                                  Response on 0 to 10 scale
      Compound or                     Eye         Vegetation
      compound class               Irritation       damage

      Internally  double-
        bonded oleflns               4 to 8           10

      Dlalkyl and trlalkyl
        benzenes                     4 to 8 -        5 to 10

      1-Alkenes                      4 to 8         6 to 8
      D1olef1ns                        10              0
      Ethene                            5              + IU

      Toluene and other
        monoalkyl benzenes              4.  .        0 to.3.
      Aliphatic aldehydes              +vU           IVU
      Paraffins                         0              0
      Benzene                           0              0
      Acetylene                         0              0

      (1)  Effect  noted  experimentally, but  data
          Insufficient  to quantltate.

          Barth, et.al.  (1971)  report a study which  Indicates  the
      eve  Irritation  threshold  as  being about 0.1 ppm oxldant.
      time to cause barely perceptible eye  Irritation, based  on
      human  panel testing of  synthetic smog, 1s  four  minutes.

Pollutant Regulatory Laws

     Federal  Standards

     The purpose of the pollutant regulatory  laws  1s  to  con-
trol emission sources so that ambient  air in  inhabited  areas
is not injurious to health,  vegetation,  or property,  nor de-
tracts from enjoyment of life.   Acceptable concentration levels
that permit attainment of  the above requirements  have been es-
tablished by the Environmental  Protection Agency  (EPA,  April
30, 1971) for six types or groups of pollutants;  namely, car-
bon monoxide, sulfur oxides, nitrogen  dioxide,  photochemical
oxidants, hydrocarbons, and  particulates.  The  established
concentration levels are given  in Table  3-1.   For carbon
monoxide the established level  is of the order  of 10  ppm,
while for the other gaseous  pollutants 1t 1s  of the order of
0.1 ppm.

     The primary standards,  in Table 3-1, are the pollutant-
concentration levels considered to be protective  of human
health, with a safety margin.  These standards  are to be
attained in three years starting April 30, 1971.   The sec-
ondary standards are more restrictive in regard to sulfur
dioxide and partlculate concentrations and are to be attained
in  27 additional months.  These standards are, however, sub-
ject to changes as more knowledge becomes available on health
and welfare effects.

     States were required  to  adopt  and submit plans to the
EPA by Jan. 30, 1972, which will maintain and enforce the
Federal ambient air  quality  standards 1n  each air quality-
control region.  The regions have  been  established by the
EPA and were determined on  the basis  of  population,  climate,
topography,  and geography.

      Control District  Regulations

      The  ambient air quality for  a  given  control  district
depends  on:  (1) the amount and  type  of  pollutants emitted
from  the  combinations  of  emission  sources, and   (2)  the  pre-
vailing  meteorological  conditions,  topography, and geography.
Since  the  latter vary,  sometimes  considerably, from  one
district  to  another, the  restrictions to be  placed on  the
pollutant-emitting Industries  should  also vary to attain the
desired  ambient air quality without being unnecessarily re-
strictive.   It is  also apparent  that, for a  given set  of
meteorological  conditions,  Increasingly more restrictive
 controls  would be  required  on  each source as:   (1) the number
 of sources increase,  (2) the  number  and quantity of reactive-
 pollutant emissions Increase.

           Many  states,  counties  and  cities have their own air-
      pollution  laws  in  addition  to the recent Federal regulations.
      These local  regulations vary to meet the local requirements,*
      although many are  patterned after those of Los Angeles County
      and  San  Francisco  Bay Area.  Any person with a pollution
      problem  should  check the  local  regulations first.

           Of  specific interest in the control of pollutants from
      small  emission  sources are  Rule 66 of Los Angeles County
      and  Regulation  3 of San Francisco Bay Area.  These two rules
      or regulations  are similar  to each other and are concerned
      with  control of solvent emissions to reduce smog, which is a
      particularly difficult-to-correct problem in these two con-
      trol  districts.  However, the smog problem is already pres-
      ent  in most high population-density areas; hence, the essence
      of Rule 66 and  Regulation 3 is  applicable elsewhere.

           To indicate the general context that regulations in
      general may have,  pertinent portions of Rule 66 are quoted
      below  as taken  from Rules and Regulations, Los Angeles
      County Air Pollution Control District, June 18, 1970.

           Rule 66a - A  person shall not discharge more than 15
      pounds of organic materials into the atmosphere 1n any one
      day from any article, machine, equipment or other contrivance
      in which any organic solvent or any material  containing or-
      ganic solvent comes into contact with flame or is baked,
      heat-cured or heat-polymerized, in the presence of oxygen,
      unless all organic materials discharged from such article,
      machine, equipment or other contrivance have been reduced
      either by at least 85 per cent overall  or to not more than
      15 pounds in any one day,

           Rule 66b - A person shall not discharge more than 40
      pounds of organic material into the atmosphere in any one
      day from any article, machine, equipment or other contrivance
      used  under conditions other than described in section (a),
      for employing, applying, evaporating or drying any photo-
      chemically reactive solvent, as defined in section (k), or
      material containing such solvent, unless all  organic mater-
      ials discharged from such article, machine, equipment or
      other contrivance have been reduced either by at least 85
      percent overall or to not more than 40 pounds 1n any one day.

           Rule 66k - For the purposes of this rule, a photo-
      chemical ly reactive solvent 1s any solvent with an aggregate
      of more than 20 percent of Its total volume composed of the
      chemical compounds classified below or which exceeds any of
      the following individual percentage composition limitations.

referred to the total  volume  of  solvent:

     1.  A combination of hydrocarbons,  alcohols,  alde-
         hydes, esters, ethers or  ketones  having  an  ole-
         finic or cyclo-olefinic type  of unsaturation:
         5 per cent;

     2.  A combination of aromatic compounds  with  eight
         or more carbon atoms to the molecule except
         ethylbenzene:  8 per cent;

     3.  A combination of ethylbenzene,  ketones  having
         branched hydrocarbon structures,  trichloro-
         ethylene or  toluene:  20  per  cent.

     Whenever any orqanic solvent  or  any constituent of  an
orqanic solvent may be classified  from its chemical  structure
into more than one of the above  groups of  organic compounds,
it shall be considered as a member of  the  most reactive  chem-
ical group, that is,  that group  having the least allowable
per cent of the total volume of  solvents.

     Regulation 3 does not include trichloroethylene or  ke-
tones having branched hydrocarbon structures.

      3.   Research and Development Recommendations


     Under the provisions of Section  103  of  the  Clean  Air
Act of 1970, the Environmental  Protection Agency is  required
to establish a national  research and  development program for
the prevention and control  of air pollution.   As part  of such
a program they shall  "conduct,  and promote the coordination
and acceleration of,  research,  investigations, experiments,
training, demonstrations, surveys and studies  relating to  the
causes, effects, extent, prevention and control  of air pollu-
tion".  The achievement  of these broad requirements  dictates
that survey studies such as this package  sorption device
study provide problem-solving research and development pro-
grams to  aid in future planning for governmental action.

     A research and development program has  been developed
based on  the results  of  a thorough review and  evaluation of
the sources and emissions,  as discussed in Chapter 1 of Part
I, and control technology relevant to application of package
sorption  devices, as  discussed  in Chapters 4  and 6 in  Part II
of this report.  Assessment of  these  results  in  terms  of pres-
ent and future technology needs, priorities,  and costs has
resulted  in the formulation of  specific and  relevant research
and development recommendations.

Sources and Emissions

     The  results of the  review  of sources emitting noxious,
toxic or  smog forming pollutants in the range  of flow  rates
amenable  to control by package  devices showed  that the majority
of significant emission  sources could be  assigned to one of  two
broad categories:  (1) solvent  users  and   (2)  chemical, thermal,
or biological processing industries.   Data for characterization
of the individual sources within these two broad categories  are
far from  complete, but,  in terms of those parameters necessary
to the selection and  design of  package control systems, the
solvent user sources  and emission are better defined and under-
stood than are the sources in the second category.

     Specific problem areas within the solvent user  category
that have been identified are primarily related  to  those sources
which emit dirty, wet, low solvent concentration gas streams,
such as paint spray booths, or  which emit pollutants resulting
from thermal treatment applied at some point  in the  process.

     Most solvent emissions may be readily characterized as to
composition by a knowledge of the composition and physical
properties of the original solvent or solvent mixture  used.

     It is  primarily when  chemical  changes  occur  through  thermal
     processing,  such as  baking  of  coatings,  printed  products, etc,
     that emission  characterization is  lacking.   The  thermal  treat-
     ment causes  partial  decomposition  of resins,  drying  oils and
     also of high boiling  solvents  to  the extent  that the effluent
     emitted can  not be  predicted or characterized on the basis of
     raw material input  to the process.  Such solvent process
     emissions  bear similarities to emissions from sources  in the
     second broad category of chemical  or thermal  process industries,
     such as rendering of  meat,  poultry, and  seafood  wastes and by-
     products;  canning and other food  processing;  and coffee  roast-
     ing.  The  emissions  are  undesirable by-products  of the process
     and are complex mixtures which vary widely according to  the
     type and conditions  of processing.  To make  the  best choice of
     type and design of  control  system  to be  applied, an  analysis
     of effluent  mixture  is necessary  to determine the molecular
     weight range of the  gaseous components and type  of particulates,

     Industry Trends

          Certain industry trends are  also  apparent from  the
     review of  pollutant  sources and emissions.   These may  be
     categorized  as  (1)  solvent changes or reformulation,  (2)
     phase-out  of traditional processes, and   (3)  increase  in
     application  of control systems.

          Solvent usage  throughout  the  country has been markedly
     affected by  the Los  Angeles County Rule  66 and its widespread
     incorporation  into  other local and regional  legislation.  The
     result is  a  growing  trend toward  decreasing  or eliminating
     the active photochemical smog  precursors from all types  of
     solvent usage.  New,  low reactivity solvent  mixtures have
     been developed which  simulate  the  desirable  solvent  properties
     of previous  reactive  systems.   Branched  aromatics and  unsat-
     urates are being eliminated or replaced. Where  feasible,
     water base coatings  are  substituted for  solvent  systems.  New
     coatings and printing processes are being developed  around
     solventless  systems  and  nonthermal curing operations.

          The indictment  against trichloroethene  as a smog  pre-
     cursor has resulted  in widespread  use  of alternative de-
     greasing and metal  cleaning solvents,  particularly inhib-
     ited 1,1,1-trichloroethane. The  expense and difficulty  in
     controlling  degreasing emissions  and the availability  of
     substitute solvents  will shortly  result  in the complete
     elimination  of trichloroethene as  a vapor degreasing solvent.

          Certain traditional processes and industries are  being
     nhased-out or  are undergoing significant process modifications.
     Notable among  these  are  the widespread elimination of  open

burning or backyard incineration of wastes,  the replacement
of meat smokehouses by chemical  treatments,  and reduction of
leather tanning.

     Many metropolitan areas have put into effect bans  or
limitations on open burning of household,  municipal  and
industrial wastes, with significant reductions  of total
emissions of organic pollutants.  These small  sources are not
amenable to control and thus complete elimination with  sub-
stitution of alternative waste disposal methods is the  only
real solution.

     The impact of the accelerating trend  toward implemen-
tation of existing control technology is difficult to assess
in quantitative terms, particularly as it  relates to the
small emission sources.  Some indication may be derived from
the organic emission reduction from all stationary sources
achieved by the stringent controls in effect in the  major
metropolitan areas of California.  According to the  1969
inventory of emissions for the San Francisco Bay Area,  the
organic emissions from stationary sources  were only  about 47
percent of the potential or uncontrolled emissions.   Since
many of the organic emission sources under control are  within
the purview of this small source study, it may be seen  that
implementation of existing control technology can effect
significant reduction in emissions.

     The proposed R&D programs are so oriented that they do
not cover pollution problems that are being corrected by
forces or trends already in effect.

Carbon Resorb Systems, Presejrt Technology

     The results of the review and evaluation of existing
commerically available package sorption devices  for control
of  emissions from  small sources  show that for the areas of
air purification  (pollutant concentrations of about 1  ppm
and lower) and solvent recovery  (recoverable solvent con-
centration above about 700  ppm), existing control technology
is  generally adequate and  control devices are available.   Of
the devices used  in these  two areas, the  technology which
forms  the basis  for the solvent  recovery  type  systems  is
more amenable to  general  gaseous air pollution  control  than
the air  purification  technology.   In order  to  apply  the  sol-
vent recovery type  systems  to the  broader application  of  air
pollution control,  modifications are required  to the systems
and  to  their  operation.   To  indicate  the  broader applica-
tion,  these type  systems  are  referred  to  in  this report as
the  carbon  resorb  systems.

          The  carbon resorb systems operate on a cyclic pattern
     consisting  of  an  adsorption phase and regeneration phase.
     During  the  adsorption phase the gas stream, containing the
     pollutant or solvent vapor, is passed through the (granular)
     activated carbon  bed, usually, until some small amount of
     the  vapors  start  to penetrate the bed.  The gas flow is then
     stopped or  diverted to another previously regenerated bed.
     During  regeneration, a gaseous regenerating agent at elevated
     temperature is  passed through the bed to desorb the pollutant
     or  solvent  and  restore the carbon to its previous capacity
     and  permit  its  reuse in  the system.

          In solvent recovery operations, low temperature steam
     is  used as  regenerating  agent because of several desirable
     properties. During regeneration a  portion of the steam con-*
     denses  in the  carbon bed and releases considerable heat at
     212F,  a  temperature at  which a great majority of pollutants
     of  interest desorb readily and a temperature which causes
     very little or  no damage to the desorbing vapor or the car-
     bon. The steam and solvent, being  condensable, can be readily
     separated from  each other by decantation or distillation.     3

          Carbon resorb systems can be classified into three type$
     depending on the  manner  in which the carbon bed is handled
     during  adsorption and regeneration.  The most commonly used
     system  utilizes two fixed beds which are alternated between
     the  adsorption  and regeneration phases of the cycle.  The
     dual fixed  bed  systems are the simplest of the systems but
     are  also  the bulkiest.   The other two systems, the moving b^d
     and  fluidized  bed, move  the carbon  countercurrent to the gas
     phases  in both  the adsorption and desorption phases of the
     cycle.  The countercurrent flows increase the carbon and
     usage efficiency, resulting in operating cost reductions,     t
     disadvantages  are the more complex  and more costly construction
     costs of  the moving bed  type systems and the carbon attrition
     losses  in the  fluidized  bed systems.

          Carbon resorb systems have been used in solvent recovery
     under conditions  which are favorable for profitable recovery
     of  the  solvent.   Conditions considered favorable are:  (1)
     clean airstream devoid of participates or containing partlc*
     ulates  that are easily filtered out,  (2) virtually no vapor*
     in  the  airstream  other than that of the solvent being recovSk
     ed,   (3)  airstream at low temperature (130F or lower),  (4\T
     low  relative humidity (below 80%) and  (5) (most important)
     vapor concentration over 700 ppm but not over 25% of lower
     explosive limit.

          Carbon manufacturers and solvent recovery system design
     engineers have  the technical knowledge to work out problems

associated with recovery of solvents  from relatively clean
airstreams having high vapor concentrations.   Solvent-laden
airstreams having these qualities  are common  in a number of
the solvent user industries; namely,  degreasing, dry cleaning,
most surface coating, flexographic and gravure printing,
rubber products and paper coating.  The main  problem is the
coordination of efforts with process  machinery designers and
plant layout men to attain the best integrated system and
thereby the most economical control of air pollution from
these emission sources.

Carbon Resorb Systems. New Technology

     When conditions and types of  pollutant emissions are not
applicable to solvent recovery type operations, the situation
becomes more complex and, also, competing control systems
such as catalytic and thermal incinerators must be brought into
consideration.  When the pollutant has no recovery value but
is combustible and is emitted at concentrations higher than
700 ppm, the incinerator systems are  more economical to oper-
ate for air pollution control than the carbon resorb systems.
For a noncombustible pollutant the control is of course limited
to the carbon resorb or absorption type systems.  At pollutant
concentrations below 700 ppm the new technology on carbon re-
sorb systems is in competitive position with the incinerator
systems for pollution control.

     At progressively lower concentrations the use of steam
as regenerating agent becomes increasingly less suitable
while noncondensable inert gas or  air at 200 to 300F, on
ba.sis of theory, shows economic advantages.  At low concen-
trations, large amounts of steam are required to maintain a
practical operating capacity for the carbon and it becomes
impractical to separate the pollutant from the large volume
of steam condensate.  Use of steam would convert the air
pollutants to water pollutants.

     With noncondensable gas regeneration, several econom-
ically feasible procedures have been worked out on sound
engineering principles for the carbon regeneration and sep-
aration or disposal of desorbed pollutant vapors.  When the
pollutant vapor is combustible and has no recovery value, it
becomes economical to dispose of it by catalytic or thermal
incineration.  In this application the incinerator need be
about one-tenth of the size that would have been required
to incinerate the air initially.  If it is necessary or de-
sirable to recover the desorbed pollutant vapor for reuse
(solvent) or disposal by other means, it can be done by
readsorption on a secondary carbon bed or by condensation

     The secondary adsorber would be small  relative to the primary
     adsorber and could be readily regenerated by the conventional
     steam method.  When the pollutant vapor is readily condensable,
     it can be partially condensed from the hot regenerating gas
     by cooling, with the cool  regenerating  gas being recycled
     through the (other) on-stream carbon bed.  These conceptual
     designs can be incorporated into the fixed bed, moving bed  or
     fluidized bed carbon resorb systems.

          Economic analyses show that the various above proposed
     carbon resorb systems are definitely more economical  to operate
     than catalytic or thermal incinerators of equivalent  air hand-
     ling capacities for air at near 77F initial temperature.   At
     elevated airstream temperatures the incinerators become in-
     creasingly more economical to operate  and could equal the
     carbon resorb systems in  operating costs.  There is  therefore
     a definite area in the air pollution control scheme where  the
     new carbon resorb systems technology should be seriously con-

     Adsorption Theories

          Two theories related to adsorption dynamics have been
     advanced which can greatly facilitate  development of  the carbon
     resorb systems new technology for pollution control.   (Chapter

          One of these is the  Polyani  potential  theory which can
     be used to calculate the  adsorptive capacities of carbons
     for various homologous series of molecules  as function of
     temperature, molecular size and concentration.

          The other theory involves mass transfer of vapor from
     the airstream to  carbon and into the pores  of the carbon.
     On the basis of mass transfer theories Klotz developed an
     equation which can be used to calculate the mass transfer  zone
     or adsorption zone length in the bed.   The  controlling para-
     meters are:   air  density  and viscosity, carbon particle size,
     airflow velocity, diffusion coefficient of  vapor in air, and
     ratio of vapor influent concentration  over  vapor concentration
     within the bed.

          The effective use of either equation requires minimal
     data on adsorptive capacities and adsorption rates on the
     particular carbon and homologous series of  the pollutant or
     pollutants under  consideration.  Data  that  would be useful
     for this purpose, particularly from the problem areas of
     pollution control, are lacking or are  not in a form suitable
     for use by the engineer that would do  the development and/or
     design of the control  system.

Monitoring Instruments

     In solvent recovery operations  the  instrumentation re-
quirements for monitoring are not complex.   The  solvent,  the
rate of emission, and concentration  are  known  to a  fair degree
of certainty permitting the installation of optimum size  and
design of carbon resorb system for the particular emission
source.  The adsorption-regeneration cycles are  preset and
automatically controlled at short and equal time intervals.
A comparison of solvent recovery with the solvent input wi11
give a measure of the efficiency of the recovery operations.
Simple instruments such as a manometer and temperature re-
corders are used to determine the time for particulate filter
change and to check the air inlet temperature  and temperature
profiles during adsorption and regeneration.

     In the problem areas of pollutant control,  the emissions
can be quite variable in quantity and type.  The adsorption
phase  time can therefore vary considerably.  Monitoring in-
struments are highly desirable to avoid overrun or large
underrun on the adsorption phase.  The instruments for this
purpose should be simple and durable  in construction,  low
cost,  and readily understood and operated  by plant operating

     Chapter 8 describes techniques  and instruments that  show
suitability for monitoring of pollutant vapor concentrations.
Some have capabilities  for identifying  pollutant vapors.
Much of the available  instrumentation are  too expensive and
also require  technically trained  personnel  for  their  oper-

Research  and  Development Programs

     Specific  programs  resulting  from the  identification  of
needs  and priorities  are described  below.   Study priority is
based  on  three  relative classifications:   high;  medium; low.
Estimated cost  is  based on  a  nominal  $50,000  per man-year
unless unusually  high  materials  or  construction  costs  are in-
volved.   For  the  latter cases,  estimated  costs  of  such mater-
 ials  or  construction  are  added.

      The  recommended  programs are divided  into  four  areas of

      Emission and Source Characterization
      Evaluation of New Adsorbent Systems
      Generation of Adsorption and Regeneration  Data
      Monitoring Instrument Development

          Emission  and  Source Characterization

          Objective  - To  provide detailed chemical and engineering
    data  necessary  for the selection and design of optimally effec-
    tive  control systems.

          Approach  - Pollutant sources recognized as being partic-
    ularly  difficult to  control due to low (below 700 ppm range)
    pollutant  concentration, highly odorous nature, presence of
    aerosols or particulates, and presence of difficultly adsorbed
    species would  be investigated in detail.  Specific tasks v/ould
    include problem emission source selections, field testing of
    emissions, engineering evaluation of overall process with
    determination  of process flows and point sources, and process
    modification required to make the most effective use of a
    potential  control system.

          The data obtained on some sources may clearly define the
    best  choice of system, i.e., catalytic incineration, thermal
    incineration,  liquid absorption, or carbon resorb.  Further
    study to determine the best suited type and optimum size of
    system may be minimal and done with readily available infor-
    mation.  It is foreseeable that commercially available carbon
    resorb systems may be suitable for some applications.  If the
    emission and source  characterization data do not indicate a
    clear choice of competing system and also leaves best choice
    of carbon  resorb system in doubt, further investigation as
    prescribed under the program to Evaluate New Adsorbent Systems
    is required.

         Priority, Duration, Effort and Cost - This program was
    assigned a high priority rating since characterization of the
    emission source is absolutely essential before a logical pro-
    cedure can be planned for its control.

         The number of point sources of emissions is large with
    new ones appearing as new products and processes are developed
    and old ones phased  out as they become obsolete.  It is fore-
    seeable that many sources will be quite similar, and the
    characterization of  one will essentially characterize the
    others.  This pattern will reduce the number of investigations
    that need  to be made to a limited number of the more critical
    ones.  It  is difficult to assess with any certainty the number
    of critical sources  existing.  The difficulty of character-
    ization also varies, but with a well equipped laboratory and
    trained personnel the cost per source can vary from $5,000 to
    $15,000 each.  The characterization would be carried only to
    the degree required  for control system selection and to estab-
    lish design, size, and operating procedure.  It is quite
    certain that a minimum of 10 sources would require character-
    ization and it appears improbable that the number would rise
    much above 30.

     Ten sources would require about 3,800 man hours at a
cost of $100,000 to be performed in less than one year.

     A more extended program of systematic characterizations
and including possible new sources would increase cost and
time.  At some point in time the need for further studies
would diminish and if 30 sources is arbitrarily set as max-
imum the man hour requirement would be 11,400 to be performed
over a five year period at a cost of $300,000.

     Evaluation of New Carbon Resorb System

     Objective - To evaluate, technically and economically,
potentially promising or incompletely proven carbon resorb
systems and to demonstrate feasibility of the selected system
for control of pollutants emitted from specific sources of
the type studied under Emissions and Source Characterization.

     Approach - On the basis of data obtained from the
Emission and Source Characterization study or from other
sources, the first step of the program would be to determine
whether a carbon resorb system should be used, and, if so,
which type.  Systems that are of foreseeable interest are
briefly described as follows:

     Dual stationary beds for adsorption phase, regen-
         eration with hot noncondensable gas, and dis-
         posal of desorbed vapors by catalytic or
         thermal incineration.

     Dual stationary beds for adsorption phase, regen-
         eration with hot noncondensable gas, readsorp-
         tion of desorbed vapors in small secondary
         adsorbed and ultimate recovery of vapors by
         conventional steam regeneration.

     Dual stationary beds for adsorption phase, regen-
         eration with hot noncondensable gas, recovery
         of vapors from regeneration gas by condensation.
     Alternatives for each of the above three systems
         would be the use of moving carbon bed or
         fluidized carbon bed.

     Polar adsorbent systems (for low molecular weight
         polar vapor).

          In order to make the choice of system,  some  fundamental
     adsorption studies may be required (laboratory scale)  to
     obtain additional data for the calculation of adsorptive
     capacities, adsorption zone lengths,  carbon  bed sizes,  service
     times, regenerabi1ity, and regeneration  gas  requirements.   On
     the basis of all of the acquired information, several  models
     of carbon resorb systems would be designed that appear  techni-
     cally and economically attractive.  From the theoretical  per-
     formance of the best model, when compared to performances  of
     incinerators, a decision is then made whether to  continue
     further to demonstrate feasibility of the system  under  test
     condi tions.

          If a system demonstration is indicated, the  second step
     would be the construction or assembly of a pilot  sized  system,
     which would be tested in the contractors pilot plant  or under
     actual conditions in industry.  The choice of location  would
     be determined by several factors.  If the operating conditions
     can be reproduced in a pilot plant, the  pilot plant would  be
     the preferred location since the tests can be better  con-
     trolled to yield data that can be treated theoretically.
     Spray painting operations would be an example of  this  type.
     A pilot plant location would be preferred when the study
     involves vapors in the explosive concentration range,  as  would
     be the case of emissions from gasoline storage.   When  the  vapor
     emission conditions cannot be reproduced in  the pilot  plant, as
     would the case with pollutants emitted from  a rendering plant,
     the preferred location is the plant of the industry emitting
     the pollutant.

          Priority, Duration, Effort and Cost - This program was
     assigned a high priority rating since it involves  the  basic
     objectives of the package sorption study.

          About 5,000 man hours over a one year operating  time
     would be required to test a control system.   The  estimated
     cost would vary between $135,000 and  $145,000 when equipment
     costs vary between $10,000 and $20,000.   Equipment costs  at
     this level would limit test control systems  to airflow
     capacities of less than 2,000 ftVmin.

          The size of the program depends  on  the  number of  system
     tests that are deemed necessary to get adequate information
     on feasibility of carbon resorb type  systems for  pollution
     control.   Four industries that should receive top  priority
     for control  study are:  meat rendering,  fish rendering, spray
     coating and petroleum distillate evaporation.  For a  limited
     study (million dollar total program level),  these  four  should
     receive first consideration.   It may  become  necessary  to  test
     more than one design of carbon resorb system at each  of the

emission sites, in which  case  more  than  four  systems  would  be
tested.  For a four system study  at $140,000  each,  the  total
cost would be $560,000,   The systems can be  tested  simultan-
eously or extended over  four years.

     Depending on the results  of  the first  four  systems
studied, the program may  be discontinued or  extended  to  in-
clude other emission sources such as waste  water treatment,
soap manufacture, restaurants, baking,  plywood manufacture,
azo-dye manufacture, crematoria,  hospital  exhausts,  etc.   The
program will ultimately  reach a point of diminishing  returns
where the results of previous tests are  translatable  to  new
situations and no further testing is economically justifiable.
The point at which further testing is not justifiable is
difficult to determine,  but it is believed  that  at the  level
where control systems for 12 or 13 emission sources  of  varied
types have been tested the point of diminishing  returns  has
been reached.

     For 13 systems at $140,000 each, the total  cost would be

     Generation of Adsorption and Regeneration Data

     Objectives - To generate adsorption and regeneration data
in forms readily usable by the carbon resorb systems design
engi neer.

     Approach  - The recommended  research program to obtain
and compile  these data would  consist of laboratory scale
studies conducted under gas flow rates  of the order of 0.1
to 0.4  cubic  feet per minute  and carbon bed diameters of
one-half to  one  inch.  These  conditions are sufficient to
give the necessary  data without  excessive expenditures in
time or equipment.

     Due to  the  large number  of  pertinent parameters and their
permutations  to  be  investigated  in  this proposed program,  care-
ful utilization  of  statistical experimental design procedures
would  be employed.  The pertinent  variables and  suggested
levels  of  experimentation  for each  variable are  shown below:

            Variable                       Levels
    Carbon  type                  2  -  significantly  different
                                    pore  structures  (source
                                    of  carbon)

         Pollutant type

         Pollutant concentration

         Gas velocity

         Bed depth

         Mesh size


3 levels for each homologous
series or chemical  group
classification of interest

3 - 10, 100 and 1000 ppm

3 - 20, 60 and 100  ft/min

3 - longer than adsorption

3 - 4 to 6, 6 to 10 and 10
    to 30 mesh

3 - 25, 50 and 100C

Spot checked only,  little
expected effect
     Regeneration runs with a noncondensable gas  (nitrogen)  would
     be conducted at two temperatures on carbon beds  spent with
     various pollutants at the varied concentrations  as  indicated
     above.  Also repeated cycles will  be carried out.   About
     forty regenerations are foreseeable.

          The product of the study would be tables,  graphs,  equa-
     tions and procedures for calculating adsorptive  capacities,
     operating capacities, adsorption zone lengths,  and  regenerating
     agent requirements.  These results would be  further utilized
     for estimating carbon bed dimensions, service time, adsorption-
     regeneration cycle timing and frequency and  for  determining
     effective disposal techniques for  the pollutant.

          Priority, Duration. Effort and Cost - This  program was
     assigned a medium priority rating  since the  information it
     would yield would be used primarily to expedite  the progress
     in the program to Evaluate New Carbon Resorb Systems.  In the
     latter program adsorption and regeneration data, if required,
     would be experimentally determined specifically  for each
     pollutant emission source.  This procedure would be less
     efficient, more costly, and produce Information  that would  be
     limited in regard to general application than if a  systematic
     study were conducted as proposed here.

     To attain a fairly  complete  coverage  of  parameters,
homologous series of compounds, and  carbons,  the  program  would
require 15,200 man hours of  effort over  a  3-yr  period.  Con-
siderable instrumentation is required  since some  adsorption
phases of the experiments would be run continuously  over  ex-
tended periods with continuous monitoring  of  the  effluent gas
stream.  The instrument  and  other equipment costs come  to
$20,000.  The total cost would be $400,000.

     Some phases of the  program are  more important than others
and the study of some homologous  series  of compounds are  more
pertinent to problem areas in pollutant  control than others.
On the basis of priorities in this  respect,  the three year
overall program can be divided  into  three  one-year programs.
The first year's program, however,  would be  the least produc-
tive since it includes a considerable  lag  time  for the  pur-
chase and installation of the instruments, adsorption and
regeneration systems.  It includes  16  weeks  equipment instal-
lation period followed by 36 weeks  of  experimental work and
final report preparation.  The nan  hours would  be 4,500 at a
cost of $130,000,  including the  $20,000  for instruments and
equipment.  A two-year program would be  an expansion of the
one-year  program with considerably larger portion of the  esti-
mated 9,800 man  hours assigned to experimental  work.  The cost,
including instruments and other equipment, would be $265,000.
     Monitoring  Instrument Development
     Objective - To develop effective and relatively inexpen-
 sive methods  for monitoring the operation of package sorbent

     Approach - Three  promising areas of monitoring instrumen-
 tation  research and development have been preliminarily explored
 and  recommended for further study.  These are:

     1.   Indicator Tubes  - Development of chemical indicator
   methods  for detecting  breakthrough or monitoring of service
   life  of  carbon sorption  systems.  This area of  research and
   development would rely  heavily on the extensive work already
   done  in  the development  and  application of selective and
   specific indicator  tubes for a wide variety of  both organic
   and  inorganic air contaminants.

     2.  Sorption  and  Incineration  Devices  - Development  of
   monitoring  devices  applicable  to  organic  contaminant  control
   by either carbon  sorption or incineration systems,  based  on
   the  principle of  the catalytic CO indicator,  widely used
   in vehicular  tunnels and other areas  having  significant
   potential for  hazardous CO  concentration  build-up.   The

       catalyst  used  in  these  indicators, Hopcalite, is an
       effective  oxidation  catalyst  for  not only CO but hydro-
       carbons.   Thus, the  presence  of hydrocarbons in the
       control system exhaust  could  be detected by a temperature
       increase  of  an adiabatically  operated Hopcalite catalyst
       bed.   A steadily  increasing temperature difference be-
       tween  the  Hopcalite  bed  and the effluent airstream would
       indicate  pollutant penetration.

          3.  Stack-Gas  Samplers - Development of a simple
       stack  sampling device by which, periodically, the control
       device users could collect samples to be quantitatively
       analyzed  at  a  centrally  located analytical laboratory.
       Such a device  could  be  similar to those used by industrial
       hygienists in  collecting work room air samples for sub-
       sequent analysis, which  use a small carbon canister and
       hand pump  to collect and retain air contaminants.  Alter-
       nately, such sampling canisters could be integrally in-
       stalled as part of the  control system and would be
       periodically removed and sent for analysis.

          Priority, Duration,  Effort and Cost - This program was
     assigned a  medium priority since the need for the relatively
     inexpensive  monitoring methods would not arise until operational
     control  systems  have been  developed, i.e., results of the pro-
     gram  to  Evaluate New Carbon Resorb Systems.

          This study  is expected to  require 12,500 man hours over
     a  2-yr period at a cost of $375,000.
         The contract specified the development of two 5-year
    programs, Plan A and Plan B, based on assumed total ex-
    penditures of $1 million and $5 million, respectively.  The
    various programs as described above have been grouped into
    Plan A of $1 million total expenditure and Plan B of $3
    million total expenditure as shown in Tables 3-1 and -II,
    respectively.  An expanded plan of $5 million was not fore-

   Table 3-1   Summary of One Million Dollar Research and Development Plan  (A)
Emissions and Sources
 (30 sources at $10,000

Evaluation of New
 Adsorbent Systems
 (4 systems at $140,000

Generation of Adsorption
 and Regeneration Data


                                                Time schedule, yr


_L_    2     3      4      5     Priority

	^>     high
                              medi urn


Table 3-II  Summary of Three Million Dollar Research and Development Plan (B)
Emissions and Sources
(30 sources at $10,000
Evaluation of New Adsorbent
(13 systems at about
$140,000 each)
Generation of Adsorption
and Regeneration Data
Monitoring Instrument
Time schedule, yr
12345 Priority
	 -v hi nh

	 > medium

                  Part II
Package Sorptlon Device Technology Handbook

Introduction to Part II

     This part of the report presents  the  results  of  the
investigations and analyses of the  packaged  systems that  are
or could be used for control of air pollutants  from those
emission sources discussed in Chapters  1 and 2.  Although
this study was primarily concerned  with sorbent  systems,  the
competing systems - catalytic and  thermal  incinerators  -  are
also discussed in order to place each  system in  proper  per.
spective in regard to its effectiveness and  operating cost.

     Chapter 6. Package Sorption Systems,  is the most Import*
ant chapter of Part II.  In it are  discussed  (1)  the present
sorbent systems technology,  (2) the new developments and
proposed systems that show promise  for  pollutant control  and
(3) the operating costs of the various  systems.  Under  present
systems technology, systems are considered that  are In  wide
commercial usage or have been proven over  a  number of years
of successful  operation.  Under new technology,  systems are
considered that have appeared very  recently  in  commercial
usage but have not yet acquired wide spread  acceptance  or
systems that have been engineered  on theoretical basis  but
have not been  tested or tested on  a very limited pilot  plant
     Since one of the important objectives  of  the  study  was
to assess potential  new technology  to  control  pollution,
Chapter 6 was written against a theoretical  background.   A
theoretical  understanding of the adsorption  and  regeneration
processes permits a  ready extension of the  operational know-
ledge on known systems toward the design of  new systems for
control of pollutants emitted under new and  more difficult
circumstances.  Chapter 4 was written  to supply  this  1nfor.
mation in considerable detail.   Chapter 5 is a considerably
more condensed review of catalytic  combustion  theory.

     Chapter 7 presents a brief .study  of catalytic  Inciner-
ation to augment the more extensive study of sorbent  systems

     When the new systems become(commercial, the need for new
testing and control  instrumentation is foreseeable.   Chapter 8
describes basic concepts suitable for  pollutant  identification
and concentration measurements.

         4,   Sorbent Types  and  Sorption  Theories

Sorbent Types

     Important Sorbent Properties

     Sorbents differ considerably  from each  other  in  their
selectivity, or preferential  adsorption, and their adsorptive
capacities.   These two factors  can vary  the  suitability of  a
sorbent for  air pollution control  usage, limiting  some  for
special conditions, while others are  generally applicable.
The selectivity and capacity are dependent on four properties
of the sorbent.  These are:

     1,  Chemical constitution  of  the sorbent, partic-
         ularly the surface,

     2,  Extent of the surface  area,

     3.  Pore size distribution,

     4.  Particle, or granule,  size.

     The term "sorbent" includes two types of solids (1) those
that concentrate vapors on the  free surface by "adsorption",
and  (2) those that concentrate vapors by "absorption", wherein
the vapor molecules penetrate into the mass of the sorbent
solid.  Penetration of hydrogen into palladium metal lattice
is an  example of the  latter.  At present, the absorption pro-
cess in solids (to differentiate it from  liquid absorption)
does not appear  to play any role in the control of pollutant
emissions of  the type discussed in Chapters  1 and 2.

     On the  basis  of  chemical constitution  and pore size dis-
tribution,  adsorbents can  be grouped  into four different types.
They exhibit  greatly  different  selectivities  for  vapors and
differences  in manner of adsorption.  The types are:

     1.  Chemically  reactive adsorbents,

     2.  Polar adsorbents*

     3.  Nonpolar  adsorbents,

     4.  Molecular sieves.

     The differences  in  adsorptive behavior of the  first
 three  are primarily  due  to differences  in chemical  constitu-
 tion  of  the solids,  with secondary differences contributed by

     extent of surface area and pore-size distribution.  The molec-
     ular sieves can be either polar or nonpolar but their dis-
     tinctive adsorptive behavior is due to the small size of the
     pores which physically block entry, of oversized molecules.

          Chemically Reactive Adsorbents

          When adsorption of vapors is accompanied by a chemical
     reaction, the process is defined as "chemisorption".  It has
     a  number of well defined characteristics which are enumerated

          1.   It is highly selective; not all vapors react
               with all adsorbents.

          2.   It tends to be Irreversible.

          3.   The rate Increases with rise in temperature.

          4.   The heat of chemisorption is of the same magnitude
               as that of chemical reactions, usually over
               10 kcal/mole.

          5.   The process is exothermic.

          6.   Chemisorption stops when all of the active sites
               on the surface have reacted, or the surface is
               covered with a unimolecular layer of chemisorbed

          Examples of chemisorption, which have been intensively
     studied by various investigators, are hydrogen on heated
     metals, oxygen on heated tungsten, carbon monoxide on iron
     catalysts, nitrogen on nickel and iron, and ammonia on alumi-
     na  (Langmuir, 1912 and 1916; Emmett, 1937; Beeck, 1950; Webb,
     1957).  Chemisorption of oxygen on carbon is well known.  At
     higher temperatures, it is one of the steps in carbon acti-
     vation; i.e., H20 vapor and C02 react with carbon to form an
     oxygen surface complex which later breaks up by releasing CO.

          Inorganic acid vapors react readily with transition metal
     oxides.  When the active metal oxides are used to catalyze
     pollutant incineration, chemisorption of add vapors plays an
     adverse role by causing catalyst poisoning.  Conversely, chem-
     isorption on reactive metals is a potential process for purify.
     ing air of traces of halogen acids and $03.

          Soda lime by itself, or in combination with activated car-
     bon is used to chemisorb vapors such as ethanoic acid, ethane-
     nitrile, propenenitrlle, 3-chloro-propene, ethenyl propyl
     dlsulflde, chlorine, cyanogen, fluorine, hydrogen bromide,

hydrogen chloride,  hydrogen  fluoride,  and  hydrogen  sulfide.
Several  of these vapors  have been  identified  with  the  rubber
products and rendering industries.
     In the chemically reactive adsorbent
impregnated polar and nonpolar adsorbents
pregnant reacts chemically with the  vapor
initial physical adsorption has occurred.
impregnant-adsorbent combinations  and the
which the impregnant is reactive.
    oroup are also the
     wherein the im-
    molecule after an
     Table 4-1 listed
    vapor molecules to
Table 4-1  Types of Impregnated Adsorbents

          Impregnant-adsorbent        Pollutant vapor
          Phosphoric acid-carbon
          Potassium permanganate-
            activated .alumina
          Cupric oxide-carbon

          Cupric oxide-pentavalent
            chromium on carbon
          G-32E-treated activated
          Ammonia-molecular sieves
             alumina (8.4 wt
           Sulfur acid-carbon
             (20 wt *)
mercury, radioact. iodine
low molecular wt
organic acids

carbonyl chloride
cyanogen chloride
hydrogen cyanide
.carbonyl chloride

sulfur  compounds
hydrogen sulfide
carbonyl sulfide
mercaptans  from
hydrocarbon gases

carbon  dioxide
hydrogen sulfide
.carbon  disulfide

 sulfur dioxide

 thiophenol and
 similar compounds
    (1) Marbon Chemical  Division, Washington, West Virginia
    (2) Chemetron Corporation, Catalyst Division,
        Louisville, Kentucky

          Polar Adsorbents

          When adsorption is not accompanied by chemical bonding,
      the active forces are referred to as van der Waal's forces
      and are the same as those that bind liquids and solids. This
      type of adsorption is also referred to as physical adsorption
      and occurs on chemically nonreactive adsorbents having polar
      or nonpolar surface properties.

          Physical adsorption varies from chemisorption 1n several
      respects In that:

          1.   It is less selective.

          2.   It is reversible.

          3.   The heat of adsorption is of the order of 10 kcal/

          4.   Adsorption can continue beyond the unimolecular
               layer surface coverage.

          5.   The amount adsorbed decreases with increasing
               temperature at a fixed vapor pressure.

          Examples of polar adsorbents are silica gel, activated
      alumina, fullers earth, bentonite, and the zeolite type of
      molecular sieves.  Of these, fullers earth and bentonite are
      usually used for liquid phase adsorption processes; the other
      three  are used in gas separation and purification processes.

          The chemical make-up of these adsorbents consists of
      silicon oxide and/or aluminum oxide groupings which impart a
      highly polar character to the surface.  Silica gel is essen-
      tially pure S10? with less than 0.3% of other metallic oxide
      impurities.  Activated alumina is over 9Q% pure A^Oj with
      other  metallic oxides as impurities making up the balance.
      Zeolites are alumino-silicate crystalline structures contain-
      ing sodium, potassium or calcium cations.

          The polarity of the surface has been found to produce
      strong affinity to polar vapor molecules or vapor molecules
      that are readily polarized, such as olefins. Table 4-II lists
      a number of air pollutants in decreasing order of affinity on
      silica gel (Trappe, 1940) which shows the effect of physical
      properties of the pollutants, particularly  the dielectric
      constant and dipole moment, on the adsorption affinity. Ex-
      cept for a few variances, the affinity decreases with de-
      creasing dielectric constant and decreasing dipole moment.
      In contrast, the affinity tends to decrease with Increasing
      V  and van der Waals' constant a_;  for definition of van
      der Waals1 a,  see Appendix D.   The latter trend

Table 4-II Effect of Pollutant
Order of

Pol lutant
Ethyl acetate
Ethyl ether
Tri chloromethane
Di chloromethane
Tetrachlorome thane
ve Affinity
Physical Properties on
for silica
Gel  Decreasing
Affinity Downward
at 67F

van der
a x 103

contradicts the concept of active van der Waals1  forces  when
dealing with polar adsorbents.   Adsorptive affinity on non-
polar adsorbents
sfib'wn in the next
with van der Waals1 a.as will be
     The strong affinity of polar adsorbents for water great-
ly limits their general use in air pollution control,   they
are used for drying of various gases; some are listed  below:
          Carbon dioxide
Sulfur dioxide
     When the gas stream is free of water vapor, polar ad-
sorbents are suitable for pollution control involving small
polar molecules such as:
          Hydrogen chloride
                Sulfur dioxide
                Hydrogen sulfide

      In each case  the pollutant  can be desorbed with a relatively
      small amount  of water,  after which the polar adsorbent can
      be dried at elevated  temperature and put back Into use.

          Nonpolar Adsorbents

          Activated carbon 1s  the primary example of a nonpolar
      adsorbent.  Others are  possible  such as metallic surfaces,
      but  the surface areas attainable with the latter are too
      small to be useful for  gas  or air purification.

          Physical adsorption  on nonpolar and polar adsorbents
      differ primarily in selectivity, or differences in affinity,
      for  different molecules,  Qji nonpolar adsorbents  adso-rptiste
      af f 1 n 1 ty increases with molecular size and yjn der WaaJ.5 '
      cjjis_iant_j^."*'"In'"TabYe"4"-ni tins Is shoM *w1 tTTTeveTPhydro-
      carbons, which are listed downward in increasing order of
      adsorptive affinity on  activated carbon  (based on data of
      Ray, 1949).   The correlation of Vm with  amount adsorbed  has
      been worked in with the Polanyi potential theory to derive
      an equation for calculating the amount of adsorption of  var-
      ious pollutants at varied vapor pressures and temperatures.
      This equation is discussed  under the theoretical part of
      this chapter.

          By impregnating  activated carbons,  polar adsorbent  and
      chemically active adsorbent properties can be imparted to
     Table 4-III  Effect of Pollutant Physical Properties on
                  Adsorptive Affinity for Activated Carbon
                  aFlOO"Ft Increasing Order  of Affinity Downward

                  Amount adsorbed
                  at one atmosphere       Vm,     van der Waals
     Pollutant    pressure. cmMiq./q cnrliq./mol    a x 103

     them.  Activated carbons generally contain ash, such  as  S10*,
     AlgOa, Fe203, NaOH, KOH and chemisorbed oxygen, all  of  which

impart polar properties.  By leaching out the ash followed
by hydrogen reduction, a wholly nonpolar carbon can be ob-
tained.  Adsorption of water vapor at low vapor pressures is
very sensitive to the presence of polar impurities.  Acti-
vated carbons that have not been specially treated are poor
adsorbents for water and can be desorbed by most of the or-
ganic pollutants emitted from the small sources.

     Molecular Sieves

     Molecular sieves are available having polar or nonpolar
surface properties.  The polar molecular sieves are complex
crystalline structures consisting of basic units or tetra-
hedra of (AT, Si)04.  The silicon atom is electrically bal-
anced by the four oxygen atoms of the lattice, but the alumi-
num atom, with a valence of three, is not.  For every alumi-
num atom in the lattice, a monovalent (or equivalent) cation
is required to attain electrical balance.  The cations used
are Na, K, and Ca.

     When first manufactured, the crystalline structure con-
tains water of crystallization, which is removed by heating
to temperatures above 570F.  The removal of the water leaves
a porous solid consisting of a network of cavities inter-
connected by round pores of smaller diameter.  The cavity and
pore diameters are uniform; hence, a pore-size distribution
in the usual sense as applied to silica gel, activated alumina
and activated carbon has no significance in molecular sieves.

     The manufacturers of polar molecular sieves (Linde.
Davison) designate over four different types of sieves.  A
group A has 11.4 A cavity diameter, a 1 to 1 Al to Si ratio
and is designated as simple cubic crystalline structure.  A
group X has a 1.0 to 1.25 or 1.5 Al to Si ratio and is desig-
nated as body-centered cubic crystalline structure.  By
changes in Al to Si ratio and type of cation, the pore dia-
meter can be varied over the increments 3, 4, 5, 8 and 9 A.
In each case  only molecules having critical dimensions small-
er  than the nominal pore size can be adsorbed.  Table 4-IV
lists the types of Linde molecular sieves and their general
chemical formulas; those of Davison are quite similar.  Table
4-V also gives the nominal pore diameters and lists some of
the pollutants that can be selectively adsorbed by each type
of molecular sieve, as determined by the critical dimensions
of the molecule.  For a given molecular sieve, those molecules
listed under sieves of larger pore diameter are screened or


     Table 4-IV  Types of Molecular Sieves

     Type    	General formula, molar ratios	

       3A     0.72  Na20 : 0.24 K20  : 1.00 A1203 : 1.92 S102 : 2H20

       4A     0.960.04Na20 : 1.00  A1203  : 1.92  0.09 Si02 : xH20

       5A     0.72  Na20 : 0.24 CaO  : 1.00 A1203 : 1.92 S10 : 2H20

     10X     0.62  N20  : 0.21 CaO  : 1.00 A1203 : 2.48 S102 : xH20

     13X     0.830.05Na20 : 1.00  A1203  : 2.48  0.03 S102 : xH20
           Superimposed  on  the sieving action of these adsorbents
      is  the  strong  affinity for small or highly polar molecules;
      see  discussion  under  Polar Adsorbents.  As was the case with
      silica  gel  and  activated alumina, the strong affinity of
      molecular sieves for  water greatly limits their uses for
      pollution control.  In a dry  alrstream, the molecular sieves
      are  more effective for the adsorption of low molecular weight
      polar or unsaturated  pollutants at low  concentrations than
      silica  gel,  activated alumina or activated carbon.  Results
      presented below, as reported by Gustafson (1961), show this
      difference  in  the  adsorption of ethene  and ethyne at 77F
      and  20  mm of Hg pressure.

                              Amount adsorbed. % by weight
               Gas          4Tsilica geTactivated carbon'

               Ethene        6        1.75              2.5
               Ethyne        7        1.5               1.8

      The  adsorption  of  low molecular weight  saturated hydrocarbons
      such  as methane and ethane are more strongly adsorbed by car-
      bons  than by molecular sieves, but too  poorly in any case to
      be  used effectively in pollution control.  Activated carbons
      will  have a  much lower equilibrium capacity for organic com-
      pounds  at low  partial pressures than 13X; but at high par-
      tial  pressures! activated carbon  adsorptlve capacity is two
      to  three times  that of 13X.  Molecular  sieves adsorb NOX
      gases at high  relative humidities with  about 5 wt % pick*
      up  at concentrations  in the 1,000 to 2,000 ppm range (Sun-
      daresan, 1967).

           Molecular  sieves are readily regenerated by passing
      water vapor  or  liquid water through the sieve bed and the ad-
      sorbed water.in turn, 1s readily removed by heating the bed
      to 570F.

Table 4-V Effect of Molecular Sieve Pore Diameter nn

selective Adsorption of Pollutan
Sieve type and Cri
pore diameter, Jl Pollutant DO!


[3A J Ethyne
3A J Carbon monoxide
4^ Water
[ Hydrogen sulflde
5A_ Ethene
5^ Ethane
Butane to
L C22H46
1 ,3-Butadiene
Isobutane to
1 s o C22^46
Tetrachlorome thane
t Vapors
tical diameter of
lutant molecule, A


     In recent years (since  1969)  developmental  work  has  been
done on carbon molecular sieves.   Because  of  nonpolarity  of
the surface, these adsorbents  would  have no greater  affinity
for small  polar molecules than for nonpolar molecules   Pre-
ferential  adsorption for small organic  molecules can  be  ex-
pected in  the presence of moisture.   Pittsburgh  Carbon Com-
pany, Division of Calgon Corporation, has  developed  a carbon
molecular  sieve which shows  very  little adsorption for tetra-
chloromethane but considerable adsorption  for Iodine   Th*
pore openings appear to be slit or rectangular shane* nf
critical dimension in the 5.0  to  5.5 X range       P

          Because of attractive forces from adjacent and opposite
     pore walls, such pollutants as ammonia, hydrogen sulfide and
     possibly methanethlol can be expected to be adsorbed more
     strongly than on conventional carbon.  As with polar molecu-
     lar sieves, the sieving action can be utilized for separation
     of molecules of differing sizes.

          Cooper (1970) describes a platinum-carbon catalyst having
     molecular-sieve properties.  The platinum is incorporated
     into the carbon structure during the molecular sieve prepara-
     tion.  With these catalytic-carbon sieves, hydrogenatlon of
     linear olefins can be carried out because of their penetra-
     tion into pores while branched olefins are not hydrogenated.

          Effect of Physical Properties on Adsorption

          Rhysical Properties - The chemical constitutions of the
     adsorbents, as discussed in the previous text, form the basis
     for classifying the adsorbents into three types: the chemi-
     cally reactive, polar and nonpolar.  A fourth type, decided
     primarily on pore structure, is the molecular sieve which 1n
     chemical constitution can also be classified polar or non-

          The physical properties such as surface area, pore size
     distribution, pore volume, granule or particle size and den-
     sity determine the adsorptlve capacity, adsorption rate and
     resistance to flow of the pollutant-air mixture through the
     adsorbent bed.  These properties are very frequently pre-
     sented in sales literature to indicate the quality of the
     adsorbent being sold.  Tables 4-VI and -VII Riresjnt ranges
     of most of these physicaT"|fn;FpliTtleT^^
     EjOBBA gels, acUvmd: aTMinas., ;4MH|ilZcirr^|^^d
     molecular sieves.  Comparative data for chemicaTTy" reactTve
     adsorbents were not available and would not have the same
     significance since chemisorptions are very specific.

          To further define the adsorptive qualifications, sales
     literature presents adsorption isotherms and/or adsorptive
     capacity data at varied conditions for various vapors for
     which the adsorbents are recommended.  For polar adsorbents
     and polar molecular sieves, the graphical presentation of
     the water adsorption isotherm or adsorptive capacities at
     varied relative vapor pressures are usually the first ad-
     sorptive capacity data given.  For activated carbons  the
     iodine number is most frequently given if the intended use
     is in aqueous phase.  When the intended use is in gas phase,
     it is the carbon tetrachloride (tetrachloromethane) activity.
     The carbon tetrachloride activity is the weight percentage
     adsorbed at 40F and 0.16 relative vapor pressure.  Details
     of this test are  given 1n Appendex D.

Table 4-VI  Physical Properties of Adsorbents. Surface Areas and Pore Volumes
                         Surface area
Silica gel
Activated alumina
Activated carbon
Molecular sieve
210 to 320
300 to 560
                                                            Pore volume



37 est




Table 4-VII  Physical Properties of Adsorbents, Densities and Mean Pore  Diameters
Silica gel
Activated alumina
Activated carbon
Molecular sieve
                   Bulk density, Particle density, True density,
                      g/cm3	g/cm3	g/cm3	
                   0.90 to 1.00
                   0.35 to 0.50
          1.5 to  1.7 est.
          0.60 to 0.80
                        2.3 to 3.3
                                Mean  pore
                                diameter, A

                                18 to 20,..
                                15 to 20<]'
                                3 to  9
(1) Carbon pore structure  consists  of  two  pore  size  ranges,  (1)  those  over  30  A
    diameter termed macrooores,  and (2)  those less  than  30 A  diameter  termed micro-
    pores.  The 15  to  20 A mean  diameters  are for  the  micropores  only;  95%  of  the
    surface area is associated with them.

          Surface Area - In the adsorption process, the pollutant
     molecules in the gas phase are attracted to the adsorbent
     surface initiating a unimolecular layer of molecules which
     can spread until complete surface coverage is achieved. This
     is definitely the case with chemisorption.  When adsorption
     occurs by van der Waals1 forces, multilayers of molecules
     can form after completion of the unimolecular layer, or
     sooner  if other factors are active.  Other factors encoun-
     tered in porous adsorbents are the attractions of the adjacent
     and opposite pore walls when pore diameters are small.

          The opposite-wall attraction, or effect, enhances the
     adsorption to the degree that molecular packing can occur 1n
     the pores, not unlike capillary condensation.  The stage at
     which molecular packing starts is dependent on (1) relative
     sizes of the adsorbed molecule and pore diameter, and (2)
     the mutual attraction of the molecules relative to the at-
     traction of the molecules to the adsorbent surface.  When
     the molecular diameter is larger or mutual attraction 1s
     strong, molecular packing starts early in the adsorption

          The extent of the surface area is then a sole measure
     of the adsorptive capacity only when the molecule is small,
     has strong affinity to the adsorbent such as water on polar
     adsorbents, and the relative vapor pressure is low.  At in-
     creased relative vapor pressures, multilayer and capillary
     condensations can still occur.  When the mutual attraction
     is greater than the surface attraction, the molecule is
     repelled as mercury from carbon.  When the mutual attraction
     is not too greatly different from the surface attraction,
     molecular packing and/or capillary condensation occurs read-
     ilyias with carbon tetrachloride and water on carbon.  In
     the latter case, the pore size distribution determines the
     adsorptive capacity of the adsorbent.

          The surface areas, as reported in Table 4-VI, are based
     on nitrogen adsorption data using the Brunauer, Emmett and
     Teller (BET) method (Brunauer 1938).  On the per-unit-we1ght
     basis, the surface areas of activated carbons are consider-
     ably larger than those of the polar adsorbents.  Activated
     carbons used for vapor adsorption range in area from 1100
     to 1600 m2/g, with all types ranging from about 600 to 1600
     m2/g.   Silica gels and activated alumina range from 800 ma/g
     downward to about 200 m2/g.

          Adsorbent Porosity - The pore size distribution, mean
     pore diameter, surface area and pore volume are parameters
     that define the pore structure.  In Tables 4-VI and -VII,
     the pore volume for each adsorbent was obtained by subtract.
     1ng  the nonporous, or Impervious volume from the volume of

the particles.   The impervious  and  particle  volumes  were  ob-
tained from the particle and  true densities  as  given.   The
mean pore diameters for silica  gel  and  activated  alumina
were calculated assuming a cylinder of  volume  and surface as
given.  The mean micropore diameter range  for  the activated
carbons was selected from pore  size distribution  curves
shown in Figures 4-1 and -2.

     For the silica gels, activated aluminas  and activated
carbons, the mean pore diameters are in the  15  to 22 A range.
Molecules of concern in air pollution range  in  diameter from
3.6 to over 8.7 A; a partial  list of pollutant  molecular
diameters is given in Table 4-V.  In view  of the  relatively
small difference betv/een the molecular and pore diameters,
the opposite-wall effects play  an important  role  in  the pol-
lutant vapor adsorption process.  Adsorptive capacity is
determined by the pore space occupied by the pollutant liquid
volume.  The probability of molecular packing  decreases with
increasing pore diameter; the largest pore diameters in which
liquid packing  can occur at low relative vapor pressures
vary  with  molecular diameter and0adsorptive affinity, but
would be in the range of 20 to 30 A.  Adsorption at high
relative0vapor  pressures proceed in the large pores (30 to
100,000 A) by multilayer adsorption and capillary condensa-
tion.  Most pollution control problems are, however, con-
cerned with emission rates at low  relative vapor pressures;
therefore, the  small-pore volume is the prime factor in
determining the adsorbent effectiveness.

      Activated  Carbon Pore Size Distribution - Because of
the prevalence  of  moisture in pollutantaTrstreams, the use
of  polar adsorbents  is  very limited in pollution control,
while  activated carbons  are successfully  used  in many  appli-
cations.   In further discussions on adsorbents,  the emphasis
is  on  activated carbons,  although  much of the  information
presented  will  also  be  applicable  to polar  adsorbents.

      In  Figures 4-1  and  -2, pore size  distribution  curves
are  presented  for  four  activated carbons  of greatly differ-
ent  adsorptive  capacities.  These  four curves  essentially
span  the gamut  of  distributions observed  in carbons.   For
discussion purposes,  the pores  less  than  30 A  diameter are
designated as  micropores and those that are larger, macro-
pores.   Dubinin (1936)  has  classified  pores below 20  A dia-
meter as micropores,  those  larger  than 200  A as  macropores
and those  between  20 and 200 A  as  transitional pores.

      Curve I  is for a  granular  Darco  carbon developed  by
Atlas Powder  Company for aqueous phase  adsorption  usaqe.
The open pore  structure (i.e.,  the large  macropore  volume)
 1s highly  desirable since it accelerates  the  liquid phase

   1.20 -





*  0.40 -
   0.20 -
              30  50   100
  500 1000
Pore diameter, A
   Figure  4-1   Pore Size Distribution Curves For Activated Carbons with
               Cumulative Pore Volume Calculated Per Unit Carbon Weight
               I  Darco Carbon, IV Superactivated Coconut, II Type BP
               Pittsburgh Activated Carbon, and III Type SXA Union Carbide

                              500  1000
                            Pore diameter, A

Figure 4-2  Pore Size Distribution Curves For Activated Carbons with
            Cumulative Pore Volume Calculated Per Unit Carbon Bulk
            Volume Basis,  I Darco Carbon, IV Superactivated Coconut,
            II  Type  BP Pittsburgh Activated Carbon, and III Type SXA
            Union Carbide  Corp.

      adsorption rate which is generally very slow.   The small
      micropore volume of 0.18 cm'/g greatly limits  the capacity
      of this carbon for vapor adsorption;  therefore, the carbon
      is not competitive with the other carbons  represented  in  the
      Figures 4-1  and -2.

           Curve IV represents the other extreme in  adsorptive
      capacity.  The micropore volume is 0.73 cm3/g  and is approach'
      ing the upper limit attainable.  This was  a highly active
      Barnebey Cheney coconut carbon that was further activated
      (superactivated) in an attempt to reach the ultimate in
      capacity.  The yield was very low, less than J^jof.starting
      material, which makes its cost of manufacture'e^cesVive."""'

           Curve II is for a type BP carbon manufactured by  the
      Pittsburgh Activated Carbon Company;  curve III  is for  a  type
      SXA manufactured by the Union Carbide Corporation.  These
      carbons occupy the middle of the capacity  range with micro-
      pore volumes of 0.42 and 0.46 cm*/g,  respectively.  The  BP
      carbon is used in considerable quantities  in military  gas
      masks and the SXA,in air purification and  solvent recovery
      systems.  These two distribution curves are characteristic
      of carbons economically best applicable to air  pollution

           The carbon tetrachloride activity test, which 1s  most
      frequently used to determine the suitability of a carbon  for
      vapor adsorption, measures the micropore volume.  Under  the
      test conditions the liquid volume of carbon tetrachloride
      adsorbed is  very nearly equal to the micropore  volume.
      Shown below  are comparative figures of carbon  tetrachloride
      test results and micropore volumes of the  four  carbons:

                   Carbon tetrachloride activity     Micropore
      Carbon            n/q        cm3 liq/g      volume. cm3/o
         I             0.30           0.20              0.18
        II             0.68           0.45              0.42
       III             0.73           0.48              0.46
        IV             1.20           0.79              0.73

      Since the micropore size distributions of different carbons
      vary, as illustrated in the Figures 4-1 and -2, the adsorp-
      tive capacities of carbons of equal carbon tetrachloride
      activity will only agree at one relative vapor pressure for
      any vapor.  For carbon tetrachloride the point of agreement
      is at 0.16 relative vapor pressure.  Carbons with smaller
      mean micropore diameters will adsorb vapors more strongly
      at the lower relative vapor pressures.

     Adsorptlve Capacity of Operating  Unit  - The various
properties of the adsoroents are reported on the per-unit-
weight basis in the manufacturers' sales 1iterature,  although
the adsorbents are used in various operating units  on the
fixed volume basis.  Unlike the weight basis, the bulk vol-
ume basis would vary with mesh size,  mesh size distribution,
particle shape  and the method of packing the adsorbent bed.
A comparison of properties on the weight basis is a simpler
and more reliable evaluation of the relative merits of the
different adsorbents.

     To determine the true effectiveness of an adsorbent in
actual usage,  the various physical and adsorptive  proper-
ties need to be recalculated on the per-unit-volume basis
as has been done in Tables 4-VI and -VII.  When comparing
the surface area in mz/g with m2/cm3  bulk volume, and pore
volume in cm3/g with cm'/cm3 bulk volume, much of the large
apparent advantage of carbons over polar adsorbents is
greatly reduced.  Similarly, the comparison of micropore vol-
umes on the weight basis with those on the  bulk volume basis
tends to nullify the large apparent advantage of Carbon IV
over Carbons II and III.

     Table 4-VIII shows the effect on carbon  tetrachloride
activity and surface area when the values  are converted to
bulk volume basis.  Less weight of the lighter Type  IV car-
bon is required to fill the adsorption operating unit and
thereby maintain cost closer to the other  heavier  carbons.
Table 4-VIII  Physical Properties of Activated Carbons.
              Densities. Carbon Tetrachloride Activities
              and Surface Areas

         Bulk          Carbon tetrachloride
         density,           activity          Surface  area
Carbon    g/cm*         g/gg/cmm*/g   ma/cm3

    I      0.38          0.30       0.11        580      220
   II      0.50          0.68       0.34       1100      550
  III      0.45          0.73       0.33       1300      580
   IV      0.33          1.20       0.40       1600      530

     Theories of Adsorption

          Forces of Adsorption

          Van der Waals '  Forces - The forces active in physical
     adsorption on polar and nonpolar adsorbents are associated
     with van der Waals1  forces as measured by constant a, in the
     van der Waals' equation of state; see Appendix D.  This
     association was first observed by Arrhenius (1912).  Various
     investigators since then have derived mathematical expres-
     sions that offer explanations of the oriains of these forces.
     These theories have not lead to useable methods for pre-
     determininq the adsorptive affinity from fundamental proper-
     ties of the adsorbed vanor and adsorbent surface.  Among the
     complicating factors are (1) the surface nonuni forml ties ,
     (2) contribution of subsurface atoms to the adsorption po-
     tential, (3) uncertainty of distance between the adsorbed
     molecule and the surface, and (4) opposite-wall effects in
     small pores.

          Orientation and Induction Effects - To account for the
     van der Waals1 attraction between polar molecules, the dipole
     moment and the polarizabili ty are regarded as the primary
     forces.  The first is often referred to as the orientation
     effect and was investigated by Keesom (1921); and the second,
     investigated by Debye (1920), is referred to as the induction
     effect.  Lorenz and Lande, in applying the orientation effect
     to adsorption, derived an adsorption potential equation on
     the assumption that an interaction occurs betweenthe dipo
     JUbJL-ifl mirror image of thfi dipoj.fi in a po1ariz_able surfac**.'
     Jaquet U9Z5J followed the same line of reasoning but also
     included the induction effect.  According to Jaquet, the
     adsorption potential, Q, at low temperatures is

                             Ny2   2       Nu*
                                         ~*o^T              (4-1)
     where n  o/(2o)3, a is the polarizabili ty, N is Avogadro's
     number, y is the dipole moment and a the distance of adsorbed
     molecule  to   adsorbent surface.  At low temperatures where
     thermal agitation is small, the dipoles are all oriented
     perpendicular to the surface.  At high  temperatures where
     all orientations are equally probable,  the adsorption poten-
     tial is
     The orientation and  induction effects do not account for  the
     adsorption of nonpolar molecules, particularly to nonpolar

Dispersion Effect - London (1937) attributed the attrac-
tion between nonoolar molecules to fJuctuz
Ky mnme.nta.rv cVanaes in electron ffTstribu
nuciei. The fluctuatina dinoles induce d^
moTecules which then fluctuate in nhase w

:ina dinoles caused
on about thP stom^c
ioles in neiahboring

resulting attraction is referred to as the dispersion effect,
and the potential between two molecules of the same type is
approximately expressed by the equation

where h is the Plank's constant and v is the characteristic
frequency of oscillation of the molecule, the same as in the
Sellmier light dispersion equation.  For simpler molecules
   is almost equal to the ionizatign finerm/: but for larger
molecules, particularly if they contain chromophoric groups,
hv may be considerably different from the ionization poten-

     With reference to Equations 4-1 and  -2, the induction
effect, determined by the polarizabil i ty  a, is  always  small,
while the orientation effect varies directly as the  dipole
moment u, or as u to the second power.  For small  polar
molecules the  orientation potential is  large;  but  for  in-
creasingly larger molecules  the dipole  moment  decreases
rapidly and thereby the orientation potential  also decreases.
For most of the organic pollutants of concern,  the orienta-
tion potential is essentially  zero,  The  dispersion  potential
i s Dominated bv the Dolarizabi 1 i ty . and,  because  poianz-
a bill ty is pjioF?15i'T^Ty?TToTair^l ume  (Vm) ,  the  dj.s^
                                                  j u t w i t h
minor  variations  produced  by  hv.mis  means  that small  non-
pniar  molecules ajis_j^o__s_trongly  adsorbed bv  anv n^Jt^S&
cnrbentS  DV  the van  HPT  U !'  ln>.,-oc    Thle *h,...i, *.>
5n  essential  agreement with observed  adsorptive affinity
variations as  presented  in Tables 4-II  and -III.

     Adsorptive Capacity

     In  the  prior text of  this  chapter, the effects of chemi-
cal  constitution  of  the  adsorbent, its  pore structure and
the chemical  nature  and  size  of the vapor molecule were dis-
cussed relative to the adsorptive affinity.  These factors
are considered 1n selecting  the type  of adsorbent that is
best suited  for the  particular  pollution control problem
under  consideration.

     One of  the parameters that determines the effectiveness
of  the selected adsorbent is  its  adsorptive capacity  under
the expected operating  conditions.  This Information  is

     necessary  before  any  sizing  estimates can be made for the
     sorption unit.   In  a  general  sense,  the capacity increases
     with  increased  vapor  pressure and decreases with increased

          Methods  for  plotting  adsorption data and adsorption
     equations  have  been developed which:  (1) make presentation
     of  data concise  and convenient,  (2) permit extrapolation of
     data  to unmeasured  pressures  and temperatures, and  (3) per-
     mit prediction  of adsorptive  capacities of unmeasured vapors
     from  few data on  those  that  have been measured,  The last
     item  is of great  importance  in pollution control work because
     of  the  large  number of  different vapors of concern that are
     emitted under a  wide  variety  of conditions and for which ad-
     sorption data are lacking  or  very meager.

          Adsorption  Isotherms  -  Adsorption at constant tempera-
     ture  is frequently  plotted on a linear graph as a function
     of  vapor pressure or  relative partial pressure (p/po).  Figure
     4-3 shows  a family  of hydrocarbon isotherms on activated car-
     bon plotted in  this manner,   The amount adsorbed increases
     rapidly with  increased  pressure at the low pressure end of the
     isotherm but  levels off at the higher pressures.  The liquid
     volume, 0.43  cm3  liq/g, at 760 mm of Hg pressure for butane
     is  approaching  the  micropore  volume; hence, the diminished
     adsorption is caused  by filling up the small pores.  At
     higher  pressures  approaching  p/po near one, adsorption would
     continue with the filling  up  also of the macropores.  This
     phase of the  adsorption would be accompanied by a considerable
     change  in  shape  of  the  isotherm.  Macropore filling can only
     occur at temperature  below the critical temperature.

          Curve III  of Figure 4-1  shows the micropore and macro-
     pore  size  distribution  of  the type of carbon used in adsorp.
     tion  measurement  plotted in  Figure 4-3.  At 760 mm of Hg
     pressure for  butane,  p/po    0.28.

       *  Only  the very  low  pressure end  of each isotherm in
     Figure  4-3 would  be of  direct interest in pollution control
     problems.   Pollutant  vapor concentrations at the sources of
     emission are  generally  below  1,000 ppm, and combustible
     solvent concentration,  for safety reasons, is maintained at
     or  below the  25%  lower  explosive limit during recovery oper-
     ations.  Table  C-I  gives LEL  concentrations for pollutants
     of  concern.   The  25%  LEL for  most organic vapors Is below
     13,000  ppm which  in pressure  units corresponds to 10 mm of

          When  it  is  necessary  to  present adsorption data over a
     wider pressure  range  than  the single order of magnitude
     shown in Figure  4-3,  the adsorption  data are plotted on a

 0.40 -
  0.20  ~
300    400   500
  p, mm of Hg
 Figure 4-3  Adsorption Isotherms  of Hydrocarbon Vapors,
             Amount Adsorbed w at  Pressure p on Type
             Columbia L Carbon at  100F, Liquid Volume
             of w at Bollinq Temperature.

      log-log  graph  as  in  Figure  4-4.  These  isotherms also level
      off  at  the  higher pressures  because  of  micropore filling at
      the  0.50 cm3/g level.   In  the  low  pressure region below 10
      mm of Hg, which  is of  interest in  pollution control, this
      manner  of nlotting clearly  indicates  the adsorptive capaci-

           Isotherms of a  different  type that are important to
      pollutant vapor  control  are  those  of  water vapor on activated
      carbon.   The  nature  of the  isotherm  gives some  insight into
      the  type of interference the presence of water  vapor will
      have on  the adsorption of  organic  vapors.  Isotherm I of
      Figure  4-5  corresponds to  pore size  distribution Curve I in
      Figures  4-1 and  -2.   Isotherm  II is  typical for activated
      carbons  best  suited  for  adsorption of vapors; i.e., of carbon
      types represented by Curves  II and III  in Figures 4-1 and  -2.

        * The characteristic  features of  the water adsorption iso-
      therm are the  pronounced hysteresis  of  the adsorption and
      desorption  branches  and  the  rapid  increase in amount adsorbed
      at  p/po  above  0.5.   At p/po  below  0.4 the moisture  in a
      pollutant airstream  will  cause very  little interference with
      the  adsorption of most of  the  pollutants of concern.  Moisture
      can  cause a further  decrease of the  already poorly  adsorbed
      organic  vapors such  as chloromethane, ethane and methane.
      Organic  vapors of Vm greater than  90  cm3/mol displace water
      from the carbon,  but at  the  higher p/po of water the adsorp-
      tion rate of  organic vapors  decreases.  In thin carbon beds,
      a high  water  P/PO can  cause  a  premature pollutant vapor
      penetration.   Moisture interferes  less  with water soluble
      pollutants  than  with nonsoluble pollutants.

           The hysteresis  can  cause  operational problems.  If
      satisfactory  operation is  possible at p/po range from 0.5
      to 0.7,  but the  moisture content of  the pollutant airstream
      should  happen  to  go  to higher  humidities such as to saturate
      the  bed, the  carbon  would  have to  be  dried to p/po  of 0.4  or
      less before satisfactory operation can  be regained.

           Equations for Isothermal  Adsorption - Various  investi-
      gators  have developed  adsorption equations that agree with
      data for specific adsorbate-adsorbent systems and over fairly
      limited  concentrations.   These equations are suitable for
      interpolation  and extrapolation purposes.

           An  equation  found applicable  for liquid phase  decolor-
      izing applications and also frequently  checked  out  on vapor-
      phase adsorption  data  is the Freundlich (1933)  equation  as
LJ f | vl J t  ^wi ****  -  _--_   _
given below in logarithmic form.

.  0.10
            j	I
                 I  I .
                      1 I I
                            p, mm  of  Hq
                                                     1  I ' ' ' H
  Figure  4-4   Adsorption  Isotherms  of  Butane  at Three Tempera
              tures  on  Pittsburgh  BPL  Type  Carbon, Amount u>
              Adsorbed  at Pressure  p.   Data From Grant  (1962)

     Figure 4-5  Isotherms of Water on Activated Carbons Showing
                 Hysteresis Between Adsorption and Oesorptlon
                 Branches, I Darco Carbon, II Typical of Pitts-
                 burgh BP or Union Carbide SXA.

               log u   log  k  +  [1/n]  log  p              (4-4)

Here k and n are constants  dependent  on molecular  structure
of adsorbate, chemical  constitution of adsorbent,  pore  struc-
ture and temperature; k 1s  equal  to to when  p  is  one.

     When the adsorption data obey the equation,  a  log-log
plot yields a straight  line graph. The type  of  curves  in
Figure 4-4 are frequently called  Freundlich isotherms;  how-
ever, in these cases, the curvatures show   deviation  from
the equation.  At lower pressures, below  1  mm of Hg,  the
tendency is for the curves  to straighten  out.

     In 1916, Langmuir  derived  an adsorption  equation for
vapors on kinetic considerations.  According  to  Langmuir,
vapor molecules colliding with  the adsorbent  surface  suffer
Inelastic collisions; the time  lag is responsible  for ad-
sorption.  He also assumed that only  unlmolecular  layer
adsorption could occur.  The Langmuir equation 1s  presented
1n two forms below; the latter 1s more  convenient  for test-
Ing against adsorption  data.
In these equations v is the volume of vapor adsorbed on a
unit weight of adsorbent, b is a constant, and vm Is the
amount of vapor necessary to form a unimolecular adsorbed
layer.  The equation fits chemisorptlon data and physical
adsorption data at low relative pressures or data of gases
at temperatures above their critical temperatures.  Essen-
tially all pollutant vapors are adsorbed below their critical
temperatures and on adsorbents of small pore diameter in
which opposite pore wall attractive forces would induce
adsorption by molecular packing rather than unimolecular
adsorption.  The equation has little or no direct application
in pollution control studies; but  it 1s an Important equa-
tion used 1n developing later equations useful to pollution

     Brunauer, Emmett, and Teller (1938) derived equations
for multilayer adsorption of vapors by generalizing on Lang-
muir's treatment of unimolecular adsorption.  On powdered
materials or adsorbents having large pores, where an un-
restricted number of molecular layers can be adsorbed; the
equation has the form

                                  1      c-1
                      v[p0-p]     vmc    vmc PO                 '

     In  this  equation  c  Is  a constant dependent on temperature,
     the  average  heat  of  adsorption E]  in the first adsorbed layer,
     and  the  he^t of liquifaction  E$.
                         c - exp - ^j                     (4-8)

     When  the adsorption  data obey  the equation, a plot of

                       Vn^TpT versus ^                    (4.9)

     gives a straight line graph from which it is then possible
     to  calculate vm and  c.  The number of molecules in VD mul-
     tiplied by the surface area covered by a single molecule
     gives the total surface area of the adsorbent.  This is the
     well  known BET equation, and, amongst many other applications,
     is  used by the adsorbent manufacturers to determine and report
     the surface areas of their products.  It has been the center
     of  considerable investigation by numerous researchers whose
     works have been reviewed by Gregg and Sing (1967).

          For surface area determinations, the most practical gas
     is  nitrogen and the adsorption data are obtained at liquid
     nitrogen temperature, i .e. , -196C.  The nitrogen molecule
     mean  diameter is 4.3 A,  and the area covered by each molecule
     is  16.2 square A.  In adsorbents having small pores, partic-
     ularly in activated carbons, the nitrogen molecule diameter
     is a  large fraction0of the pore diameter; as for instance in
     pores less than 17 A diameter (the mean diameter for SXA and
     BP Figure 4-1) the ratio is over 0.25.  In these cases, the
     plot of relationship (4-6) above gives a very short straight
     line graph or one that has a continual but slight curvature;
     hence, the surface areas reported for these adsorbents are
     subject to some error.  The BET equation developed for a
     flat surface deviates when applied to small-pore-sized adsor-
     bents.  When the bulk of the pores are 30 A and larger, the
     plot of relationship (4-6) gives a straight line graph that
     can extend to 0.35 p/po, and thereby lend a greater confidence
     level  in these applications.

         Where the number of molecular layers is restricted to
     some finite number n, as might be the case in small-pore-
     sized adsorbents, the BET equation has the form

                   vmcx l-(n + l)x"  +  nx"
                   1-x  1  + (c-l)x -  cxn + '

where x = p/p0.  This equation, however, has  not found accep-
tance among adsorption investigators  as an equation  for de-
termining the surface area.

     Pore Structure Determination

     Based on Nitroqen Adsorption Data - The  pore size dis-
tribution is one of the most important factors that  deter-
mine the overall shape of the isotherm, and conversely the
isotherm should give a measure of the pore size distribution
when correctly interpreted.

     According to multimolecular layer and capillary conden-
sation theories, the adsorption of a gas such as nitrogen at the
nitrogen boiling point (-196C) commences as  a unimolecular
layer followed by multimolecular layers until the pore space
is filled.  This mechanism determines the shape of the ad-
sorption branch of the isotherm.  On desorption, the nitro-
gen  is released by a different mechanism that gives a hys-
teresis  loop  to the  desorption branch of the isotherm.
The  general form of a nitroqen isotherm is illustrated in
Figure 4-6 for an activated carbon suited for vapor phase

     When an incremental volume of adsorbed nitrogen (liquid
volume)  is desorbed, a part of it is vaporized from the
multimolecular  layer on the pore wall  and a part from  the
meniscus of the condensed  phase  in the  inner  capillary space.
The  p/Po bears  relationships  simultaneously to the meniscus
radius and multimolecular  layer  thickness.  The  Kelvin e-
quation  (Thompson  1871) given  below  expresses  the function
between  p/Po  and meniscus  radius, r^.
                      rk  =  Y   c?s                       (4-11)
                       k    ftt  ln(po/p)                   VH  IU

 Here  Y  is  the  surface tension  of  the  adsorbed  nitroqen,  V
 the  liquid molar  volume  of the adsorbed  nitrogen, and    the
 contact angle  of  the  miniscus  with  the multimolecular ad-
 sorbed  layer.   (For nitroqen  at -196C,    is  zero).   The
 multimolecular layer  thickness, T,  can be  determined  ex-
 perimentally for  nonporous adsorbents  and  then applied  to
 porous  adsorbents assuming the function  still  applies.   The
 pore  radius at any p/p0  on the desorption  branch of the

             -  0.40
             t 0.30
               0.10 -
                            I	I
     I	 I    I    I    I
0.6     0.8     1.0
          Figure 4-6  Nitrogen Isotherm on Activated Carbon
                      at -195C  Showing Hysteresis Loop
                      Extending from p/p0  0.5 to 1.0

isotherm is then

                      r * rk + T                        (4-12)

Barrett, Joyner and Halenda (1951), on these basic assump-
tions, developed a method for determining pore size distri-
bution curves for adsorbents from the desorption branch of
the nitrogen isotherm determined  at -196C.   The calculations
are quite involved  requiring tables of precalculated T vs
p/p0 values (American Document Institute).   The method, on
theoretical grounds, is applicable only to the portions of
the isotherm above the lower p/p0 at which  the hysteresis
loop closes.  The lower limit can range from 0.4 to 0.5 p/pfl;
the 0.4 p/p0 corresponds to 40 A  pore diameter.

     Because of difficulties in measurinq p/p0 accurately
above 0.97, the practical upper limit on diameter determina-
tions is 670 A.  The method has,  however, been extended by
many investigators to pores smaller than 40 A.

     Based on Hater Adsorption Data - An alternative method
applicable only to activated carbons  involves the direct
application of the Kelvin equation (Equation 4-11) to the de-
sorption branch of the water isotherm (Juhola and W11g,i949).
The pore size distribution Is derived by measuring desorbed
liquid water volume increments at each reduced p/p0.  In
Equation 4-11,molar volume V varies between  19.3 and 21.2 cm8/
mol with changinq p/p0 (Wiig and Juhola, 1949) and cos 
      penetration  at  each  pressure  increase.  This technique was
      first  suqqested by Washburn and first used by Ritter and
      Drake  (1945).   They  found  that <> was (qua! to 140 on a
      variety  of solids although some uncertainty exists in its
          The  largest pores measurable are 100,000 A at a pressure
      of  17.4  Ib/sq in. absolute.   A pressure of this magnitude is
      required  to  outline  adsorbenj granules; or conversely,
      "pores"  larger  than  100,000 A diemeter are in reality inden-
      tations  in thu  granule exterior surfaces.  The smallest
      pores  measurable depend on maximum pressure attainable  and
      the  effect the  pressure has on the pore structure.  A theo-
      retical  pressure of  60,000 lb/in2  is required for mercury
      penetration  into pores of  30  A diameter.  At these pressures
      compression  of  the carbon  can be expected resulting in
      erroneous volume measurements.  Measurement up to 60,000
      lb/in2 are,  however, made  as  standard practice with commer-
      cially available equipment.

          Comparison of the Three  Methods - The pore size dis-
      tribution, as determined by the above described three meth-
      ods, overlap in the  30 to  760 A diameter which makes possi-
      ble  their comarison  over a limited range of application.

          Joyner, Barrett and Skold (1951) compared the pore size
      distribution curves  obtained  by the nitrogen isotherm method
      and  mercury  porosimetry over  the diameter range 60 to 600 A,
      Their  studies were with bone  chars and synthad, a material
      chemically similar to bone char.  These adsorbents were
      chosen because  they  had an abundance of pores in the range
      where  the two methods overlap.  They found a good agreement
      between  the  two methods on some of seven adsorbents tested.
      Where  there  was more than  one peak in the distribution curve
      both methods indicated their  presence, shape and size with
      good agreement  in some cases, although in others.there was

          Pore distributions in the 10 to 600 H range for acti-
      vated  carbons,  as determined  by nitrogen and water adsorp-
      tion methods, yield  curves of the same overall shape but
      disagree  to  some extent on magnitude of diameter at any
      given  accumulative volume.  The most significant  variance
      starts at the 40 A diameter and is carried down to the
      smallest  pore measured (Juhola, Palumbo and Smith, 1952).

          Comparisons of  pore size distributions as determined
      by  the water adsorption and mercury porosimetry methods al-
      ways show a  qualitative agreement in shape of distribution
      curve, but vary in agreement  on magnitude of pore diameter
      at  any given accumulative  volume.  With some carbons, the
      overlapping  section  by the two methods coincide; while with
      others,  the  varUnce can be either way.

     The conclusion is that the three  methods  are  in  agree-
ment in indicating the type of pore  structures  for diverse
carbons.  The type of distribution has an  important bearing
on the adsorption of pollutant vapors  at low concentrations.
Carbons with an abundance of small pores have  the  largest
capacity at low concentrations, a  desirable  property  for
pollutant control; however, an abundance of  small  pores  is
an undesirable property in solvent recovery  carbons oper-
ated at high solvent concentrations.

     Adsorption Thermodynamics

     Heat of Adsorption - Adsorption is accompanied by two
energy changes:(1) a decrease in free energy AF  measured
by the decrease in vapor pressure, and  (2)  a  decrease in
entropy AS due to decrease in mobility of the  adsorbed vapor.
These changes decrease the heat content AH of  the  system in
the manner expressed by the equation

                    AH - AF + TAS                        (4-14)

All adsorption processes are exothermic.

     AH can be determined calorlmetrically,  or more conven-
iently* from adsorption isotherms  measured at separate tem-
peratures by application of the Clausius-Clapeyron equation.

                    In p - M + k                         (4-15)

This equation applies at a constant amount of adsorbed vapor
and applies over a  limited temperature  range because of  a
gradual variation  in AH with change in  temperature.

     Its application can be demonstrated with data taken
from Figure 4-4.   Table 4-IX presents  the pertinent data
and the calculated  AH and  k.   Sample  calculations  are given
1n Appendix E ,  page E-l.

Table  4-IX  Adsorption Data and Calculated AH and  k for
Butane on Activated Carbon Type BP
M| cm3 liq/9
mm of Hg

-AH .cal/mol
[ca 5.500]

     AH in these cases are the differential  or  isosteric  heats  of
     adsorption.  They decrease as the amount  of  adsorbed  vapor
     is increased.  When the pore space 1s  almost completely  filled,
     AH will  approach the heat of 1iquifaction, which  varies  from
     5,600 at -9 to 5,400 cal/mol at 170eF.

          The heat of adsorption in the initial stages  is  the
     highest  because the vapor first becomes attached  to  the  most
     active sites on the surface or first becomes located  in  the
     smallest pores where the opposite wall  attractions are the
     strongest.  In the final stages of adsorption,  where  the ad-
     sorption mechanism can be multimolecular  layer  or  capillary
     condensation in large pores, the heat  of  adsorption  approaches
     that of  liquifaction.

          The heats of adsorption increase  on  the molar.weight
     basis, but when converted to unit weight  or  unit  liquid
     volume basis, they are in a fairly narrow  range,  i.e., 100
     to 200 cal/g or 200 to 300 Btu/lb of vapor adsorbed.  At
     low pollutant vapor concentrations, below  100 ppm, the heat
     release  rate 1s slow and is greatly moderated by  the  air-
     flow.  When the vapor concentrations are  at  the 25%  LEL,
     considerable heating of the carbon bed can occur.  Auto-
     ignition of carbon beds used for ketone recovery  have been
     reported.  This problem is avoided by  leaving the  bed wet
     from the prior steam regeneration.

          Figure 4-7 shows the data of Table 4-IX plotted  as
     isosteres.  When pressure data are available for  a given u
     at more  than two temperatures, a plot  of  this type measures
     the degree of agreement with theory.  The  three pressure
     points for u - 0.30 cm3liq/g yield a straight line curve
     indicating that AH is quite constant over the temperature
     range -9 to 170F.  Carbon adsorbent  systems for pollutant
     control  and solvent recovery are operated most effectively
     in the 60 to 130F temperature range.

          Free Energy Change - The free energy change  AF,  as  part
     of Equation 4-14, is a function of the relative vapor pressure
     given by the following relationship.

                        AF - -RT ln[po/p]                      (4-16)
     On the basis of the data in Table 4-IX AF values were cal-
     culated to show their characteristics; these AF values are
     qiven in Table 4-X.  For these calculations, po at -9*F 1s
     300 mm, p0 at 77F is 1820 mm and p0 at 170F 1s 7000 mm of



 	 , 	 p 	 , 	  	 T
\ 170*F K 77F -9^
X \
\ \
\ \

       r- N


   ft  10
.40 cm'Hq/g     
                         . . .   .  I   .  . N
3.0         3.5
      1/T x 10J
Figure 4-7  Isosteres  for  Butane  on  Activated  Carbon.
            Type BP

     Table  4-X   Free  Energy  Change AF  During Adsorption of
                 Butane on Activated Carbon

                                    -AF, cal/mol	
          to,  cm3liq/g        at -9F      77F       170F

              0,20            	      3,300       3,200
              0.30            2,400      2,300       2,100
              0.40            1,400      1,200       	
     AF  decreases as  the amount of vapor adsorbed on the carbon
     is  increased.  At complete filling of the pores, AF approach-
     es  zero and AH = TAS.

           Entropy Change - Table 4-XIgives the entropy changes for
     the butane adsorption obtained by substitution of AH and AF
     from  Tables 4-IX and  -X into Equation 4-14.
     Table 4-XI  Entropy Change AS During Adsorption of Butane
                 on Activated Carbon

                              	-AS. cal/mol	
          a), cm3 liq/q        -9F       77F       170F

             0.20            	       18.6        16.0
             0.30             25.2       21.6        18.8
             0.40             26.0       22.6        	
           [liquid]          [22.4]     [18.4]      [16.0]
     AS measures the change in mobility of the molecules in trans-
     fer from the vapor phase to the adsorbed phase.  A large
     -AS indicates a large decrease in mobility while a +AS in-
     dicates a large increase.  As the micropores are filled -AS
     increases and is maximum at the point where the micropores
     are completely filled, i.e., at about 0.45 cm*/q.  During
     macropore filling -AS decreasesand ultimately reaches the
     value at normal condensation.

          Potential Theory Equations - Two properties of AF use-
     ful in pollution control work are:  (1) AF is fairly constant
     over a moderate temperature range; and (2) AF decreases with
     increased amount of adsorbed vapor.  Equations have been
     developed on the basis of these two properties that permit
     estimating adsorptive capacities for an unmeasured vapor-when
     meager data are available on some other.but similar, vapor.

     Tables 4-X and -XII  and Figure  4-8  present  numerical
figures that indicate the degree of  constancy  of AF with
temperature change, and show the decrease  of AF  with in-
creased 0).

     For propane (Table 4-XII)  at u  less than  0.20 cm3  liq/g,
AF increases gradually with increased temperature, while
above 0.20, it decreases  with increased  temperature.  At
high adsorption pressures p or  high  vapor  pressures p0, the
use of fugacities, f and  f0, in place of p and PA 1n Equa-
tion 4-16 improves the constancy  of AF.  Sample  calculations
showing the use of fugacity and its  effect on  AF are worked
out in Appendix E, page E-3.

Table 4-XII  Variation of AF or ej with  Temperature and
             TTmount u> of Propane Adsorbed  on Type L Carbon

                	AF or e-j, cal/mol	

                                                 Average of
u>. cm8 liq/g    100F   150F   200F   250F    100F & 150F

    0.10        3,800   4,000   4,200   4,500      3,900
    0.15        3,300   3,400   3,600   3,600      3,350
    0.20        3,000   3,000   3,100   	      3,000
    0.25        2,600   2,500   	      2,550
    0.30        2,100   2,000		      2,050

Equation 4-16 is the basis of the Polanyi potential  theory
of adsorption, the symbol ci being used in  place  of AF
(Polanyi,  1914).   He observed that at any fixed amount to
adsorbed,  el is constant; hence,

                      ~ --InCpo/p]                     (4-17)

If PQ/PI and TI are  known at a  fixed u>i, P0/P2  can  be  esti-
mated for  ui at T provided T^  and T are not widely sepa-
rated.  Carbon adsorbent systems  are most effectively  used
in the fairly  narrow temperature  range  of 60 to  130F;
hence, AF  or e-j will be fairly  constant in  the  temperature
range of interest.

     The curves for  propane, butane  and pentane  1n  Figure
4-8  show graphically the variation of AF with molecular size
and  w.  When AF is divided  by the molecular weight,  I.e.,
by converting  AF  to  cal/g,  the  three curves are  drawn  closer
together giving a  common curve;  this effect 1s  shown  in

           1.00 r
                      1,000   2,000   3,000   4,000
                               -AF or -ei,cal/mol
5,000   6,000
      Figure  4-8  Variation  of  Free  Energy  AF  or  Polanyi  Potential
                  ei  with  Amount  of  Vapor w Adsorbed on Activated
                  Carbon.  u> Measured  per Gram of Vapor and AF  and
                  ei  per Mol of Vapor

Figure 4-9.  The amount adsorbed is  now a  function  of  M,  T
and po/P as shown below,

              w  function of j^- In  (p0/p)               (4-18)

where M is the molecular weight and  w is measured  in g vapor
per g of carbon.

     If a few adsorption points are  known  for  an activated
carbon so that a curve can be drawn, the curve can  now be
used to estimate isotherms of other  members  of a homologous
series or of other similar vapors.

     Because of packing of molecules in the  micropores,  ad-
sorptive capacity is determined volumetrically rather  than
gravimetrically.  By dividing AF by  the molar  volume,  Equa-
tion 4-18 is greatly broadened to include  other homologous
series and a wider range of higher density molecules.   The
proper choice of molar volume has been an  area under consid-
erable investigation amongst different adsorption  investiga-
tors.  Dubinin, et al. (1947) suggested using  the  liquid
molar volume of the adsorbed vapor at the  adsorption temper-
ature.  Lewis, et al. (1950) and later Grant,  et al. (1962)
experimented with the molar volume at the  temperature  where
vapor pressure equals adsorption pressure.   Robell, et al.
(1965) used the liquid molar volume  at the normal  boiling
point.  Lewis and Grant used fugacities in place of adsorp-
tion and liquid vapor pressures where a significant differ-
ence existed.

     In view of the packing of relatively  large molecules of
over 4 A diameter in pores of less than 20 fl diameter, the
normal concepts of liquid molar volumes lose some  of their
meaning.  Two effects can occur:  (1) steric hindrance which
tends to promote packing of lesser density than in normal
liquids, and  (2) stronger attraction toward the pore  walls
which tends to promote packing of greater  density  than 1n
normal liquids.  The actual volume of adsorbed liquid  can be
expected to vary with molecule-to-pore diameter ratio  and
compressibility of the molecules.
     For purposes of estimating w, the right side of Equation
4-18 is now modified so that it has the form

                  [T/V] log [po/p], or                  (4-19)

                    [T/V] log [f0/f]                    (4-20)

where V is the liquid molar volume of best choice for the
particular system.  From adsorption data (at least two adsorp-
tion points) a generalized adsorption curve 1s drawn wherein u
In liquid volume per gram of carbon 1s plotted on a logarithmic


                               10       15       20
                               -AF  or ei,  cal/g
      Figure 4-9   Variation of Free Energy AF or e^,  with Amount
                  of Vapor CD Adsorbed on Activated Carbon,  w,
                  AF and et Measured per Gram of Vapor

ordinate scale as function of the quantity RgJ or \20j above  on
a linear abscissa scale.   Whatever density  is  used  to  calcu-
late molar volume V is also used to calculate  the  liquid
volume of w; the effects  of density selection  thereby  tends
to cancel out.  For future reference,  the quantity  in  if 9? or
    s designated by A when it is more  convenient  to do so.
     The adsorption results for propane,  butane  and  pentane
as presented in Figure 4-8 are replotted  in  Figure 4-10,
using quantity /IsQand molar volume Vm as  measured  at the
normal boiling point,  w was also converted  from g/g to cm3
liq/g, using the liquid density at the normal  boiling point.
To obtain Vm, the method of LeBas was used;  see  Appendix  D
for table of additive volume increments.

     The method gives the best generalized adsorption curves
for homologous series.  On activated carbons the generalized
curves of different homologous series tend to  coincide; the
largest variance comes at the large A values.   Grant, et al.
found that the generalized curve for sulfur  compounds was
considerably above that for hydrocarbons.  On  silica gel  the
variances between curves of different homologous series are
larger; adsorption data on one homologous series do  not lead
to an acceptable prediction regarding adsorption of  members
of other homologous series.

     Effect of Pore Structure on Generalized Adsorption Curve
. Figure 4-11 shows three generalized adsorption curves that
may cover the range of this type of curve for commercial carbons
suitable for vapor phase adsorption.  S154 is an experimental
superactivated peach pit carbon of 154% carbon tetrachloride
activity having micropore volume near 1.00 cm3/g.   It  is
similar to carbon IV of Figure 4-1 but with larger  micropore
volume and also larger mean micropore diameter.  It approaches
the upper limit on high capacity carbons.  GI is a  high activ-
ity coconut carbon of 0.65 cm3/g micropore volume.  The gen-
eralized adsorption curve for BPL was determined with  adsorp-
tion data of sulfur containing vapors such as methanethiol ,
ethanethiol, propanethiol , hydrogen  sulfide and carbon disul-
fide.  These vapors are more  strongly adsorbed than  other
organic vapors of equivalent  A value.  Generalized  adsorption
curves for other carbons  are  expected to  be below this BPL
curve at the larger A values.

      Figure 4-12 shows generalized adsorption curves  that  are
typical of solvent recovery and  gas  mask  carbons.   The pore
size  distribution of  SXA  is number  III in Figure 4-1  and  that
of the upper BPL curve is  number  II  in Figure 4-1.   The  lower
BPL curve, designated here  as Type V, was obtained  from
studies carried out  by Robell  (1965), and is  presented here
because a large number of  regeneration calculations discussed
In Chapter 6 are based on  this curve.


-~ C4H10

                         i   i
 J	L
                                      log  [p0/p]
      Figure  4-10   Generalized  Isotherms  For Propane, Butane
                   and  Pentane

      S154, Superactivated Peach Pit Carbon,
            MSA Research Corp. experimental  carbon

          GI, Barnebey Cheney Coconut
                                BPL, Pittsburgh
                                     Activated Carbon,
                                     curve for sulfur
                                     compounds, Type VI
| 1  i i  i I  M I i  I I  I I  h I I  t  I I  I I  I I   I  I I  I M I  I
                    10     15      20

                     A, [T7YB]loq[p0/Pl]
Figure 4-11  Generalized Adsorption Curves For Three
             Widely Different Types of Activated Carbon

                             SXA,  Union  Carbide
                                  Type  III.  Figure 4-1
BPL, Pittsburgh
     Carbon, Type V
                                             BPL, Type  II,
                                                 Figure 4-1
            I i  I i  I I  I I  I I  i I  I I  I I  M I  ll i  i  i I  I I  i i  I i  t
10      15     20

A, [T/Ym]lofl[p0/p]
      Figure  4-12   Generalized  Adsorption  Curves Typical of
                   Activated  Carbons  Used  in  Gas Masks and
                   Solvent  Recovery

     Under the conditions of the carbon tetrachloride activ-
ity test described in Appendix D,

                [T/VJ log [p0/p] = 2.5 .               (4.21)

At 2.5, CD  0.45 cm3 liq/g for BPL(II).  From Curve  II in
Figure 4-1, w  0.45 cm3/? corresponds  to  a pore diameter of
40A.  This means that all adsorption at A  larger than 2.5
occurs in pores smaller than 40 A diameter, and  also  the
curvature of the generalized adsorption curve is determined
by the distribution of pores in this small-diameter  range.
For SXA Curve III, the volume of pores  smaller than  40 A
diameter is 0.48 cm3/g.  Above 0.05 cm3/g  pore volume, the
pore diameters of SXA are smaller than  those of BPL.  This
accounts for the stronger adsorption by SXA at u> laroer than
0.05 cm3 liq/g.  At A  17 or u>  0.05  cm3 liq/g, the SXA
and BPL curves cross each other; and likewise  the pore-size
distribution curves cross each other at 0.05 cm3/ci pore vol-
ume.  At u or pore volume less than 0.05, BPL is the stronger

     In macropores larger than 40 A* diameter or, stated other-
wise, at A less than 2.5, adsorption proceeds very slowly,
although for SXA and BPL the available  macropore space is
about 0.40 cm3/g.  At A less than 2.5,  AF, the driving force,
for the adsorption approaches zero, and at the theoretical
point where all the pores are filled

                        AH  TAS                       (4-22)

By accurate measurements of adsorption  at p/p0 above 0.99,
the generalized adsorption curves of SXA  and BPL would ap-
proach A  0 gradually and stop at u near 0.85 cm3 Hq/g.
By a straight extrapolation of the curves to A  0,  as done
in Figure 4-11, and -12, the w intercepted is aemeasure of
pore volumes usually of diameter less than 100 A.

     The operating capacity in solvent recovery applications
1s in the low A value range (1 to 15);  hence, a large ad-
sorptive capacity is desirable 1n this  A  value range but, to
ease regeneration,a low capacity is desirable in the higher
A range.  To adsorb weakly adsorbed pollutants or pollutants
at very low concentrations, strong adsorption is required at
large A values  (above 15); the capacity at low A values In
this case makes very little or no contribution.

          Dynamic  Adsorption Processes

          Vapor  Distribution in Adsorbent Bed - Sorptlon devices
     for  solvent recovery  or air purification most frequently
     employ a  fixed adsorbent  bed.   It is the simplest and most
     trouble free  in  its operations.  Other systems have been
     developed in  which the adsorbent is moved countercurrent to
     the  vapor   airstream  or the adsorbent is fluidized in the
     vapor airstream.  When the adsorbent bed is fixed or moves
     as a  unit,  the vapor  concentration distributions in adsorbent
     beds  follow definite  patterns that can be analyzed and the
     knowledge so  gained used  to determine the adsorbent device

          Curve  1  of  Figure 4-13 illustrates the vapor concentra-
     tion  distribution at  the  start  of the adsorption process in
     a  fresh carbon (adsorbent) bed.  The vapor concentration de-
     creases from  Ci  to a  concentration approaching zero in a
     relatively  short distance.  Theoretically, zero concentration
     is attained only at infinite bed length, but for practical
     purposes  concentrations of 0.01 Ci to 0.001 C-j are considered
     essentially zero, hence the finite bed lengths.

          At a later  adsorption time, the vapor distribution
     curve has lengthened  as illustrated by Curve 2.  If adsorp-
     tion  is continued, the vapor distribution curve ultimately
     stops lengthening or  reaches a  steady state.  Curve 3 shows
     the  distribution curve when it  has just attained the steady
     state; Curve  4 shows  it when it has moved down the bed some
     distance  in its  entirety.  The  length of the distribution
     curve is  the  adsorption zone length, LZ.  The bed length,
     Ls,  following the adsorption zone is the saturated bed
     length.   In this section  of bed, the adsorption and desorption
     rates are equal; the  bed  is at  equilibrium with the influent
     concentration Ci.  If adsorption is continued, the adsorption
     zone  ultimately  starts to penetrate the bed as illustrated by
     the  lower end of Curve 5.  Length LZ has remained constant
     but  Ls has  increased.  If it is undesirable to permit any
     appreciable vapor penetration,  adsorption in the bed is dis-
     continued by  diverting the vapor-airflow to a fresh carbon
     bed.  It  is apparent  that the full capacity of the carbon
     has  not been  utilized since L2  is only partially saturated.
     If it is  not  objectionable to allow continued and increased
     penetration  as when a fresh carbon bed can be connected
     temporarily in series with the  nearly spent bed  adsorption
     can  be continued advantageously until LZ has penetrated the
     bed.  At  this stage of the operation, the vapor concentration
     at the effluent  end of the bed  has risen to the level illus-
     trated by Curve  6, i.e.,  Cb  C^.

Figure 4-13  Movement of Vapor Concentration Distribution Curve in Carbon Bed
             with Increased Adsorption Time

          As  the  adsorption  2one  depicted  by Curve 5 penetrates
     the  bed,  the effluent vapor  concentration changes with time
     in  the manner shown  in  Figure 4-14.   (This is an experimen-
     tally determined  curve  for butane on  activated carbons; con-
     ditions  of adsorption and sample calculations are given in
     Appendix  E.)  The service time, tb, can be determined from
     this  curve when the  breakpoint  concentration, Cb, is defined.
     The  curve in Figure  4-14, indicates a service time of 600
     min.  at  Cb B 0.01  Ci and 1100 min. at Cb a Ci.

      * In  air  pollutant control,  the carbon bed can be operated
     under conditions  indicated by Curves  2 to 6 inclusive.  When
     the  pollutant is  strongly adsorbed, and the concentration,
     Ci,  is very  low (1 ppm) and  large volumes of air are purified
     the  resistance to  airflow becomes an  overriding factor.  Under
     thjie_ circumstances  it  is economical  to sacrifice capacity,
     as when  bed  length is limited to those depicted by Curves 2
     or 3.  A  short or  thin  bed length is  used, varying from 1/2
     to 2  inch, and pressure drop is maintained below 0.5 inch of
     H20.   Because the  concentration is low and the pollutant 1s
     strongly  adsorbed, the  vapor distribution curve moves down
     the  bed  very slowly.  Penetration through a fresh bed may be
     1% to 53!, and, after a year's time, the penetration has in-
     creased  to 102.   At  that time the bed is exchanged for a
     fresh  one.

          In solvent recovery the carbon bed is operated under
     conditions depicted by Curves 5 and 6.  In these cases, Ci
     may  be over  three magnitudes larger than in odor removal,
     but  total airflows are considerably smaller.

          Determination of Service Time - To attain the desired
     air  pollution control, or economic solvent recovery, or air
     purification, the service time  is one of the prime factors
     to be considered  in sizing the  sorbent device.

          When deep or long carbon beds are used,  as in solvent
     recovery, LS  is large relative  to Lz and essentially deter-
     mines the service time.  With usual solvent recovery prac-
     tice, the solvent-air velocity  through the bed is maintained
     at near 100  ft/min.  In these cases LZ is of the order of
     3 in. and L  can range from 16 to over 36 in.   The capacity
     of the saturated  bed length, LS, can be calculated with con-
     siderable degree  of accuracy if an adsorption isotherm has
     been  determined,  or it can be estimated if a generalized
     adsorption curve  (Figures 4-11 and -12) is available.  Pol-
     lutant vapor  Vm values to be used in conjunction with the
     generalized  adsorption curves are listed in Table C-I.

 g  40
                     tb = 600
                     Cb * 0.01 C
                                         1	L
  800    1000
t, min.
Figure 4-14  Effluent Concentration Curve of Butane Vapor From An
             Activated Carbon Bed as Function of Time

           The  capacity  of  the  adsorption zone, besides varying
      with  C-j and  T,  also varies with  the carbon mesh size and
      flow  velocity.   The service  time, t^, to penetration concen-
      tration, 05, is dependent on the saturated bed and adsorption
      zone  adsorptive capacities as expressed by the equation,

                                      + w2L2],               (4-23)

      where A-  is  the area  of the  carbon bed, F the solvent-air*.
      flow  rate, us tne  adsorptive capacity per unit carbon bed
      volume in  the saturated zone  and  wz   the mean adsorptive
      capacity  per unit  bed volume 1n  the adsorption zone.  In
      this  equation,  w2L2 is a  fixed amount wherein the L2 1s de-
      fined as  the bed length In which C decreases from C-f to
      essentially  zero,  Ls varies with the C^/C^ ratio; at CK 
      Ci, Ls   L while at Cb s  0,  Ls * L-L2.

           In thin beds  of  2 1n. and less, as used 1n air purifi-
      cation, flow velocities   near 40 ft/m1n. are usually employ-
      ed which  shortens  Lz  down to the 2 in. level.  In these
      applications, the  service time 1s essentially

                                    zLz '                   <4"*4)

           To attain  the desired service time, the dimensions of
      the carbon volumes

                        AC  [Ls  + Lz] or ACLZ

      are optimized with consideration given to resistance to a1r
      flow  that  can be tolerated and the hardware to contain the
      carbon bed.  When  dealing with cylindrical beds, the para-
      meter L/D  1s frequently referred to.

           Characteristics  of the  Adsorption Zone - The adsorpt1v
      capacity  and length of the adsorption zone can be calculated
      if the effluent concentration curve has been determined for
      the flow  system.   This 1s possible only if the adsorption
      rone  has  reached a steady state  before the vapors penetrate
      the bed.   If the bed  is shorter  than L2, the vapor distribu-
      tion  curve  is still changing, and the effluent concentration
      curve reflects  this change.  By  procedures and sample calCul
      tions given  in  Appendix E, page  E-5, the effluent concentratiV.
      curve of  Figure 4-14  was  converted to the vapor distribution
      curve given  in  Figure 4-15.  At  C-f  66.8 ppm, P  5 lb/1n*
      absolute,  T   111F,  28 to 60 carbon mesh size, and 14.5 ft/
      min.  face  velocity, the following values were calculated for

    5  40
                               L, cm
Figure 4-15  Butane Vapor Distribution In Activated Carbon
             Bed at Time When Vapor Has Started to Penetrate

     the unknown parameters of Equation 4-23:
                 1.27 cm
                 0.00738 g/cm3;
                 1.13 cm
                 0.00363 g/cm3;
0.0182 g/g

0.0089 g/g
          wz  is about one-half as large as u>s and on comparing
      the  last two Items about one-third of the service time for
      this particular vapor-adsorbent system is contributed by
      the  adsorption zone.  Since AcuzLz stays constant, all of
      the  adsorption would occur in the adsorption zone if the bed
      were shortened to 1.13 cm and the service time would be re-
      duced to 180 from the 600 min.  On lengthening the bed, an
      increasingly larger portion of vapor will be adsorbed in the
      AcwcLs bed.  At higher gas velocities and with coarser mesh
      carbons, Lz increases in length.

          Critical Bed Depth - A useful concept for determining
      the extra amount of carbon required to accommodate the ad-
      sorption zone is the critical bed depth.  This concept was
      first expressed by Mecklenburg (1925) in the following ser-
      vice time - bed length equation,
                                     - Lc]
where Lc is the critical bed depth.  When experimentally de-
termined tb are plotted against L, a graph of the type shown
by ABC in Figure 4-16 is obtained.  At short bed lengths
the graph is curved as indicated from L  0 to some point B,
and from B to longer bed lengths the graph is straight.  When
the straight portion is extrapolated to tb = 0, 1t intercepts
the L axis at some point now referred to as Lc.  When tjj is
determined at low penetration concentrations, i.e., Cb  0.01
       is positive, when tb is determined at Cb near 0.5 C-j,
       and at Cb u C-j, LC has a negative value.  Equation 4.25
has utility in adsorbent bed size, ACL, determinations at
any penetration concentration although the most frequent
interest is at Cb ^ 0.01 C-j .  An estimate of us as a function
of CT  is obtainable if a generalized adsorption curve is
available by procedures discussed under Potential Theory
Equations.  Lc varies with the air velocity, F/AC; hence,
tb must be known for one L at each F/AC of interest as the
minimum data requirement to use the equation.
        * 0
          The particular service time - bed length graphs in

Figure 4-16  Service Time - Bed Length  Curve  For  Adsorption of Butane on Activated

     Figure 4-16 were determined from data derived from the efflu-
     ent concentration curve in Figure 4-14 and data qiven 1n Appen
     dix E, i.e., at L - 2.4 cm, tb  600 min, F  0.223
5.06 x 10
           0.00738 g/cms  and  Ac    0.502
Lc as calculated with Equation  4-25 and by  extrapolation
Figure 4-16 is 0.57 cm.
          A relationship exists between Lz and Lc which makes it
     possible to calculate Lc from LZ, if Lz and the shape of its
     vapor concentration curve happens to be known.  For low CD ,
     I.e., when very little vapor penetration has occurred and Lz
     is also measured to the same Cfc, the following relationship
     holds with which Lc can be calculated:
[us -
     The quantity /Ci over 0.01,  the  calcu-
     lated Lc would be  considerably larger than the correct Lc.

Table 4-XIII  Lc Calculated from Lz  Using  Equation  4-27

                                         Lc,  ctn
             0.7          0.01             0.57
             1.3          0.02             0.49
             4.0          0.06             0.34
            11.3          0.17             0.19
            27.2          0.41             0.04
            33.2          0.50            -0.01

     Klotz Equation - Klotz did theoretical studies to es-
tablish a relationship between  the vapor concentration dis-
tribution in the adsorption zone and the physical properties
of the gas and carbon.  The equation involves two parts:

                      Lz  Lt + Lr                      (4-28)

Lt 1s a function of the diffusion rate  (mass transfer rate)
of the vapor from the gas stream to the exterior surface of
the  carbon particles.  The equation consists of known vari-
ables and has the form
 L-  1s  the  function  of  processes  occurrinq within  the  carbon,
 such  as  diffusion rate of  the  vapor molecules  alonq the  pore
 surfaces,  accommodation  of the molecules  in  the pores  and
 chemical  reactions.  A factor  that can  affect  both parts is
 the heat of adsorption, which can produce  considerable  rise  in
 the carbon temperature.  While the adsorption  1s  proceeding
 at  Its maximum rate, the increased temperature reduces the
 adsorptive capacity, butafter the rate has  subsided, the
 capacity increases  with the falling  temperature.  The  overall
 effect 1s  a lengthening of the adsorption zone.

      Lr has not been mathematically  related  to the  properties
 of  the system but has  been found to  vary linearly with the  air
 velocity.   The equation has the  following form

                 Lr    2.30  k Um log  [C^Cb]               (4-30)

      Experimentally,  1t has been found  that  Lt 1s very small
 for the smaller molecules  of low boiling points while Lr is
 relatively large.  For example,  Lt  for  the proceeding butane
 adsorption is 0.08  cm, while Lz  is  1.13 cm.   Lr,  by differ-
 ence, Is then 1.05  cm.  Therefore,  the  equation (4-28)  has no
 utility for predicting Lz  of the low Vm vapors.  For  vapors

      of  large  Vm  when  adsorbed  on  coarse mesh  carbons at high
      vapor-air velocities,  the  mass  transfer becomes rate con-
      trolling, and  Lz  is  nearly equal  to Lt.

          The  shape of the  vapor concentration distribution  is
      determined by  log (C-j/Cb), as  Ct  is varied from essentially
      zero to C-j.  For  these  curves

                     C   X  -       AX TT
     where  X    log  [Ci/C[j].  At  C^/C^ = 0.1, 0.01 and 0.001,  the
     values for the  above  factors are 0.61, 0.77 and 0.85 respec-
     tively.  For the  above  discussed butane adsorption, the
     factor at  Cb/C-j =  0.01  is 0.50, considerably below the theo-
     retical  0.77.   This difference is due to the fact that the
     Klotz' theoretical curve, as shown by the broken line curve
     in  Figure  4-15, is not  sigmoid as is the one for the butane
     adsorption.  The  Klotz1 theoretical shape curve is, however,
     frequently observed in  the  adsorption of higher molecular
     weight vapors.

          Equation 4-29 (jives a  good measure of the effects various
     operating  parameters  have on Lt.  The effects produced by
     the parameters  are:
1.    Effect of particle diameter, Dp,
                   Lt - k    DP"*>  k      g               (4-3!)

               where a   C1'6]6.  Lt increases rapidly with  in-
               creased Dp.  Dp

               Effect of  vapor-air velocity. Um cm/sec

                  Lt " k2 um"*1 - k2 }'03 u'"1             (^

               where Um is  the  interparti cle velocity in cm/sec
               and U is the face velocity in ft/min.
Effect of Temperature

       L I
                 , , 0  2 6 -" 0  6 7 r\ - 0  87
                 V      P      o

               The parameters produce opposing effects on Lt, but
               the overall effect  is a decrease  in Lt with in-
               creased temperature; y increases  with temperature

          while p decreases, both producing an increase in
          Lt.  Dv increases with a rise 1n temperature pro-
          ducing a decrease 1n Lt.

     4.   Effect of Pressure

                 Lt " k4 p"*67Dv"0-67                 (4-34)

          An increase in total pressure decreases p and Dv,
          thereby Increasing Lt.

     5.   Effect of molecular size. V cma/mol

                    Lt - k5 Dv--7                    (4-35)

          The molecular size of the vapor affects only the
          coefficient of diffusion Dv.  Dv decreases with
          increase in molecular size; hence, Lt increases.

     Figure 4-17 presents three sets of curves which give Lz
1n Inches as a function of U measured in ft/m1n as calculated
with the complete Klotz Equation 4-28. In the fiqure, Dn 
0.35 cm is the mean particle diameter for 4 to 10 meshpcar-
bon, Dp  0.24 cm for 6 to 16 mesh and DD  0.12 for 12 to
30 mesh.  V is the molar volume determined at 77F.  The
span, Including the 160, 110 and 60 cm3/mol V, covers the
molecular size range of pollutant vapors of interest.  The
calculations were made for adsorption at 77F and Ch/C< was
set at 0.01.                                         1

     In order to use the equation for Lr, a numerical value
was required for k.  In studies performed for NASA under
Lockheed Subcontract 28-5175, 1t was observed that k varies
W1th the vapor pressure of the liquid.  For pollutant vapors
of Interest, an average value for k is 0.008,  The contribu-
tion of Lr to Lz is small; hence, the accuracy of the numeri-
cal value of k is not significant.

     When all fixed quantities are substituted Into Equation4-
28, the equation has the form

      Lz * 0.80 Dp1-*1 U^^Cl/Dy]0'67 + 0.016 U        (4-36)

     To get approximate values for Dv, Equation 10.34, paqe
421 from Jost (1960) was used.  When all necessary constants
or parameters for air are substituted Into the equation   1t
has the reduced form

               7 ofi      1       |HL_*2i--.
               '  <  r-_ j m""i~n*tf I '        I             (4-37}

     1.0 -
     0.0 -
                             40        60
                                 U, ft/mln
Figure 4-17  Effect of Gas Velocity, U, on Adsorption-Zone Length,Lz, for Carbons of
             Mean Particle Diameter. Dpt and Vapor Molar Volume. V, 4 to 10 Mesh *
             0.35 cm.  6 to 16 Mesh * 024 cm and 12 to 20 Mesh * 0.12 on.

where Mi is the molecular weight of the pollutant and a  is
the collision diameter of the pollutant molecule.  a-| is
proportional to V as expressed by the equation

                     a} * 1.328 V1/s.                  (4-38)

     Table 4-XIV gives V and Dy values for some of the most
frequently encountered pollutant or solvent vapors.  Equation 4-
37 is a function of T1'*; 298K has already been substituted.

     Details of the procedure for calculating Lz are given
1n Appendix E, page E-10.

     The curves in Figure 4-17  show that for the 4 to 10 mesh
carbons, L2 is longer than 2 inches at U over 40 ft/min,  Since
coarse mesh carbons are used in thin beds, velocities are
maintained below 40 ft/min to avoid too great a loss of ca-
pacity.  Since the factor

                      h  "zl - 0.77
when Cfc/C-j  0.01, 
           Adsorption of Vapor Mixtures

           The  theoretical  treatment of the adsorption process 1n
      the  prior text of this  chapter deals with adsorption of a
      single  vapor.  In actual practice, the gas phase invariably
      consists  of  two or more components; the mutual effects of
      the  separate  components requires further theoretical con-

           The  simplest system consists of a carrier pas, usually
      air,  and  the  solvent  (or pollutant vapor) as in degreasing
      and  dry cleaning.  Air  is so poorly adsorbed relative to the
      solvents  of  concern that these gas mixtures can be considered
      single  vapor  systems  as far as adsorptive capacity is con-
      cerned.   This holds true with both polar and nonpolar adsor-
      bents.  However, water  vapor is usually present, and in these
      cases the gas phase is  a ternary mixture.  When the adsorbent
      is activated  carbon and the relative humidity is low, the
      gas  mixture  can again be treated as a single vapor in re-
      gard to  adsortive capacity.  At high relative humidities
      water vapor  causes interference although not in the same
      manner  as  would be caused"by a mixture of organic vapors.
      The  presence  or absence of water vapor is the deciding factor
      in the  choice of adsorbents.  When water vapor is present,
      adsorption of organic vapors cannot be effectively carried
      out  with  polar adsorbents.

           In air purification and recovery of solvents used in
      surface coating, graphic arts, etc., mixtures of solvents are
      used.  These  solvents are strongly adsorbed relative to each
      other and  thereby compete for the adsorptive capacity.

          Displacement of Adsorbed Vapors - When an airstream
      containing a mixture of organic vapors is passed through an
      adsorbent  bed, the vapors become separated along the bed
      length.    Vapors most strongly adsorbed are separated out
      near the  influent end of the bed  and  the less-strongly-
      adsorbed  ones separate out in successive bed length segments
      in order  of decreasing adsorbability.  On continued adsorp-
      tion, the  more-strongly-adsorbed vapor displaces the less-
      strongly-adsorbed ones.  The least^strongly-adsorbed vapor
      penetrates the bed first.

          For  polar adsorbents, the ranking of vapor adsorbabiH.
      ties is given in Table 4-II.  Water vapor heads the list,
      hence would displace the less polar vapors lower in the 11st.

          For  nonpolar adsorbents, the ranking of vapor adsor-
      abilities  is  more dependent on polarizability or molar volume
      Table C-I  lists most of the pollutant vapors in decreasing
      order of  molar volume, Vm, as measured at the normal boiling

point.  When there are large differences in Vm within the
mixture of vapors, each vapor with the larger Vm virtually
causes complete displacement of each vapor of lower Vm
When Vm are close together, displacement still occurs but
is accompanied by a considerable amount of coadsorption.

     Profile of Adsorbed Vapors with no Coadsorption - The
profile of the adsorbed vapors in the carbon bed can be
readily calculated on the basis of material balance, esti-
mates of adsorption zone lengths from Figure 4-17 and esti-
mates of adsorptive capacities using the potential  Equation 4-
19 and generalized adsorption curves of the type presented
in Figure 4-11 and -12.  When no coadsorption of adsorbed
vapors is assumed to occur, the adsorption process  is treat-
ed as though each vapor is adsorbed in a separate carbon  bed
but with the beds placed in series.  Figure 4-18 illustrates
the type of adsorbed-phase profile that is derived  with this
assumption for a ternary solvent mixture (Table  1-XX) used
In spray coating operations.

     The mixture selected for the calculations consists of
2-ethoxyethyl acetate, toluene and propanone at  a total con-
centration of 500 ppm.  This is the most probable concentra-
tion used in this operation.  A 3,000 ft'/min flow was used
and bed area then sized to give a 110 ft/min air velocity.
Additional parameter data for the vapor-carbon system are
given in Table 4-XV.  Details of the calculations are given
1n Appendix E, page E-12.

     As shown in Figure 4-18, zones 1, 3 and 5 are saturated
bed lengths.  In zone 2, 2-ethoxyethyl acetate is being
adsorbed while previously adsorbed toluene is completely
desorbed.  Similarly, in zone 4, all of the influent and
desorbed toluene is adsorbed while previously adsorbed pro-
panone is completely desorbed.  In zone 6, all of the influent
and desorbed propanone is being adsorbed.   The zone lengths
and amount of adsorption in each zone are  summarized in
Table 4-XVI.

     By this procedure of estimating, an activated carbon
bed of 63 in. diameter and 28 in. length is required to com-
pletely adsorb the ternary solvent mixture at 3,000 ft'/min
flow and 500 ppm total concentration for an 8 hr service  time
The diameter becomes fixed by the flow rate and  the decision  '
to operate at 110 ft/min velocity.  The bed length is de-
termined by the type of vapors adsorbed, their concentrations
gas temperature and length of service time.  This procedure is
the simplest to work out but gives an overestimate of the bed
length because of two factors which would  have qiven a short
er estimated bed length had they been considered.  These are
(1) coadsorption occurring 1n both the saturated and adsorption

Table 4-XV  Selected  Conditions  for  Adsorption  of  a Ternary
            Solvent Mixture  Used 1n  Spray  Coating  Operations
Total concentration, C-j
     2-Ethoxyethyl acetate

Solvent properties

     2-Ethoxyethyl  acetate

Total  vapor-airflow,  F
Vapor-air  velocity, U
Vapor-air  temperature

Carbon bed area,  Ac
Carbon particle  size
 Carbon density
 Service time, tk
 t  determined at
                        500  ppm
                         50  ppm
                        181  ppm
                        268  ppm

                        3,000 ft3/roin
                          110 ft/min
                           25 C

                        3,930 in.2
                        4 to 10 mesh; 0.35 cm
                        0.50 g/cm3;  0.0181 lb/1n.3
                        8 hr
                        Cb  - 0.01  Ci
 Table 4-XVI   Summary  of Adsorbed-Vapor Profile  Data  of
              Figure 4-18
Zone 1,

Zone 2,

Zone 3,
Zone 4,
       Zone  5,
1 wz
L, in.





  (1)  Molar  volume  V  measured  at  25C.

     Table 4-XVI   (Continued)

                                    L, 1n.     td. Ib/in.
          Zone 6,
L.       2.9
                          Totals    27.6
     zone bed lengths, and  (2) the rise 1n adsorptive capacities
     in zones 3 and 5 because of desorptlon in zones 2 and 4
     causing rises in vapor concentrations in zones 3 and 5. When
     Vm are greatly different and the adsorbed liquids are quite
     incompatible with each other, the estimated bed length 1s
     not too greatly in error.

          Profile of Adsorbed Vapors with Coadsorptipn - When co-
     adsorption and the effects of desorptlon are taken Into con-
     sideration, the adsorbed-phase profile can be considerably
     different as is apparent when Figure 4-19 1s compared with
     Figure 4-18.  The zone lengths and amounts of adsorption in
     each zone are summarized in Table 4-XVII to permit a compari-
     son of the results of the two procedures.  Details of the
     calculations are given in Appendix E, page E-15.

          Coadsorptlon is assumed to obey the following relation-
     ships between adsorption vapor pressure and mol fraction of
     the adsorbed phase:

     For  2-Ethoxyethyl  acetate[E]  CT  *]C0-\ C^d
          Toluene [T]               C2  X2Co2 [wtJ
          Propanone [P]             C3 
     Here Ci , C2 and C, are the respective vapor concentrations
     or partial pressures in the carbon bed.  In Zone 1 they would
     be the influent concentrations, but are different in the
     following zones.  X1 , X2 and X3 are the respective mol frac-
     tions in the adsorbed phase.  Col is the calculated equilib-
     rium vapor pressure of E when the weight of adsorbed E 1s
     assumed equal to the sum, [o>t3 of the actual adsorbed
     weights of E, T and P.  This Is also true for CQ2 and Co3.
     Ci , Co and Co are known for zones 1, 3 and 5.  C ., Cft? and
     c'3 aPe calculated by use of Equat1on4-19 andthegenerSlized
     adsorption curves of type given 1n Figure 4-11  .  CQl, Cft~
     and Cn-5 are determined at several [<>*] until one 1s found c
     that yields XT + X2 + X3  1.0.  X times [wtJ then determines
     the amount us of each vapor adsorbed.


2 -H 3 1*- 4 -*!
         2-Ethoxyethyl acetate
Fiaure 4-19  Adsorbed Vapor  Profile  1n  an  Activated Carbon Bed After
Figure 4 .*                ^  Established with Coadsorptlon

     Table 4-XVII  Summary of Adsorbed-Vapor Profile Data of
     Zone 1 -
     Zone 2-
     Zone 3-
     Zone 4-
     Zone 5-
     Zone 6-
Figure 4-19
, in.
'Ls -6
us for E
ws for T
- us for P
Ws for E
Ws for T
,W$ for P
"Lz 3.6
wz for E, average
wz for T, average
- wz for P, average
Wz for E
W2 for T
,W2 for P
ws for T
 us for P
Ws for T
.Ws for P
"L2 3.3

     The length, Lz,  of zones  2,  4  and  6  are  determined  from
Figure 4-17 and the average  amount  uz of  each  vapor  adsorbed
1s determined by the  factor,

as worked out for Equation 4-26.  The  amount  uz  remaining  on
the carbon in zones 2 and 4 during desorption  is  determined
by the above factor when equal  to 0.23.

     The procedure for carrying out the  many steps  of  the
calculation, as demonstrated in Appendix E,  can be  programmed
and carried out rapidly by a computer when many such calcula-
tions are required.

     The Equation (4-39-41) was proposed by  Grant and  Manes
(1966).  It is most accurate when applied to ideal  solutions
such as might be approached by homoloaous series.  Increased
deviation occurs with mixtures of dissimilar components.

     The difference In bed length calculated by the two pro-
cedures is due primarily to the much  larger  wz when coad-
sorption is taken Into consideration.  Coadsorptfon and de-
sorption produce a small increase in  tos.

     Under the conditions of these calculations,the coadsorp-
tion decreased the bed length from 27.6  in.  to 20.6 in.,  or
255J decrease.  When the mixture consists of  very similar
vapors, such as the naphthas used in surface  coatings,  and
where the differences in Vm are small, coadsorption occurs
to the greatest extent.  If the service time is shortened,
the Ls bed lengths decrease first and thereby the bed lengths
predicted by the two procedures diverqe more.   Conversely,
for long beds or low air velocities the predicted bed lengths

     Single Versus Mixture of Solvents -  In surface coatings,
graphic arts, rubber products and other solvent  user Indus-
tries, solvent mixtures  are used  to attain  the desired vis-
cosity, solvent power and evaporation rate  of the reactant
mixture.  When solvent  recovery  is an important  economic
factor, consideration should be  given to  the possible sub-
stitution of a single solvent for the mixture.   Frequently
the single solvent vapor can be  recovered in a smaller carbon
sorbent unit.  To  illustrate this, carbon bed  lengths were
calculated to show the  bed  length required  to  adsorb each of
the previously discussed solvent  vapors each at  500 ppm  in-
fluent  concentration.   As  shown  below,  the  calculated  bed
lengths required for single solvent  vapors  are  compared  with
ternary mixtures:

          Solvent	^

          2-Ethoxyethyl acetate, singly at 500 ppm       13
          Toluene, singly at 500 ppm                     14
          Propanone, singly at 500 ppm                   24
          Ternary mixture, no coadsorption, 500 ppm      28
          Ternary mixture, coadsorption, 500 ppm         21

          Resistance to Airflow

          An important  parameter that enters into the determina-
     tion of the optimum L/D ratio is the airflow resistance of
     the adsorbent bed.  This information is generally available
     in the sales literature of adsorbent manufacturers, but, 1f
     not available, it can be estimated with the following equa-
     tion (CEP, 1952).

     By substitution of the fixed variables for air at barometric
     pressure and 25C, the equation 1s reduced to the form:

          AP  2.66xlO's Hllll  + 1.02xlO- Hill !il .   (4-43)
                           e'   Dp*              e'   Dp

     AP is in inches of water per inch of carbon bed depth, U  1$
     in ft/min and Dp in cm.   The pressure drop is sensitive to
     the intergranular void volume fraction,  E, which varies with
     the way the  carbon is  loaded into the adsorber and  with the
     mesh  size distribution.   For a tightly packed bed, e can be
     as low  as 0.30  and for a loosely packed  bed as high as 0.45.
     To show the  variation  the tightness  of packing has  on AP,
     AP were calculated for e between 0.30 and  0,45, when U 
     40 ft/m1n and mesh size  fraction Is  6 to 16 (0.24 cm Dp).
     These results are as follows:

                   _e	         AP, 1n. H^O/in.  carbon

                   0.30                0.51
                   0.35                0.28
                   0.40                0.18
                   0.45                0.10

     The AP  can vary five fold.   If e decreases  due to particu-
     late  matter  lodged between  carbon  particles,  the  increase
     In AP would  be  considerably greater.

     Regeneration of Activated Carbons

     The preceding text of this chapter discusses methods
for analyzing the adsorption phase of the adsorption-regen-
eration cycle.  In this part of the chapter a  procedure Is
described for estimating the amount of regenerating  agent
required to regenerate a spent activated carbon.   In this
area considerable knowledge has been accumulated  by  the sol-
vent recovery systems designers and activated  carbon manu-
facturers.  This knowledge is quite limited to operations
that are profitable, i.e., concentrations generally  above
1,000 ppm and recovery of the more common and  more expensive
solvents.  Air pollution control operations, however, cover
the concentration range from below 1 ppm to over  1,000 ppm
and include vapors that have no recovery valve.   In  the con-
centration range 10 ppm and over,regeneration  of  the carbon
1s an economic necessity in order for the adsorption systems
to be competitive with thermal incineration, catalytic In-
cineration or liquid absorption.

     When no past experience data are available,  the esti-
mating procedures for determining the carbon bed  size and
regenerating agent requirements perform a valuable service
1n regard to economic analyses and preliminary choice of
type of control system best suited for the control problem
under consideration.

     Desorption of Adsorbed Vapors - Figure 4-20  Illustrates
the adsorbed-phase profile of a single vapor at the service
time when the vapor has just started to penetrate the bed at
some predesignated concentration Ct>.  This is  the usual or
desired status of the system just prior to regeneration.
Figure 4-19 Illustrates an adsorbed-phase profile for a ter-
nary mixture.  Its regeneration will follow the general pat-
tern observed in single vapor regeneration.

     Figure 4-21 illustrates the progression of the adsorb-
ed-phase profiles as increasingly larger amounts  of regen-
erating agent are passed through the bed in reverse flow
direction.  The numbers do not Indicate equal  time Intervals
nor equal amounts of regenerating agent.  Adsorbed vapor at
profiles numbered 1 and 2 are easily desorbed requiring rel-
atively small amounts of agent.  Increasingly larger amounts
of agent are required to desorb to profiles 4, 5 and 6.
With reference to the generalized adsorption curves, Figure
4-11, desorption at 1 and 2 is occurring at smaller A
values, while at 6 it is occurring at larger A values.   In
terms of pore structure desorbed vapors at profile 1 and 2
are coming from micropores of larger diameter, while vapors
at 6 are coming from micropores of small diameter.  For
reference, pore size distribution curves are presented  1n



             Direction of
             air-solvent flow
                  I  I  I  I J	I  J^ I  I  I  I	I  II  I  1  lit  ^Vj
                                 L. in.
20 21.3
    Figure 4-20  Adsorbed-Phase Profile for Trlchloroethene
                 at Service Time when Vapor Starts to
                 Penetrate Bed, BPL V Type Carbon
                             i  i  i  I  i  lit
                                 L,  in.
20 21.3
    Figure  4-21   Adsorbed-Phase Profiles  for Trlchloroethene
                 at  Various  Stages  of Regeneration,  BPL  V
                 Type  Carbon

Figures 4-1 and -2.   Very large  amounts  of  agent are  usually
required to attain complete  regeneration.
     Procedure for Estimating Regenerating  Agent  Usage
"theoretical plate" approach  1s  used  In  these  estimates.
The carbon bed length is  divided into theoretical  segments
with thermal and mass transfer equilibria  assumed to  occur  in
each segment as the gas mixture  passes through.   The  hot  gas
mixture entering the segment  imparts  heat  and, according  to
the modified Polanyi potential (4-19), I.e.,

                   [T/Vm] log [C0/C]

or as further defined

                 A - [T/Vn] log [C0/C],                     (4-44)

the temperature rise increases the vapor pressure C.   The
released vapor is swept out of the segment causing an in-
crease in A.  The rise in A registers the  amount  of vapor
desorbed.  By a summation procedure the  vapor transfer over
successive segments is determined as  a function of the input
of regenerating agent.

     Because of the complexity of the calculations, they  are
done most efficiently  by computer.  To illustrate the pro-
cedure and supply data for economic analyses presented 1n
Chapter 6, regeneration calculations  were done for three
solvents and four carbons after equilibration at four dif-
ferent influent concentrations.  The  steps for preparing  the
computer programs and  for carrying out the computations are
described 1n Appendix  D.

     Results of Regeneration  Estimates - The calculations
cover steam and air as regenerating  agents at the following

                   Steam  at 212F
                   Steam  at 260F
                   Air    at 302F  (150C
                   Air    at 392F  (200C

The  results of  the  air regeneration  are applicable to other
noncondensable  gases.  The three  solvents  Investigated are,


They represent a  wide  range  of  adsorbabiHtles,  essentially
bracketting most  of the  pollutants that are  of concern.  The
four carbons,  listed  below,  were  chosen to represent four
diverse  types  of  generalized adsorption curves.

            Barnebey Cheney GI Type, density 24 lb/fts
            Pittsburgh BPL II, density 31 lb/fts
            Pittsburgh BPL V
            Pittsburgh BPL VI

     The GI represents a highly active carbon with pore structure
     similar to S154, Type IV In Figure 4-1.  The BPL Types II, V
     and VI are the same carbon, but represent different gener-
     alized adsorption curves (Figure 4-11 and -12) by different
     investigators on greatly different types of compounds.  The
     Generalized Adsorption Curves II and V are representative of
     most vapor phase carbons.

          In regard to flows  and bed dimensions the following
     additional conditions are stipulated:

          Influent flow,  F              3000 ft'/m1n, stp
          Influent flow velocity,  U      110 ft/min
          Bed  area, Ac                    27.2 ft2, 3930 in2
          Bed  length,  L                 21.3 1n. BPL,  23 1n. GI
          Adsorption-zone length,  Lz     3.3 in.
          Carbon weight                 1500 Ib
          Mesh  size                     4  to 10

          Figures  4-22, -23 and  -24 present  the calculated  results
     selected to illustrate some of the  regeneration  situations
     that  can arise during pollution  control  operations.  Each
     curve was  calculated for  a  1500  Ib  carbon  bed  (I.e., L 
     21.3  in. for BPL and  23 In.for  GI) that  had been  exposed
     (equilibrated)  to  the given concentration  until  the  bed was
     saturated  to  its entire length.  The  abscissa  gives  the mean
     weight  of  adsorbed vapor  remaining  on the  carbon  after the
     amount  of  regenerating agent,as  given on  the ordlnate,  has
     passed  through  the bed.

         Necessary data  are presented in  the  curves  to  permit
     calculation of: (1)  the  total  regenerating agent  through-
     put,  (2)   total amount of vapor  desorbed,  (3)  total amount
     of adsorbed vapor remaining on the carbon, and (4)  amount
     of regenerating vapor required to desorb a unit weight of
     vapor.  A  typical problem would  be the estimate of the re-
     quired  amount  of 212F steam to  regenerate the carbon bed
     on a one-hour  cycle  for trichloroethene at 3,000 ppm Influ-
     ent concentration.   From  Figure  4-22, the  saturation capa-
     city is 0.363  Ib solvent/lb carbon or for  1500 Ib the total
     amount  of  solvent 1n the  carbon  Is 545 Ib.  To operate on
     a 60 min cycle,180 Ib of  the solvent must  be desorbed on
     each regeneration leaving 365  Ib on the carbon.  The 365 Ib
    gives an average 0.243 Ib solvent/lb of carbon.  From Figure
    4-22, the  amount of steam required to desorb to 0.243 1s
    0.30 Ib/lb carbon.  Total  steam  requirement Is then 450 Ib

-o  1.0
             10 ppm
             	 212F steam

             	 260F steam
100 ppm
                       I  I I  I
     1000 ppm
                    I  .
                     0.1            0.2

                      u,  Ib  solvent/lb  carbon
   Figure  4-22   Amount  of  Steam  Required to Regenerate BPL V
                Type Carbon  Equilibrated with Trlchloroethene
                at Varied  Concentrations

                                              	1	TT	
         - 212F steam

         - 260F steam
                                           3000  ppm
                                     1000 ppm
                                         I,.. I
                   L.  I    I
                          (i),  1b  solvent/lb carbon
     Figure 4-23  Amount  of  Regenerating Agent Required to Regen-
                  erate BPL  V  Type  Carbon Equilibrated with Pro-
                  panone  at  Varied  Concentrations


             0.1    0.2     0.3    0.4     0.5

                     CD, Ib solvent/lb carbon
Figure 4-24  Amount of Regenerating Agent Required  to  Regen-
             erate GI Type Carbon Equilibrated with 4-Methyl-
             2-Pentanone at 10 and 3000 ppm Concentrations

and the pounds of steam used per pound of solvent recovered
are then

               0.30/CO.363-0.243] - 2.5.

For deep beds of  16 to 36 in.,this method gives a quick
but approximate method for estimating regenerating agent

     Since most solvent recovery operations are carried to
service times at low vapor penetration, some adjustments are
required to take into consideration the effect of the adsorp-
tion zone.  This can be done by programming the adsorption
zone Into the computer, or it can also be estimated from the
adsorbed-vapor profiles.   The profiles shown 1n Figure 4-21
are for trichloroethene when the adsorption phase of the cy-
cle is at 3000 ppm concentration.  They were read-out when
the computer was operating on the 3000 ppm curve 1n Figure
4-22 for 212F steam.  The computer calculated profile ex-
tends from L  18 1n., In reverse direction, to L  0 and
theoretically to L - -3.3 in.  The parts of the profiles 1n
LZ are extropolations of  the computer calculated curves.
The effect of Lz is  that  between 10 to 12 Ib additional
steam 1s required.   The steam requirements for the L  0 to
18 1n.  profiles  lengths are still the same as originally
calculated for L  21.3 1n. I.e., the efficiency of steam
utilization increases as  the bed length 1s increased.  To
determine the amount desorbed, the area under each profile
1s subtracted from the pounds adsorbed at the time of the
Initial  penetration, I.e., when CD * 0.01 Cj.  The total
pounds  adsorbed  comes to

                 Ls  x ws  + Lz x wz  W,

                           0.77; 03, * 0.23 u>s
        and  18 x 25.6 + 3.3 x 0.23 x 25.6 - 480 Ib

     Table 4-XVIII summarizes the pertinent quantities calculated
     from the profiles.  The general pattern as established for
     this solvent-adsorbent system appears in other systems.   The
     regeneration time, as given in Column 2, 1s calculated for a
     steam flow rate of 250 fts/m1n measured at standard temper-
     ature and pressure.

          When Lz is taken Into consideration, 2.8 Ib steam/lb
     solvent is required to desorb  180 Ib of trlchloroethene,
     Profile 3.  The regeneration time Is established at 40 m1n
     and steam requirement at 500 Ib to recover the 180 Ib of
     solvent per cycle.

Table 4-XVIII  Regeneration of Trichloroethene-Activated
               Carbon System Based on  Figure  4-21  Profiles

         Regeneration     Steam     Solvent        Lb  steam
Profile   time, min     input, Ib  desorb.,lb per Ib solvent

   1          17           210         79          2.7
   2          33           410        160          2.6
   3          40           500        180          2.8
   4          65           810        240          3.4
   5         145          1820        310          5.9
   6         640          8040        410          20
     Operating Capacity - An examination  of Figures  4-22,
-23 and -24, and the results of Table 4-XVIII  show that  the
amount of regenerating agent required increases  rapidly  as
desorption is carried closer to completion on  each cycle.
In order to maintain a profitable solvent recovery operation,
the extent of desorption, now termed the  operating capacity,
must be closely gauged to minimize the steam-to-recovered
solvent ratio.  Table 4-XIX gives steam-to-solvent ratios
and operating capacities for the three solvents  when the
vapor influent concentration is 3000 ppm  and the adsorption-
regeneration cycle time is one hour.  Two to four pounds of
steam for each pound of solvent recovered are  considered
optimum ratios; 4-methyl-2-pentanone and  trichloroethene fall
in this range with the particular sized system under discus-

     The steam-to-solvent ratio of 6.1 for propanone recovery
is on the high side,  It can be lowered by decreasing the
operating capacity, but to do this will require an increase
in bed size.  There is, however, a limit to the extent the
bed size can be increased since steam requirements to heat
the carbon and associated hardware  increase.

      In these particular recovery operations the  steam re-
quired to heat the bed to 212F and  supply the heat of de-
sorption is  80 to 100 Ib.  This steam condenses and
remains in the bed.  The balance of  the steam, 200 to 400
Ib, serves as a sweep gas.

      As the  influent  concentration  decreases the  service time
increases and the amount of  regenerating agent required also
increases.   In Table  4-XX this  type  calculated data are pre-
sented when  the influent concentration is  10 ppm.  The  steam.
to-solvent  ratio  for  each vapor comes to a minimum  at some
operating capacity.   For 4-methyl-2-pentanone the minimum is  at
0,025 Ib/lb  operating capacity, for trichloroethene at 0.025 lb/
 Ib and  for  propanone  at  0.005  Ib/lb.  The  service  time  1n  each

Table 4-XIX  Estimated Operating Capacities and Steam Requirements for Recovery of
             THchloroethene,  4-Methy1-2-pentanone and Propanone for a One-Hour
             Adsorptlon-Regeneratlon Cycle Time, Carbon Equlllbrated at 3000 ppm

used, Ib

Lb steam per
 Ib solvent


(1)  4-Methyl-2-pentanone adsorbed on GI type carbon and the other two solvents
     on BPL V type carbon.

            Estimated Steam Requirements to Regenerate Carbons*  *  Equilibrated with
Solvent (Pollutant)



pentanone, Trl
on and at Vari
cap., Ib/lb
chloroethene and Propanone at 1
ed Operating Capacities
time, hr
desorb. ,1b.
, 212*F
used, 1
"  *"
m m~
0 ppm Influen
Steam Used

Ib steam per
b Ib solvent

"* *

(1)  All calculations are for BPL V type carbon


      case  1s  sufficiently  short  so  that steam regenerations, even
      though  large  amounts  of  steam  are required, are much less
      costly  than carbon  replacement and reactivation.  The longest
      service  is for  4-methyl-2-pentanone at 0.16 Ib/lb saturation
      capacity.  The  610  hr at 24-hr days is equivalent to a one-
      month service time.

           When  the influent concentration is down to 1 ppm
      the  pollution control  operations change.  At this concentra-
      tion  it  would take  about one year to equilibrate the carbon
      with  4-methyl-2-pentanone.  The cost of steam regeneration
      is  at the  same  cost level as carbon replacement and reacti-

           Selection  of  Regenerating Agent - As the results pre-
      sented  in  Table 4-XX  show,  the amounts of steam required for
      regeneration  are large relative to the amounts of solvent
      desorbed.  Miscible solvents (or pollutants) as represented
      by  4-methy1-2-pentanone  and propanone are completely dis-
      solved  in  the large amount  of  steam condensate.  Their re-
      coveries by distillation are not economical, thereby creating
      a disposal problem  if water pollution is to be avoided.  In
      these cases the solution to the problem is to regenerate the
      carbon  with a noncondensable gas and burn the released vapors
      immediately 1n  a small thermal  incinerator.  If the amount
      adsorbed is small  so  that the  release rate never exceeds 25%
      LEL,  then  air can  be  used as a regenerating agent; otherwise*
      an  inert gas  such  as  flue gas  would be safer.  By the pro-
      cedures  already demonstrated,  the noncondensable agent re-
      quirements can  be  calculated.   The cost of using nonconden-
      sable gases is  lower  than that of steam since no fuel is re.
      quired  for latent  heat of vaporization.

           When  the condensed  vapor  1s Insoluble 1n the steam
      condensate, or  slightly  soluble as is trichloroethene, com-
      plete or partial separation of the organic liquid 1s then
      possible by decantation.

           When  the desorbed organic liquid 1s soluble, but on
      thermal  Incineration  would  repollute the air, as by release
      of  SOg.  HC1 or  N02i the  problem of pollutant disposal becomes
      more  difficult.

           At higher  influent  concentrations where the steam-to-
      solvent ratio decreases, the advantages of using nonconden-
      sable regenerating  agents decrease.  In the concentration
      range where solvent recovery is practiced, steam is the  only
      practical  regenerating agent.

     Some calculations were carried out to  determine the
performance of 260F steam.  The general  conclusions are
that in the high concentration range,  above 1000 ppm, it
offers very little advantage over the  use of 212F steam.
At times the use of 260F steam serves a useful  purpose for
the periodic removal of high molecular weight molecules,
which when present in the  airstream,  accumulate in the in-
fluent end of the carbon bed.  When regeneration with 260F
steam is carried to low solvent contents, the steam condensed
in the early part of the regeneration  is vaporized leaving
a dry carbon bed.  It then behaves as  a noncondensable gas
but is more costly to use than air or  flue gas.

     Generalization on Adsorbent System - The results of
the calculations on adsorption and regeneration  discussed
in this chapter were used to work up the costs of operating
adsorbent systems.  These costs were compared to those for
the competing systems, thermal incineration and  catalytic
incinerations.  A number of generalizations were arrived at
in regard to the circumstances under which each  system might
be used advantageously.  These generalizations are presented
1n Chapter 1, pages 20 to 24.

           5.   Theories  and Types  of Catalysts



     The general characteristics of the process  termed
"catalysis" were outlined by Berzelius.  In 1835,  he codified
most of the chemical reactions developed to that time and ob-
served that many reactions were affected by the  presence of a
nonreacting  substance.   The process appeared to be initiated
by a "catalytic force" which separated or broke  down the re-
actants prior to the chemical reaction.  He used the term
"catalyst" to define that substance whose presence affected
the rate of a reaction while remaining essentially unchanged

     Modern scientific technology has significantly altered
the "catalytic force" concept proposed to explain catalysis.
The fundamental principles of catalysis are being gradually
unveiled through the combination of modern analytical instru-
mentation, continued research and the correlation of previous
data.  Mechanisms and theories describing the process of ca-
talysis continue to evolve as the result of these investiga-


     Although catalytic  reactions have been  used for over  a
thousand years, theories  describing these reactions  have been
actively developed only  during  the  last sixty years.  The
works  of Freundlich  (1915),  Langmuir  (1918)  and Sabatier (1920)
have been  used  as the foundation for  most of these  theories.
Catalysis  has a dual nature  involving  both physical  and  chemi-
cal aspects which have stimulated much of  the research  to-
wards  a description  of its  complex  nature.


     Three factors  have  been  established  as  criteria  for de-
scribing  catalysis:  (1)  the  catalyst  remains chemically  un-
changed,  (2)  the  reaction rate  can  be correlated  to the  a-
mount  of  catalytic  reaction  surface,  and  (3) the  catalyst
does not  affect the  equilibrium concentrations.   Examples  of
catalysts, which  conform to  these  criteria,  are  available  for
solid, liquid and gaseous states, with both  positive and
negative  influence  on the chemical  reaction.  A  concise defi-
nition and evaluation of the theories and  types  of catalysis
can  be found  in Kirk-Othmers'  Encyclopedia  of Chemical  Tech-

          Air pollution control systeirs are almost exclusively
     confined to solid-gas, or heterogeneous contact catalysis.
     The definitions, formulae and cursory examinations of theo-
     retical models presented in this chapter pertain to hetero-
     geneous catalysis.

          Reactions of gases at a solid catalyst surface present
     one of the more important alternatives available for air
     pollution control.  However, it is necessary to understand
     and use some theoretical models and mechanisms to discern
     its prominence in pollution control systems.

     Theoretical Models


          The prominence of surface and surface effects in hetero-
     geneous catalysis is evident in the theoretical models de-
     veloped to describe these reactions.  Theories of catalysis
     fall into four general categories:  (1) geometric, (2) elec-
     tronic, (3) boundary layer and (4) crystal field.  Research
     and development of theoretical models for catalysis have
     lagged behind the applications of catalysis, but have been
     instrumental in its recent growth.

          The general idea in each of the theoretical models has
     been a more thorough elucidation of the "chemisorbed inter-
     mediates" proposed by Sabatler (1920).  The successes attain-
     ed by these various models have increased the use of catalysts
     and opened such fields as air pollution for their application.

          Geometri c

          Geometric models for describing catalysis rely heavily
     on Langmuir's (1918) adsorption theory and surface area
     measurements by Emmett and Brunauer (1937) and Brunauer,
     Emmett and Teller (1938).  Correlations between molecular or
     atomic volume and surface area or pore sizes were made to
     evaluate and predict catalytic activity.

          Thlele (1938), Weisz (1954) and Wheeler (1951) did ex-
     tensive work 1n this area.  Wheeler (1955) derived the

                           h - 6\/K/rDv                       (5.!)

     to relate the reaction kinetics, pore size, surface area,
     catalyst specificity, apparent reaction order and poisoning.
     Balandin (1929) studied dehydrogenation and Beeck (1950)
     studied hydrogenation in relation to the geometry of various
     metal  catalysts which engendered useful correlations.


     Dowden and Reynolds  (1959)  are  primarily  responsible
for the development of the  electronic  theory of  catalysis.
Catalysts are classified  as conductors,  semiconductors  or
Insulators.  This  theory  proposes  the  slow transfer  of  elec-
trons between catalyst and  adsorbate as  the rate determining

     Boundary Layer

     Recent work in catalysis has  concentrated on the geo-
metric and electronic states of surface  layers only, bringing
it into the realm of the  boundary  layer  theory.   The influence
of surface defects, dislocations,  holes  and terminal surface
atoms has been suggested  by many researchers.   These reactive
centers have been used successfully to describe many types  of

     Crystal Field

     The partial success  of electronic and boundary layer
theories has been used as  the basis for applying the crystal
field  theory.  As proposed by Dowden and Wells  (1961)  and
elaborated by  Mango and Schacktschnelder  (1967), this  theory
has  proven satisfactory for those few systems for which the
detailed quantum mechanics of surface atoms can  be  calculated.



     Adsorption of  gases on  solid surfaces  is one of the
primary  steps  involved in  catalysis.  Its  fundamental  and
unique character requires  some  explanation before attempting
any  evaluation of  catalytic  processes.

      Heterogeneous  catalysis  requires at  least  one  of  the
components to  be adsorbed  on the  catalyst surface before
chemical  reaction  can occur.   This  adsorption concentrates
atoms  at the interface between  the  two  phases and decreases
the  free energy of the surface.  The  free energy decrease  or
entropy loss makes  the process  exothermic. The energy re-
 leased during  adsorption  is  available for the catalytic re-
 action (activation energy).

      Theoretical  Principles

      Since adsorption 1s  a basic  part of catalysis, some dis-
 cussion of theory  is applicable.   The adsorption isotherm
 developed by Langmuir (1918) has  traditionally  been used as

     the basis for adsorption models.  His premise of kinetic and
     surface limiting factors has been established by many re-
     searchers.  The equation is derived from the kinetic theory
     of gases:

                dn - KMS6, to give e   
for catalyzed systems  have been reviewed by several  authors.
Kougen and Watson (1943) developed mathematical  expressions
to describe catalytic  reaction mechanisms.   Yancr and Hougen
(1950) presented a summary of experimental  methods and mathe-
matical expressions for determining rates and mechanisms.
Similar work has been  reported by Corrigan  (1954), Hwang and
Parravano (1967), Otake, Kumiqlta and Nakao (1967) and Lin
(1969).  Brodski et al. (1967) have also discussed methods
for determining the accuracy of theoretical mechanisms on
the basis of kinetic data.

     Mathematical descriptions of reaction rates have been
developed empirically by Arrhenius:

                      K  A exp  E                      (5-6)
and from statistical mechanics by Eyring:

               K  A RT exn AS* exn '-'AH*                (5-7)
               K   A NTT exp IT  exp FT  '               *    ;

The accuracy of these  absolute reaction  rates  can  always be
improved by more complex  notations Involving more  constants
and requiring  sophisticated mathematical  treatment.

      The general results  obtained from these equations  pre-
sent  an  interesting  paradox in the evaluation  of  catalytic
systems.   If these  equations  are used  to calculate the  ratio
of  heterogeneous and homogeneous reaction rates,  the cataly-
tic reaction rate  appears considerably smaller.   By making
similar  assumptions  in the calculations  of reaction rates,
1t  can be  shown  that the  ratio

        Kheterogeneous  ^ 10-io to  10-is  exp AE/RT.

      In  actual practice  this theoretical deficit is sur-
mounted  by the decrease in activation  energy and the use of
 large surface  area catalysts.  This  rati.o denotes the sig-
 nificance  of the relative activation enerqies  for catalyzed
 and uncatalyzed reactions.  It follows, therefore, that cata-
 lysts which strongly adsorb the reacting species perform best.

      Extraneous interferences may occur  in three areas which
 render these rate calculations inoperable for some  applica-
 tions of catalysis.  Reaction chains, extraneous  chemical
 interactions and catalyst attrition affect the reaction rates
 but  do not conform  to simple time functions.


          Thermal calculations require data from four general
     areas:   (1) heats of reaction,  (2) heat transfer coefficients,
     (3) reaction rates and  (4) thermal properties of system. These
     data permit the calculation of  heat release as a function of
     the catalytic reaction.

          A volumetric heat  release, for the reaction in the cata-
     lyst bed, can be described as:

                    q/Vcat  X'Q.t.Ppollutants               (5-8)
     Hlavacek  (1970), Wilhelm, Johnson and Acton (1943) and Lin
     (1969) have published different mathematical solutions to
     this proportionality.

          By assuming isothermality over small integrated sections
     and a linear variation of the reaction rate, Wilhelm, Johnson
     and Acton  (1943) have generated the equation:

                                           ld"                (5-9)

     for small, cylindrical sections of the catalyst bed.  In this
     calculation, HQ is the value (q/Vcat) for the initial re-
     action of  the "clean" catalyst in equation (5-8). Hj Is the
     rate of change of (q/Vcat) as the catalytic reaction pro-
     gresses; r 1s the radius of the catalyst bed and Al 1s an
     incremental length.

          Heat  Is lost from these sections of the catalyst by two
     mechanisms: (1) radial conduction equal to


     where K Is the thermal conductivity, and (2) volumetric con-
     vection equal  to


     where Kf is the film heat transfer coefficient.

          Equating  the heat generated and lost through this cylin-
     drical  section generates the Identity:
               -2irrAA K ^1 + J  2irrAl Kf (Tsond"Tgas )dr     (5-10)

Equation (5-10) may be used to describe any thermal change
occurring in the system.

     Solutions to these general  equations form a series of
curves or parabolas depending upon the system parameters.  If
appropriate data are available for the system, solutions can
be obtained for the rate  of heat release and reaction temp-
erature.  Additionally, residual or equilibrium concentrations
can be predicted.

Types of Catalysts

     Although many substances can be prepared to act as cata-
lysts, only two types are prominent in the heterogeneous re-
actions prevalent in air pollution control.  Noble metals,
such as platinum, are the most efficient and commonly used
metal catalysts.  Transition metals, such as vanadium, are
the second type.

     Noble Metals
     The noble metal catalysts (platinum, palladium, gold
and silver)are usually deposited on an inert support to in-
crease the exposed surface area.  They are most efficient for
catalyzing the oxidation reaction of hydrocarbons and maintain
their reactivity to high temperatures.

     These catalysts are susceptible to poisoning reactions
with acid gases and erosion by particulates in the gas stream.
Depositing them on metal carriers does, however, increase
their resistance to abrasion.

     Transition Metals

     The most  common transition metal  catalysts include:
vanadium, nickel,  iron,  cobalt, manganese and  copper.  Both
the pure metals and thin oxides have catalytic properties
which may be  used  in air pollution  control systems.

     Transition metal  catalysts may be poisoned by  absorption
of  inert gases, chemical reaction and  erosion.  Although  they
are easily poisoned, their lower cost  and greater versatility
are an advantage for air pollution  control systems.

Methods  and  Parameters  Used  to  Evaluate  Catalysts

      Catalysts vary in  their  reactivity  and  efficiencies  for
oxidizing  the various  air pollution emissions.   Many of  the
parameters used to evaluate  catalysts  are dictated  by the
operating  conditions of a particular  process.

          Optimum properties of a catalyst include:  hiqh deactl-
     vation temperature, resistance to poisoning, low initiation
     temperature and high efficiency.  Although these criteria
     remain unchanged for all catalytic reactions, the means for
     obtaining the desired characteristics change with reaction
     type, reactants and reaction media.

          Various catalysts have been studies by Anderson (1961),
     Stein (1960) and Dmuchovsky (1965) to define the "best"
     catalysts for effective oxidation of pollutants in terms of
     the above parameters.  In general, the noble metals are the
     most efficient for hydrocarbon oxidations.  The transition
     metals follow in the order:  Co>Cr>Mn>Cu>Ni>V>Fe>Ti>Zn.

          Catalytic activity is usually insensitive to the amount,
     chemical composition or the type of support used.  The ac-
     tivity of metals deteriorates at elevated temperatures, and
     any catalyst selection must be made with thought of both
     efficiency and attrition at the operating temperature.  After
     the selection of the catalyst, only the surface area and
     temperature are variable.

          Unfortunately, the development of an optimum catalyst
     for a given set of conditions is not as yet relegated to the
     realm of scientific selection.  Most of the synthetic work
     is proprietary with manufacturers and has not received much
     scientific correlation.  In addition, the codification of
     catalysts is complicated by the combination of physical and
     chemical aspects which usually make the rate and products
     unique for each catalysts-reactant system.

              6.   Package  Sorption  Systems


     The study of sorbent  types  and so'ption  theories,  as
presented in Chapter 4,  demonstrated that,  of the  several
types of adsorbents available,  activated  carbons are  better
suited for gaseous air pollutant control  than the  others.
Polar adsorbents  such as silica  gel, alumina  gel and  mole-
cular sieves, adsorb water preferentially to  the organic
vapors while, conversely,  activated carbons adsorb organic
vapors preferentially to water  vapor.   Since  moisture is
invariably present in pollutant  airstreams, only activated
carbons can function effectively over any practical  length
of time.  Activated carbons have an added advantage  in  that
their adsorptive  capacities are  also larger for the  organic
vapors.  The quest for adsorbent systems  suitable  for air
pollution control thereby  narrowed  down to systems using
only activated carbon.  Polar adsorbents  can  be of limited
Interest in the adsorption of inorganic add  vapors  or
where the airstream 1s dry.

     Activated carbon adsorbent systems,  as presently manu-
factured for industrial  and commercial uses,  are oriented
toward "air purification"  and "solvent recovery".   In the
air purification  application, the objective is to  remove
pollutants at trace level  concentrations  (1 ppm or less)  to
make the air suitable for breathing purposes.  In  the solvent
recovery application, the  adsorbent system is used to recover
a solvent from a  process exhaust stream.   Profitable recovery
is attained at concentrations above 500 to 1000 ppm, the
lower level depending on the value  and adsorptive  properties
of the solvent.  The technology associated with these two
applications is designated as "present technology" in this
handbook and is discussed in considerable detail  1n this

     Pollutants are, however, emitted into the atmosphere
at concentrations ranging from less than 1 ppm to  over
10,000 ppm.  The  concentration range from 1 to 500 ppm (the
neglected range under present technology) is  an important
range in regard to pollutant emissions.  Much of the "new
technology" is concerned with development of  control systems
and procedures for this concentration range.   These develop-
ments are also discussed in considerable detail in this
chapter with emphasis on theoretical support  for the various
concepts proposed.

          Both the air purification and solvent  recovery  systems
     are also suitable for applications where the  primary objective
     is the purification of air exhausted to  the atmosphere.   Air
     purification systems would be effective  where the  pollutants
     are emitted at very low concentrations  but  need  to be controlled
     because of their highly malodorous or toxic nature.   These
     systems are cost effective only at low  concentrations because
     of their nonregenerative design.   At concentrations  above a
     level somewhere in the range 1.0  to 10  ppm, the  cost of  more
     frequent carbon bed replacements  would  raise  the air purifi-
     cation costs to levels above those of competing  systems.  Sol-
     vent recovery type systems in basic design  are regenerative
     systems wherein the activated carbon bed can  be  regenerated
     within the system.  At concentrations above 10 ppm,  these
     systems purify air more economically than the nonregenerative
     type and their development,  strictly for pollution control,
     forms the basis of the "new  technology".  Activated  carbon
     regenerative systems whether for  pollution  control or solvent
     recovery are designated In this handbook as carbon resorb

     Basic Functions

          Table 6-1 presents the  overall  pattern of air purifi-
     cation and pollutant emission control in regard  to pollutant
     concentration, odor level  and emission  sources.  The types
     of pollutants encountered  at the  various  levels  that require
     control vary considerably  and thereby the mode of  control
     also varies.

          In the concentration  range below 1.0 ppm, the primary
     concern is with pollutants that are  highly  odorous or to
     some degree toxic or those that are  oxidants  as  In smog.
     Pollutants of this type originate  from  combustion  of fuels,
     incineration, and process  industries such as  meat  rendering.
     Some industries emit reactants such  as  carbonyl  chloride
     which are toxic at very low  concentrations.   Occupied spaces
     such as kitchens, bathrooms, offices, conference rooms,
     theaters and hospitals quickly develop malodorous  atmo-
     spheres if not properly ventilated or the air purified by
     some means.

          Solvents used by the  solvent  user  Industries  do not
     contribute significantly to  the low concentration  odor prob-
     lem except when the solvents are  reactive to  solar radiation
     and thereby contribute to  smog formation.   Except  for 4-
     methyl-2-pentanone and xylene, solvents  have  odor  recogni-
     tion threshold concentrations that are over 5.0  ppm  and  TLV
     over 25 ppm; see  Table 2-IV  for odor and  Table 2-III for  TLV
     Toluene and xylene, the solvents  emitted  1n largest  quantl-  *
     ties, are highly  reactive  smog precursors,  see Table 2-VIH.

 Table 6-1   Correlation  of Pollutant Concentration  with  Odor  level.  Emission  Source,
            and Type of  Adsorbent System Required
 Pollutant cone.,

  0.05 to 0.1
   0.5 to 1.0

     5 to 10

    50 to 100

   500 to 1,000

 5,000 to 10,000

50,000 to 100,000
Qualitative odor

Moderate odor
Heavy odor

Heavy odor

Moderate odor

Heavy odor

Intolerable for
continuous ex-

Intolerable and
deadly even for
Intermittent ex-
     Emission sources
Seldom used space, non-
airconditioned offices

Highly polluted outdoor
air, conference rooms

Emissions from process

Industrial plants

Solvent emissions from
Industrial operations

Solvent emissions from
industrial operations
Safe only in closed cycle
operations with inert
Control device
                                                                       "  purification


          The basic function of control devices applicable in the
      low concentration range is to purify the indoor air and/or
      impure air brought in from outdoors so that it will be safe
      for breathing.  The control devices used for this purpose
      are designated in this handbook as air purification systems.

          When the would-be pollutant is a valuable solvent and
      is emitted in the 500 to 10,000 ppm concentration range,  it
      is usually profitable to recover it for reuse with a solvent
      recovery system using steam for regeneration.  Although the
      motivation is solvent cost sayings, the solvent recovery
      system also functions as an air pollution control  device.
      However, many solvent using operations utilize no control
      devices.  In such cases the odor level can be very strong
      near the emission sources.  On dilution with air mass move-
      ments, the concentrations become low enough to become non-
      detectable as odors.  This point is reached with many sol-
      vents at 100 ppm and almost all solvents are nondetectable
      at 5 ppm.

          A third concentration range extends from about 5 to
      500 ppm.  At the lower end of the range, the detectable and
      also offensive odors are those from meat and fish rendering
      plants, from incomplete incineration of refuse, from waste
      water treatment plants or any process where thermal or bio-
      logical degradation takes place.  At the upper end of the
      range, the pollutants are primarily solvents within the
      plant or neighboring areas to the plant.  Pollutants In this
      (5 to 500  ppm) concentration  can reach inhabited areas and
      cause damage to humans and property.  Prevention of pollutant
      emissions  here Is a direct cost to the pollutor with the
      benefit of purer air going primarily to the neighboring
      inhabited  areas; hence, their reluctance to spend funds for
      this purpose.   This is the pollutant control area that re-
      quires considerable study to  solve the many problems associ-
      ated with  it.

      Air Purification Systems.  Present Technology

          General  Discussion

          Air purification is generally carried out in a partially
      closed system, i.e., the carbon-purified air is redrculated
      into the occupied spaces.   When the air purification device
      is an Integral part of the air conditioning system, make-up
      air from outdoors is added to the recirculating airstream
      in quantities  ranging from 0% to 25% of the total  volume.
      To maintain  the C0 concentration below 1.0% - the upper
      acceptable  concentration level  - outdoor air (with CO? con-
      tent of 0.03%) is required at make-up rate of 2 ft'/mln per

person at rest and at 5 ft'/roln ptr person at work.   This
establishes the minimum make-up air Input rate when  the
space per person ratio 1s relatively small.   In large spaces
where occupancy Is small and 1nterm1ttant, make-up air may
not be required to maintain a low C02 concentration  level.

     A make-up air Input rate of 5 ft'/wln per person 1s
also sufficient to maintain the oxygen concentration above
the 17* lower allowable limit.

     If humidity and temperature are controlled by dryer and
cooler outdoor air, the make-up air rate would be larger than
the 5 ft3/m1n per person as required for C02 control.

     When the heating and/or cooling of the occupied spaces
are a function of the ventilating system, the air redrcula-
tlon rate would be determined by their requirements.  The
reclrculatlon rate thus established also determines  carbon
bed area.  For low pollutant emission rates the bed  area
thus determined may be  larger than necessary.

     Recommended fresh  air or purified redrculated air
requirements for general offices are  15 ft'/m1n per person;
for private offices, 25 ft*/m1n per person; and for factor-
ies, 10 ft3/ro1n per person.  The minimum  requirement  for
factories  1s 0.10 ftvmln  per square  foot of  floor  area  or
as governed by local codes  (Perry Chemical  Engineering  Hand-
book, 1963).  According to  ASHRAE  (ASHRAE Handbook  of Funda-
mentals,  1967), the recommended fresh air or  purified reclr-
culated air to an occupied  space with considerable  air  pollu-
tion as by tobacco smoke 1s  15  ft'/mln  per  person*

          The above analysis establishes the redrculatlon  rate
     at a minimum of 15 ftj/m1n which Includes  at least  2  ftj/m1n
     of outdoor air.  Large auditoriums seating 2,000  people
     would then require an air purification device that  can
     purify 30,000 ft'/min of air.  For a small room 1n  a  home,
     the other extreme 1n size, the required air handling  capacity
     would be about 50 ft'/min.

          Figure 6-1 presents a flow diagram of a central  ventila-
     tion system that incorporates the carbon bed and  particulate
     filter elements as integral parts of the overall  system.  Not
     shown are the heating and/or cooling elements that  are usually
     required.  The ventilation systems may be  one of  earlier
     installation to which the particulate filter and  carbon  bed
     are attached.  It is then necessary to find Installation
     space for the air purification elements and also  install  a
     larger blower to overcome the added airflow resistance.

          When no central ventilation system exists or It  is
     not feasible to make the Installation of the air  purification
     elements, air purification for the occupied space can  still
     be accomplished with a self-contained unit or true  package
     sorbent device.  These consist of a cabinet with  Influent
     and effluent airflow grilles, a blower, particulate filter
     and the carbon bed.

          Small air purification devices consisting of one  or
     two radial flow canisters and a blower are available  for
     very small rooms.

          Carbon Bed Designs

          Common Characteristics - The carbon beds 1n  air  purifi-
     cation systems have very low L/D or L/Ac ratios.   L/D 1$  of
     the order of 0.01 where D is expressed as the diameter of a
     circular bed of area Ac.  For comparison L/D for solvent
     recovery systems are of the order of 0.1.   L 1s very  short
     or more appropriately stated as being thin; depths  of 0.5,
     0.75, 1.0, 1.5, 2.0 and 3.0 1n. are 1n frequent use.   Bed
     areas are sized to control the airflow velocity 1n  the 20
     to 60 ft/m1n range with most units operated at about 40 ft/
     min.  Coarse mesh carbons of 4 to 6, 4 to 8 and 4 to 10,
     fractions are used.  The thin bed, the relatively low air-
     flow velocity and coarse mesh all contribute to low airflow

  Occupied spaces

          Outdoor make-up  air
Figure 6-1
          Central Ventilation System Showing Particulate
          Filter and Carbon Bed for A1r Purification
resistances which are generally below 0.25 1n. of HoO pressure
drop.  An exception to the above are the air purification
systems used 1n chemical-bacteriological-radiological (CBR)
warfare service.  Beds are thicker, usually 1.0 In. and over,
and  mesh size ranges are 6 to 8 and 12  to 30.  Consequently
the  pressure drop  1s over 0.25  1n. of HgO.  Safety against
toxic  agents 1s  the  overriding  factor.

     The thin, expanded-area carbon beds are attained by
retaining the carbon between reinforced metallic screens or
perforated  sheets  that are spaced apart to give the desired
bed  depth.    The  retainer   can be  shaped  Into  various
forms   to   provide   advantages  or  economtes   in  the
assembly, operation  and replacement of  the carbon bed.  The
retainer can be  1n the form  of  a rectangular panel,  also
designated  as a  cartridge or tray.  A number of the  panels
are  Inserted into  a  cell  having a designed air-handling
capacity of some definite flow  volume such as  500,  1000  or
2000 ftVmln.   The retainer  can also be one  continuous  piece
folded back-and-forth  or  pleated  across the  depth  of the cell.

     The fold  at  each end  can be rounded OP v-shaped.  Another
     approach,  a  carry-over from the gas mask, 1s the use of
     radial  flow  canisters.  A number of canisters are attached
     to a manifold within  the cell.

          The  cell is a unit of fixed or recommended capacity
     which can  be combined with others to form a multiple cell
     unit of any  desired total capacity.  The multiple cell unit
     may be  held  together  in a frame and inserted into the ven-
     tilating  duct,  thus appearing as an enlargement of the duct.
     In large  installations, the multiple cell units are usually
     located in large walk-In cabinets or rooms.  A very small
     air purifier may only consist of a small blower and one or
     two radial flow canisters.

          Different  manufacturers or fabricators specialize 1n
     one or  more  of  the types of cell constructions described
     above.  A  list  of carbon adsorbent systems manufacturers
      and/or  vendors  1s  given  later  in  this  chapter.

          The  following text briefly describes the various types  of
     carbon  beds  commercially available.

          Panel and  Pleated Carbon Beds - Figure 6-2 Illustrates
     a nine-cell  carbon filter, a cross-sectional side view of a
     cell and  two carbon panels.  The dimensions of the panels
     produced  by  different fabricators vary some, but are gen
     erally  in  the ranges  2 by 2 ft, 2 ft by 8 in and 1 ft by
     8 1n.   The cell front Is usually 2 by 2 ft and depth varied
     from about 2 ft to 8  1n.  The number of panels per cell can
     vary from  10 to 24.   Carbon loading for a cell varies be-
     tween 45  to  over 95 Ib per 1000 ft'/min airflow  rate.
     Ease of removal and replacement of the carbon bed 1s an
     important  consideration in regard to operating costs.  With
     this type  cell  the carbon panels can be readily pulled out
     and replaced by simply removing the front grille which holds
     the panels in place.  Seals are not a critical problem ex-
     cept for  the CBR filters.  In some carbon filters, such as
     used in some toxic agent air purification systems, the panels
     are parallel rather than slanted toward each other.

          Figure  6-3 shows one type of CBR filter for a small
     protective shelter.   The carbon bed area is one square foot,
     hence at  the 48 ft3/tn1n flow the air velocity is 48 ft/m1n.
     This filter  meets the shelter requirements of up to 12
     people, which gives a minimum purified air change of 4 ft1/
     min per person.

   Carbon filter  /
          Side view cross
          	 section of
                    Carbon panels
  Moure 6-2  Multiple Cell Activated Carbon Filter for A1r
    8         Purification

      Figure  6-3  A Chemical-B1ological-Radiological Filter   S1?P
                   12 by 12 by 8  1n'..  Operated at 48 ft9/mfn'for
                   Small Protective  Shelters.   Courtesy MSA Co.
               BPACINO MMOER
               TOCAIMKXM BCD
                                               PERFORATED METAL
                                               (2* GAGE IN CARBON STEEU)
                                               (2 OAOE IN STAIIslLESS STEEL)
                                               NON PERFORATED METAL
                                               (I" MARGIN)
                                               SEPARATE PLEATS TO
                                               RETARD SETTLING
                                               NON PERFORATED BAFFLE
                                               FLAIsKjC TO SEAL AGAINST
                                               NEOPRENE PAD
                                               OED DEPTH (MIN. I")
                                               SPONGE NEOPRCIS4C PAD TO
                                               PREVENT BYPASSING OF AIM
                                          SPONGE NEOWteNC FACCOASKET
      Flqure 6-4  Pleated Type Carbon  Filter  for  Control of
                   Radioactive Pollutants,  Size  24 by
                   11.5  In.,  Operated at
                   Courtesy  of MSA Co.
1000 ft'/mln
24 by
Fl ow.

     Dimensionally the  pleated  carbon  beds  are  not qreatly
different from the panel  type.   Fiqure 6-4  shows  the  qeneral
exterior appearance and a cutaway  section  shows  the  inter-
nal construction.   This carbon  filter  or  cell  is  of partic-
ularly ruqqed desiqn for  use in nuclear reactor facilities
to control  radioactive  pollutants that miqht  be released
accidently.   Its operating capacity is 1,000  ft3/min  at 1.0
in. of water pressure drop and  pollutant  removal  efficiency
qreater than 99.99%.  The cell  size is 24 by  24 in. front
dimensions  and 11.5 in. deep.   With 12 horizontal  sections
the carbon  bed area is  approximately 20 ft2  and air flow
velocity 50  ft/min.  Bed  depth  L is 1.0 in,  or  deeper.   For
non-CBR use, pleated bed  cells  are simpler  in construction,
L is usually less  than  1.0 in., and coarser  carbon is used.

     The pleats may be  horizontal as in Fiqure  6-4 or
vertical.  To unload the  spent  carbon, the  cell is removed
from the multiple  cell  filter and one  of the  plates at the
end of the  pleats  removed,

     Canister Carbon Beds - Fiqures 6-5, -6 and -7 show:
(1) the construction of one type of canister, (2) their
assembly into the  cell  and (3)  a small self-contained air
purifier.  Specifications are qiven for two canisters, a
Type 28 with 0.5 in. bed  thickness and a Type 42 with 0.75
in. bed thickness.  The mean area, between outer and inner
retainer screens,  for Type 28 is approximately 0.73 ft2
and for Type 42, approximately  0.76 ft2.  The"recommended
maximum capacity for Type 28 canister  is 30 ftVmin.  At
this flow the oressure drop is  0.15 in. of H20 and flow
velocity approximately 41 ft/min.  For Type 42 canister
the recommended maximum capacity is 25 ft'/min which qives
a 0.16 in.  of HgO pressure drop  and 33 ft/min  flow velocity.
For a filter or cell of 1,000 ft'/min  flow capacity, 33
canisters of Type 28 are needed  and 40 of Type 42.

     Fiqure 6-6 shows  a filter where  the manifold  plate is
in  a slanted position.  Vertical and  horizontal plate  posit-
ions are also used.  In this filter construct!on,the canister
is  held to the  plate by means of a bolt  passed throuqh the
center.  By  looseninq  a wing nut,the  snent canister  can be
easily removed  for  replacement with a  fresh  canister.  The
flow direction  is  from the  outside to  the center  of  the
canister.  This flow direction  is  preferred  to a  flow  in  the
reverse direction  because  it qives a  shorter adsorption zone
     Materials  of  Construction

     The carbon-cartridqe  frames,  cell frames  and other
enclosing panels  and supporting  structures are mild  steel
when  pollutant  vapors  are  noncorrosive.   Under  operations

       Figure 6-5
   Radial Flow
Design for Air
       Courtesy of Connor Engi-
       neering Corp.
Figure 6-6   Arrangement of
Canisters on a Slanted Mani
fold Plate for Medium and
Large Air Purification Sys-
where considerable moisture  can  enter  the  cells  or  corrosive
vapors are present, special  coatinns  are  applied to the
steel.  Zinc-nlated and chromate finishes  are  mentioned  by
some manufacturers.  Some attention  is  also  qiven to pre-
vention of air leakage around the cartridqes.   This can
be done by interlockina construction,  felt seals, or caulk-
ing.  The place where corrosion  is most likely to occur  is
at the interface between the carbon  and retainer screen.
Moist carbon or carbon containing corrosive  pollutants  es-
tablish galvanic cells at the interface and  corrosion pro-
ceeds rapidly.  Stainless steel  or plastic retainer screens
are required under conditions where this  can happen.

      Operation of Air Purification Systems

      Operating Parameters - To  attain the most effective and
economical service from  air  purification  systems,  various
parameters need to be optimized.  These parameters are flow
velocity, carbon mesh size,and  bed thickness.  They determine
the resistance to flow,  the  adsorptive capacity  (assuming
selection of carbon  has  already  been made) and  carbon bed
area.  Another important factor  is the determination of  the
optimum  service time  at  which the carbon  bed  should be re-
placed.   Temperature  is  not  a controlling factor in air-
purification as it very  often is  in other pollutant control
applications.  Occupied  spaces  are generally  in  the 50  to
90F  temperature  range,  a range in which  activated  carbons
perform  effectively.   Pollutant concentration  is also an
important factor  that  should  be recognized  because of its
effect on capacity and  service  time.

      Flow Velocity - An  increase in  flow  velocity  produces
an  increase  in  pressure  drop through  the  bed  and an  increase
in  adsorption  zone length.   Both increases  are  undesirable.
An  increase  in  pressure  drop raises  the  power reguirements
to  operate  the  blower, and  an increase  in  adsorption  zone
length decreases  the overall  adsorptive  capacity.   Converse-
ly  a  decrease  in  flow velocity  produces  a favorable effect
in  these two properties  of  the  system,  but  the decrease  in
flow velocity  is  accomplished by increasing the bed area.
The added panels, pleated  cells or  canisters  represents  an
 increase in  investment,   When the pollutant is not extreme-
 ly  toxic, a  balance  in cost factors  has  been worked out so
 that most systems have bed  thicknesses of 0.5, 0.75 and
 1.0 in,  and  are operated at about 40 ft/min flow velocity,
 A coarse mesh size falling  in the range  4 to 10 mesh is
 used,giving pressure drops  that are  below 0.25  in. of
 water.  Some sacrifice is  made  on complete purification;
 the efficiency specification is about 9856.

          Filters for toxic agents contain larger mesh  carbons,
      12 to 30, and are thicker, 1,0 1n.  and over.   Efficiencies
      over 99.99% are required.  The pressure drop  is  higher  be-
      cause of the smaller carbon granules and thicker  bed.

          Pressure Drop - The pressure drop through the carbon
      bed is a function of the carbon granule size, the size  dis-
      tribution, the packing of the bed and flow velocity.  Equa-
      tion 4-43 shows the relationship between the  pressure drop
      and flow velocity, mean particle diameter and void volume,
      e, between the granules,  e is dependent on the  mesh size
      distribution and degree of packing  of the bed.   e can be
      calculated if the particle density  of the carbon  and bulk
      density of the bed are known, but usually e is the unknown

          Figure 6-8 presents several curves that show  the effect
      of mesh size fraction and flow velocity on the pressure drop.
      These curves were recalculated from curves given  in Union
      Carbide Corp. Catalog Section 3-6450, Data Sheet  No. 1.

          Service Time - The service time of the filter 1s a
      function of the flow velocity, influent pollutant concentra-
      tion and the adsorptive capacity of the selected  carbon for
      the particular pollutant under consideration. Equations
      4-23 and -24 establish the relationships between  these
      variables for flow systems in which the bed is equal to or
      deeper than the adsorption zone length at steady  state.  Be-
      cause of the coarse carbon granule  size and thin  beds that
      are used in air purification systems, the bed depth is
      usually considerably shorter than the adsorption  zone.  Equa-
      tion 4-24 does not then apply, except to indicate an upper
      limit on service time.

          From Figure 4-17, Lz at U - 40  ft/mln is  1.7  in. for low
      molar-volume odor molecules In coarse 4 to 10 mesh carbon
      beds.  Lz increases to 2.3 in. for  high molar-volume odor
      molecules.   When carbon beds of 0.5, 0.75 and 1.0 in. thick-
      ness are used, pollutant penetration of the bed occurs  long
      before the adsorption zone has attained steady state.   The
      adsorption zone is in a transitory  state during  the whole
      service time as depicted by Curves  1 and 2 in Figure 4-13.
      Pollutant penetration of the bed occurs early in  the service
      time but because of the low Influent concentration the
      effluent concentration rises very gradually with  time.  High
      99.99% pollutant removal efficiencies are not attainable,  but
      at 98% efficiency, practical service times are attainable.


2  o.io:
                             10     20

                             U,  ft/m1n
100   200
  Figure 6-8  Pressure Drop Through Granular Carbon Beds of
              Varied Mesh Size Fractions as Function of Flow

           When a finer 12 to 30 mesh carbon  is  used,  Lz  range
      is from 0.6 to 0.8 in (at Cb/C-j  0.01).   In  carbon beds of
      1.0 In.and over,the Lz has attained steady state and is
      contained within  the bed depth.  This  state is  depicted  by
      Curve 3 in Figure 4-13.  Pollutant removal efficiencies  of
      99.99% and over are attainable because  the adsorption zones
      of length greater than 0.6 to 0.8 can  be  contained  in the
      bed, i.e., zones  measured to Cb/Ci <0.01.   See  Chapter 4,
      Dynamic Adsorption Process for further  discussion.

           The service  times attained vary from  6 mo%  when the
      odor concentration is described as heavy,  to  much longer
      periods of up to  2 yr, when odor concentration  is light.
      Odor concentrations are not often quantitatively defined 1n
  Jr   terms of ppm since odors are not always identified  with
      specific compounds and a direct quantitative  measure made.
      Operating efficiencies of 98% (1.e.,Cb/Cj   0.02) are con-
      sidered acceptable and when the efficiency has  decreased
      to about 90% (i.e., Cb/C-j  0.10), the  carbon bed 1s re-
      placed .

          Equation 4-24 was used to estimate tb at  L  Lz,
      U   40  ft/mln  and  Ci - 0.5,  0.1  and   0.05 ppm  for
      4  to  10 mesh carbon.  4-Methyl-2-pentanone was  used to repre-
      sent the odor and BPLV the vapor phase  activated carbon.  For
      this system and conditions, Lz  2.1 in. when Cb/C^   0.01.
      Since the bed depth 1n common use are considerably  less than
      Lz, the true tb for L
the first 23 pollutants listed in Table 2-IV.   To this  list
may be added smog-created vapors, odor vapors  qenerated in
washrooms and two additional  odor vapors emitted  from ren-
dering plants.  The odor vapors are:

Pollutant                                   Source

Skatole, 3-methyl-l-benzo(b)pyrrole        Washroom
Indole, l-benzo(b)pyrrole                  Washroom
Putrecine, 1 ,4-diaminobutane               Washroom
Cadaverine,  1,5-diaminopentane             Washroom

Propanoic acid                             Rendering
2-Methyl propenal                          Rendering

     The recognition threshold concentrations of the above
23 designated pollutants range from the lowest, 0.0002 ppm,
for trimethylamine to  the highest, 1.0  ppm, for nitrogen
dioxide.  The recognition threshold concentration for ska-
tole has been reported at 7.5  x  10"8  ppm and 0.019 ppm by  two
different investigators.  Threshold concentrations were not
found  for the others listed above,but  they are expected to
be low  although  higher than 0.0002 ppm.

     The odor recognition threshold places an upper  limit
to the  odor-vapor  concentration  level  that the air  purifi-
cation  system can  deodorize.   With intermittent  release of
odor vapors  and  redrculatlon  of the  air,  the air  purifi-
cation  system need not remove  the  odor vapor  concentration
all the way  down to  the  recognition threshold on  a  single
pass.   The  repeated  passage through the carbon bed  will
bring  the  odor concentration  down  gradually  to and  below
the recognition  threshold.  Also,  odor vapors  are  not
equally obnoxious  so that some may be tolerated  consider-
ably  above  the threshold concentration.

      An air purification system  operating  at  98% efficiency
can deodorize air containing  trimethylamine  in  a single
pass  if the concentration does not exceed  0.01  ppm.   With
recirculation the initial  concentration level  can be higher
but  the odor will  be detected for.a  longer time.  For skatole
the  initial odor vapor concentration  can be  at  about 1.0  ppm,
 For  nitric oxide the Initial  concentration can  be 50 ppm,
 but  at this concentration,  and down  to at  least  10 ppm,  the
 carbon resorb system would  be more economical to operate.
 Odor  vapor concentrations  in  the range 1.0 ppm and down are
 realistic for the airflows  to be deodorized with the above
 described air purification  systems.

          Carbon Bed Replacement - To attain the maximum but un-
      interrupted service from a carbon filter, 1t would be desir-
      able to monitor the progress of the adsorbed-phase profile
      through the carbon bed so that the service time could be
      determined at which the carbon filter should be replaced.
      The other-wise generally low capital cost of the cell (about
      $60 to $80 per 1000 ftj/min capacity at manufacturing cost)
      does not warrant high expenditures for monitoring instru-
      ments that can determine this service time precisely.  How-
      ever, a determination can be made of the service time by
      means of a test element attached to the Influent face of one
      of the cells.  From prior experience with the same purifi-
      cation system or with other similar systems and operating
      conditions, an approximate service time is already known.
      Prior to the estimated service time, the test element 1s
      removed from the cell and sent to the system's manufacturer
      for analysis and determination of the remaining service

          The spent panels, cells or canisters are shipped to the
      system's manufacturer where they are emptied and reloaded
      with reactivated or virgin carbon.  Special procedures are
      used in the reloading process to Insure tight packing of the
      carbon and thereby avoid later settling of the carbon during
      transit or usage.  Poorly packed beds can develop open areas
      through which the Impure air can pass.  The carbon is re-
      turned to its manufacturer where It 1s reactivated with high
      temperature steam to its original capacity.  Some handling
      and activation losses occur on each cycle, which can be as
      high as 5%.  Over a number of cycles, the used carbon is com.
      pletely replaced with virgin make-up carbon.  The air puri-
      fication systems user gets credit for the returned spent
      carbon; the cell replacements are less costly than the Ini-
      tial purchase.

      Solvent Recovery Systems, Present Technology

          General Discussion

          The solvent recovery systems have been developed to a
      high degree of efficiency and reliability.  The technology
      on solvent recovery is also well documented 1n literature.
      The purpose of this part of the chapter 1s to summarize
      this technology and in the next part of the chapter to ex-
      plore ways whereby this technology can be expanded to appU.
      cations that are unprofitable but concerned strictly with
      decreasing pollutant emissions.

     Process engineers  and  equipment  designers  have  directed
their efforts to improve  the  economics  of  solvent  recovery
operations.   These  efforts  have  been  directed  toward decrease
of power or  fuel requirements for  regeneration,  reduction of
pressure drop during the  adsorption phase,  increase  of  ad-
sorptive capacity by utilizing residual  capacity of  adsorp-
tion zone, and improved methods  for recovery of  the  desorbed
solvent.  The result has  been the  development  of three  types
of systems differentiated by  the manner in  which the carbon
bed is maintained or handled  during both phases  of the  ad-
sorption-regeneration cycle.   These are:

     Fixed or stationary  bed
     Moving  bed
     Fluidized bed

The selection of the system depends on  the type of solvent
vapor and the circumstances under  which it is  to be  recovered,

     The most efficient and economical  service is  obtained
from the solvent recovery system,  regardless  of type, if the
airflow is kept at  a minimum but in  keeping with proper
ventilation.  In a  plant with well defined operating units
that emit pollutant vapors, the airflow can be kept  low by
use of closely fitting hoods  over  the operating units and
the hoods connected by way of ducts  to the pollutant control
system.  By use of close fitting hoods, sufficient draft is
created to draw the pollutant away from the point of emis-
sion with use of minimum airflow.   Such situations can be
taken advantage of in degreasing (Figure 1-1), dry cleaning
(Figure 1-2), rotogravure and flexographic presses,  the
various surface coatings applications, rubber products
dipping operations, paper impregnating operations  and sur-
face coating and ink drying ovens.  When a new plant is
being contemplated, the pollutant control  system should be
considered as an integral part of the plant.  This would
involve an effort on the part of the  machinery designers
and plant layout men.

     Figure 6-9 presents a flow diagram of a dual, sta-
tionary bed, solvent recovery system with auxiliaries for
collecting the  vapor-air mixture and  for solvent  separation
from steam condensate.  The  vapor-air mixture,  collected
from the  various point sources, passes  through  the  particu-
late filter and into the on-stream carbon adsorber,  in  this
case the  upper  one.  The effluent air, which is virtually
free of vapors, is usually vented outdoors.  In this dia-
gram the  blower precedes the  particulate filter;  in  many
applications  the blower follows the particulate filter.  The
lower carbon  adsorber  1s regenerated  during the service time
of  the  upper  adsorber.  A steam generator or other  source of

          Hoods or direct connections to point emission sources
                     on stream
                                         Carb^ont _bed

        Steam  or  hot
       gas  generator
      Figure 6-9
Stationary Bed Carbon Resorb System with
Auxiliaries for Vapor Collection and Solvent
Separation from Steam Condensate or Incin-
eration of Pollutant Vapors

steam is required.  The effluent steam-solvent mixture  from
the adsorber is directed through the  condenser and  the  liqui-
fied mixture then run into the decanter and/or distillation
column for separation of the solvent  from the  water.

     Other carbon resorb systems can  be used  1n place of  the
dual, stationary bed system as the circumstances  warrant.

     Stationary Bed Systems

     Figures 6-10 to -13 show two designs of  stationary bed.
solvent recovery systems.  The type shown in  Figures 6-10  and
-11 employs vertical cylindrical beds wherein  the solvent-
laden air flows axially down through  the bed.   This  particular
design is advertised for use 1n recovery of solvents used  in
degreasing and dry cleaning although  equally  well suited  for
recovery of solvents from other Industries.  Solvents men-
tioned are trichloroethene, tetrachloroethene, toluene, Freon
TF and dichloromethane.  Regeneration 1s with  steam  up-flow
through the bed and, since the above  mentioned solvents are
Immiscible in water, decantatlon 1s used to separate the
condensed solvent from the steam condensate.   The valves  are
disk type, open and close by pistons.  Water  Is used as
coolant 1n the condenser.  Steam, electric power to  drive  the
blower, and cooling water are three operating cost Items.
Cost of steam and cooling water Increase with frequency of

     Figure 6-11 shows the external features  and arrangement
of the component parts of a two-adsorber system.  This  par-
ticular unit 1s a V1c Model 572 AD, with two  adsorbers  used
alternately, I.e., while one adsorber Is on-stream the  other
1s regenerating.  The recommended airflow capacity Is  3,000
ft'/min for each adsorber of approximately 5  ft diameter  and
bed length near 20 in.  This adsorber model was used as a
basis for the calculation of bed depths and adsorbed-phase
profiles, Chapter 4, page 59, and regeneration of activated
carbons, Chapter 4, page 67.

     Dual adsorber systems are also operated  with both  ad-
sorbers on stream 1n parallel when the solvent concentration
is low.  Regeneration Is less frequent than every hour  and
may be done during off-working hours.  Operation In parallel
almost doubles the air handling capacity.  The capacity of
Model 572 AD Increases to 5,500 ftVmln.

                                 & air
          Adsorption phase

  Desorption phase
     Figure 6-10  Cut-Away Diagram of Solvent  Recovery Adsorber
                  Showing Vapor-Air Flow Pattern  During the Ad-*
                  sorption and Steam Flow During  Desorption.
                  Courtesy of Vic  Manufacturing  Company.
     Figure  6-11   Stationary  Bed  Solvent  Recovery  System of 3000
                  ft3/min  Flow  Capacity for  Each Adsorber.  Used
                  for  Uater  Immiscible Solvents. Courtesy of  Vic
                  Manufacturinq Company.




^7r 	 \ '
\L- - J/
f ^^ y ^'
J :

-7, 	 A
-Wv %J

, r


                        LOW PRESSURE STEAM
Figure  6-12   Flow Diaqram of a  Solvent Recovery System
              Courtesy of Vulcan-Cincinnati, Inc.

Finure 6-13
Solvent Recovery  Svsten  for Recovery  of  Manhtha
fron a Rubher  Products  Plant, 12.0DQ  ftVnin
 low Capacity.  Courtesv of Vulcan-Cincinnati
I nc.

           For a one-hour adsorption-regeneration cycle time, the
      sales  literature reported recovery of trlchloroethene 1s 225
      Ib  and steam requirement for regeneration 1s 675 Ib, giving
      a steam.to-solvent ratio of 3.0.  This Is typical for a high
      solvent concentration  level.  The solvent recovery and steam
      requirements as given  in Table 4-XVIII for Profile 3 are 180
      Ib/hr  of solvent, 500  Ib/hr of steam giving a steam.to-sol-
      vent-ratio of 2.8 at solvent concentration of 3,000 ppm.
      The  good agreement between the steam-to-sol vent ratios 1s
      an  Indication of the reliability of the methods used to cal-
      culate the theoretical 2.8 ratio.

           The largest vertical or up-right adsorber constructed
      by  Vic Manufacturing Company has an airflow  capacity of
      5,000  ft3/min with carbon bed diameter of about 6 ft.  The
      overall length and width dimensions are 18 and 11 ft and
      overall height 11 ft.  The smallest has a 700 ft'/mln air-
      flow capacity and overall dimensions 7 by 5 ft, length and
      width, and 11 ft height.  All of these units fit Into the
      general concept of packaged sorption devices.

           The horizontal tank type adsorber as Illustrated 1n
      Figures 6-12 and -13 are structurally more amenable to large
      capacity solvent recovery systems than the vertical type, al-
      though Vulcan-Cincinnati, Inc. engineers systems as small
      as  2,000 ft3/m1n.  The 2,000-ft3/m1n system, designated as
      Model  22, has two adsorbers, overall length and width di-
      mensions are 18 and 7  ft and overall height 1s 13 ft.  The
      largest system, designated as Model 403, ha.s 40,000 ft'/mln
      air  flow capacity, overall length and width dimensions are
      52  and 27 ft and overall height is 26 ft.  Their largest
      skid mounted system, Model 102, has design capacity of
      10,000 ft3/m1n, overall length and width dimensions are 38
      and  12 ft and overall  height 1s 16 ft.  Units of size Model
      102  are basically off-the-shelf packaged sorbent devices
      while  units of size Model 403, fall Into the broader defi-
      nition of a packaged unit in that they can be transported
      and  later assembled from smaller components.

           This type of solvent recovery system 1s successfully
      being  used 1n a variety of Industries.  To follow the In-
      dustry category system presented in Table l-II, the Industries
      1n which they are 1n use are, plastics, rubber products,
      adheslves and paper coatings.  Two other Industries not
      Included 1n Table l-IItalthough mentioned on page  1-67,
      are  Pharmaceuticals and chemical manufacturing.  Solvents
      that  are being recovered are listed below with their Vm,

molar volumes, to indicate their relative  adsorbabilities:
cmM iq/mol    Solvent
Heptane                 163
Naphtha, 240-340F   140-207
Xylene                  140
4-Methyl-2-pentanone    140
Hexane                  140
Freon TF(1)
2-Ethoxy ethanol
Dlethyl ether
             Ethyl  acetate

                             3! iq/mol
     1,1 ,2-Trichloro-l ,2,2-trif luoroe thane

These  solvents are frequently used also in surface coatings
and  graphic  arts  industries, particularly toluene, xylene
and  4-methyl-2-pentanone.  The solvent recovery systems
would  also be effective  in these two industries although not
specifically identified  amonq the systems' users.  Table C-I
Identifies these  and  other solvents with various  pollutant
emission  sources.

     The  flow diagram of Figure 6-12 shows one  horizontal
carbon bed in each adsorber.  This type  of  adsorber  is  also
fabricated with  two  beds, one above the other.  The airflow
enters the adsorber  between  the beds with up-flow through
the  upper bed and down-flow  through  the  lower.  This  reduces
the  adsorber shell size  for  a given  flow, thereby reducing
construction costs.

      The  solvent recovery system  illustrated  in Figure  6-13
 is the Vulcan-Cincinnati, Inc.  Model  102 designed to  operate
 at 10,000 ft3/min airflow.    In  this  particular application,
 it is  operated  above design  capacity,  12,000  ftj/min, to  re-
 cover naphtha  from a rubber  products  plant.   The  bed  area
 in each adsorber is  about 110  ft2  and  bed  length  is  16 in.
 At the 12,000  ft3/min flow, the  linear  velocity through the
 bed is approximately 110 ft/min.   It is  operated  on  a one
 hour adsorption-regeneration cycle, recovering on  the average
 293 lb of naphtha per cycle.  Steam  consumption is about
 1770 lb per cycle which, on  an  average,  gives a 6 Ib steam/
 lb naphtha  ratio.  The carbon  bed volume is  approximately

      145  ft3  and.when  loaded  with  carbon  of  31  lb/ftj  density  the
      bed  weight  is  4,500  Ib.   For  the  293  "ib/cycle  recovery  the
      operating  capacity  is  then  0.065,  with  no  correction  beinq
      made for the  adsorption  zone.

           The naphtha  being recovered  is  an  aliphatic-naphthenic
      petroleum  fraction  boilinq  in  the  240  to  340F  range.  Nor-
      mal  and  branched  chain aliphatic  hydrocarbons,  cyclooctane,
      and  substituted  cyclohexanes  and  cycl oheptanes  of  Cj  to C-JQ
      fall into  this boiling ranqe.   Most  of  these solvents are
      included in Table C-I.  Their  odor recognition  threshold  and
      TLV  concentrations  are relatively  high  and  they  are not
      particularly  reactive  to solar radiation.   They  can pass  into
      the  atmospnere unnoticed at low concentrations.

           The theoretical  steam-to-naphtha  ratio for  the above
      recovery operation  is  7.2 (compared  to  the  measured 6.0)
      when cyclooctane  is  used to represent  naphtha  in  the  calcu-
      1ati ons.

           Figure 6-14  illustrates  a small,  cabinet-enclosed,
      single adsorber,  solvent recovery  unit.   Hoyt  Manufacturinq
      Corp. advertises  four  models;  the  one  shown is  Model  4.   It
      is designed for  use  in plants  using  up  to  3 drums  (165  qal)
      of solvent  per month  at  airflow  capacity  of 700  ft3/min.
      Its   overall  dimensions  are 49 by  45  in.,  length  and  width,
      and  81 in.  height.   These units are  pined  directly to the
      production  machine  emitting the solvent (or pollutant).   In
      an existing plant where  the use of vanor collection ducts
      to a central  solvent  recovery  systemarenot feasible, the
      use  of a number  of  small directly  connected units  may be
      the  solution  to  an  otherwise  difficult  problem.

           Figure 1-2  presents a  flow diagram of  a solvent  recovery
      system as  an  integral  part  of  a coin-operated,  dry cleaning .
      machine.  The  solvent  used  is  Freon  TF, 1,2 ,2-trichloro-
      1 ,2 ,2-trif1uoroethane, which  has  a Vm  a 120 cm3l1q/mol
      and  118F  boiling point.  It  is relatively volatile yet
      strongly adsorbed.   It is not a strong  pollutant in regard
      to odor, toxicity or  as  a smog precursor.

           Two sizes of dry  cleaning machines of the above  type
      are  manufactured.  The smaller size  machine cleans an 8-lb
      load of  clothing  and  the larger commercial  size a 25-lb load.
      One  gallon, or 13.3  Ib.of Freon TF is required per pound  of
      clothing in the washing  and rinsing  phasesof  the cleaninq
      cycle.  After the rinsing phase the  Freon TF  is drained from
      the  washer-extractor and clothing tumbled while air at  about
      85F is  passed through the tumbler.   The vapor-laden  air-


Finure 6-14
Small Cabinet-Enclosed Solvent Recovery Unit
    ftVmin Capacity. Steam Reaeneratl'on
Courtesy of Hoyt Manufacturina Corporation

    stream then passes through one of the carbon adsorbers.
    The overall cycle time is 14 mln, but programmed to operate
    different lengths of time on extracting and tumbling depend-
    ing on clothing to be cleaned.

         In the 8-lb machine the clothes drying airflow is
    about 100 ftVmin,  The adsorbers are 16 in. diameter and 28
    In. in length and contain up to 100 Ib of 4 to 6 mesh pel-
    letlzed activated carbon.  Steam regeneration of the carbon
    usually starts after the fourth clothes cleaning cycle and
    proceeds over several cycles.  The Freon TF-steam vapor Is
    condensed, separated by decantation and the Freon TF run
    Into storage for future use without any additional  treatment

         Moving Bed Systems

         The concept of the moving bed system is illustrated In
    Figure 6-15.  The rotary component of the system consists of
    four coaxial cylinders.  The outer cylinder is impervious to
    gas flow except at the slots near the left end.   They serve
    as Impure air 1njet ports where they are shown uncovered and
    as steam-vapor outlet ports  as shown at the lower left end
    of the rotary.  The carbon bed is retained between  two cylin-
    ders made of screen or perforated metal  and also segmented
    by partitions placed radially between the two cylinders.  The
    inner cylinder is again Impervious to gas flow except at the
    slots near the right end of  the rotary.   These slots serve as
    outlet ports for the purified air and Inlet ports for the
    regenerating steam.   On each rotation of the rotary, each
    segment of the bed undergoes adsorption and regeneration.
    The desorbed solvent can then be  separated from  the  steam by
    decantation or distillation.

         Because of the  continuous regeneration capability of the
    rotary bed,  a  more efficient utilization of the  carbon is
    possible  than  with stationary bed systems.   In most  solvent
    recovery  operations  the adsorption zone  is  between  3 and 4
    in.  The  unsaturated carbon  bed 1n front of the  adsorption
    zone and  the saturated bed behind 1t are idle  but add to the
    bulk of the system and Increase airflow resistance  through
    the bed.   In deep beds of 12 to 36 in.,  as  required  in the
    stationary bed systems for regeneration  time,  a  large portion
    of the bed is  idle at any one time.   By  continuous  regeneration
    the regeneration time for each segment of the  bed Is shortened
    and thereby shorter  bed lengths can  be used.   This  leads to
    two advantages:   (1)  more compact system and  (2) reduced
    pressure  drop.

              Vapor-laden  air  in
           Purifiea air out

       Steam in
                                         Activated  carbon
                                    Cross-sectional view
         Steam  and  vapor  out
        air in
        Steam  and
        vapor  out
Horizontal exterior view
Figure 6-15  Continuous Rotary Bed, Based on Sutcliffe
             Speakman Co., LTD Illustrations, 66 Palmer
             Ave., Bronxville, N.Y.

     The disadvantages are  those associated with wear on
moving parts and maintaining seals 1n contact with moving
parts.  The use of shorter  beds also decreases the steam
utilization efficiency,

     Fluidized Bed Systems

     Figure 6-16 shows a flow diagram for a fluidlzed bed
solvent recovery system, based on illustrations from Avery
(1969).  The carbon  is recirculated continuously through the
adsorption-regeneration cycle.  The spent carbon, saturated
with solvent, is elevated to the surge bin and then passed
down into the regeneration  bed where it is contacted to an
upward flow of steam.  The  regenerated carbon is then metered
into the adsorber where the carbon traverses nine beds while
fluidized with the upward flow of the vapor-laden air.  Air
velocity of 240 ft/min is required to cause fluidization of
the bed.  In both the regeneration and adsorption phases the
carbon is moved countercurrent to the gas or vapor.

     The countercurrent movement Increases the .efficiency
of, regeneration, i.e., the  steam-to-sol vent ratio Is less
'than for a stationary bed under otherwise comparable condi-
tions.  For a number of solvents, Avery reports 402 to 50%
less steam required  for their recovery.  In addition to the
beneficial effects of the countercurrent movement, the bed
length can be increased to  further Improve the steam utili-

     The countercurrent movement also increases the effective
    use ftf thft Carbon' more solvent can be recovered with less
    carbon than with stationary or rotary bed systems.  By adjust-
    ments or balance of the carbon and solvent input rates, the
    total carbon in the adsorber can be made part of the adsorp-
    tion zone but have the carbon reach saturation in the lowest
    bed just before it is discharged Into the elevator.  Very
    little of the carbon is then Idle, hence the maximum utili-
    zation is made of the carbon.
         The fact that the jcju&MLJbJUur.fijU^
    de.Lly.e.r.&cLto ,th&gAAer.AjLian jxhasfi. 1s another factor 1n the
    reduced steam requirement.  In this respect the fluidized bed
    system offers the most favorable conditions for solvent de-
    sorptlon while the rotary bed offers the least favorable,  in
    the notary bed, operations., the carbon moved Jnto_tbe regener-
    ation phase has reached the . loweflTTtinti of saturation TrHEhe
    three systems.                 "           ~

 for vapor-
 laden  carbon
                             Surge bin
                                  Desorbed vapor and
                                  regenerating agent

                                   Regeneration bed
                                   Regenerating agent
        f Carbon metering  valve
       Impure air in
                                  Purified  air out
                                Fluid bed adsorber
Figure  6-16  Fluidlzed Bed  Solvent Recovery System

          In the operation of this  particular  unit,  Avery  reported
     penetration of 100 ppm when  carbon  disulfide was being  ad-
     sorbed at 1,000 ppm and 100  ppm when  acetone was being  adsorbed
     at 1,700 ppm.   This was a continuous  penetration and  not  the
     temporary penetration that might occur  from a stationary  bed
     just before it is  taken off  stream.   By increasing  the  number
     of beds, the penetration from  the fluidized bed system  can  be
     decreased to any desired level  but  with an increased  pressure
     drop penalty.

          When large air volumes  are treated and available space
     for the installation is at a premium, the smaller size  and
     also the lower initial cost  are definite  advantages over  the
     stationary bed system.

          A serious disadvantage  is  high attrition l.os$&|j>f the
     carbon caused by the fluidization of  the  beds".  Because of
     the attrition, filtration of the effluent airstream may be
     required.  Influent airstream  need  nojt  be filtered  since  no
     plugging of the fluidized beds  can  occur.
     -ji- -~ 

     Carbon Resorb Systems, New Technology

          General Discussion

          Most of the present adsorption technology  has  been
     directed toward goals other than the  reduction  of  pollutant
     vapor emissions into the atmosphere.   In  complying  with one
     of the objectives  of this study, attention  has  been directed
     toward developing new means  or adapting the  present technology
     for pollutant emission control.  To attain  this goal  a  detailed
     study was made to determine the capabilities  of available ad-
     sorbents and to explore possible modifications  to  presently
     used adsorption systems to adapt them for pollution control
     under the wide variety of conditions  under  which  pollutants
     are emitted.

          The area open to technical advances  in  package sorbent
     systems for pollution control  appears to  be  in  modification
     of the systems and operating procedures used  in solvent re-
     covery.  Since the primary objective is pollution  control
     rather than solvent recovery,  the modified  systems  are  re-
     ferred to in this  report as carbon  resorb systems.  As  in
     solvent recovery,  two phases of the operating cycle are ad-
     sorption and regeneration, but also includes  an additional
     operation, the disposal of the desorbed pollutant.  An  import-
     ant requirement of the disposal procedure is  that 1t  does not
     cause repollutlon  of the air or waterways.

     The main departure in the regeneration of carbon resorb
systems is the use of heated noncondensable gases in place of
steam.  As the influent concentration decreases, the amount
of steam required per regeneration also increases.   At low
concentrations the amount of steam required relative to the
amount of pollutant desorbed is so large that separation of
pollutant from the steam becomes impractical.  The  extent of
pollutant dilution in steam condensate was calculated for a
3,000 ftVmin capacity solvent recovery system when used to
clean an airstream containing pollutants at 10 ppm  concentra-
tion and regenerated with steam.  The results of the calcula-
tions are summarized in Table 4-XX.

     The three organic compounds listed in the table were
selected as representatives for the  wide variety of solvents
or pollutants that require control.   4-Methyl-2-pentanone is
representative of strongly adsorbed  compounds which are
completely oxidizable to C02 and HzO and are slightly soluble
in water.  Propanone is representative of weakly adsorbed
compounds, completely oxidizable, and completely miscible
with water.  Trichloroethene is intermediate in adsorbabi1ity,
is slightly soluble in water, but on oxidation yields HC1,  a
potential air pollutant.

     The high steam-to-solvent ratios, as given in  the last
column of the table, for each of these compounds mean that
many pollutants cannot be economically separated from the
steam condensate.  Those that are soluble or even slightly
soluble will be completely dissolved and separation by dis-
tillation would be too costly.  Disposal of polluted steam
condensate   into the sewer can also lead to further pollution
of the waterways.  Means are proposed by which this problem
can be avoided.

     Compounds such as tetrachloroethene, aliphatic and
naphthenic hydrocarbons above Ce are reported to be insoluble
1n water.  Separation of compounds of this  type appears econom-
ically possible by decantation.

     When a noncondensable gas  is used as regenerating agent,
the desorbed pollutant may be disposed of by several means.
If it is completely oxidizable  to C02 and HpO it can then  be
immediately destroyed by incineration upon  desorption  from the
carbon.   If it has recovery value, or is  noncombustible  or
would repollute the air on  incineration,  the desorbed  pollut-
ant may be readsorbed on a  small secondary  carbon  bed  and
recovered  from it by  conventional steam  regeneration  or  it
may be recovered from the  regenerating  gas  by condensation.

          These  three  procedures  for  immediate disposal or recovery
     of  the  pollutant  can  be  incorporated  into any of the three
     solvent recovery  type  systems described earlier in this chap.
     ter,  i.e.,  stationary  bed, moving bed and fluidized bed systems.

          Other  aspects  of  the adsorption  and regeneration phases
     have  also been  examined  to determine  possible means for im-
     proved  utilization  of  the carbon and  regenerating agent.

          The various  concepts as incorporated into various oper-
     ational  systems can be identified according to the mode of
     regeneration  and  disposal of the desorbed pollutant.  These

          Heated air or  inert gas regeneration followed by
            thermal  or catalytic  incineration of the de-
            sorbed  (completely combustible) pollutant,

          Heated air or  inert gas regeneration of primary
            bed  followed by readsorption of desorbed pol-
            lutant on  a  secondary bed and  steam regenera-
            tion of  secondary bed to recover pollutant,

          Heated air or  inert gas regeneration with desorbed
            pollutant  (or  solvent) condensed at lowered
            temperature  and recycle of uncondensed vapors
            through  carbon bed,

          Two-stage  adsorption with stationary bed system
            and  single-stage regeneration,

          Two-stage  regeneration with stationary bed system
            and  single-stage adsorption,

          Regeneration by  pressure reduction,

          Continuous or  intermittent removal of spent
            carbon from  adsorption system  followed by
            high temperature steam reactivation of the
            carbon on  plant site.

          These  systems  are described in the following text with
     statements  made regarding advantages  and disadvantages.  The
     descriptive portion is followed by economic analysis of
     solvent recovery and some of the proposed systems.  Cost com*
     parisons are given with the competing systems:  catalytic and
     thermal incinerations.

     Gas Regeneration with Incineration

     For pollutants that are completely oxidizable,  regener-
ation with an inert, noncondensable gas followed  by  inciner-
ation of the desorbed pollutant is a procedure  that  offers  a
relatively low cost regeneration and simultaneously  effects
complete disposal of the pollutant.  In this  case the  carbon
bed acts as a concentrator.  The incinerator  can  then  be
relatively small requiring little additional  expense to

     To show the manner in which the system would operate,
the following sample calculations have been worked-up  for
4-methyl-2-pentanone from data given in Table 4-XX.   At
0.025 Ib/lb operating capacity, 37.5 Ib of the  solvent are
adsorbed on the carbon from 1,500,000 ft3 of  air  passed
through the bed over a period of 82 hr at 3,000 fts/min  flow
and 10 ppm concentration.  The regeneration can be accom-
plished with 37,000 ft3 of inert gas (near 28.9 lb/mol)  at
302F regeneration temperature, yielding an effluent mixture
of 4,000 ppm mean solvent concentration.  (The  regeneration
data are obtained from predetermined curves of type  shown
in Figures 4-22, -23 and -24.)  The size of the incinerator
depends on the regeneration time.  For a two-hour regener-
ation, the incinerator would be sized for a 300 ft3/min  flow
rate.  This is 10% of the 3,000 ft3/min flow capacity of the
carbon resorb system or incinerator if the vapors were in-
cinerated directly at the 10 ppn.  In addition, direct in-
cineration would require a considerable amount of fuel to
heat the air to the incineration temperature.  Incineration
of the carbon bed effluent would require very little fuel
since the air volume is much smaller and the high concen-
tration of the pollutant supplies  the major part of the fuel

     By use of a noncondensable gas, rather than steam,  the
heating cost of the regenerating gas is considerably less.
The above regeneration requires 2,400 Ib of steam, which at
O.H/lb costs $2.40.  Noncondensable gas regeneration re-
quirement is 2,700 Ib, which at 0.005*/lb, costs $0.13.  Steam
cost is about $0.01/hr per 1000 ft'/min flow.  At the 10
ppm pollutant concentration level  where regenerations are
not frequent, the steam and definitely the noncondensable
gas costs are not significant relative to the other operating
costs which can add up to at least $0.25/hr per 1000 ft*/min
or to $0.55/hr per 1000 ft'/min, the more probable  cost level

           As  the pollutant  concentration  increases,  the  advan-
      tages of using  a  noncondensable  gas  regeneration-Inciner-
      ation combination decrease.   When  the  pollutant 1s  a  sol-
      vent of  recovery  value,  the  Incineration  of  the solvent
      would constitute  an  Increasingly greater  loss of a  valuable
      product.  At higher  pollutant concentrations and more  fre*
      quent regenerations,  increasingly  larger  incinerators  are
      required.   At concentrations  approaching  25% LEL, the  re-
      quired incinerator capacity  is the same as for  direct  In-
      cineration.  At concentrations between 700 to 1000  ppm,  the
      cost of  the carbon resorb  with incineration  approach  those
      of direct  Incineration;  see  Figure 6-31.

           The general  concept of  carbon resorb and Incineration
      systems  combination  has  been  Investigated by Mattla  (1969,
      1970).  A  flow  diagram of  the proposed process  1s presented
      in Figure  6-17.  Adsorption  proceeds in the  same manner  as
      1n solvent recovery,  but regeneration  1s  accomplished  by
      recycle  of gas  through the carbon  bed  and incinerator, which
      can be either thermal  or catalytic.  The  regenerating  gas
      is Initially air  but  on  passing  through the  adsorber  and
      incinerator it  becomes converted to  a mixture low in  oxygen
      with increased carbon dioxide  and water vapor.   Make-up air
      is added to the Incinerator  to supply  enough oxygen  to burn
      the pollutants  desorbed  from  the carbon bed.  To avoid
      damage to  the carbon  from  over-heated  regenerating  gas,  a
      cooler is  required in  the  recycle  line.   To  maintain  proper
      regeneration temperatures  and flow ratesInstrumentation
      and closer controls  are  required than  needed for solvent
      recovery.   The  process 1s  recommended  for fairly large air
      volume control,  10,000  ft'/nHn  and  larger.

           In  the operation  of this process, a  balance is main-
      tained between  the time  period for the adsorption phase  and
      regeneration phase to  attain  the best  overall efficiency.
      As an example of  operating conditions, a  process may  be
      designed to adsorb a  pollutant at  200  ppm concentration  in
      90 min and flow velocity near 100  ft/min.  Regenerating  gas
      flow and temperature  are adjusted  to regenerate the  carbon
      in 75 min  allowing 15  min  for bed  cooling.   Regenerating gas
      flow rates are  variable  from  2%  to 10% of the adsorption
      flow rate. Temperatures  are adjusted  to produce  the  desired
      rate of  desorption.   A 300F  temperature  was considered
      maximum.  The carbon  bed depths  are  4  to  8 In., much  shorter
      than in  solvent recovery,  to  bring the service  time  to the
      one hour level  at the  lower  concentrations.

           It  should  be noted  that  the regenerating gas flow 1nd1
      cated in Figure 6-17  is  In the same  direction as the  ad-
      sorption flow,  Instead of  reversed as  normally  done 1n



                                             Exhausts  to
                Regen. gas
              Make-up air(""""
                 Nat. gas 	
Figure 6-17
A1r Pollution  Control System Utilizing Carbon-
Resorb with  Gas  Regeneration and Incineration
of Desorbed  Pollutants,   (from Mattla, 1970)
Figure 6-18
Adsorption-Phase  Profile  After Carbon Bed Has
Been Regenerated,  When  L  1s  4 to 8  1n.t and

     solvent  recovery operations.  This procedure leaves an
     adsorbed-phase  residue  in  the carbon bed that extends to
     the  effluent  end.  This  is  shown  in Figure 6-18.  Because
     of  the residual capacity,  immediate penetration occurs in
     the  adsorption  phase  at  a  low concentration.  The penetra-
     tion concentration is the  concentration in equilibrium with
     the  residual  capacity.   By  increasing the operating capacity,
     Aoi   tos-u)r the amount  of  penetration can be decreased to
     lower levels.   The process  is also operated with reverse
     flow (Mattia, 1969).  In this case the adsorbed phase pro*
     files are  similar  to  those  illustrated in Figure 4-21.

          Another  operational characteristic that should be
     noted is that the  pollutant concentration in the regenerating
     gas  stream is not  constant  but  varies with regeneration time
     in  a manner shown  in  Figure 6-19.  This variation in concen-
     tration  produces effects in the disposal operation which must
     be  compensated  by  corrective measures.  In incineration, It
     is  basically  a  variation in fuel  input; natural gas input rat
     must be  adjusted to maintain the  Incinerator temperature at
     the  effective level over the total regeneration phase.

          Gas Regeneration with  Secondary Adsorption

          When  the pollutant, adsorbed from a low concentration
     airstream, contains halogens, NOx and SOx groups, Inciner-
     ation of the  desorbed pollutants  would repollute the air with
     acids of these  groups.   Other means are required to dispose
     of  the desorbed pollutants  than the incineration method des-
     cribed in  the preceding  text.   One of the methods that 1s
     feasible on theoretical  grounds utilizes a secondary carbon
     adsorber operating as a  solvent recovery unit for the efflu-
     ent  regeneration gas  stream.

          The manner in which the system would operate can be
     demonstrated  with  the sample calculations used  in the pre-
     ceding text on  Gas Regeneration with Incineration.  In the
     adsorption-desorption process the 4-methyl-2-pentanone is
     removed  from  an air volume  of 1,500,000 ft3 at  10 ppm con-
     centration and  concentrated Into  an air volume  of 37,000 ft3
     at  a mean  concentration  of  4,000  ppm.  The secondary adsorber
     is  then  required to recover 37.5  Ib of 4-methyl-2-pentanone
     every 82 hr from 37,000  ft3 of  gas volume at 4,000 ppm mean

          Since the  desorbed  solvent concentration varies over
     the  regeneration time period in a manner shown  1n Figure
     6-19  the  estimating procedure for determining the carbon
     bPd  size would  be  quite  Involved  1f the changes in concen-
     tration  were  taken Into  consideration.  A very  approximate
     estimate of the bed size may be attained if the 4,000 ppro
     mean concentration Is assumed constant over the regenera-
     tion time  period.

                       Regeneration  time

Figure 6-19  Pollutant Desorptlon  Rate  Variation with
             Regeneration  Time
                                              Exhaust to '
                                      Regen. gas
Regen. gas
 Figure  6-20  A1r Pollution Control System Utilizing  Carbon-
             Resorb with Gas Regeneration and  Secondary  Ad-
             sorber for Desorbed  Pollutants,(from  Mattia.

          The  parameter  that  greatly  affects  the required bed
     size  is  the  choice  of  steam-to-solvent ratio  that the ad-
     sorber  is  to operate at.   If  operations  were  chosen at steam-
     to-solvent ratio  of 3.0,  the  required operating capacity is
     0.08  Ib/lb for  the  saturated  layer.  The estimated carbon
     bed weight is approximately  500  Ib or 332 of  the 1500 Ib for
     the primary  bed.  By increasing  the  steam-to-solvent ratio
     to 6.0,  the  required operating capacity  is increased to 0.14
     Ib/lb.   The  estimated  carbon  weight  reduces to 300 Ib or 20%
     of the  primary  bed  weight.   The  details  of the calculations
     for the  latter  estimate  are  given in Appendix E, page E-24.

          By  reducing  the primary  bed thickness and thereby, also,
     the adsorption-regeneration  cycle time,  the secondary bed
     size  can be  further reduced.  When the adsorption and regen-
     eration  phases  for  the primary adsorbers are  balanced, two
     secondary adsorbers may  be used, also, operating alternately
     on adsorption and regeneration.

          Figure  6-20  presents  a  flow diagram of the proposed
     system  as  engineered by  Mattia (1969, 1970).  Inert gas
     (could  be flue  gas) is fed into  the  system at the start of
     the regeneration  and recycled through the primary and secon-
     dary  adsorbers.   The regenerating gas from the secondary ad-
     sorber  passes through  the  surface heater where it is heated
     to an elevated  temperature.   On  blending with a side-stream
     being recirculated  through the primary adsorber, the temper-
     ature of the combined  gas  stream on  entering  the primary ad-
     sorber  is  near  300F.   A part of the  vapor-laden gas stream
     from  the primary  adsorber  passes through the  cooler where
     it  is cooled down to 100F,  or lower, before  being directed
     through  the  secondary  adsorber.

          At increasingly  higher  pollutant concentrations  the
     advantages in using secondary adsorbers  decrease and  direct
     regeneration of the primary  adsorber increase.

          Gas Regeneration  with Condensation  and  Recycle

          A third proposed  method utilizes  a  condenser  to  recover
     the  desorbed pollutant vapor from the regenerating  gas.U)
     By  operating the  condenser at a  very low temperature  an  almost
     complete condensation  of most pollutants could  be  realized,
     but  to avoid costly refrigeration a  more economical  approach
     is  to operate the condenser  at  a readily attainable  higher
     temperature  and recycle the  uncondensed  vapor through  the
     (1)  Proposed by B.B. Carr, MSA Research Corp., Evans City, p*

on-stream adsorber.   As the condenser temperature 1s raised
the amount of recycle becomes larger,  The limiting factor
1s the adsorptive capacity of the carbon bed,  hence opera-
tion of the process  at practical  condenser temperatures Is
only attainable with the less volatile pollutants.   The In-
crease 1n regenerating gas to handle the recycled vapor 1s
small and 1s not a significant factor 1n the operating cost.
     Figure 6-21 presents a flow diagram that Illustrates
     eneral scheme of this type of operation.  An inert re-
generating gas is prepared and heated by burning natural
gas or any other completely oxidizable fuel.   The regener-
ating gas is then cooled to near 300F and passed through
+ ho A-f f - c t ioam arl e/>vKe *  in + 1*4* *   A A ,1  j<*U * o   TU~
ating gas is tnen cooled to near 300T and passed through
the off-stream adsorber, in this case Adsorber 2.  The
vapor-laden hot effluent from Adsorber 2, on passing through
the condenser, is cooled and, if the dew point is reached,'
a portion of the vapor condenses.  The uncondensed vapor and
regenerating gas continue their flow into the Impure air
stream and to the on-stream Adsorber 1 where the recycle
vapor is readsorbed.  During the desorption phase of Adsorb-
er 1, the vapors from the Impure air and recycle gas streams
are both desorbed but at a higher concentration and in a
slightly larger regenerating gas volume than during the
Initial start-up cycle.  On each successive adsorption-re-
generation cycle the amount of vapor recycled increases
until either the adsorptive capacity of the carbon bed 1s
exceeded or the condensation rate equals the pollutant vapor
Input rate.

     To gain further insight Into the manner in which this
system operates, calculations were made using the carbon-
resorb system as described in the prior text under Gas
Regeneration with Incineration and Gas Regeneration with
Secondary Adsorption.  In the Initial cycle of the adsorp-
tion-desorption process the 4-methyl-2-pentanone is removed
from an air volume of 1,500,000 ft* at 10 ppm concentration
and concentrated into a gas volume of 37,000 ft3 at a mean
concentration of 4,000 ppm.  On the second and successive
cycles, these gas volumes and concentrations change until a
steady state is reached.  At steady state operation the
condenser removes 37.5 Ib of 4-methyl-2-pentanone from the
regenerating gas stream with the uncondensed balance return-
ed to the on-stream adsorber.

     The regeneration of the off-stream adsorber is started
on the 80th  hr and continued to the 82nd  hr over  a 2-hr
regeneration period.  By regenerating the off-stream ad-
sorber near the end of the service time,  the recycled  vapor
is readsorbed 1n the influent end of the  on-stream  adsorber

Partlculate f1

I"! Blower
f /^\
Iter Exhaus
?i >M
V ; VIs
-+ _-r .---.
*x ^^
*? 2:?
! I U
i f
r~^ I Cooler
	 	 _J a 300F
/\ ( l^
*1 _

S ^ 	 I-K
ower L_J
^ ^ Regeneratlng-
                                                    Liquid solvent
                                         9as           pollutant
     Figure  6-21   A1r  Pollution Control System Utilizing Carbon
                  Resorb with Gas Regeneration, Condensation,
                  Recycle of Uncondensed Vapor.

The readsorbed vapor is
ease of desorptlon when
reverse direction.
                        now 1n  a  favorable  position  for
                        regenerating  gas  flow Is  in  the
     The concentration of the desorbed  vapor  in  the  regen-
erating gas stream varies 1n a manner indicated  in  Figure 6-
19.  This concentration variation has an affect  on  the  oper-
ation of the condenser and also on the  readsorption  of  the
recycled vapors.  To simplify the calculations  the  concen-
tration was assumed constant at the mean level.

     Two estimates were made, one with  the condenser oper-
ated at 77F and the other at 32F.  In each  case the initial
flow to the adsorber is 3,000 fts/min at 10 ppm  concentration
When the condenser 1s operated at 77F, the steady  state is
attained on the 6th  cycle.  At the steady state,the follow-
ing operating condltionsprevall during  the 2-hr regenerating
time period:

         F  to adsorber
         C-j to adsorber
         F  to condenser
         Ce to condenser
         Weight of vapor
         Weight of vapor
                             3,330 ft
                               940 ppm
                               330 ft3/m1n
                            13,000 ppm

                                95.2 Ib/cycle

                                37.5 Ib/cycle
When  the  condenser  1s operated at 32F, the steady state  1s
attained  on  the  2nd  cycle.  At the steady state  the  following
conditions prevail  during  the 2-hr regenerating  time period:
          F   to  adsorber       3,310
          C-j  to  adsorber         160
          F   to  condenser   *     310
          Ce  to  condenser      5,500
          Weight of  vapor
            recycled             15
          Weight of  vapor

 During the  service  time  up to 80
 3,000 ftVmin and C^  1s  10 ppm.

                                   ,1  Ib/cycle

                                 37.5  Ib/cycle

                                  hr,  F  to  adsorber is  still
      For the system operating at 77F condenser temperature,
 132.7 lb of vapor are adsorbed by the carbon every 82 hr of
 Wh1ch 95.2 lb are recycle vapor and 37.5 lb are from the
 atmosphere.  For the system operating at 32*F condenser

     temperature,  the  carbon  adsorbs  52.6  Ib of vapor of which
     15.1  Ib are  recycle  vapor  and  37,5  Ib are from the atmo-
     sphere.   Detailed calculations for  the 32F condenser oper-
     ation are  given  in Appendix  E, page E-27.

          Figure  6-22  presents  adsorbed-phase profiles showing
     the probable  status  at  the start of the regeneration phase
     and at the end of the adsorption phases for the process
     when  operated at  the two condenser  temperatures.  The ad-
     sorptive  capacity is measured  in Ib/in. of bed length; hence
     the product  of capacity  times  bed length pives a direct
     measure of pounds adsorbed under any  curve.

          The  broken-line profile represents the status just
     after regeneration has  been  completed.  The mean capacity
     	     lb/1n.   The area under the  curve, Wr - 189.5 Ib,
- -   -  - -^ ^ -.. -   .  .. -. _ v ~-  v  i wT'"*r*i^*vww  i 11 v 111 % w 11 V M I-' Vk \ I b Jr
wr  8.88 lb/1n.  The area under the curve, Wr  189,5 Ib,
is the total amount remaining on the carbon after any cycle.
The profile designated by 10 ppm represents the status at
the end of the adsorption phase when no recycling 1s done.
In this case, us  10.65 Ib/in. and the area under the curve
represents the adsorbed amount, Ws - 227 Ib.  The difference,

           Ws-Wr  227-189.5  37.5 Ib/cycle,

is the amount of pollutant vapor desorbed or removed from
the atmosphere.

     When the vapor 1s recycled from the 32eF  or 77"F con-
denser,  the profile as indicated 1s superimposed on the  10
ppm profile.  The area between the 10 and 940  ppm profile
represents 95.2 Ib recycle vapor and the area  between 10
and 160  ppm, 15.1  Ib.

     Theoretically, the process is feasible for pollutants
of Vm near or larger than 140 cm311q/mol.   Pollutants of
lower Vm would require lower condenser temperatures to
avoid exceeding the carbon capacity with the recycle vapor.
When the concentration is high enough, recycle 1s no longer
necessary; direct steam regeneration again becomes feasible.
Very low molecular weight pollutants cannot be recovered by
this process unless the Input concentration 1s quite high
or the condenser is operated at a very low temperature.

     Two-Stage Stationary Bed System
         This system is intended for control of pollutants ad-
    sorbed in the 10 to 100 ppm concentration range.  It require;
    two adsorbers or carbon beds on-stream In series.  The first'
    carbon bed, on the airstream Influent end, 1s designated as
    the first stage and the second bed the second stage.  For

.160  ppm,
                                    940  ppm,  recycle  from
                                        77*F  condenser
                    recycle from
                   10 ppm
            	 Adsorption phase

            	 Regeneration phase
20 21.3   25
Figure 6-22  Adsorbed-Phase  Profiles  at  the  End  of  Adsorp-
             tion and  Regeneration  for a Gas Regeneratlon-
             Condensatlon-Recycle Process,  Lt "  21.3  1n.
best efficiency the bed lengths should be equal to or slightly
longer than the adsorption zone length.  When adsorption zone
has moved into and completely occupies the second stage bed,
the first stage bed has become completely saturated and is
taken off-stream for regeneration by any one of the means
previously described.  The second stage bed is moved Into the
first stage position and a third carbon bed, previously regen-
erated, is moved into the second stage position.  The three
carbon beds are moved consecutively through the second and
first stages of the adsorption phase and then through the
regeneration phase.

          The flow diagram in Figure 6-23 shows  the  essential
     components of the two-stage system.   At the particular  time
     period depicted by the diagram, the  airstream is  passing
     through Adsorbers 1  and 2 in succession while Adsorber  3  1s
     being regenerated.  When the vapors  begin  to penetrate  Ad-
     sorber 2, Adsorber 1 has become saturated  with  vapor  and  1s
     taken off-stream for regeneration and Adsorber  3  becomes  the
     second stage and Adsorber 2 the first stage.

          This system approximates the countercurrent  contacting
     of the carbon and gas phase, as occurs in  the fluidized bed
     system, and thereby  also the effective use  of the carbon  1s
     increased.  However, with the two-stage stationary bed  system
     no carbon loss occurs because of attrition  as it  does in  the
     fluidized bed system.

          The disadvantage is the cost of the additional ducts,
     valves, more elaborate adsorber construction and  controls to
     operate the valves.

          Two-Stage Regeneration of Stationary  Bed

          Figure 6-19 shows the change in desorption rate  with
     regeneration time for a stationary bed.  As the bed becomes
     heated with passage  of regenerating  agent,  the  desorption
     rate increases to a  maximum and then decreases  as the pollut-
     ant becomes depleted in the bed.  In a countercurrent regen-
     eration, the desorption rate does not go through  a maximum
     with regeneration time but settles to a constant  rate.
     Countercurrent regeneration is thereby contributive to  better
     utilization of the regenerating agent; the regenerating agent-
     to-desorbed pollutant ratio is decreased.

          Completely countercurrent regeneration is  only possible
     with systems such as the fluidized bed or  moving  bed  of the
     "hypersorption-type" described later in this chapter.  Much
     of the benefit of countercurrent regeneration could be
     gained by multistage regeneration where many stationary beds
     in series are 1n the regeneration gas stream.  Saturated  beds
     would be intermittently added to the gas effluent end while
     regenerated beds are removed from the gas  influent end.  The
     complexity of the system would economically over  balance  the
     benefits from the countercurrent regeneration.   It Is doubtful
     that the use of stages greater than two would be  economical.

          For a two-stage regeneration system,  as Illustrated  by
     the flow diagram in  Figure 6-24, three adsorbers  are  required
     At the particular time period depicted by the diagram,  Adsorber
     1  is on the adsorption phase while Adsorbers 2 and 3  are  on the
     regeneration phase.   Adsorber 2 1s 1n the Stage 2 position and

     air  in
  agent in
                                              to  solvent
                                            _ _ ^
Figure 6-23  Two-Stage Stationary Bed System
 Impure air
    agent in
rs *--^
ng iv
1 ?
J} ExhaustA
X air '
. i /""K.
^ A
' "gr ^
-it J
/^x A .x \
i 4-
1 A
^M*i v
	 1 t
--T- |SOlV
_ rprr
/ V i 1
Ji i
. i
* i I
V V s
 Figure 6-24  Stationary Bed with Two-Stage Steam

     is  in  the  influent  end  of  the  regeneration gas stream and
     Adsorber 3  is  in  the  Stage  1 position.  When vapors begin
     to  penetrate Adsorber 1  it  fs  taken  off-stream and connected
     to  Stage 1  position on  regeneration.  Adsorber 3 is moved
     from Stage  1 to Stage 2  position  and Adsorber 2 goes on-stream
     on  the adsorption phase.

         The desorption rate  from  a two-stage regeneration will
     fluctuate  between maxima  and minima  with regeneration time.
     The various  pollutant disposal methods will operate more
     efficiently  than  when operated 1n conjunction with a dual
     stationary  bed system.

         The disadvantage 1s  the cost of the additional ducts,
     valves, more elaborate  adsorber construction and controls  to
     operate the  valves.

          Regeneration by  Pressure  Reduction

         When  a  process is  operated at above atmospheric pressure,
     advantage  can  be  taken  of  the  higher pressure in the adsorption*
     regeneration cycle.  The  adsorption  increases with Increased
     total  pressures since the  pollutant  concentration  Increases  1i>
     the same proportion.   Conversely, on releasing the pressure
     the vapors  adsorbed at  the  higher pressure are released.   A
     small  amount of sweep gas  at ambient or elevated temperature
     is  then required  to accelerate the removal of the  vapors  from
     the carbon  bed.   Table  6-II  presents estimated capacity  figures
     that show  a  ws "  0.029  Ib/lb at 8 atm pressure during the
     adsorption  phase  and  when  pressure 1s released to  1 atm,  ws
     reduces to  0.014  Ib/lb.   The operating  capacity  is then  0.015
     Ib/lb.  The  large disadvantage 1s the Initial cost of the
     heavy  construction  required for the  adsorber tank  and valves.
     Table 6-II   Effect of Pressure on  Adsorptive  Capacity  of
                 Carbon for Prooanone.  10  ppm Concentration

Pressure, atm


us, Ib/lb

     Multistage  with  Bed  Displacement

     This type system is  Illustrated in  Figures  6-25  and  -26.
Figure 6-25 shows  the inverted  conical shaped  lower bed of
the first stage  and the upright conical  shaped upper  bed  of
the second stage.   Additional  stages can be  added  above the
second, although advantages  accruing from stages beyond the
second diminish  rapidly.   Impure airflow is  maintained  through
the first stage  bed until it becomes completely saturated.  I.e.,
In equilibrium with the influent concentration.  The  carbon  of
the first stage  is then removed through  the  bottom valve  and
simultaneously the carbon of the second  stage  lowers  into the
first stage position.  New or regenerated carbon from the
hopper moves into  the second stage position.  The  spent carbon
can now be regenerated or reactivated  1n a separate unit.

     The two-stage system Illustrated  in Figure 6-26, has
an operating capacity of 10,000 fts/min.  The  upright tank
containing the carbon beds 1s approximately 8  ft in diameter
and 14 ft in height.   Each stage contains about 3,400 Ib  of
carbon, mean flow area is about 82 ft2,  and bed thickness is
16 in.  The carbon is a pelletized type  of 4 to 6  mesh.   In
this particular application, at a waste  water  treatment plant,
the carbon is displaced in each of the two stages  every 134
da.  The spent carbon is then activated  with steam at a high
temperature (over 1600F) to its original activity.   This
means that the operating capacity 1s  the total  capacity of
the saturated layer rather than a fraction of the saturated
layer capacity as 1s the case in solvent recovery operations
where steam Is used for regeneration.

     Service  times were estimated for the above described
system at  1.0, 10, and 100 ppm  concentrations for 4-methyl-
2-pentanone and propanone, as representatives of widely
divergent  pollutants in  regard  to adsorbabl11ty.  The details
of the calculations  are  given  in Appendix E,  page E-31   and
the results in Table 6-III.

Table 6-III   Service Time for  Pollutant  Vapors  at Varied
              Concentrations  in  a Two-stage  System with
              Displacement of Beds

                4-methyl-2-pentanone         propanone	
Ci,  ppm         MS, Ib/lb       t,  hr     ms,  Ib/lb      t,  hr

     1              0.11          2450       0.0047       180
    10              0.16           360       0.013         50
   100              0.20            45       0.034         13

                             Drum  elevator
                    Perforated  metal
                    conical  shell


                            Clean  air
                           Impure  air
     Figure 6-25  Multistage
     Pollutant Control  System,
     Internal  Construction  of  a
     Two-Staqe System,  Courtesy
     T.  Melsheimer Co., Inc.
                                      Figure  6-26   Two-Staqe
                                      Pollutant  Control  System
                                      10,000  ft3/min  Caoacity,
                                      Courtesy T.  Melsheimer  Co.

     This type system offers  the  most  benefits  when  the
pollutants are high  molecular-v/eight  types  and  are emitted
at 1 ow concentrations.   High  molecular-weight pollutants,
having Vm larger than 190  cm3!iq/niol,  ere  not completely
desorbed by the regular steam regeneration  procedures  as
practiced in solvent recovery.   Incomplete  desorption  also
occurs when the pollutant  can polymerize while  in  contact
with  the carbon.  Reactivation  would  then  be a  frequent
necessity and would  be a costly operation  with  the design
of equipment used in solvent  recovery.  The design of  this
system minimizes the cost  of  transfer of  carbon between  the
adsorption and activation  phases  of the operation.

     The longest service time calculated  for Table 6-III
is 2450 hr or 102 da occurring  at 1 ppm concentration  when
the pollutant has the adsorptive properties of  4-methyl-2-
pentanone.  For the  control of low molecular weight  pollut-
ants, as represented by propanone, or control  of the heavier
molecules at 10 ppm and higher concentrations,  the service
times are much shorter thereby requiring more  frequent re-
activation of the carbon.   The frequency and cost of re-
activation is the controlling operating cost which determines
the suitability of the system for any given application.
Cost estimates of carbon reactivation relative to the amount
of air  purified show the trend to be  expected.   The  equation
used for  the estimates is

  Wc reactivated x $/lb for  react.                    ,  .
  -	1 x [F/1000]	 '  $/hr  Per  10 ft>/min.

The estimated costs  for three service  times selected  from
Table 6-III  are given  below  when  cost  of reactivation, re-
activation  losses and  handling are $0.25/lb:

           t.  hr         $/hr per  1,000 ftVmin

           2450.                  0.034
            360                   0.23
            180                   0.47
 For  the  2450-nr  service  time  the  $0.034  reactivation  cost  is
 not  significant  relative  to  the  fixed  operating  costs,  which
 are  generally  near  $0.50/hr  per  1000  ft'/min  for carbon-
 resorb  systems.   Carbon  reactivation  costs  of $0.47/hr  per
 1000 ftVmin would  make  the  multistage system noncompeti ti ve
 with catalytic incineration  and  possibly with thermal incin-
 eration  if the airstream is  already  at an elevated temperature.
 See  Figure 6-31  for estimated operating  costs of the  different*
 types of systems.

         A disadvantage of this system is that adsorbed pollutants
    containing halogens, nitrogen or sulfur would repollute the
    atmosphere when desorbed and decomposed during reactivation.

         "Hypersorption" System

         Figure 6-27  illustrates the "Hypersorption" system as
    developed by Berg  (1947) for the separation of the light
    hydrocarbon gases:  hydrogen, methane, ethane and propane.
    The hot carbon from the steam regeneration, at the lower end
    of the column, is  gas lifted to the hopper at the top.  On
    passing through the cooler, the carbon 1s cooled with recycle
    gas used in the gas lift.  This is the least-strongly-adsorbed
    gas of the mixture.  In the adsorption section, the downward
    moving carbon flows countercurrent to the gas mixture entering
    the system at the  inlet port.  The least-strongly-adsorbed gas
    is partially exhausted from the system at the discharge port
    and partially as  the recycle gas.  The more-strongly-adsorbed
    components of the  mixture are carried downward with the carbon
    through the rectifying section and to the stripper.  In the
    stripper the most-strongly-adsorbed gas is desorbed and leaves
    the system at the  make-gas port.  The reflux of the most-
    strongly-adsorbed  gas displaces the next-strongly-adsorbed gas
    upward in the carbon bed.  This gas leaves the system at the
    side-cut port.  The reflux of the next-strongly-adsorbed gas
    displaces the most-weakly-adsorbed gas upward in the rectifying
    section and thereby forcing It Into the adsorption section.

         The adsorption of methane, ethane and propane are quite
    small even at atmospheric pressure, being of the order of
    0.001, 0.01 and 0.20 Ib/lb, respectively.  The differences
    are significant so that separation is possible, but, because
    of the low capacities, frequent adsorption-regeneration
    cycles are required to attain a practical production rate.

         By only moving the carbon and by the countercurrent
    operation, three beneficial effects are realized:

         1.  In the regeneration phase only the carbon
             and adsorbate are heated; no heat Is wasted
             in reheating hardware;

         2.  Because of the countercurrent flow the full
             adsorptive capacity of the carbon is utilized;

         3.  The transfer of carbon between the adsorptions
             and regeneration phases is carried out con-
             tinuously in an efficient manner.

section """""""
C + a a m  ~ .

S >-
j 	 ;* -

k x
) :>\


r J^ u i a vn u 1 y c
> ^ ej j . _.. j.

I ii  *^
1 Gas lift
< for carbon
Figure 6-27  Moving-Bed  System  for Separation of Low
             Molecular-Weight Gases  (Berg, 1947)

          The  disadvantages  relative  to other carbon-resorb
     systems are  the  higher  structural costs and the loss of
     carbon by attrition  because  of movement through the column
     and  the gas  lift.

          This system was described because the basic concept
     is  a potential means for  removal of  low molecular-weight
     pollutants from  air.

          Rotating Axial-Flow  Bed with Vapor Recycle

          This system,  proposed  by Carr,  MSA Research Corp.,
     utilizes  a rotating  axial-flow carbon bed combined with a
     condenser and recycle scheme for recovery of  the desorbed
     vapors.   The carbon  bed,  illustrated in Figure 6-28, consists
     of  a horizontal, segmented  bed which is relatively shallow
     and rotates  at about one  revolution  per hour  or slower.  The
     pollutant vapor  recycle is  accomplished by a  purge airstream,
     which is  circulated  within  the unit  by a small blower.  This
     recycle  is illustrated  in Figure 6-30.  The main pollutant
     airstream is drawn down through  the  carbon bed and exhausted
     through  the  center duct by  a blower  located on the top of
     the unit.  Part  of the  purified  air  forms the purge stream
     (regenerating gas) which  is  first drawn up through the
     cooling  section  of the  carbon bed in order to cool the carbon
     and heat  the purge stream.   The  purge stream  is further
     heated  by an electric heater and forced up through the regen-
     eration  section.  The purge  airstream, now loaded with the
     desorbed  pollutant vapor, is cooled  by heat exchange with  the
     exhaust  air  and  further cooled to a  lower temperature  1n a
     small refrigeration  unit  where the contaminant is condensed
     and stored as a  liquid.  The purge airstream  from the  refrig-
     eration  unit is  warmed by heat exchange with  the purge air
     to  the  refrigeration unit and recycled by the blower above
     the carbon bed.

          The  rotation  of the  bed may be  intermittent, i.e.,
     rotated  a distance of one or more  segments at a  time and
     held for  a predetermined  time  period or  it may be rotated
     continuously.  When rotated intermittently,  the  desorption
     rate varies  over the time period in  a manner  similar  to
     curve given  in Figure 6-19.  When  regeneration  1s performed
     over two  segments  simultaneously but the  bed  rotated  a
     distance  of one  segment at each  time period,  the desorption
     rate is  considerably steadier  than  indicated  in  Figure 6-19.
     However,  any variation in desorption rate  1s  undesirable
     because  of adverse effects on  any  of the  subsequent  pollut-
     ant vapor disposal schemes that  might  be  used.   In  this
     respect  continuous rotation has  Its  advantages.

                           Cool ing
                               Impure  air
                              Carbon beds
                                             1ng gas
Figure 6-28  Rotating Axial-Flow Bed, Plan and Elevation
Carbon bed

                                            Lower  hood
                                            regen.  section

          (1)  Solenoid releases pressure during
              bed movement.

 Figure 6-29  Rotating Axial-Flow Bed, Showing Seal Detail
              and Mechanism  for Raising and Lowering Bed
Upper hood
Regen. and cool,
 section gasket

0-ring seal

    air  In

                                  air  out
                   Refrlger-  -*"![
                   ator unit
                                               Heat exchange
                                       Regenerating air
   Figure 6-30  Rotating Axial-Flow Bed with Condenser and
                Recycle for Uncondensed  Vapors, 3,800 ft'/mln
                Airflow  Capacity, 6 ft Bed Diameter

     The advantages of this  system relative  to  the  stationary
bed, are compactness and efficiencies  produced  by  continuous
operation.   The continuous discharge  permits  minimum size
and utility requirements for the auxiliary  recovery or  dis-
posal facilities.   The disadvantages  are  those  of  maintaining
seals, more elaborate and costly construction,  and  wear of
moving parts*  In  regard to seals, continuous wear  will occur
when rotation is continuous.  When rotation  is  intermittent,
the bed can be lowered during rotation to lessen  pressure
against the seals  and again raised to tighten the  seals for
the stationary time period.   In this  way  maintenance on seals
Is lessened.  Figure 6-29 Illustrates the seal  arrangement
and mechanism for  pressure release on the seals.

     The effect of vapor recycle on the adsorptive capacity
of the carbon is discussed in the previous text under Gas
Regeneration with  Condensation and Recycle.

     Materials of Construction

     As was the case for the air purification systems dis-
cussed in the prior text, one of the main concerns for
carbon resorb systems is corrosion.  Activated carbons in
contact with a metallic surface establish galvanic cells
when moisture or acidic vapors are present and thereby
cause rapid oxidation at the contact interfaces.   Many of
the pollutants are directly corrosive  to metals and some are
decomposed during adsorption or regeneration forming corrosive
acidic vapors.  The latter can cause deterioration, not only
of the retainer holding the carbon, but  also of structural
surfaces although  the most rapid  deterioration would be at
the metal-carbon interface.  To slow down or prevent cor-
rosion, structural members subject to  corrosion are con-
structed of  stainless steel or Hastaloy, or  if constructed
of mild steel, are coated with corrosion resistant plastics.
The  surface  coatings must also be  resistant  to elevated
temperatures  required during regeneration.   When the carbon
is  continuously or  intermittently  moved  through the bed,
stainless  steel construction is recommended  since  carbon  is
quite abrasive and would  rapidly  wear  away any coatings that
might be applied.

          A u x i 1 i ajry _Eq uj_pjn_e n_t

          The emphasis  in the prior discussions  of  air  purifi-
     cation and  carbon  resorb systems  has  centered  on the  adsor-
     bent bed proper since it is  the characterizing component
     of the overall system.   Important to  the  effective  function-*
     ing of the  adsorbent bed and ultimate disposal  of  the pollu-
     tants are  the other components of the overall  system.   The
     other components,  listed in  Table 6-IV, are considered
     auxiliaries  to the functioning of the adsorbent bed although
     they may also be  considered  as integral parts  of the  overall
     system depending  on one's point of view.   It is, however,
     outside the  scope  of this manual  to describe the auxiliaries
     in any detail but  within the scope to describe their  tasks or
     Table 6-IV  Auxiliary Equipment  for  Sorbent  Type Air
                 Pollution Control  Systems
          Hoods  and ducts
          Filters  and  scrubbers
          Air coolers
          Steam  generators
          Inert-gas heaters
          Refrigerating  units
Thermal incinerators
Distillation columns
Heat exchangers
Carbon reactivators
          Hoods  and  Ducts  -  Hoods  to  collect  the  pollutant  vapors
     from the  point  sources  with minimum  airflow  and  ducting  the
     collected vapors  to  the sorbent  system are an  economic neces-
     sity.   By ducting,the quantity of  air to be  purified 1s  kept
     smaller and the  pollutant  concentration  maintained  higher.
     Both act  to reduce  the  size of the sorbent system required
     and  to  reduce  the  operating costs.

          In existing  operations,  the retrofitting  of ducts and
     the  sorbent system  can  be  a problem.  Industrial processes
     are  often designed  to conserve space with little room  re-
     maining for later  installation of  pollution  control equip-
     ment.   Carbon  resorb  systems  are frequently  Installed  out-
     doors  necessitating  extra  long ducts.

          Filters and  Scrubbers -  When  partlculate  matter 1s
     present in  the  pollutant airstream,  1t should  be removed from
     the  airstream  prior  to  the removal of the vapors.   Partlculates
     tend to form coatings on surfaces  they contact.  When  the
     particulate deposition  occurs in air coolers and blowers their
     operating efficiencies  decrease  and  cleaning becomes necessary

Particulates plug adsorbent beds and thereby necessitate a
costly replacement of the bed.   Some particulate matters are
toxic and should be completely  removed for this  reason.

     Particulates may be removed from the  alrstream with
large area fabric filters made  of organic  and/or glass
fibers.  These may be in a pleated form or bags.  Bag  filters
have a beneficial feature in that they may be cleaned  and
used repeatedly.

     Fabric filters of several  efficiencies are  available.
The high efficiency filter is used when finely divided  toxic
material is present such as chemical warfare aerosols,  bacteria
or radioactive dust.  These filters have an efficiency  of
99.97% in the capture of 0.3 micron diameter particles.   When
the concern is more with protection of the adsorbent bed, an
intermediate or roughing filter may be used.  The intermediate
filters are considerably less efficient than the high  effic-
iency filters and are generally rated according  to the  National
Bureau of Standards (MBS) atmospheric air  stain  test at over
30% efficiency.  The roughing filter takes out the coarsest
dust and are not rated according to NBS test.  The cost and
resistance to airflow increase  as the efficiency is raised.
As the filter becomes coated with partlculates,  Its efficiency
Improves as also the resistance to flow.  The initial  resis-
tance of the Intermediate filters 1s between 0.25 to 0.55 1n.
of water pressure drop and when it rises to 1.0 in. during
usage the filter is replaced.                            *

     Water scrubbers may under certain conditions be more
effective than a dry filter.  This is the case when the
pollutant mixture contains highly soluble but poorly ad-
sorbed vapors.  When a scrubber is used a demlster should
also be used and the alrstream heated to lower  the relative

     Air Coolers - Exhaust air or gases from industrial
operations are generally at elevated temperatures.  To main-
tain the adsorptive capacity of the carbon at an effective
level and also to attain more uniform adsorption-regeneration
cycles, the exhaust air is cooled down to at least 95*F.

     Blowers - One or more blowers are required in the carbon
resorb systems as described in the previous text of this
chapter.  For the main airflow stream the blower must over-
come the resistance primarily of the carbon bed which can
vary from 2 to 18 1n. of water pressure drop.   Additional
resistance comes from the particulate filter which can

      vary from virtually zero to 1.0  in.  of  water.   Resistances
      to flow in the ducts, valves and  air  cooler  are  small  rela*
      tive to the carbon bed resistance.

           Figure 6-8 presents graphs  givlnq  the  pressure drop
      through granular carbon beds of  varied  mesh size  fractions
      and as  function of flow velocity.

           Additional, but smaller, blowers  are  required  for  qas
      regeneration,  recycle of uncondensed vapors and circulation
      to secondary adsorber as described in the various  proposed
      systems.   In some cases the blower is operated  at  an
      elevated  temperature near 300F,

           For  the air purification systems, the total resistance
      to airflow  can vary from less than  0.5  in. of  water  when
      a filter  1s used for nontoxic vapors, to over 1.5  In.  of
      water,  when a  filter Is required  for toxic  chemicals.

           The  power consumption to operate the blower  is one of
      the operating  costs and varies with  air  volume  to  be  moved
      and resistance to flow.

           Steam Generators - Unless steam Is  available  from
      other operations 1n the plant, a  steam  generator  1s re-
      quired  for the carbon regeneration phase.   A low  pressure
      steam at  near  212F temperature  has  been found  most effec-
      tive in solvent recovery operations. Super heated steam
      has not been found to be much more effective than  212F
      saturated steam.  The enthalpy of the vapor at  tempera*
      tures above 212F vaporizes the  condensed steam and there*
      by the  latent  heat of condensation is lost  to the  regen-
      eration process.  High temperature,  high pressure  steam
      could reduce the steam-to-sol vent ratio  but the benefit
      1s offset by the Initial  cost of  heavier constructed  ad-
      sorber  and valves.

           Steam generation is one of  the  higher  operating  costs
      1n solvent recovery operations.

           Inert Gas Heaters - When steam  regeneration  is not
      feasible, the  carbon bed can be  regenerated with  air  heated
      indirectly by  an electric or gas  heater or  the  flue gas or
      products  of combustion from the  gas  burner  can  be  used
      directly  as an inert gas.  With  Inert gases, regeneration
      temperatures of 200Q to 300F can be safely used.  Higher
      temperatures begin to have drawbacks 1n that cost savings
      are nil and decomposition of the  adsorbed vapors  are

     Heating of the regenerating gas  is  another operating
cost but has been found to be small  and  does  not appreciably
affect the air purification cost,

     Condensers - When steam regeneration  is  used in  solvent
recovery, a water cooled surface condenser is  usually used  to
condense the steam and desorbed solvent  vapors.  The  cooling
water is one of the operating costs;  the usual  ratio  of  cool-
ing water weight to condensed steam  weight is  30 to  1.0.

     In the system utilizing gas regeneration  and vapor  re-
cycle, a condenser is used to liquify the  vapors from the
regenerating gas.  Where lower than  ambient temperatures are
required this condenser would be cooled  with  a  refrigerant.

     Refrigerating Unit - Required to cool condensers.

     Decanters - A virtually complete separation of  the  con-
densed solvent and steam condensate  can  be effected  by de-
cantation when the solvent is insoluble  in water.  Examples
of much used solvents of this type are toluene, xylene,  oc-
tane and naphtha.  Theoretically,  the method  is effective
even at large steam-to-solvent ratios of over 30.

     However, a large portion of solvents, or pollutants of
concern, are soluble to varying degrees.  Such  commonly  used
solvents as 4-nethyl-2-pentanone and  trichloroethene  are
slightly soluble.  The steam condensate  phase will always
contain a small concentration of the  solvent  and thereby be
a source of water pollution.  When the steam-to-solvent  ratio
is small (less than 6.0) the solvent  loss  is  not significant
costwise and the decantation method  is effective.  As the
steam-to-solvent ratio increases,  the loss of solvent in the
steam condensate increases and, ultimately, a point is
reached where other methods of recovery  become  more  economical,
such as condensation with recycle  or recovery with a  secondary
adsorber; see prior text on carbon resorb  system.  Often two
phases are obtained, one rich in solvent and  the other poor
in solvent.  This usually poses a  disposal problem.   In  some
cases the solvent rich phase can be  distilled to recover the
solvent and thereby defray the cost of disposal of the solvent
poor phase.

     Distillation Columns - Solvents  that are soluble in the
steam condensate can be separated  from the steam by direct
fractionation of the steam-solvent vapor mixture.  This  pro-
cedure is economical only when the steam-to-solvent ratio   is
low and the solvent quite valuable.   Liquid phases are also
obtained from decantation, which can be economically dis-

         Thermal  Incinerators - When the desorbed pollutant vapor
     cannot be effectively recovered by steam regeneration and
     decantation or distillation, a more economical approach is to
     regenerate with a noncondensable gas and burn the desorbed
     vapor immediately.  For this purpose a small thermal inciner-
     ation is well suited.  The size of the incinerator can be less
     than one tenth of the size normally required to incinerate the
     vapors from the primary airstream.  See prior text under Gas
     Regeneration  with Incineration.

         Heat Exchangers - The proposed carbon resorb systems as
     illustrated by flow diagrams in Figures 6-17, -20 and -21
     simultaneously heat and cool gas streams.  By transferring
     the heat from one stream to another, savings on cooling water
     and fuel are  realized.

         Carbon Reactivators - In carbon resorb systems such as
     the multistage with bed displacement, Figure 6-25, and the
     fluidized bed, Figure 6-16, the carbon is removed from the
     portion of the system where the adsorption phase 1s performed.
     When the carbon shows signs of activity loss it can be readily
     passed through a high temperature steam activator operating at
     temperatures  over 1600F.  This is the standard procedure with
     the multistage system.  For large installations where a small
     activator can be operated continuously the cost of reactivation
     would be greatly reduced.

         Monitors and Controls - In solvent recovery operations,
     the solvent to be recovered is known and its concentration
     determined from solvent usage rate and exhaust airflow rate.
     The service time to be expected may be known from past ex-
     perience with other similar installations or can be estimated
     by using the  procedures as described in Chapter 4.  It then
     becomes a matter of setting a time schedule for change of
     adsorbers for the adsorption and regeneration phases.  This
     can be automatically controlled with a timer and mechanically
     operated valves.

         Manometers, to measure the pressure drops across the
     filter and carbon bed, signal the need for change of filter
     and/or carbon bed when a pressure rise occurs.

     A temperature recorder before the  bed  can  be  used to
determine the efficiency of the air cooler,  if  one is  being
used, and a temperature recorder on the effluent side  of
the bed can be used to monitor the time of  adsorption  zone
penetration.  The approach of the adsorption zone  to the
effluent end of the bed is accompanied  by  a  temperature rise
in the effluent stream.  This can be used  as a  signal  to
change adsorbers.

     Steam input during regeneration should  be  metered to
avoid under or over regeneration and also  timed to fit the
regeneration phase within the overall  cycle.

     The auxiliary equipment, the condensers, distillation
columns, incinerators, etc, also need to be  monitored  and
their operations controlled to integrate them into the over-
all operations.

     When the carbon resorb system is involved  in  less well
defined areas of pollution control, the monitoring and con-
trol functions are more involved and difficult.  It is
necessary to identify the components and determine their
concentrations in the pollutant airstream.   These  analyses
are required before a decision can be made  regarding the
type and size of control system best suited for the partic-
ular emission source.  At low pollutant concentrations the
adsorption zone temperature can be so slightly different
from that of the rest of the bed that it is not a  reliable
indication of the adsorption zone position  in the  bed.  In
cases of this type, a gas analyzer is required to  signal
the adsorption zone penetration of the bed  and thereby avoid
an overrun.

     Chapter 8 discusses instruments that are suitable for
monitoring the adsorbent systems operations.

     Operation of Carbon Resorb System

     Operating Parameters - Unlike the air purification
systems, the carbon resorb systems are subject to a much
greater variety of conditions which demand modifications to
the equipment and modes of operation in order to attain the
most effective and economical service for any particular
application.  The prior text, Carbon Resorb Systems, describes
various proposed and successfully used systems which show
the variety of equipment modifications that are possible
This part of the chapter discusses the operating parameters
which form the rationale for these modifications.   The

      operating parameters that are controlHnq to varying degrees

           Pollutant type            Flow velocity
           Pollutant concentration   Adsorption-regeneration cycle
           Carbon type                 frequency
           Temperature               Regenerating agents

           Pollutant T.yne -  In assessing the pollution control
      situation for an  adsorbent system, the most Important prop-
      erty  of  the pollutant  that should be considered 1s Its molar
      volume,  designated by  Vm 1n this handbook.  Vm 1s used 1n
      the adsorption potential equation, Equation 4-44 to calculate
      A as  used in the  generalized adsorption isotherms, Figures
      4-11  and -12.  The equation 1s reproduced in a modified
                    A -  [T/Vm] log
                                                        and A
                                                   nverse ratio
                                                        See Chap.
The adsorptive capacity decreases with Increased  A
increases with decrease 1n Vm because of the  1nver
and also because C0 Increases with decreasing Vm.   see Chap
ter 4, Potential Theory Equations, for further discussion.
Appendix C-l gives the Vm values for pollutants of interest
In control of small emission sources.

     If Vm is between 80 and 190 cmjlig/mol,  there generally
are no real problems encountered 1n the adsorption and de-
sorptlon phases of the cycles.  Exceptions  occur  when the
adsorbed vapor polymerizes and Vm thereby increases by a
large amount.

     Vapors with Vm greater than 190 cmjliq/mol tend to be
difficult to desorb by the commonly used method,  steam at
212F, or by a noncondensable gas at temperatures  up to
300F.  If a carbon resorb system is used and vapors of
this type are dominant 1n the pollutant mixture,  high tern-
perature steam reactivation (over 1600F) becomes  a neces-
sary part of the operating procedure and also becomes a
substantial operating cost.  The multistage system (Melshlmtr
system) 1s designed for these conditions in that the trans-
fer of carbon bed between the reactlvator and adsorber can
be done rapidly and with minimum of labor.   The continuous*
moving bed (Hypersorber) and fluldlzed bed  systems are also
readily adaptable to outside reactivation.

     When the pollutant Vm are less than 80 cm3!Iq/mol, each
pollutant has to be evaluated separately because the effect
of the functional group becomes pronounced, but the general

trend is  for these  pollutants  to  be  poorly  adsorbed.   The
alternatives are to;  (1)  operate with  a  system  that  can be
economically regenerated  frequently  or  continuously,or
(2) use an Impregnated adsorbent  that  is  specific  for the

     Many pollutants have no recovery  value.   This creates
a disposal problem  after  the pollutant has  been  desorbed
from the carbon and alters the mode  of operation,  which may
be different than used in solvent recovery.  When  the pollu-
tant is completely  combustible it can  frequently be disposed
of with the least cost by immediate  incineration after de-
sorption.  When the pollution is  noncombustible or contains
groups or elements  such as NOX, SO*  or Cl that will repollute
the atmosphere on incineration, it again becomes necessary
to recover the pollutant and hold it in storage until some
other means is used for its disposal.

     Pollutants, or solvents, vary in their stability when
in contact with the carbon, and the presence or lack of
stability can considerably alter the operation and economics
of the control system.  For example, cyclohexane  and tri-
chloroethene are quite stable and carbon beds used in  their
recovery  last over  10 years before replacements are  necessary.
4-Methyl-2-pentanone  undergoes a small amount of  polymeriza-
tion causing a shortening of useful  bed life to 3  to 5 yr.
1,1,1-Trichloroethane decomposes forming an acid  as  one of
the  decomposition  products.  To avoid corrosion when recover*
ing  this  solvent,  Hastaloy construction 1s used.   2-N1tro-
propane  decomposes  yielding N02; carbon-bed fires  have been
reported  when  this  solvent has been adsorbed.  2-Butanone
decomposes  to  a small extent  and carbon bed fires  have also
been reported  in this case.

      In  any  new  situation,  it  is necessary to know the ad-
sorbabilities  of the  pollutants  and also their stabilities
before the  best  suited control unit can  be designed  and its
operating procedures  worked  out.

      Pollutant Concentration  - It is  desirable  to have the
 pollutant,  or  solvent, vapor  concentration high.  I.e., up
 to the 2555  LEL,  since the adsorptive  capacity increases with
 concentration.   A  smaller carbon bed  is  then  required to
 control  the vapors emitted from  any given  source. The higher
 concentration  allows  operation with a larger  operating capac-
 ity for a given regenerating agent  usage.   Equation  4-44  shows
 the effect of  change  1n  Ci on A  and thereby  on the adsorptive

          The use of hoods, covering the point sources,  and  duct*
     ing of the pollutant vapors to the control  unit,  as a means
     for attaining higher concentrations, are  important  first
     considerations in assessing the pollution control situation.
     When the circumstances are such that the  concentration  is
     low and cannot be raised by such means,  the other components
     of the system need to be designed so that the  air can be
     purified as economically as possible.   This may mean the  use
     of an inert, noncondensable gas for regeneration  rather than
     steam.  The desorbed vapors may then be  incinerated, or if
     incineration is not feasible, the desorbed  vapors may be
     readsorbed in a secondary adsorber and recovered  for later
     disposal as a liquid.  The desorbed vapors  may also be  re-
     covered by the condensation-recycle procedure  as  described
     earlier in this chapter under Gas Regeneration with Conden-
     sation and Recycle.

          The heat evolved during adsorption  can under conditions
     of high concentration create problems  that  lower  the effec-
     tiveness of the carbon resorb system.   The  heat is  evolved
     in the adsorption zone and most of the heat is Initially
     absorbed by the carbon in the adsorption  zone. The adsorp-
     tion zone and heated-carbon zone move  down  the bed  at the
     same rate.  At high concentrations the temperature  rise can
     be sufficient to lower the adsorptive  capacity of the ad-
     sorption zone and thereby lower the bed  capacity.   The
     elevated temperatures can also cause vapor  decomposition
     which may be especially undesirable in the  presence of  air.
     In such cases internal bed cooling is  necessary as  done by
     cooling coils.  At low concentrations, the  temperature  rise
     is not sufficient to cause any problems.   In solvent recovery
     operations, when temperature sensitive solvents are being
     recovered, the adsorbers are operated  in  such  a manner  that
     at least 12% moisture remains on the carbon after the steam
     regeneration.  The desorption of the water  during the ad-
     sorption phase moderates the temperature  rise. See Chapter
     4, Adsorption Thermodynamics, for discussion on heats of

          Carbon Types - Commercially available  carbons  offered
     for vapor phase adsorption by the manufacturers fall Into
     fairly narrow ranges in regard to density and  to  adsorptive
     capacity as measured by the CC14 activity.   The density
     range is 28 to 31 Ib/ft* and CC14 activity  range  1s 50X to

          A variety of mesh sizes are available  but those of
     greatest interest in carbon resorb systems  utilizing fixed

beds are the coarsest ranges,  4 to 6  and  4 to 10 U.S.  sieve
size.  The coarsest mesh sizes are used  to keep the resis-
tance to  airflow low although the coarse mesh size also
lengthens the adsorption zone  length.  The economics favor
the reduced pressure drop in this  mesh size range.   Carbons
coarser than 4 to 6 become more expensive to manufacture,
are more difficult to activate uniformly, are more  friable
and also produce a considerable increase  in the adsorption
zone length,

     To know the response of a carbon  under the many varied
applications, a generalized adsorption isotherm for the
carbon is necessary.  This information is seldom offered as
part of the sales literature.   The CC14  activity gives the
adsorptive capacity at only one point (A  2.5) and is at
near saturation capacity.  The slope  of  the generalized
adsorption isotherm (Figures 4-11, and -12) at A greater
than 2.5 has a bearing on whether the carbon is best suited
for solvent recovery or air pollution control at low vapor
concentrations and for poorly adsorbed vapors.  The isotherm
with the steep slope and high  CC14 activity is the better
carbon for solvent recovery while the carbon with a small
slope is better suited for adsorption at the low concen-

     At times the ash content and ignition temperature of
the carbon become important parameters.   The type of ash
present can affect the ignition temperature and can cata-
lyze the decomposition of the adsorbed vapors.

     Hardness becomes an Important parameter in applications
where the carbon 1s moved frequently as  in the multistage,
continuous moving bed or fluidized bed systems.  In this
respect the pelletized carbons show a greater resistance.

     Additional Information on carbons is presented at the
end of this chapter under Adsorbents and Adsorbent  Systems
Manufacturers and Venders.

     Temperature - Relatively small differences in  the carbon
bed temperature cause large differences  in the  adsorptive
capacity.  This is an important factor in regard to the ad-
sorption and regeneration phases  of the  control process.
The adsorption phase is  carried out at as  low  a temperature
as  is practical and the  regeneration  at  an elevated temper-
ature at such a level where minimum damage  is  done  to  the
carbon, the adsorbed vapor  and equipment.

           The  temperature  effect  can  be demonstrated in an
      approximate  manner  by use  of Equation  4.44 and a  generalized
      adsorption  isotherm for  any  selected pollutant vapor.  Table
      6-V  presents estimates on  the  effect of temperature on the
      adsorptive  capacity for  4-methylf2-pentanone on BPL V-type
      carbon,  isotherms  in  Figure  4-12,

      Table 6-V   Effect  of  Temperature on Adsorptive Capacity
                 and  Equilibrium Concentration for 4-Methvl-2-
                 pentanone  Adsorbed  on DPI V-Type Carbon     *

      Temp. ,F   u, Ib/lb when C^lOO  ppm   Cj .  ppm when u-0.20
         77                0,20                          100
        105                0.17                          300
        120                0.15                          640
        212                0.07                       13,000
        300                0,03                      110,000

          As  is  apparent from the capacity figures in the second
      column,  the capacity decreases with increased temperature.
      A  temperature is ultimately reached where it is more econom.
      ical to  cool the airstream  than to operate at the reduced
      capacity.   The generally accepted upper temperature limit
      for the  adsorption phase is near 120F.  At this temperature
      the capacity is already down to 75%  of the capacity at
      77F,  Some solvent recovery systems are operated at 105F,

          To  regenerate the carbon, it is heated with steam or a
      noncondensable gas which also serve to sweep out the released
      vapors.  Referring to Table 6V, if the carbon bed, saturated
      to w  0.20 Ib/lb, were sealed off and heated to 212F the
      4-methyl-2-pentanone concentration would rise to approximate**
      ly  13,000 ppm and at 300F to 110,000 ppm.  On releasing the
      pressure a  considerable amount of the vapor would desorb and
      with a sweep gas the desorption rate would be accelerated,
      To  remove all of the adsorbed vapor, it would, however,
      require  a large amount of sweep gas.  At 212F, when 
bed area that will give a practical  flow velocity.  The
decision on the practical flow velocity takes into consid-
eration the adsorption-regeneration  cycle frequency, the
bed dimensions and resistance to flow.   For a given flow
rate and concentration, the adsorption-regeneration cycle
frequency is determined by the bed volume, or weight, and
operating capacity.  In solvent recovery the cycle time is
near 1.0 hr while in pollution control  the adsorption phase
may last the duration of an 8-hr work day and the regener-
ation performed during off-working hours.  The dimensions
of the bed may now be adjusted for the  fixed bed volume
that will give the best compromise on the resistance to
flow and the structural design of the bed.  A bed with a
large area and short length will operate at lower flow re-
sistances but for structural reasons may not be the best

     The air velocities commonly used are in the range 30
to 110 ft/min in the various types of carbon resorb systems.
The trend is toward the higher velocities, particularly in
the stationary bed solvent recovery systems.  Bed lengths
range from 8 to 36 in.  In order to hold the flow resistance
as low as possible, coarse, 4 to 6 mesh size, pelletized
carbons are frequently used.  High flow resistance, aside
from consuming more power for the blowers, also cause oper-
ational problems with the valves.

     When pressure drop measurements are not available for
the particular carbon being considered, an estimated pressure
drop can be read-off from predetermined curves such as shown
1n Figure 6-8, or calculated with Equation 4-42.

     Flow velocity also affects the adsorption zone length,
although in carbon resorb systems this  is not a critical
factor.  Even at the 110 ft/min flow velocity and for coarse
mesh carbons, the adsorption zone length is less than 4.0
in., compared to total bed length of 12 in. and longer.
Equation 4-32 gives the relationship between flow velocity
and adsorption zone length.

     Adsorption-Regeneration Cycle Frequency - The optimum
cycle frequency is attained when the combined cost of the
adsorption and regeneration phases is at a minimum.  For a
fixed operating capacity, the adsorption phase cost increases
with service time because of the cost of additional carbon
required and the increase in the per unit-time power con-
sumption for the blower.  When the service time is shortened
I.e.. regeneration frequency 1s Increased, the regeneration  '
cost rises because of the additional fuel consumption for

      the more frequent heat-up of the carbon bed and associated
      hardware.  Steam utilization is also less efficient for the
      shorter carbon bed.  The adsorption phase time period and
      the regeneration phase time period need not be equal although
      equal phase time periods have worked-out best for stationary
      bed solvent recovery systems.

          To avoid underrun or overrun on the adsorption or re-
      generation phases, monitoring devices are desirable to signal
      the completion of each phase.  For the adsorption phase, the
      monitoring device would be required to detect the penetration
      of the adsorption zone lead edge.  For the regeneration phase
      the monitoring device would be required to measure the vapor
      concentration in the regenerating agent and, when the con-
      centration has dropped to-a predetermined  level,   signal
      the end of the regeneration phase.

          Regenerating Agents - The regenerating agent used in
      solvent recovery is steam.  It has several properties that
      make  it uniquely suited for this purpose.  Its saturated
      vapor temperature is at a temperature level where many of
      the solvents of recovery value are poorly adsorbed or become
      readily desorbable, but the temperature is not too high to
      cause extensive damage to the solvents.  The heat of con-
      densation is high, hence saturated steam rapidly delivers
      a large amount of heat to the carbon at virtually a con-
      stant and moderate temperature.  A great many solvents are
      fairly insoluble in water; condensation of the steam-vapor
      mixture and decantation effect a satisfactory separation
      and liquid solvent recovery.  Steam is fairly Inactive with
      most solvents although some decomposition is reported either
      because of direct action with the steam and/or catalytic
      help from the carbon.  The steam also provides an atmosphere
      in which combustible vapors can be handled safely at high

          In solvent recovery with steam regeneration the carbon
      acts as a collector or transfer medium.  It collects the
      solvent from the noncondensable air and transfers 1t to the
      condensable steam from which it can be readily separated
      by decantation or distillation.  Usually the volume of
      steam used is not much less than the vapor-laden air and
      the concentration of solvent Initially 1n the alrstream
      is not much less than the solvent concentration 1n the

          In air pollution control applications where the
      pollutant concentrations are  low  and, in addition, the

pollutants may have no recovery value,  steam regeneration
is frequently not the best regenerating agent.   Because of
the adsorption at low concentrations, the vapors are held
tightly by the carbon and, relative to  the amount of pollu-
tant adsorbed, the amount of steam required is  large.   Under
these conditions even slightly soluble  organic  compounds are
completely dissolved in the steam condensate.   Under these
circumstances the more economical approach is  to regenerate
with a noncondensable gas such as air,  or, if  there is  danger
of explosion, regenerate with an inert  noncondensable  gas
such as flue gas.  As already described earlier in this
chapter the use of noncondensable gas entails  considerable
changes in equipment and operating procedures.

     In the adsorption processes at the low pollutant
concentration, the carbon acts as a concentrator.  The  car-
bon collects the pollutant vapor from a large  volume of air
over a long period of time during the adsorption phase.
During regeneration the vapors are released from the carbon
at a high concentration, in a relatively small  volume  of
regenerating gas, and in a relatively short period of  time.
Disposal  of the pollutant now becomes an easier task.   The
effluent mixture of regenerating gas and pollutant vapor may
now be treated in a small incinerator to burn  the vapor, or
be passed through a condenser to liquify the vapor for
separation, or readsorbed in a secondary adsorber and  re-
covered by steam regeneration.

Polar Adsorbent Systems

     No installations appear to be in service  for air-
pollution control which utilize the polar adsorbents:  silica
gel, activated alumina or molecular sieves.  They are, how-
ever, being investigated for this application  under very
specific conditions.  The molecular sieves have a potential
application for control of SOX and NOX  from exhaust gases.
Because of the polarity and water solubility of these  mole-
cules, they can be effectively adsorbed from a humid air-
stream.  Silica gel and activated alumina have a potential
application in the control of high molecular weight vapors
(of Vm greater than 190 cm'/mol) from dry  airstreams
Their greatest usefulness is as desiccants and also for the
separation of dry organic liquids and vapors.   See Chapter
4, Adsorbent Types, for discussion of polar adsorbents.
     For air pollution control the systems design would
not be greatly different from the carbon resorb type   5
modification would be required because of differences in
mode of regeneration.  Steam or water would be the most

      probable  regenerating  agent.   Prior to start of the adsorp.
      tion  phase,  the  water  would have  to be removed from the
      adsorbent.   Molecular  sieves  require  fairly high drying

      Impregnated  Adsorbent  Systems

           Iodine  Impregnated  carbons are used  1n atomic enerqy
      plants  to prevent  escape of radioactive methyl Iodide  1n
      case  of a breakdown  of the reactor.  This  1s an exchange
      reaction  of  radioactive  Iodine 1n  the methyl Iodide with
      nonactlve Impregnant on  the carbon.  The  type filter used
      1s  Illustrated  1n  Figures 6-3  and  -4.  These are considered
      nonregeneratlve  although In time  the radioactivity would
      decay completely.

           For  toxic  chemicals such  as  phosgene, cyanogen chloride
      and hydrogen cyanide,  carbon  Impregnated  with copper oxide
      and hexavalent  chromium  are very  effective.  The type  filter
      used  1s Illustrated  in Figure  6-3.  These  are nonregenera-
      tlve.  When  the  chemicals have reacted with the stolchlo-
      metrlc  amount of pollutant, the carbon 1s  replaced.

      CatalystImpregnated Carbon Systems

           The  use of  catalytic Impregnants to  permit 1n-place
      reactivation of  a  spent  carbon bed was proposed and labora-
      tory  tested  by  Turk  (1955).   The  carbon 1s Impregnated with
      2%  to 3%  of  an  oxide of  chromium,  molybdenum, tungsten,
      cobalt  or copper.   After the  carbon has become saturated
      with  organic vapors  by normal  adsorption,  the reactivation
      1s  Initiated by  raising  the temperature of the bed  to  effect
      catalytic oxidation  of the adsorbed organic compounds  or
      their vapors as  they are being desorbed.

           As 1n catalytic Incineration,  the oxidation rate
      Increases with  Increased temperature, I.e., complete oxida-
      tion  becomes more  probable as  the temperature 1s raised.
      In  catalytic Incineration, operating temperatures  at which
      combustion efficiencies  are 90% and better with reasonable
      sized units  are 1n the range  600 to 900F,   Ignition  temper.
      atures  of carbons  vary but are between 600 to 700F,  hence
      the optimum  temperatures for  combustion catalysis  cannot  be
      used.  Some  sacrifice  1n regard to completeness of  combustion
      1s  necessary unless some corrective measures  are taken.   A
      corrective measure that  can be taken to avoid air  pollution
      by  the  combustion  products  1s to  recycle  the  gases  through
      the bed or Install a small  thermal  Incinerator to  complete
      the combustion.   Discussions  on catalytic Incineration are
      given In Chapter 7 and on the use of thermal  Incinerators

for combustion of desorbed pollutants  are given 1n this
chapter under Carbon Resorb Systems.

     The potential  advantage over an  unlmpregnated carbon
system would occur  when the pollutants are of type that
polymerizes on adsorption, thus making their desorptlon
Impossible by the more conventional  regeneration procedures,
or are nondesorbable because of high  molecular weights.  In-
place catalytic oxidation of these pollutants to lower mole-
cular weight components would effect  their desorption.

     The construction of the adsorbent system would be
similar to those already described for carbon resorb but
modified so that regeneration at temperatures up to 600F
are possible.

Economic Analyses

      General  Discussion

      In the  evaluation of  the  adsorbent  systems relative to
the  catalyllc  and  thermal  Incineration  systems,  the  final
decision  as  to  which  type  of system should  be  used  for  a
given control  problem  is  based  on the cost  of  cleaning  a
unit  volume  of  air.   When  a cost analysis  clearly  indicates
that  a carbon resorb  system should be used, further analysis
involving  the operating  parameters can Indicate the optimum
operating  procedure.   An  approach Is  taken  in  this  section
of the chapter that supplied answers  in  both of these areas.

      To work up the estimated  costs  of air pollution control,
data on various operating costs are  required.   Common to all
 three types of systems are the costs  of:

      Operating labor
      Electric power

 The  catalysts used in catalytic incinerators frequently
 deteriorate  faster than the rest of  the system.  Its re-
 placement thereby becomes  a direct major operating  cost and
 should be added to the above list when  referring to  cataly-
 tic  incineration.  Frequently,  activated carbons also  de-
 teriorate faster than the  rest  of the system;  the  replacement
 cost should  be  added  to  the above list  when referring  to
 carbon resorb  systems.   An additional cost in  the  operation
 of  some  carbon  resorb systems  1s that of  cooling  water used

      for  vapor  condensation.   In  order  to determine the depre-
      dation  cost,  the  initial  equipment and-installation costs
      (fixed  capital)  should  be  ascertained.

           Cost  estimates  for carbon  resorb  systems were prepared
      for  4-methyl-2-pentanone,  trichloroethene, and propanone at
      influent flow  concentrations  of 10, 100, 300, 500, 1000 and
      3000 ppm and flow  rates of 800  and 3,800 ft'/min.  The es-
      timates  are presented  on a $/hr per 1000 ft'/min flow rate
      basis.   The systems  analyzed  are:

           1.   Stationary  bed with  steam regeneration and
               solvent separation  by  decantation

           2.   Stationary  bed with  air regeneration and
               thermal Incineration of combustible desorbed
               pollutant vapors

           3.   Rotating  axial  flow  bed with  air regeneration,
               desorbed  vapor condensation and vapor recycle.

           Comparative costs  of  air cleaning  at varied pollutant
      concentrations are presented  for carbon  resorb and the
      incineration systems.   The effect  of pollutant molar volume,
      pollutant  concentration, flow rate, and  adsorption-regeneration
      time on  the cost of  air cleaning are also Investigated.

           Operating Costs

           Fixed Capital - Cost  Information  for the catalytic  and
      thermal  incineration processes  was collected from several
      sources: ASKRAE  (1966  and  1967), Meinhold (1963), UOP, -and
      Phillip  (1971).   Equipment and/or  depreciation costs from
      these sources  were generally  for larger  capacity systems than
      the  3,800  ft3/min   airflow.   In these  cases  the 0.6 power
      factor  was used  to estimate  the cost for the system at the
      3,800 ft'/min  flow capacity,  thus:

      Cost for 3,800 system   [3,800/F]   [cost of larger system]

      where F  is the flow  capacity  of the larger system.

           Installed costs for the  catalytic incinerator vary
      between  $3,500 and $6,000 per 1,000 ft'/min  air handling
      capacity.   The higher  end of  the range includes cost for
      heat exchangers, filters for  partlculate removal and also
      reflects the costs of  more expensive catalysts.  As the
      size increases,  the  cost per 1000  ft'/min flow capacity
      decreases  to some  extent, but for  the  range  2,000  to  10,000

ft /min flow capacities, the costs per unit flow volume were
assumed essentially constant.  For these estimates, the
higher $6,000 figure was used at concentrations below
the self-sustaining concentrations (below about 3.000 pom)
since heat exchangers give a decided operating cost reduc-
t10n%o^rrtJhe 3800 ft /m1n system the installed cost is
then $23,000*

     The installed cost for a thermal  incinerator of 3,800
ft'/min airflow capacity, with heat exchanger is approxi-
mately $29,000 or $7,600 per 1,000 ftVmin capacity.

     For custom designed solvent recovery systems, the
installed cost will usually range from $10,000 to $15,000
per 1000 ft'/min airflow capacity.  Package type systems
are available at a considerably lower cost.  The equipment
cost at the manufacturer's plant can range from $2,000 to
$3.500 per 1,000 ft'/min capacity.  Freight and installation
costs must be added to these figures (Cooper, 1967).

     For the air pollution control cost estimates presented
here, the base prices on stationary bed systems are those
provided by Vic Manufacturing Co. and Hoyt Manufacturing
Corporation.  Table 6-VI presents the approximate systems
costs and sizes over a range of airflow capacities from
700 to 5,000 ft'/min per adsorber.  Single and double ad-
sorber systems of 3,800 ft'/min capacity cost $19,480 and
$32,500. respectively.  Installation costs raise the fixed
capital costs to $21,500 and $35,800,  respectively.  A cost
estimate was also prepared for airflow of 800 ft'/min   A
dual adsorber system at this capacity level costs $6.200
and when Installed, $7,000.

     For the stationary bed utilizing air regeneration and
thermal incineration, the above costs were used for the
adsorber section of the system and cost added for scaled
down sized thermal incinerators.  The installed cost of a
single adsorber, 3,800 ft'/min capacity system with desorbed
VJPSnn !ClJnna^s/fer?t1n9 at re9enet1ng gas flow range
of 200 to 500 fts/min 1s approximately $26.000.

     For the axial flow bed continuous recycle system  it
was necessary to work-up the cost from basic component
parts.  The size and cost for a 3,800 fts/m1n capacity
system varies some with type of pollutant and concentration.
The installed cost range Is $16,500 to $19,800.       atlon-

     Depreciation - For the carbon resorb. catalytic and
thermal incineration systems, a 10-year depreciation rate

Table 6-VI



Equipment Cost and Slze^1^
c'e, $
9 "
| ~f %* W
Auto, or Motor per tank





1 500
vreiqht ,
w x 1











x 1

x h
x 81
x 81
x 81
x 81

x 95

8 X 114

x 137

82 x 140

 (1)  Based on Information
     Manufacturing Company.
 (2)  Tank or adsorber.
provided by:  Hoyt Manufacturing Corporation and Vic

1s used in the estimates  prepared  in  this  handbook  although
systems lives  of up to 20 yr are  claimed under  various  cir-
cumstances.   The control  system life  can vary  considerably,
before major repairs are  required, depending on corrosive-
ness of the  pollutant vapors, frequency of the  adsorption-
regeneration cycles, and  continuity of operation,  I.e.,
whether intermittent or continuous.

     For the estimates in this handbook, an 8-hr da,  2080-hr
yr, operating  basis was used.  For carbon  resorb systems, the
depreciation cost on the  8-hr basis is approximately  one-half
of the total operating cost.  For the Incinerators, the de-
preciation cost on the 8-hr basis is  about one-third  of total
cost at low concentrations but approaches  one-half  at 2555 LEL
concentrations.  By adjusting the operating time closer to
continuous operation, the total operating  cost can  be reduced
by 30%.  The estimated cost figures are on the high side in
this respect but retain their correct relative positions for
the three types of systems.

     Operating Labor - Low labor costs are generally claimed
for the operation of the carbon resorb and incineration
systems.  A low level of operating cost has been estimated
on the basis of 5 min of labor per 8-hr shift for each 1,000
ft3/min capacity (Hein, 1964).  For the estimates for this
handbook a considerably lower labor rate 1s used, $300/yr
at $10 per man  hour.

     Maintenance - Yearly maintenance  costs for catalytic
Incinerators have been estimated by some analysts on the
basis  of 10% of the  initial  incinerator cost (Hein,  1964).
Other  methods  can consist of  periodic  catalyst  replacement
cost combined with  a  low percentage rate of the initial
incinerator cost.   For the  estimates  in this handbook,  the
overall yearly maintenance  cost combines a $1,200/yr cata-
lyst replacement cost  and a  2% of  initial   capital.

     For  the  thermal  Incineration  a straight 2% of Initial
capital 1s  used for  the yearly maintenance cost.

      For  the  carbon  resorb  systems, the overall yearly
maintenance cost  combines  a $300/yr carbon replacement
cost  and  a  1%  of  Initial  capital.

      Electric  Power - The main electrical  power consumer is
the  air circulating blower.   The  kilowatts to  move 1,000
ft'/min of  air through the Incinerator or carbon resorb bed can
vary  from  1.5  to  6.0 depending on resistance  to flow.   Local
electrical  costs  per kilowatt-hour will determine  the power

           For  the  stationary  bed  carbon resorb systems, which
      are  operated  at  bed  lengths  of 30 1n., the blower power
      consumption  is about 6 kilowatts per  1,000 ft*/min.  Addi-
      tional  but small  power requirements are needed to operate
      monitoring and control devices.  At a rate of $0.015/kwh,
      the  yearly cost  of  electrical power for 2080 hr of blower
      operation is  $710.   An additional $60/yr for the other power
      costs brings  the  total to  $770/yr.  For the rotating axial
      flow bed  system  with condensation and vapor recycle, electric
      power is  also required for refrigeration and regenerating
      gas  heating.  For the incinerators  which operate at lower
      airflow  resistances, the power cost is $500/yr.

           Fuel -  Fuel  in  the  form of  natural gas is required  in
      steam and noncondensable gas regenerations, and to preheat
      the  air-pollutant mixtures for the incineration processes.
      Natural  gas  costs vary for different  locations but a gen-
      erally accepted  average  cost figure is $0.60 per 1,000 ft*
      or $0.60  per 106  Btu. On  the basis of gas consumption,
      steam generation costs $0.60/1000 Ib, but due to heat losses,
      the  generally accepted cost  is $1.00/1000 Ib.  Noncondensable
      regenerating gas  heating costs are considerably less than for
      steam because no latent  heat is  involved.  To heat air to
      300F, the  cost  is  about $0.05/1000 Ib and to heat air to
      400F, $0.07/1000 Ib. The amount and type of regenerating
      agent required varies with the influent concentration and
      pollutant type,   see  Chapter  4, Regeneration of Activated
      Carbons for  further discussion.

           Fuel costs  to preheat the pollutant airstream for the
      incineration processes vary  considerably with the concen-
      tration and  to  a lesser  extent with type.  Preheat require-
      ments are discussed in Chapter 7, under Pollutant Concen-

           Cooling Water - Cooling water  is required to condense
      the  steam-pollutant vapor mixture.  The ratio of cooling
      water at ambient temperature to  steam is 30  to 1.0.  Water
      cost varies  with the locality but the generally accepted
      cost is $0.10/1000 gal.   Cooling water may also be used  for
      condensation of  vapors from  noncondensable regenerating  gas,

           Pollution  Control  Costs

           Comparative Costs of Competing  Systems  -  Relative
      pollution control costs  for  the  competing  systems  are
      graphically presented in Figure  6-31. It  is  recognized  that
      actual operating costs can vary  markedly depending  on  local
      utility rates,  operating time,  choice of  operating  conditions.
      and sophistication or completeness  of the  system.   The  cost


  *  0.80
     JTTTT]     I   '  i  | ""I
                                  Thermal incineration
             Catalytic Incinerati
+ Incinerati
                no solv. credit
                or disposal cost
              solvent  credit
      it           \i     '
    i  I i 1111	i   i  i \ i iii
                                                i   i
                           1000   3000
Figure 6-31 Comparative Costs of Pollution Control with
            Carbon Resorb, Catalytic Incineration and
            Thermal  Incineration, 3,800 ft*/m1n Capacity
            System,  Calculations Based on 4-Methyl-2-pen-
            tanone and Propanone, 2080 hr/yr Operating
            Time .

      factors  that were  used  are detailed 1n Tables 6-VII and -VIII.
      In  some  cases,  it  may be  possible to ratio the cost factors
      to  reflect  actual  local conditions.

           Minimum pollution  control costs result where credit can
      be  taken for the recovered solvent as indicated by the low-
      est curve in the figure (for 4-methyl-2-pentanone).  Although
      this solvent is slightly  soluble in water, it was assumed 1n
      making  the  solvent credit calculations that complete re-
      covery  had  been effected  (by decantation) even at operations
      at  the  10 ppm  influent  concentration level.  The curve illus*
      trates  the  results achievable where solvent recovery 1s
      feasible and the solvent  has recovery value.  In this case
      the cost breakeven point  is at Ci " 500 ppm; recovery oper-
      ations  at { greater  than 500 ppm would be profitable.  The
      breakeven point varies.   For some solvents it 1s at Ci over
      1,000 ppm,  but  a good average is at Ci * 700 ppm.

           Below  about 200  ppm  and assigning neither credit nor
      disposal cost  to the  recovered solvent, the lowest pollution
      control  cost is still indicated for the carbon resorb system.
      In  this  concentration range, a single adsorber unit operating
      on  an eight-hour working  cycle with regeneration at night
      was assumed.   Here the  cost level 1s just above $0,50.  The
      sharp Increase  in  operating cost, to over $0.80, shown
      between  200 and 500 ppm,  results from the switch to a double
      adsorber unit,  because  of the necessary increase in fre-
      quency  of regeneration.   The cycle time here is one hour.

           The next  lowest  operating cost, below 200 ppm, 1s
      shown for a combination of carbon resorb and thermal incin-
      eration.  In this  system  the resorb unit is used to concen-
      trate the pollutant to  25% of LEL before incineration.  Air
      heated  by the  combustion  unit exhaust is used for carbon
      regeneration.   The pollution control cost is slightly over
      $0.60 but begins to rise  rapidly at concentrations above
      200 ppm  because of the  increase in cycle frequency and need
      for an  Increasingly larger back-up incinerator.

           As  shown  by the  upper two curves, catalytic and thermal
      incineration provide  the  lowest disposal costs at pollutant
      concentrations  above  about 1000 ppm.

           Pollutant Molar  Volume - The effect of pollutant molar
      volume on the  air  cleaning cost was Investigated  for  the
      three vapors under study. The results are  summarized  1n
      Table  6-IX.    They show  no significant differences  although
      the Vm range brackets a considerable  portion of  the  pollutants
      of  interest.   These cost  figures  are  for steam regeneration
      with no credit for pollutant  recovery or disposal  cost,  1.e,.

Table 6-VII Cost Factors for A1r Pollution Control by Carbon Resorb Systems
Design basis
             3800 ftVmin at 77F
             4-methy1-2-pen tanone
             4 to 6 pelletized, 3000 Ib/adsorber
             8 hr/da, 2080 hr/yr
 Influent  concentration,
 Regenerating  agent
 Regenerating  agent  temp.,
 Number  of carbon adsorbers
 Installed equipment cost.
 Airflow to  Incln.,  ft3/
 Operating costs, $/yr
 Steam at  $1.00/1000 Ib
 Water at  $0.10/1000 gal
 Power at  $0.015/kwh
 Gas at  $0.60/1000 ft3
 Operating labor at  $10/hr
 Carbon  replace (5-yr life)
 Depreciation  (10-yr life)
 $/hr per  1000 ftj/m1n
Pollutant recovered. Ib
Pollutant value at 50.1
Operating cost - rec. v
 $/hr x 1000 ftVminl1)

ppm 10
,F 212
rs 1
, $ 21,500
V ~ w

r 300
e) 300
) 2,150
yr 1,090




_ _


+ incineration



Cost of separating pollutant from condensate not Included

Table 6-VIII Cost  Factors  for  Afr  Pollution  Control  by  Catalytic  and Thermal

Design basis

Airflow: 3800 fts/m1n at 77F
Pollutant: 4-methyl-2-pentanone
Operation: 8 hr/da, 2080 hr/yr
Influeht cone., ppm
Installed equipment
cost, $
Operating cost, $/yr
Power at $0.015/kwh
Gas at $0.60/1000 ft3
Operating labor at
Catalyst replacement
Maintenance, 2%
Depredation, (10-yr
$/hr x 1000 ft'/mln
  (1) 50* heat recovery

the pollutant and  steam condensates  are  sewered.   When  re-
covery or disposal  (by nonpolluting  means)  is  carried out,
the costs become variable depending  on concentration, com-
bustibility and water solubility.

Table 6-IX    Effect of Pollutant  Molar  Volume and Con-
              centration on Air Cleaning Cost  When No""
              Pollutant Recovery or  Disposal  Cost is

                   2-pentanone  trichloroethene  nropanone

Vmi cm'liq/mol       140             94             74

	 Costs when Cj  3,000 ppm, $/hr per 1,000 ft'/min

Steam and cooling

	 Costs when Cj  1,000 ppm, $/hr per 1,000 ft'/mln 

Steam and cooling

	 Costs  when Ci    100  ppm,  $/hr per 1,000 ft'/min

Steam and  cooling
      Pol 1utant Concentration  -  The  effect of pollutant con-
 centration on air cleaning  cost is  shown 1n Figure  6-31  and
 Table  6-IX.    The general  pattern  is  for the air cleaning
 cost to rise  with increased concentration.  With reference
 to Table 6-VIII, the cost at C^   3,000 ppm 1s at the $0.90
 level while at Cj - 100 ppm it  is at the $0,50 level.  When
 disposal of pollutant of no recovery value Is added to the
 process,the costs rise but  the  cost pattern remains.  Only
 when a pollutant of value is recovered is the pattern re-
 versed, as shown by the lower curve 1n Figure 6-31.

           Flow  Rate - The effect of flow rate, or more correctly
      the  flow capacity, on air  cleaninq cost does not follow the
      usual  pattern observed,  i.e., decreased cost with increased
      size.   Cost  figures worked up for Table  6-X show a small
      Increase in  air cleaning cost when a larger unit Is used.

      Table  6-X    Cost Factors for Air Pollution Control by
                  cTrbon-Resorb  Systems at Two Flow CapaciTies

      Design basis

           Pollutant:     4-methyl2-pentanone
           Concentration: 1000 ppm
           Carbon:        4 to 6 pelletized, vapor phase
           Operation:     8 hr/da, 2080 hr/yr

        Flow rate                    800 ft'/min     3,800 ft'/mln
        Installed  equipment cost       $7,000           $35,800
        Pounds carbon/adsorber            350             3,000

      Operating  costs. $/vr

        Steam at $1.00/1000 Ib
        Water at $0.10/1000 gal
        Power at $0.015 kwh
        Operating  labor at $10/hr
        Carbon replacement (5-yr life)
        Maintenance, 1.0% of invest.
        Depreciation, (10-yr life)

           $/hr  per 1,000 ft3/min           0.83              0.87

      Two  operating costs, the operating labor and maintenance, are
      subject to adjustment which could reverse the air cleaning
      cost  figures.  The general impression is that at least in the
      flow  capacity range from 800 to 10,000 ftj/m1n the air clean*
      ing  costs  do not vary greatly with size change.

           Adsorption-Regeneration Cycle Time - Every air pollution
      operation  can be optimized with respect to the adsorption*
      regeneration cycle time  length.  For a given control situation
      the  cycle  time is dependent on the operating capacity select**
      Operating  capacity is the  difference between the saturation    *
      capacity,  us, at the particular Influent concentration, Cf,
      and  the mean amount of adsorbed pollutant, urt remaining on
      the  carbon after regeneration.

     The primary cost factor that varies with the cycle time
is the steam consumption, hence a study of steam consumption
as function of cycle time or operating capacity leads directly
to the cycle time effect on air cleaning cost.  Results of
this type calculation are presented graphically in Figure
6-32.  For these calculations, the carbon resorb system con-
sisted of two beds with 3,800 fts/min airflow capacity for
each bed.  4-Methyl-2-pentanone at 500 ppm concentration was
used as the representative pollutant.  As shown in the figure,
the steam consumption for heating decreases as the cycle time
increases.  This reflects the decrease in number of times the
carbon and associated hardware have to be heated.  The steam
consumption for sweeping the pollutant vapors out of the bed
increases rapidly with increase in cycle time.  This reflects
the decrease in vapor pressure as the pollutant is progress-
ively removed from the carbon.  (See Chapter 4, Regeneration
of Activated Carbon for further discussion.)  For this par-
ticular system, the total steam consumption passes through a
minimum at the 2-hr cycle time,

     Table6-XI gives the operating capacity, ws-ur, the number
of cycles per year and the air cleaning cost ($/hr per 1,000
ft /min) for the above discussed pollution control conditions
when no credit for recovered vapor or vapor disposal cost is
taken into consideration.  The minimum cost is $0.83 at the

Table 6-xl Effect of Cycle Time on Cost of Air Cleaning for
           3.800 ftVmin Carbon-Resorb System. 500 ppm
           Pollutant Concentration with Steam Regeneration

     Cycle time               1 hr    2 hr    5 hr
     us-wr, Ib/lb             0.012   0,024   0.061
     Cycles/yr                2080    1040    416
     $/hr per 1000 fts/min    0.85    0.83    0.89
2-hr cycle time or 0.024 Ib/lb operating capacity.  Over the
cycle time range investigated, the cost differences are not
large, but the cost difference would become very large at
cycle times over 5 hr.  At concentrations larger than 500 nnm.
the minimum cost shifts to cycle times approaching the 1-hr
level and at smaller concentrations the minimum cost shifts
to cycle times longer than the 2 hr.

     Rotating Bed with Condensation and Recycle . To determine
whether some of the proposed innovations to the more con-
ventional single or double stationary bed solvent recover*
systems can be more economical for air pollution control  cost

                                    Sweep steam
                                  (75* efficiency)
                                2       3

                             Cycle time, hr
      Figure 6-32  Steam Requirement for Regeneration vs
                   Pollutant Loading and Cycle Time, Carbon*
                   Resorb System on 4-Methyl-2-pentanone,
                   3,800 ft'/mln Flow, 2080 hr/yr Operating

estimates were worked  up  for  the  continuous  vapor  recycle
system.   This  system utilizes hot air  regeneration,  recovery
of desorbed vapors  by  condensation and recycle  of  unadsorbed
vapors to the  carbon bed.   Figures 6-28,  -29  and -30 describe
the system.  The cost  factors and air  cleaning  costs are
given in Tables 6-XII and-XIII for a 3,800 ft'/min  system
operation at several concentrations of 4-methyl-2-pentanone
or propanone.

      The results show a  lower air cleaning  cost for 4-methyl-
2-pentanone (the higher boiling liquid)  by the  continuous-
recycle system than by the stationary  bed systems.   At 100  and
500 ppm concentrations, the costs are  $0.46  and $0.54/hr  per
1,000 ft'/min, respectively.   For propanone, the air cleaning
costs are the same, i.e., $0.55/hr per 1,000 ft'/min.  In  each
case, the continuous-recycle  system produces a  separated
liquid pollutant (or solvent) which may then be reused or
disposed of by any means  deemed best.   The air  cleaning cost
by the stationary bed system  leaves the pollutant  dissolved
in the steam condensate.   When the stationary bed  system  is
used with noncondensable gas  regeneration plus  incineration
of desorbed vapors, the air cleaning cost rises to $0.63/hr
per 1,000 ft'/min.  The continuous recycle system  appears
more economical at  the lower pollutant concentrations, al-
though, admittedly  there are also more uncertainties in the
cost factors used in working up the estimates.

Adsorbents and Adsorbent Systems' Manufacturers and Vendors

     Activated Carbons

     Table  6-XIV  lists under  each manufacturer's name the
trade name, properties and selling price  range of gas-nhase
carbons  that  are suitable  for  air  pollution control  applica-
tions.   For most applications, the interest is In the  coarse
mesh  sizes, 4  to 6  or  4 to 10, although  in some toxic  chemi-
cal,  biological  and radiological  applications  the finer 12
to 30 mesh  is  also  used.   The  density range 0.45 to  0.50
or 28 to  31 lb/ft3  appears to  be  fairly  standard.   In  regard
to adsorptive  capacity, this  density  range  is  optimum  when
measurements  are made  on  the  unit carbon volume basis.  Some
carbons  are pelletized and others granular.  The pelletized
carbons  tend  to  be  slightly  harder and,  because of  the
narrower mesh  size  range,  usually offer  less resistance to
airflow  than  granular  carbons  of equivalent  mean  granular
diameter.   In  regard  to  adsorptive capacity, as measured  by
the  CCU activity,  most  carbons  fall  in  the range  50% to
6035;  see test procedure  in Appendix D.   The CC14  activity
essentially measures  the  micropore volume of pores  less  than
30  X in diameter;  see  Chapter 4, Activated  Carbon Pore Size
Distribution  for significance of CC14 activity in regard to

      Table 6-XII  Estimated  Operating  Cost.  Continuous Recycle^
                  Unit  Including  Solvent  Recovery
      Design  basis

                    3800  ftV"Hn  at  77F
                    Granular  BPL  V
                    8 hr/da,  2080 hr/yr
           Regeneration:   Continuous,  Internal  air  recycle

                          250  ftVmtn  at  250F  for  100  ppm
                          600  ftVmln  at  250*F  for  500  ppm
           Solvent re-
                    Condensation  via  refrigeration  at  0*F
     Installed equipment cost

Operating costs. $/yr

  Power at $0.015 kwh
  Operatlnq labor at $10/hr
  Carbon replacement (5-yr life)
  Maintenance, 2% of Invest.
  Depreciation, (10-yr life)
                                         100  ppm
 500 ppm
           $/hr per 1,000 ft

        Solvent recovered, 90% efficient,

        Value of solvent recovered
         (3 $0.135/15, $
1 .760

        Estimated Operating Cost, Continuous  Recycle  Unit v/lth  Concentration
        of Pollutant for Subsequent Recovery  or Disposal
 Design basis
                     3800 ft'/mln at 77F
                     4 to 10 mesh granular BPL  V  type
                     8 hr/da, 2080"hr/yr
      Regeneration:   Continuous,  Internal air recycle,  250  ft3/nnn
                     at 250F

 Concentration,  ppm

 Purge  vent,  ft'/mln
 Purge  concentration,  ppm
 Purge  vent  concentration,  %  LEL
 Installed equipment cost,  $

 Operating cost,  $/yr

 Power  at  $0.015/kwh
 Operating labor  at $10/hr
 Carbon replacement, (5-yr  life)
 Maintenance, 2%  of Investment
 Depreciation, (10-yr  life)


$/hr per  1000 ftj/m1n

Table 6-XIV   Activated  Carbon  Suitable for Air  Pollution Control Applications
Trade name

Norit RB I
Norit RB II
Norit RB III
Norit RB IV
Informal on on
(coconut shell)
U.S. s

ieve No.
Am A ri r
grades not
rttt. f r h
an Norit Compan
Barnebey Cheney
Surface CC14
area, activity,
yT n r _....,,
, 1 1 1 1> . ---ii
1150-1250 60
1050-1150 60-65

CA 90-100

 (coal base)        "     6x16

Table 6-XIV    (continued)
Trade name

Columbia NXC
Columbia 3LXC
Columbia SXV1C
Columbia ACC
Columbia JXC
Nuchar WV-H
Hltco 235
U.S. s




1eve No,
on Carbide Corp
Uae t-ua rn
W1tco Chemical






           Very  significant  information regarding a carbon would
      be  its  generalized  adsorption  isotherm or isotherms; the
      type  of isotherm  referred  to are shown in Figure 4-11 and
      -12,   Cost estimates made  on different types of carbons show
      that  a  carbon with  a steep  sloped isotherm is better suited
      for solvent recovery or  adsorption-regeneration when adsorp.
      tion  occurs at  high concentrations.  Carbons with a small
      slope are  better  suited  for poorly adsorbed pollutants or
      adsorption at very  low concentrations, i.e., adsorption at
      large A values.   In Figure  4-11, the BPL VI carbon would be
      more  effective  for  adsorption  at a very low concentration
      than  the S154.  On  carbon weight basis S154 would be more
      effective  than  BPL  V for solvent recovery.  However, because
      of  the  low density  of  S154  much of the advantage of S154
      cancels out when  comparisons are based on equivalent volumes*

           The addresses  of  the  carbon manufacturers are as

           American Norlt Company, Inc.
           Jacksonville,  Florida

           Cassidy at 8th
           Columbus,  Ohio  43219

           Pittsburgh Activated  Carbon
           Division of  Calgon  Corporation
           Calgon Center, Box  1346
           Pittsburgh,  Pa.   15230

           Union Carbide  Corporation
           Carbon Products Division
           270 Park Avenue
           New York,  N.Y.  10017

           West  Virginia  Pulp  and Paper
           Chemical Division
           299 Park Avenue
           New York,  New  York  10017

           Witco Chemical Company, Inc.
           Activated  Carbon  Department
           277 Park Avenue
           New York,  New  York  10017

           Two specialty  items that  are of Interest to air pollution
      control are, (1)  a  bonded  granular carbon panel manufactured
      by  Pittsburgh Activated  Carbon  and (2) a partlculate web

manufactured by C.H,  Dexter Corporation, One Elm Street,
Windsor Locks, Connecticut  06096.   The Pittsburgh procedure
for bondinq carbon qranules qives a porous  riqid structure
that can be shaped into other forms aside from the panels.
The bondinq procedure does not siqnificantly lower the carbon
adsorptive capacity.   The C.H. Dexter web consists of a thin
flexible sheet of orqanic fibers to which are bonded powdered
activated carbon.  Fifty percent of the web weight is carbon.

     Chemetron Corporation, Catalysts Division, P.O. Box 337,
Louisville, Kentucky  40201, prepare specially treated carbons
formulated for removal of sulfur compounds  from synthesis and
natural gas streams.   The impregnated carbon is identified as
G-32E catalyst.

     Noncarbon Adsorbents
     The following list presents the names of manufacturers
of silica gel, activated alumina, molecular sieves and im-
pregnated polar adsorbents.  See Chapter 4, Types of Adsorbents
for discussion of adsorptive properties.

     Aluminum Company of America
     1501 Alcoa Building
     Pittsburgh, Pa.  15319

     Product:  Granular activated alumina, grades F-l,
               F-5,  F-6 and H-181.

     Davison Chemical
     Division of W.R, Grace &  Co.
     101 North Charles St.
     Baltimore, Md.   21203

     Product:  Granular silica gel  and  molecular sieves,
               types 3A,  4A,  5A,  2-100  and  13X,

     Linde  Company
     Division of Union Carbide Corp.
     Pleasant Valley Road  & Rt. 38
     Moorestown, New Jersey

     Product:  Molecular  sieves,  types  3A,  4A, 5A,  10X
                and  13X.

     Marbon Division
      Borg-Warner  Corp.
      Washington,  West Virginia  26181

      Product:   Potassium  permanganate Impregnated
                alumina,  Purafll trade name.

           Sorption  Systems

           The  following  list  presents the names of firms engaged
      in  engineering  and/or  fabrications of various types of sorp.
      tion  systems used for  air purification, solvent recovery, or
      specifically for control of air pollution.

           Cassldy at 8th
           Columbus,  Ohio  43219

           System:   Air purification and solvent recovery.

           Cambridge  Filter  Corporation
           7645 Seventh North  Road
           Syracuse,  New  York  13201

           System:   Air purification, pollution control -
                    carbon pleated and flat panel types.

           Conner Engineering  Corporation
           Division  of Jenn-A1r
           Danbury,  Connecticut

           System:   Air purification  pleated, panel, and
                    canister type carbon beds.

           Day  & Zimmermann, Inc.
           Philadelphia,  Pa.

           System:   Carbon resorb with air or Inert gas
                    regeneration and Incineration of
                    desorbed vapor or readsorption in
                    secondary  adsorber.

           Detrex Chemical Industries, Inc.
           Box  501
           Detroit,  Michigan  48232

           System:   Solvent  recovery - carbon stationary-
                    beds  with  steam regeneration, hood
                    ducts, freeboard chillers for degreasers,

           Farr Company
           2301 Rosecrans Ave.
           El  Segundo, California  90245

           System:   A1r purification, pollution control.

Howard S. Caldwell  Company
Caldwell Air-Park
P.O. Box 27
Baltimore, Ohio  43105

System:  Air purification, canister and pleated
         panel types carbon beds.

Hoyt Manufacturing Corporation
251 Forge Road
Westport, Mass.  02790

System:  Solvent recovery, stationary carbon bed
         with steam regeneration.

Marbon  Division
Borg-Warner Corp.
Washington, West Virginia  26181

System:  A1r  purification, potassium permanganate
         Impregnated alumina panel type beds.

T.  Melshelmer Co.,  Inc.
22610  South Western Ave.
Torrance,  California  90501

System:   Pollution  control, multistage  beds
          displaced  at Interval  countercurrent  to
          airflow, bed reactivated,

Mine  Safety Appliances  Co,
201 North  Braddock  Ave.
Pittsburgh, Pa.   15208

System:  A1r  purification,  pollution  control-
          carbon  pleated and  flat panel types
          carbon  beds.

RSE Incorporated
34875-23 Mile Rd.
New Baltimore,  Michigan  48047

 System:  A1r  purification - carbon panel  filters,

 Steffey Metal Products, Inc.
Division of MacOonald Mfg. Co.
 23-Mile at Sass Road
 New Baltimore, Michigan  48047

 System:  A1r purification  carbon panel filters.

          Sutcllffe, Speakman and Co., Ltd,
          P. 0, Box 366
          Bronxville. New York  10708

          System:  Solvent recovery, rotary bed
                   continuous adsorber and dual
                   stationary bed.

          V1c Manufacturing Company
          1620 Central Avenue. N.E,
          Minneapolis, Minnesota  55413

          System:  Solvent recovery, dry cleaners with
                   integrated solvent recovery, stationary
                   bed with steam regeneration,

          Vulcan Cincinnati, Inc.
          1329 Arlington Street
          Cincinnati, Ohio  45225

          System:  Solvent recovery, stationary carbon
                   beds with steam regeneration.

           7.   Catalytic Incineration  System
Basic Functions  and limitations

     Oxidation and Reduction Functions

     Catalysts can perform two basic functions  useful  1n air-
pollution control.  The function already being  used to con-
siderable extent 1s the catalysis of oxidation  reactions, as
of carbon monoxide to carbon dioxide and hydrocarbons  to
carbon dioxide and water.   The other function,  which may
have limited application to small sources control, 1s  the
reduction of nitric oxides to free nitrogen and oxygen.

     Since the Interest in air pollution control 1s the com-
plete combustion of hydrocarbons, catalytic oxidation  systems
are regarded as incinerators, to emphasize a comparison to
the competing thermal Incinerators.  A flow diagram of the
basic catalytic incinerator is shown In Figure 7-1.,  The
pollutant collection and blower system may be an existing
ventilating system, to which the incinerator 1s attached,
or Installed to accommodate the incinerator.  Preheat is
required to raise the pollutant-air stream to the combustion-
Initiation temperature, which can vary from ambient tempera-
ture for oxidation of hydrogen to, 750eC for oxidation of
methane on a platinum alloy catalyst.  Most pollutants of
concern have oxidation-initiation temperatures  in the range
400 to 600F.
 Figure  7-1   Basic  Catalytic  Incinerator Components

          The types  of catalysts  best  suited  for  catalytic  com-
     bustion and reduction  of pollutants  are  discussed  1n Chap-
     ter 5.

          With correctly  sized catalytic  incinerators   90%  to
     98% of  the pollutants  are burned  to  the  ultimate combustion

          To Improve the  operating  economics, more complex  In-
     cinerators incorporate.heat  exchangers  whereby the heat
     content of the  effluent stream 1s  used for preheating. When
     the combustible pollutant concentration of the Influent
     stream  1s high, sufficient heat 1s generated for the Incin-
     eration to become self sustaining.   At low concentrations
     additional fuel 1s required.                             '


          The advantages  (and disadvantages) of catalytic Incin-
     eration are usually  compared with  thermal Incineration.   In
     this  manual  of  package sorptlon devices  Us merits are also
     related to sorptlon  devices  and to a lesser extent to  ab-
     sorption devices  or  scrubbers.  When the pollutant 1s  a'sol-
     vent  with recovery value,  the  (nondestructive) $orpt1on-tyoe
     solvent recovery  systems  have  a large economic advantage;
     see Chapter 6 - Comparative  Costs  for details.  When the
     pollutant has no  recovery  value but  is combustible to  car-
     bon dioxide and water  vapor, the  catalytic and thermal In-
     cinerators perform the  same  function.  When incineration
     is  advantageous over control by sorptlon devices,  the  ad-
     vantages  of catalytic  Incineration are related to  thermal

          The  important advantage of catalytic incineration 1s
     the decreased reaction  temperature,  For efficient thermal
     Incineration, furnace  temperatures of 1200 to 1500F  are
     required  while  catalysts operate effectively in the 600
     to  900F  range.   From  this single  source, the economic  assets
     are distributed through  equipment, fuel and maintenance/
     Less  refractory containers and supports are required which
     in  turn,  reduce construction material and fabrication  costs
     The largest  economic factor  1s the savings 1n fuel.  Cost  '
     analyses  of  He1n  (1964)  favor  catalytic oxidation 2 to 6
     times over  the  thermal equivalent.  The analyses as sum-
     marized  in  Figure 6-31  Indicate that catalytic systems are
     favored  by  a factor  of approximately 1.5.

         An  Important advantage to catalytic Incineration, and
     also of  thermal  Incineration   relative to sorptlon methods,
     is  that  they can  control fumes from baking, roasting and
     varnish  cooking operations which yield, 1n addition to sol-
     vents, high molecular weight polymeric decomposition products

(considerably above 190  cm3/mol  Vm).   These  high  molecular-
weight pollutants  cannot be desorbed  from  the  carbon  on re-
generation, thereby shortening  the  useful  service life of
the carbon.

     Many processes emit the pollutant exhaust at elevated
temperatures, as for example, coffee  roasting.  This  factor
is advantageous for catalytic and thermal  incinerations
when combustible pollutant concentration is  low since it
reduces or may eliminate the fuel requirements and reduces
equipment costs, since heat exchangers are not required.


     The disadvantage of catalytic incineration relative to
the thermal equivalent is the deactivation of the catalysts
due to "poisoning".  Deactivation or "poisoning" may occur
through preferential adsorption of an undesirable or unre-
active compound, physical erosion of the active surface
sites, chemical reaction, or destruction of active surface
sites by excessively high operating  temperatures.

     Most of the heavy metals,  halogens, sulfur, phosphates
and silicates will  poison metal  catalysts.  The  poisoning
mechanisms vary among these  examples, but each  reduces  the
catalytic  activity  and  results  in higher operating costs.
Trace concentrations of  any  of  these poisons  will reduce
the catalyst life  significantly.  Operating life may  be
limited  to months  instead of years,  which might  be realized
with  a nonpoisoning  atmosphere.

      Particulates  present  in the pollutant  airstream must
be  removed by  filtration  for both the  catalytic  incinerators
and sorption systems  to  prevent possible poisoning of  the
catalysts  or clogging the  catalyst or  adsorbent  bed.   In
this  respect some  advantage  can be credited to  the thermal
incinerator, since it can  continue operations  and, in  some
cases, realize  complete  combustion of  the particulates.

      A short-coming of  both  the catalytic and thermal  in-
cinerations  is  the effects  produced  by incomplete  combus-
tion.  Oxidations  occur through a  series  of steps  producing
various  intermediate oxidation  products which may  be more
undesirable  than  the original  pollutant.  An  example of the
possible range of Intermediate  products can be illustrated
with  the oxidation of butane,  as follows:

      Butane  +  0?            Butanol,  slight  odor
      Butanol  + 02	- Butanal, malodor
      Butanal  + Qg* Butyric acid,  nauseating
      Butyric acid + Og      * 4C02 +  4H20, odorless.

     Temperature  control  must  be maintained and the Incinerator
     sized  to  minimize  or avoid the  production of Intermediates.

         Pollutant  airstreams containing considerable amounts
     of n1trogenated, sulfurated, and/or halogenated hydrocarbons
     can produce  unacceptable  large  amounts of NOX. SOx and hy-
     drogen halides  on  Incineration.  Unless provisions are made
     to control  the  effluent acids,  the alternative would be
     the sorption methods.   In a relatively dry effluent free
     of SOX and  hydrogen  hallde the  NOX can be catalytlcally

         Some precautions  must also be taken In the selection
     of oxidation parameters to prevent explosion or flashback
     in gas streams  having  high pollutant concentrations,  Com-
     plete  combustion 1s  more  probable 1n these systems, and It
     1s also possible to  extract some heat from the excess heat
     of combustion.  However,  when  temperatures are high and gas
     velocities  low, conditions may  be established for auto-
     Ignition, flashback  or explosion.  As recommended for the
     adsorption  systems,  the upper  concentration limit 1s 25X of
     LEL for general operations.  In some specific cases where
     the concentration  1s stable (I.e., does not fluctuate) or
     where  protective monitoring Instruments and controls are
     used,  the allowable  concentration can be Increased to 40X
     of LEL.

         Large  variations  1n  pollutant concentration produce
     operational  problems which, however, can be corrected, at
     a  cost, with Instrumentation and other auxiliary equipment.
     Increases 1n concentration above the concentration for
     which  the catalytic  Incinerator was designed may cause the
     catalyst  temperature to rise to the point where the catalyst
     becomes deactivated.  Conversely, a decrease 1n concentra-
     tion can  lower  the combustion  efficiency below the design
     level  unless additional fuel 1s added.

     Potential Applications

         A number of applications  of catalytic Incineration 1n
     air pollution control  are noted 1n Chapter 1, - the emission
     sources being the  spray booths, surface coating baking ovens,
     drying ovens used  in metal coating and lithographic opera-
     tions, coffee roasting, meat smokehouses and meat rendering.
     Incineration systems based on  noble metal coatings, which
     will correct air pollution from a number of Industrial pro-
     cesses, are  being  advertized (Universal 011 Products),  The
     list contains forty-one specific processes, most of them
     fitting Into the general  Industries listed above or In the
     list of Other Emission Sources  1n Chapter 1.

     Trace contaminant removal  is  also  a  significant factor
1n the preparation of controlled  environments.   Air and
space vehicle compartments,  medical  and biological  test
stations and many welding and cleanina  operations  require
use of gases cleaned of trace impurities.   Catalysis is
used in many of these applications to  provide  fast, effi-
cient purification,

     The choice of control system depends  on  the balance of
advantages and disadvantages of the  systems for the pollu-
tant emission under consideration,

System Design Parameters

     As in the application of adsorption systems for pollu-
tion control (Chapter 6), the engineer will be required to
adapt the catalytic Incinerators  in most cases to processes
already in operation.  These processes have been developed
and also optimized for better production efficiency without
much consideration given to air pollution.  As air pollution
control regulations have been coming into effect one approach
has been to increase the exhaust volume to decrease pollutant
concentration and another has been to make minor changes in
solvent mixtures by leaving out components that are specially
objectionable,  The result 1s that the conditions are not
optimum for pollution control.   When a new installation is
being considered, the opportunity exists for the machinery
manufacturer and process engineer to optimize the overall
operations to include pollution control and thereby attain
its most economical usage,

     The first source of data for design and operating spec-
ifications is the properties of the pollutant exhaust gas
from the process,  The Important properties to be considered

     Flow rate
     Type of pollutant
     Pollutant concentration
     Presence, type and concentration  of poisons
     Presence, type and concentration  of partiqulates.

     An important decision 1s the efficiency that should be
sought in the incineration process.   Complete combustion of
the pollutant is not practical; however, 99% efficiency is
attainable under favorable conditions,   Efficiencies durino
continued operation gradually decrease, hence hiqh efficien-
cies are usually attainable only  with  a fresh catalyst

          Important factors  concerning  the  catalyst and  the  In-
     cinerator operation  to  meet  the  desired  or  required  goals of
     efficiency are as  follows:

          Chemical  make-up of  catalyst
          Catalyst  surface area
          Catalyst  support
          Operating temperature
          Reaction  kinetics
          Resistance of catalyst  bed  to  gas flow
          Residence time  of  gas in  catalyst bed.

     In this  area  the mathematical  expressions,  3, 4,  5,  6 and
     10 presented  in Chapter 5 are  used  for the  Incinerator  de-
     sign purposes.  Specific  installations may  provide  additional
     data or  operating  experience useful  to finalizing the design.

          The effects of  the operating variables on the  incinera-
     tor operation  are  discussed  further  in the  following text,

          Type  of Catalyst

          Catalyst  properties  and classifications are  reviewed
     in Chapter 5.   In  general, the noble metals are the  most
     efficient  for  hydrocarbon combustion.  The oxides of tran-
     sition metals  follow in the  order Co>Cr>Mn>Cu>N1>V>Fe>T1>
          Transition metal oxide  catalysts are more easily poi-
     soned than the noble metals, but their lower cost and greater
     selectivity could  be advantageous.

          Catalysts of  tested  performance and reliability are
     available  from various  manufacturers, but their compositions
     and  preparations are either  patented or  proprietary  Infor-
     mation.   Air pollution  control systems usually require  the
     treatment  of heterogenous  mixtures  (gas mixtures,  catalyst
     poisons  and particulates) which make catalyst selection
     difficult.  The selection can  best be made by the catalyst
     manufacturer after all  the other operating parameters have
     been determined for  the particular application.

          Type  of Pollutant

          The chemical  nature  of  the gaseous pollutant determines
     not  only the catalyst selection but  also the ease with which
     it can be  oxidized.  According to Stein  (1960), the  ease of
     oxidation  followsthe patterns  given  below:

          Branched  hydrocarbons are more  difficult to  oxidize
          than  normal hydrocarbons.

     Ease of oxidation  increases  with  molecular  weight
     in a homologous  series.

     Unsaturated aliphatic  hydrocarbons  are  more easily
     oxidized than the  corresponding  paraffins,  and
     susceptibility to  oxidation  increases with  in-
     creasing unsaturation.

     Open-chain compounds  are  more  easily oxidized  than
     cyclic ones having the same  number  of carbon atoms.

     Aromatic compounds, regardless of substitution,  appear
     to be easily oxidized  over most  catalysts,

     Oxygenated aliphatics  and aromatics are equal  to
     or more readily  oxidized  than  the parent compounds.

     Organic nitrogen compounds,  such as pyridine,  mor-
     phaline and monoethanolamine,  produce essentially
     theoretical carbon dioxide and water with small
     amounts of nitrous oxide.

     The minimum oxidation-Initiation temperature or, more
important, the temperature  at  which 90% conversion  occurs
measures the ease of  oxidation of a vapor in air.  Table  7-1
records these temperatures  and also autoignition temperatures
for some of the known air  pollutants  emitted by small station-
ary sources.  These temperatures  are approximate and apply to
a noble metal catalyst at  flows maintained  to give  comparable
residence times.

     For a pollutant airstream at concentrations high enough
for the catalytic oxidation to be self sustaining,  a portion
of the catalyst bed must be heated to the oxidation initia-
tion temperature, as  by preheating the airstream.  The com-
bustion of the pollutant then supplies the  heat to raise  the
combustion process to more efficient temperature levels  and
preheat may no longer be required.   At low  concentrations
essentially all the heat must be supplied by added  fuel  to
maintain the catalyst at temperatures where  catalytic effi-
ciency is maintained at 90% or better.  Ninety percent and
higher efficiencies would  be maintained at  800F for most
air pollutants of concern.   Some of the acetates and ketones
would require temperatures at the 900F level,

     In the incineration of flammable vapors operational
safety is an Important consideration.  When  the  autoignitlon
temperature is lower than  catalytic operating temperature
as for some of the normal  hydrocarbons and ethyl acetate
(in Table 7-1) a flash-back or explosion would  occur  if  th<
concentration were allowed to rise to the lower  explosive

     Table 7-1   Minimum Temperature  for  Initiating Catalytic
                Oxidation  and   Autolgnltlpn  of Organic  Pollutant
                Vapors  In  Air,  Degrees Fahrenheit     ""

                           Catalytic                     90%
                           oxidation       Auto-      catalytic
     Pollutant  vapor        TTTT7T     ignition     conversion

     Methane               760    800         999          933
     Ethane                680               959
     Propane               650    660         870          1065
     Butane                570    570         806          750
     Pentane               590               495

     Hexane                630               433          725
     Heptane               490               433          520
     Octane                490               428
     Decane                500               406
     Dodecane               540               399

     Hydrogen                      32       1076          250
     Carbon monoxide              300       1204          500
     Ethyne                      280         365
     Benzene               440    500       1076          575
     Toluene               460               896          575

     2-Butanone             550               960          660
     4-Methyl-2-pentanone   560                           860
     Ethanol                540               793          575
     Ethyl  acetate          730               800          880(80t^
     o-Xylene               470    570         924          575
1)  Suter (1955)
2)  Romeo (1962)
     limit  (LEL).  Table C-II 1n Appendix C  records the auto*
     ignition temperatures, lower explosive limits and also the
     heats  of combustion  and molecular weights of some of the
     identified small source  pollutants.  Dangerous concentra-
     tion levels for propane, hexane, heptane and ethyl acetate
     are 21,000, 11,000, 12,000 and 22,000 ppm respectively.

     The presence  of COg  and  HgO  in  the  pollutant  stream
have an inhibiting effect on  hydrocarbon  oxidation, while
CO promotes their  oxidation  as  was  demonstrated  by Klimisch
(1968) with propene and hexane  on alumina supported copper
oxide catalysts.

     Chemical  composition of pollutants  also  affects the
nature of combustion products.   This is  important  in cata-
lytic systems  where partial  oxidation is  always  a  possibility
The desired products would have no toxic or odiferous  nature.

     Pollutant Concentration

     The effects produced by the pollutant concentration
are closely involved with the heat content of the  air and
heat of combustion of the pollutant.  These three  factors
determine the amount of added fuel and/or recycled heat re-
quired  to maintain the optimum operating temperature level.
The main heating load is the heat content of the air, after
steady  state has been attained.  Lesser  heat losses occur
through the incinerator shell.   To  raise the air  temperature
from  60F base  temperature to 600  to 900F  (the  effective
catalyst operating temperature range) requires  10.5 to 16.5
Btu/(std) ft1.  (For thermal incineration operated  in the
1200  to 1500PF range, the required  heat  input is  22.5 to
28.5  Btu/(std)  ft3, or approximately  twice as much.)

      Table  7-II presents estimated  concentrations  at which
some  of the more  frequently  encountered  air  pollutants would
be  self sustaining  at  the 90% efficiency level.   These esti-
mated concentrations do  not  cover heat  losses through the
incinerator shell  and  also apply to  incinerators  without
heat  recycle.   In  each case  the  concentration 1s  near the
25% LEL, hence  self sustaining  catalytic combustions  are
feasible at safe  concentration  levels.

      When  the  pollutant  concentrations  are low, additional
heat  input  is  required  to maintain  high  conversion efficien-
cies. . Table  7-III  presents  the  calculated amounts  of
heat  input from pollutant combustion at varied  concentra-
tions, assuming 100%  conversion  to  C02  and MpO.  The  dif-
ference between the total heating  load,  in the  10.5 to
16.5  Btu/ft3  of air range,  and the  Btu/ft3 given  1n Table
7-HI at each  concentration  is  the  amount of additional
heat required.

      At concentration  levels up to 100  ppm,  95% or more
 of the required heat input  comes from the preheat source.
 Fluctuations  in concentrations, which normally occur 1n
 many industrial processes,  would have very little effect
 on the steady operation of  the  incinerator.   Fluctuations

Table 7-II
Pollutant Concentration
Combustion at 90%
t of
Pollutant vapor Btu/ft3
4-fie thy 1 -2-pentanone
Ethyl acetate
Natural gas(methane)
Table 7-III Heat Input from
Cone. ,
Pollutant vapor
4-Me thy 1 -2-pentanone
Ethyl acetate
Natural gas (methane)

--. ..


temp. F

900 (80%)

800 (es

Heat 1

ed Concentrati

25% of


in pollutant concentrations at levels higher than 1000 ppm
can cause catalyst overheating or underheating unless com-
pensating adjustments are made to the preheat input rate
At the low concentration, the catalytic incinerator opera-
tion is the most stable but the more costly in fuel consump-

     Operating Temperature

     The operating temperature - i.e., catalyst bed tempera-
ture - influences both the efficiency and life of the cata-
lyst,  If this temperature is below the initiation tempera-
ture, no oxidation occurs; but above the initiation temp-
erature, the oxidation efficiency increases as the operatinq
temperature increases.

     The two major catalyst types - noble metal and metal
oxide catalysts - exhibit wide differences in this operating
parameter,  Transition metal oxide catalysts usually operate
in the temperature range 200 to 600 Farenheit, while noble
metal catalysts operate in the temperature range 400 to

     The optimum operating temperatures reflect these dif-
ferences (300 to 400F for metal oxides and 750 to 1050F
for noble metals).  Deterioration and deactivation of the
catalysts also follow this pattern.  Operating transition
metal catalysts above 600F limits their effective catalytic
life to the point where replacement costs may become a
limiting factor.  Similarly, noble metal catalyst lives are
limited when operating above 1200F.

     Figure 7-2 illustrates the general pattern of tempera-
ture effect on conversion efficiency.  An important charac-
teristic of the temperature effect is that the efficiency
approaches the 1002: level gradually; a large temperature
![1sf! of$?n 1nd1cated to increase the efficiency from 90%
Jf Miller (lien6 presented 1n F19ure 7-2 is based on data

     The operating temperature also has an influence on the
products formed.  Partial oxidation occurs to a greater
degree at low temperatures, producing undesirable combustion
ElSiJfJl'n*S!Jh Jr??UJjS would comPletely nullify the desired
benefits of the pollution control device since, in most cases.
these products are more photochemically reactive and more
odiferous than the original pollutants.

     In the operation of the catalytic incinerator, the
     *o nf rno nnlin4>_. -. j  _ j	 .     .   w i  wuc

        I   I       I       I    I    I       I    
              Residence time (arbitrary units)

Figure 7-3  Effect of Residence Time on Catalytic
            Combustion Conversion Efficiency
temperature and residence time become the two Important
variables to be optimized for minimum operating cost of the

     As discussed under Operating Temperature*  the conversion
efficiency increases at a diminishing rate with temperature
rise above the 90% level.  In a similar manner,the conver-
sion efficiency also increases at a diminishing rate with
residence time.  At lower operating temperatures -those
safest from temperature deactivation - the conversion effi-
ciency can be maintained by increasing residence time (i.e..
Increased bed size for a fixed flow rate). Figure 7-3 shows
this effect in arbitrary residence time units.   The limit-
ing factor 1s the cost of the catalyst.  As temperature 1s
increased the residence time can be decreased.   In this
case, the limiting factor is the deactivation of the catalysts
at the higher temperatures.

          The  curve  1n  Figure  7-3  follows  the equation;

                        %  conversion    100  - fn             (7-1)

     where f 1s  the  percentage  of  Influent concentration pene-
     trating a catalyst layer  and  n  1s the number of layers.
     The  residence time 1s  proportional to the number of layers

          When the catalyzed reaction  1s first order with respect
     to  the pollutant species,  Equation 5-5 reduces to,

                        C/C0   e'Kt                         (7-a)

     where C/C0  1s equivalent  to f 1n Equation 7-1,  Then the
     efficiency  of a catalytic  Incineration can be represented
     by  the equation:

                  fraction converted  1 - [e'Kt]n,        (7.3)

          The reaction  rate constant, K, and residence time, t,
     for  layers  or Individual cells  of catalysts, are generally
     known by the manufacturer.  Equation 7-3 shows that the
     conversion  efficiency Increases exponentially with K and t,
     Similar theoretical  correlations can be made using Equations
     5-6  and 5-10.

          When Equation 5-10 1s used, a catalyst bed length-to-
     diameter ratio 1s  also shown to Influence the efficiency.
     As seen 1n  Equations 7-2 and -3 above the conversion
     efficiency  Increases with  the residence time per Individual
     cell and the number  of cells.   Efficiency Increase, by en-
     largement of the bed, can  be done by adding the cells 1n
     series to Increase bed length or 1n parrallel to Increase
     bed  diameter, or a combination of the two.   When length
     alone Is Increased,  the flow velocity remains constant,
     while an Increase  in diameter produces a decrease 1n flow
     velocity.   K decreases slightly with decreased flow
     velocity, but because 1t 1s 1n  the exponent, this small
     change 1s magnified  1n Its effect on efficiency.   Equation
     7-3  then strictly  applies when  the bed 1s enlarged by
     adding cells 1n series.  The limiting factor 1n this case
     1s the resistance  to airflow.   In practice  the pressure
     drop 1s usually held below 12 1n, of HgO.  If the bed volume
     Is enlarged by Increasing  the diameter to avoid excessive
     pressure drop, the limiting factor 1s the reduction 1n K,

Operating Costs

     The economic analysis of the catalytic  incinerator
operations Include the following costs:

             Fixed capital
             Operating labor
             Electrical power

To establish the operating costs on a basis  that can be
compared to those of the thermal incinerators and carbon
resorb systems, they are worked up as $/hr per 1,000 ft3/m1n
of polluted air that are cleaned,


     Installed costs vary between $3,500 and $6,000 per
1,000 ftVmin (stp) air handling capacity,  The higher end
of the range includes costs for heat exchangers, and filters
for participate removal, and also reflects the costs of more
expensive catalysts.  As the size is increased, the cost per
unit air handling capacity decreases to some extent but, for
the range 2,000 to 10,000 ftVmin flow capacities, the cost
per 1,000 ftVmin was assumed to be essentially constant for
the comparative operating cost estimates reported in Chapter
6.  The higher $6,000 figure was used at pollutant concen-
trations below the self sustaining concentrations (below
about 3,000 ppm), since heat exchangers give a decided oper-
ating cost reduction,


     A 10-year depreciation rate is often used by analysts
making cost comparisons with thermal incineration.  Some
variation 1n deterioration rate will occur depending on
whether the Incinerator 1s 1n Intermittent or continuous
service, i.e., 8 or 24 hr/da.

     Operating Labor
     M^HMHM*^W^^V>^^^^^^^^^^^^^^           .   fc

     Operating labor costs for catalytic Incinerators are
relatively small after the incinerator has been put into
operation; this Is also true of thermal incinerators and
carbon resorb systems,  A low level of operating cost has
been estimated on the basis of 5 min of labor per 8-hr
shift for each 1,000 ft3/m1n capacity (Hein  1964).


         Yearly maintenance  costs have been estimated by some
    analysts on the basis of  10% of the Initial Incinerator cost
    (He1n  1964),  Other methods can consist of periodic catalyst*
    replacement cost combined with a lower percentage rate of
    the  Initial Incinerator  cost,

         Electrical Power

         The main electrical  power consumer 1s the air circulating
    fan  or blower.  The kilowatts to move 1,000 fts/m1n of air
    through an Incinerator or carbon resorb bed can vary from
    2.5  to 4 1n the control  systems of 2.000 to 10,000 ft'/mln
    airflow capacities.  Local electrical costs per kilowatt
    hour will determine the  power cost.


         Fuel costs, using natural gas for preheat purposes, are
    approximately $0.60 per  10* Btu.  At pollutant concentrations
    below  the self sustaining level, preheat 1s required at varied
    amounts depending on the  type of pollutant and pollutant con-
    centration,  The amount of preheat required can be estimated
    from the calculated data  given In Tables 7.II and 7*111 for
    several frequently encountered pollutants.  The self sus-
    taining concentrations of Table 7-II were calculated for a
    60F base temperature for the- pollutant airstream,  At higher
    pollutant airstream temperatures, 1f heat exchangers are
    utilized, the preheat requirements can be considerably !$$.

    Auxiliary Equipment

         Before the catalytic Incinerator proper can be put Into
    operation, a number of auxiliary pieces of equipment are
    required for Its efficient and safe operation.  These are
    listed 1n Table 7-IV.
    Table 7-IV  Auxiliary Equipment for Catalytic Incinerators

              Ducts and supports          Preheaters
              Blowers                     Heat exchangers
              Filters        "             Monitors
              Scrubbers                   Controls

In the previous text the blowers,  filters,  heaters  and  heat
exchangers, of necessity, have been  brought Into  the discus-
sion, but more can be said of their  functions.

     Ducts and Supports

     Ducts to direct the pollutant from the emission source
to the Incinerator are an economic necessity.   Incineration
of the total ventilation alrstream 1s invariably  too costly
to be considered.  By ducting, the quantity of  air  to be
treated is kept smaller and also the concentration  of pol-
lutants 1s maintained higher, thereby reducing  the  required
fuel costs.

     In existing operations, the retrofitting of  the ducts,
the catalytic unit and Its supports  can be a problem.  In-
dustrial processes are often designed to conserve space, with
little room remaining for later Installation of pollution
control systems,

     Filters and Scrubbers

     Filters are usually necessary to protect the catalyst
bed from partlculates in the alrstream because  particulates
may:  (1) poison the catalyst,  (2)  cause deactivation  by
erosion or coating or  (3) in the case of a granular catalyst
bed cause increase in flow resistance.  Removal of  particu-
lates from the alrstream 1s also done by scrubbing  with
water.  This procedure is feasible at ambient temperatures,
at the 77F level, where the water vapor content  will then
be established at about 4%.  Efficient demisters  would  also
be required to keep the water content low, since  water  vapor
at higher concentrations interferes with the catalysis;
higher temperatures would be required.

     Secondary filtration, after the  catalyst,  may also be
required for some catalysts such as packed or fluid beds.


     Air movement produced by convection or ventilating fans
cannot effectively overcome  the resistance of the catalyst
bed and auxiliaries such as  heat exchangers or preheaters.
Pressure drops at the  12  in. of HzO  level  are encountered
with some  catalyst beds.  Blowers are required for  the  ex-
clusive operation of  the  incinerator.


     Preheaters,  auxiliary  fuel and  external heating may
all  be  used  Intermittently.   Unless  the  pollutant  alrstream

    temperature is already at the initiation temperature,  extra
    heating is required to initiate the combustion.   Subsequent
    heat or auxiliary fuel requirements depend  on  the  fuel  value
    of the pollutants.

         Preheaters may be of the heat exchanger  type  where the
    pollutant airstream does not come in direct contact  with
    the gas burner, or it can be a direct flame type,  The  latter
    raises the water content of the pollutant airstream, which
    with some catalysts may reduce their activity,

         Heat Exchangers

         The effluent gas stream from the Incinerator  1s usually
    at temperatures ranging from 600 to 900F,   By  use  of  a
    heat exchanger a part of this heat energy becomes  available
    for preheating the influent air to the  incinerator or  for
    other purposes such as steam generation,  space heating  or
    process heating.  When the  pollutant control  is  a  steady
    operation, the heat recovery for secondary  uses  may  make
    the pollutant control a profitable operation.  This  would
    be the case when the pollutant concentration  is  at the  self
    sustaining level or at lower concentration  when  the  Influent
    1s at elevated temperatures.

         The cost of the exchanger can be 25% to  40% of  the total
    cost of the incinerator,  but a heat exchanger operating at
    50% recovery used for preheat reduces the fuel cost  by  about
    50% at low pollutant concentrations.   The estimated  operat-
    ing cost, given 1n  Chapter  6, is $0.94/hr per 1,000  ft3/m1n
    with a heat exchanger, while without an  exchanger  it 1s $1 30/
    hr per 1.000 ft'/min,

         Figures 7-4,  7-5 and 7-6 show diagrams of incinerators
    in the various combinations  of preheaters and heat exchangers.

         Monitors  and  Controls

         Some type of  measurement is necessary  to determine
    the efficiency level  at which the catalyst  is working,  A
    number of alternative methods are available which  may be
    placed in seven general categories:

         Component Identification -  Although  it 1s usually  a
    temporary arrangement,  1t 1s  advisable  to run complete  anal*
    yses  of  the exhaust  stream  to assure  the  removal of  all  pol-
    lutant species.   This will  require the  use  of hydrocarbon,
    CO,  SOX  and NOX analyses, which  can  be  done by gas chromato-
    graphy.   Usually this 1s  done at the  beginning of  operation
    or  during a pilot  run to  assure  the  unit  1s performing  tp

_ -- . . Preheater
air 	 L
- ^ _ Wt

\ o r
Vliy Fuel
(natural gas)


^t _^<* ^X""^C


Pa 1*3 1 ue
ca to i ys

Figure 7-4  Basic Catalytic Incinerator With  Gas  Burner to
            Supply Preheat


Figure 7-5  Catalytic Incinerator With Heat Exchanger to
            Supply Preheat
(natural  gas)

                                               Secondary heat
 Figure  7-6   Catalytic  Incinerator With  Burner  to  Supply
             Preheat  and  Secondary Exchanger  to  Recover
             Heat  for Process  Use

         Total Hydrocarbon Analyzers  - These monitors may be
    used to supply continuous information concerning the con-
    centration of hydrocarbons  In either the process exhaust
    or catalytically treated gas streams,

         Dew Point Measurement  - Either hydrocarbon pollutant
    or water-product dew points may be used to follow the changes
    1n concentration through the catalytic burner.

         Temperature Measurement - Increased temperature with
    passage through the catalyst bed  1s an adequate test for
    determining 1f the catalyst Is working.

         Addition of a Tracer Compound - By addition of a hydro-
    carbon having a characteristic; odor, cursory measurements
    are possible using olefactory sensitivities to the influent
    and exhaust gas streams.

         Pressure Measurement -  Pressure drop through the cata-
    lyst bed can be used to measure the attrition of the catalyst

         CO and COg Measurement  - Instrumental  analyses are  avail
    able for the continuous measuring  of CO and CO? concentra-
    tions 1n flue gases.

         Each  of the  analytical  tools  mentioned previously  can
    be used to  determine  the  deterioration  of a catalytic burner
    and the time  for  maintenance.   Here again,  the degree of
    change  necessary  to  Indicate poor  operation and efficiency
    depends upon the physical  measurement  and  the  catalyst used.
    Additional  information  on  Instrumentation  1s  given  in Chan-
    ter 8.

         Equipment  suppliers  provide the  control  parameters  and
    the amount  of change  that  can  be tolerated  while  retaining
    the 90% or  higher efficiency as  required,

    Care  of Incinerator

         Some care must be  exercised with catalytic Incinerators
    especially  during start-up,  shut-down and storage,   The  cata-
    lytic elements, or beds, must  be  kept clean and  relatively  dry"
    in  order to maintain  their activity.  The degree  of  sensi-
    tlvlty  and  required care will  depend on the type  Of  catalyst
    Specific care and maintenance  procedures should be  provided
    by  the  manufacturer.  In general though, exposure to  hydro-
    carbon  solvents and hot, moisture  laden gases  should  be


     An initial  heat-up period  will  usually  precede  the
actual  operation of any catalytic  burner.  A flow of clean
air is  passed through the system until  it  is equilibrated
at the  desired operating temperature.

     This heat-up period provides  a  time  to  clean the system
and check all analytical instruments used  to control opera-
tion.  It also assures the maximum efficiency of the catalyst
from the initial hydrocarbon exposure  which  eliminates prob-
lems of partial  oxidation of pollutants and  possible deacti-
vation  by absorption or poisoning.


     Little or no maintenance is required  for normal opera-
tion of catalytic incinerators.  When  pollutant alrstreams
contain poisons  or large variations  in the pollutant con-
centration, both maintenance and operating problems  increase.
Catalytic activity must be rejuvenated at regular intervals
by washing with  water and/or heating while passing a non-
polluted airstream over the catalyst bed.


     Shutdowns and storage also require the  use of non-
polluted gas streams to cool the system and  maintain a
relatively dry atmosphere during storage.   The catalyst-bed
temperature is slowly reduced by decreasing  the pollutant
concentration and recirculating hot effluent so the bed  will
not experience a thermal shock.

Catalytic Equipment Manufacturers

     Several manufacturers offer  complete engineering tech-
nology  for the analysis of the pollution problem and for
the design and fabrication of the catalytic incinerator to
correct the problem.  Some manufacturers supply only cata-
lysts.   In either case  the equipment supplier requires an
analysis, as complete as possible, of the pollutant airstream
to make the correct selection of  the catalyst and to design
and fabricate the incinerator.  The following text  describes
some of the products sold by these manufacturers.

     Universal Oil Products of Darien, Connecticut, manu-
factures eight basic types of pollution control equipment
that can be modified to treat most process  streams  accord-
ing to the results of analyses.   The catalysts  used are
activated noble metals  deposited  on nickel  alloy  supports.

         Oxy-catalyst.  Incorporated of West Chester, Pennsyl-
    vanla, markets catalytic oxidation systems for Internal
    combustion engines  and  Industrial processes.  Their catalyst
    1s a platinized alumina  which resists thermal shock, poisons
    and clogging.  Five different sizes are available for use as
    odor eliminators or catalytic Incinerators.

         Matthey-Bishop.  Incorporated of Malvern, Pennsylvania,
    manufactures a new  platinized ceramic honeycomb, designated
    THT.   It 1s claimed to  offer the advantages of low pressure
    drop,  higher activity and attrition resistance 1n the com-
    bustion of many organic vapors.

         Mine Safety Appliances Corporation of Pittsburgh, Penn-
    sylvania, markets several sizes of catalytic oxidation units
    for the purification of aircraft cabins, liquid air separa-
    tion plants and removal of hydraulic fluid vapors.  The
    catalyst used is MSA  "Hopcalite", which is an activated mix-
    ture of CuO and Mn02.   Catalyst life 1s designed for greater
    than 500 hours, but depends upon the type of service and
    pollutants.  Economical replacement is facilitated by de-
    signing a replaceable  catalytic unit and using this rela-
    tively inexpensive catalyst.

         The Harshaw Chemical  Company of Cleveland, Ohio, manu-
    factures catalysts of  all  types for catalytic incinerators.
    Both metallic and oxide catalysts are made from V, Cu, Mo,
    Pt, Pd, Co, N1, Mn, Sn, and Pb.  They do not, however, de-  '
    sign or fabricate complete incinerator units.

         Chemetrpn Corporation, Catalysts Activity Division, of
    Louisville, Kentucky,  manufactures tiircner catalysis.  Their
    G-43 catalyst combines Pt and N1  on an alumina support and
    is used commercially for removing methane, ethane, ethylene,
    carbon dioxide and nitrogen oxides.

         Atlantic Research Corporation of Alexandria, Virginia,
    1s a fabricator of catalytic  incinerators  but does not manu-
    facture catalysts.

         E. I.  DuPontde Nemours  & Co..  Industrial  and Biochemi-
    cal s Department of Linden, New Jersey and"'1-ngTehard Mineral's"
    and Chemicals of Newark, New  Jersey,  manufacture catalysts"
    and also design and fabricate incinerators.

         FEC, of Cleveland, Ohio 44105, custom designed cataly-
    tic unTtT using catalysts supplied by duPont and Matthey-

                8.  Pollutant Detection
     This section on pollutant detection  presents  the  prac-
tical aspects of monitoring the input and effluent from  a
packaged sorbent device.  The intent is  to provide guidance
for the designer and users of these purification devices  in
the selection of appropriate monitoring  systems.   It  is
obviously impossible to deal with each specific  pollutant
source due to the many different types of pollutants,  pollu-
tant emissions rates and emission standards.   Therefore,  the
Information which is presented is of a generalized nature
although specific details are given on detection capabilities,
sensitivities, interferences, costs and  so on.   With  this
Information, the designer or user can evaluate his specific
emission problem and select an optimum monitoring  system
based on the information presented herein.

     This section on detection covers three primary areas
relating to monitoring of emissions:

     1.  Methods of sampling gases

     2,  Techniques and instruments suitable for
         monitoring emissions

     3.  Recommended methods for specific classes
         of impurities

     Basically, the method of sampling, the technique or
Instrument chosen and the reason for selecting a specific
approach to emission monitoring will depend upon many factors
including the type and quantity of emission, the size of the
emission control device,  the rationale for using the  device  .
and cost of monitoring equipment.  All of these  factors  are
Interrelated and will require technical judgment on the
selection of a pollutant monitoring system.  In  the case of
large corporations, in-house capabilities may exist and  can
be called upon to make the necessary decisions.   Conversely,
the small businessman may have to resort to an outside con-
sulting  service in order  to  select an appropriate monitoring

     Pollution monitoring is required first to determine
whether  an emission control  device is necessary and secondly,
if an emission control device is needed, to determine when
the emission control device  requires regeneration or replace-
ment*  The need to  install  an emission control device 1s

     dependent upon two primary criteria:

          1.  As an economic measure to recover and
              recycle solvents

          2.  As a required measure to meet federal,
              state or local emission standards.

     Therefore, an organization or business must identify  what
     emissions exist and the extent of these emissions.

          In addition to actual analytical  data,  other information
     can be used to determine emission rates.   Source  evaluations
     for commercial processes rely on process  flow sheets, mass
     balance data and specific process engineering experience  to
     identify possible sources.  This requires the examination
     and assimilation of data from several  disciplines and tech-
     nical personnel with appropriate backgrounds for  evaluation
     and correlation.  This type of approach to estimating emission
     rates has been described by Jacobs (1949) and Stern  (1968).
     Small businesses can estimate emission rates in  the  same
     fashion.  For example, the operator of a  dry cleaning estab-
     lishment keeps an inventory of the quantity of cleaning  sol-
     vent which is purchased over a period  of  time.  By equating
     the volume of solvent used with the time  of operation, an
     emission rate can be estimated.

          The character of a source may be  significantly  changed
     by certain mixtures of pollutants or  secondary reactions,
     which are not readily apparent in the  initial evaluation.
     Therefore, regardless of the accuracy  with which  the  basic
     variables can be estimated, analytical characterization  of
     the effluent will be required.

     Methods of Sampling Gases

          Sampling is generally divided into two basic categories:

          1.  Grab sampling

          2.  Continuous sampling

     Regardless of which method is used, the objective is  to
     acquire a sample which 1s representative  of the  emission  1n
     terms of both types and quantity of pollutant 1n  the  exhaust

     Grab samples are instantaneous  samples  which  are  gener-
ally returned to the laboratory for  chemical  or  instrumental
analysis, while continuous sampling  is  usually associated
with on-site instrumental  analyses.   Grab  samples  can  be
taken in evacuated glass*  metal or plastic containers,  with
the restriction that the pollutant neither reacts  with  nor
is absorbed by the container wall.  Grab  sampling  is  an
acceptable means of acquiring a sample  when  the  type  and
concentration of pollutant is relatively  constant.

     When the type of pollutant or pollutant concentration
varies with time then grab sampling  is  of  questionable
value.  In such cases, an  integrated sample  can  be taken
where a sample of the effluent gas is drawn  through a  suit-
able sample collector over a period  of  time.   This type of
sampling provides an average concentration of the  pollutant
over the scheduled time period.  Integrated sampling  has  an
inherent advantage in that trace contaminants are  concen-
trated and thus the problem of measurement of pollutant
concentration may be minimized.  The disadvantages include
the necessity of accessory equipment such as flowmeters,
pumps and so on.

     The time variable in grab or integrated sampling tech-
niques is an important part of the entire analytical  pro-
cedure.  Most commercial manufacturing is done on a cyclic
schedule or Includes "batch processes"  where the emissions
and rate' of release vary.  The schedule of the process must
be known and all samples Identified concerning both time
and location for accurate Interpretation.

     Continuous sampling is usually associated with direct-
reading, instrumental systems.  A sample  of  the effluent
gas is drawn through the instrument with  a pump, and a con-
tinuous  readout is  provided by the  instrument.  This type of
sampling system has  the advantages  of  continuous data dis-
play and minimization of direct labor  costs  in procurring
a  sample.  Many instruments have  a  sampling  system as an
inherent component  or a sampling  system can  be purchased
as an accessory.   For example, portable battery powered In-
struments  to measure LEL have  sampling pumps  as part of the
instrument  package.  Most industrial chromatographs are
purchased  with  continuous sampling  equipment.

     The  sampling  procedure which is ultimately selected
will  be  influenced  by a number of factors.   Samples will  be
taken  to  determine  whether  or  not an emission control  device
is required.   In  this case,  sampling and  subsequent  analysis

      on  a  contract  basis with  an  analytical  services laboratory
      1s  frequently  the  best  route  to  follow.   If the evaluation
      Is  to be  performed  in-house,  the type of  sampHnq system
      used  should  be simple and  inexpensive.  Grab samples will
      generally suffice  except  in  those  cases where the rate and
      quantity  of  pollutants  vary  to an  extreme degree over short
      periods of time.

           If an emission control  device is deemed necessary, then
      a judgment must be made on the type and complexity of the
      sampling  system.   This  will  depend upon many variables, as
      stated earlier, and the selection  of a  specific sampling
      system will  have to be  made  on the basis  of these factors.

      Techniques and Instruments for Monitoring


           A discussion  of the various  pollutants amenable to
      removal by either  sorption or catalytic combustion was pre-
      sented in Chapter  1.  These  pollutants  cover a multitude of
      chemical  groupings exhibiting various physical and chemical
      properties.  In most cases,  the  pollutants which have been
      categorized  for potential removal  by sorptive or catalytic
      devices are  hydrocarbons.  The specific type of hydrocarbons
      rangesfrom straight chain configurations  to those with
      various functional groups attached to the hydrocarbon chain.
      Selection of an appropriate monitoring  device or technique
      will  depend  upon the specific characteristics of a given
      pollutant.   For example, straight  chain hydrocarbons will
      likely dictate that a hydrogen flame ionization or thermal-
      conductivity detector be used, while chlorinated hydro-
      carbons may  be identified more selectively by the use of
      an  electron  capture detector.  Functional groups containing
      phosphorous  or sulfur can be measured at  low levels using a
      flame photometric method.

           Instruments and analytical  techniques are classified
      according to the specific physical or chemical character-
      istics which are measured.   In general, these can be clas-
      sified into the four major categories shown in Table 8-1;
      subdivisions  within these categories are shown  also.  This
      listing is not meant to be all inclusive, but does represent
      the most  likely means of detecting and  measuring pollutants
      in  gaseous effluents.

           The  analytical requirements  of various pollution sources
      have  encouraged the development  of a myriad of analytical
      methods of varying degrees of relevancy for any particular
      application.   Compilations such  as published by Jacobs (1949).

Table 8-1  Classification of Analytical  Methods
       I.  ThermometHc Instruments
              A.  Thermal conductivity
              B.  Catalytic combustion
      II.  Electrometrlc Instruments
              A.  Flame 1on1zat1on
              B.  Ion1zat1on
              C.  Electrical conductivity
      III.  Spectrometrlc Instruments
              A.  Visible spectrum
                   1.   Interferometrlc
                   2,   Spectographlc
              B.  Ultraviolet spectrum
                   1.   Ultraviolet  photometric
                   2.   Ultraviolet  Spectrometrlc
              C.   Infrared  spectrum
                   1.   Infrared  spectrophotometrlc
                   2.   Infrared
       IV.   Chemical  Methods
              A.   Gravimetric
              B.   T1tr1metr1c
              C.   Detector tubes

      Stern (1968)  and  Ruch (1970)  are a few of  the  exceptionally
      good references  available for selecting  an analytical method
      or instrument capable of measuring an  air  pollution  sample
      accurately.   The  selection of a method or  combination of
      methods  required  to obtain a  desired  accuracy  should be made
      on the basis  of  best estimations for  several parameters.
      These complex mixtures present individually unique problems
      compounded by the large variations in  concentrations and
      possible interferences among  the components.

           In  general,  all pollution detection methods  are judged
      on the basis  of  seven parameters (I.e.,  specificity, range,
      sensitivity,  reliability, accuracy, maintenance and  costs)!
      Selection of  a preferred method is made  by optimizing the
      performance for  each parameter and selecting the  best com-
      promise  which is  available for the individual  gas sample.

           Methods  of  Analysis

           Detailed reviews of analytical methods for specific
      compounds have been published  by Jacobs  (1949, 1960, 1967)
      U.S.  Public Health  Service (1965), Stern (1968) and  Ruch
      (1970).   These authors provide excellent descriptions of the
      accuracy, interferences  and actual procedures  to measure
      specific compounds.

           Table 8-II  lists the general  classes  of pollutants
      which are most likely to require control devices,  along with
      possible methods  of measuring  these pollutants, and  the de-
      tection  range.  Costs are not  included  in  this table since
      this  factor varies  among manufacturers  and according to the
      accessories acquired with the  monitoring device.  Direct
      contact  with  a manufacturer will  frequently provide  valuable
      information on the  instrumentation which is available for a
      specific application and the  cost  of a  complete instrument

           Thermal  Conductivity - Automatic  recording instruments
      are  commercially  available for the determination of  gases
      and  vapors based  on their thermal  conductivity.  The absolute
      values of the  thermal  conductivity are seldom  used for
      analytical purposes, but several  selected  values have been
      reported (Willard,  1958).  The thermal conductivity  of a
      gas mixture to be analyzed is  usually  compared with  that of
      some  reference gas, such as hydrogen,  nitrogen, or wet or
      dry  air.

           A fine wire, usually platinum or  some other metal or
      alloy with a  high temperature  coefficient  or resistance, 1$
      stretched along the axis of a  metal or glass cylinder.

Table 8-II  Common Analytical  Methods Used for A1r Pollutant Detection
Inorganic oxides
                 1-1000  ppm
                 1-100%  LEL
                 1-100 ppm
                 1-100*  lL
100-10,000  ppm
100-10,000 ppm

100-10,000 ppm

100-10,000 ppm
100-10,000 ppm   1-1000  ppm
                 1-100X  LEL
0-1 ppm
1-10,000 ppm
1-10 ppm
10-1000 ppm
0-10 ppm
10-1000 ppm
1-100 ppm
                               0-1  ppm
                               1-10,000 ppm
    0-10 ppm
    10-200 ppm
    0-10 pom
    10-200 ppm
                    0-10 ppm
                    10-200 ppm
                    0-10 ppm

Table 8-II  (Continued)

Oxygen  containing


 Inorganic  oxides


0-1000 ppm

1-1000 ppm
1-1000 ppm
0-1000 ppm
0-100 ppm
0-100 ppm
100-500 ppm

0-100 ppm
100-1000 ppm
0-10 ppm
10-1000 ppm

0-10 ppm
10-1000 ppm
                                                 0-10 ppm
                                                 10-1000 ppm

Table 8-II  (Continued)
                           Emission            Mass
                         spectrometer      spectrometer      Scantometer
Hydrocarbons                                 0.5-100%

Oxygen-conta1n1nq                            0.5-100%

Halogenated               0-500 ppm          0.5-100%

Inorganic oxides                             0.5-100%

Odors                                                         1- LTV

Aromatic                                     0.5-100%


     Either two similar cylinders  and two resistances  or  four
     similar cylinders  and resistances are used  to  form  the  arms
     of a Wheatstone bridge circuit.   The final  equilibrium  tem-
     perature of the wire depends  upon the thermal  conductivity
     of the gas surrounding it.   Lower temperatures  produce  higher
     conductivity,  whereas higher  temperatures produce lower
     conductivity.   If  the gases  in  the cells are  the  same,  the
     wires will reach the same temperature and therefore  will
     have the same  resistance; thus  the bridge is  balanced.  If
     the gases are  different,  the  final temperatures and  resis-
     tances of the  wires  will  be  unbalanced.  The  extent  of  un-
     balance of the bridge is  measured by a galvanometer  or  a
     potentiometer  and  can be  calibrated in terms  of the  com-
     position of the gas.

          The thermal conductivity method is usually operated  as
     a continuous recording method,  since the off  balance  condi-
     tion can be measured by a recording potentiometer.

          The method is easily applied to the determination  of
     the composition of binary mixtures,  provided  that the two
     gases have different thermal  conductances (Willard,  1958).
     The greater the difference between the two  gases, the greater
     the sensitivity of the method.   A multlcomponent mixture  can
     be treated as  a binary mixture  if all  components but  one
     remain constant or if all the gases  except  one have  nearly
     the same thermal conductance  (Eckman,  1950).

          With multlcomponent  mixtures, the thermal conductivity
     detector is  frequently used  in  conjunction  with gas-liquid
     chromatography.  The components  are  separated by a packed
     column with  each component exhibiting  a characteristic  re-
     tention time.   Quantitative measurements are made by  com-
     paring peak  heights  or areas  against standards.

          Catalytic Combustion -  The  measurement of combustible
     gases or vapors is based  on  the  principle of  catalytic  com-
     bustion (Mine  Safety Catalog  No.  8B3659-3,  1971).  A  mixture
     of gases may be either flammable  or not, according  to the
     concentration  of the vapor in air 1n which  a  flash will occur
     or a flame will  travel if the mixture is Ignited.  The  lowest
     percent at which this occurs  1s  the lower explosive  limit
     (LEL) and the  highest percent is  the upper  explosive  limit
     (DEL).  The LEL is considered 100% and the  concentration
     below are considered as percent  of the LEL.

          If propane and  air are  confined and Ignited, an  explo-
     sion results.   The LEL of propane 1n air 1s 2.2% and  the  UEL
     1s 7.3%.  Mixtures above  7.3% are too rich  to support com-

bustion, and those less than 2.255 are too lean.   Since air
is the diluent and is readily obtainable, all  concentrations
above 2.255 are dangerous; therefore the instrument would be
calibrated from 0 to the 2.2% LEL for propane.

     The catalyst is usually a platinum wire filament (Mine
Safety, 1971 and Scott Aviation, 1969) maintained at a con-
stant temperature.  When the combustible gases or vapors
come in contact with the catalytic filament, they are oxidized
or  burned  catalytically.  Two identical  filaments are used
in the bridge circuit.  One filament, known as a  reference
filament, is sealed in air 1n a separate well  of  the same
chamber as the active or exposed filament.   Two fixed re-
sistors are used as the legs of the bridge  connected together
by a variable resistor.  The bridge is balanced with both
filaments in air by adjusting the variable  resistor or zero
control.  The increase in resistance is proportional to the
amount of flammable gases or vapors in the  sample and 1s
measured by a meter, the scale of which 1s  calibrated to give
direct readings in percent of the LEL,

     Portable combustible gas Indicators may sample auto-
matically or manually for the detection of  combustible toxic
gases or vapors.  The ranges and sensitivities of the indi-
cators depend on the model and the gases or vapors for which
the instrument is calibrated.  The instrument may operate on
one full scale, a dual scale, or as high as five  scale ranges
which may read 0 to 1.0%, 0 to 10% or 0 to  100% of the lower
explosive limit.  For the gases or vapors calibrated for
those indicators with two or more scales, ranges can be pro-
vided with multiple calibrations, whereby readings on any
gas or gases for which 1t has been calibrated can be made by
turning a selector switch.  There are more  sensitive models
that read 0 to 1.0% of the lower explosive  level  on one
scale and 0 to 1,000 ppm of vapor on a second scale (Johnson-
Williams Products, 1970),  Other models, while not scaled
off in ppm designation, may be calibrated to measure in this
low level area.  These models provide a check on  the toxicity
of combustible vapor concentrations and furnish ppm measure-
ments for comparison with threshold value limits.

     A partial listing of selected gases and vapors that
are reported to be measurable by catalytic  combustion devices
in the ppm range are listed in Table 8-III  (Johnson-Williams
Products, 1970).

     Continuous combustible gas detectors (General  Monitors,
1970; Mine Safety, 1970j Erdco Engineering, 1970) consist  of
two components: the control and indicating unit  and  a  remote
detection unit.  The control unit is Individually powered

     Table 8-III  Partial Listing of Gases and Vapors
                  Detectable 1n the ppm Range*
           Acetone              Hexane
           Amyl acetate         Methane
           Benzene              Methyl alcohol
           Butadiene            Methyl butyl ketone
           Butyl acetate        Methyl ethyl ketone

           Butyl alcohol        Methyl Isobutyl ketone
           Cyclohexane          Methyl propyl ketone
           Cyclohexanol         Octane
           Cyclohexanone        Pentane
           Ethyl acetate        Perchloroethylene

           Ethyl alcohol        Propane
           Ethylene             Isopropyl alcohol
           Ethyl ether          Propyl alcohol
           Formaldehyde         Stoddard solvent
           Heptane              Toluene

                      1 ,1 ,l-Tr1chloroethane

     *Most aromatic hydrocarbons are reported to be measurable
      1n the ppm range by sensitive combustible gas Indicators.

and contains the Indication and warning devices.   The  remote
detection unit designed for area monitoring,  houses  the cata-
lytic sensing device.  Each control  unit may  be coupled with
a single or a dual detection unit for monitoring  locations
as far as 3000 to 5000 ft away.  Several control  units may
be housed in modules to facilitate centralized  monitoring
from multiple locations.

     Combustible gas or vapor is detected at  the  sensory
probe by means of a low temperature  catalytic reaction on
a heat sensitive surface.  Matched elements are heated and
exposed to the gas or vapor.  One element in  the  pair  1s
catalytically active; the other 1s Inactive and acts as

     In the presence of a combustible gas, oxidation takes
place on the active element, slightly raising Its temperature.
This Increases the resistance of a leg of a  circuit  bridge
and gives a direct current signal which drives  the meter and
accessory circuit.  The gas concentration for which  the unit
1s calibrated 1s displayed 1n percent of the  LEL.

     The assembled units offer continuous monitoring  of
areas in multiple locations.  The units may be calibrated
to give direct readouts of combustible gases  and vapors In
one or two ranges, 0% to 100% or 0% to 10% of the LEL.  The
continuous sensory filaments operate at about one-half the
temperature of the conventional platinum wire filaments;
therefore the life of the circuit bridge Is relatively long.
A bridge life of  10,000 hr 1n air has been reported.   The
control unit contains the alarm circuitry for Indicating
failure of the sensing device or an Increase of vapor con-
centration, readout  circuitry for powering the meter or
potentiometric records, and all operating controls.

     The assembled units are adaptable  for use with any
gas or vapor whose molecular weight and explosive range are
known.  Their application, therefore, extends to all  Indus-
tries where combustible fluids and resultant gases and
vapors are employed.

      lonization  - The flame 1on1zat1on  detector operates by
burning hydrogen  1n  air *o produce a flame that  contains a
negligible amount of ions.  When  hydrocarbons are Introduced
Into  the hydrogen flame an  lonization  process ensues,  where-
by a  large number of Ions  are  produced.  An  electrostatic
field  1n the  vicinity of the flame causes an ion  migration
in which positive  Ions  are attracted to  a collector and the
negative ions to  a  burner  Jet.   The  small lonization  current
established between  the  two electrodes  1s measured by an

      electrometer  and  the  signal  indicated  by  a meter or by a
      potentiometric  recorder.   The magnitude of the signal is
      an  indication of  the  number  of  carbon-hydrogen bonds passing
      through  the flame  and  this is directly proportional to the
      hydrocarbon concentration.   The  sampling  system is an inte-
      gral  part  of most  commercial total hydrocarbon analyzers,
      designed to measure continuously total hydrocarbon and
      other  organic compound  concentrations  in  a gas stream or
      ambient  air (Davis Instr.  Div.(  1965;  Bendix, 1970; Beck-
      man  Instr. , 1969}.

           Trace quantities  of such organic  compounds as hydro-
      carbons  (aliphatic and  aromatics), aldehydes, amines, ketones,
      acetates and  alcohols  in part-per-million range in air may be
      detected.  Carbon  atoms bound to oxygen,  nitrogen, or halo-
      gens  give  reduced  or  no response, whereas there is no response
      to  nitrogen,  carbon monoxide, carbon dioxide or water vapors
      (EPA,  1971).

           The analyzers are  usually calibrated using methane and
      results  recorded  in terms  of methane equivalents.  A number
      of  other gases  such as  butane and propane may also be used
      depending  on  the  sensitivity desired and  the type of organic
      to  be  monitored.

           The specification  for a given total  hydrocarbon analy-
      zer will vary as  to model  or manufacturer.  They may be
      portable or used  as a stationary unit  adaptable to many
      industrial monitoring applications and as elaborate as the
      addition of optional support equipment.   The instrument may
      offer  one full  scale range or 5  to 6 ranges with a 1:10 gain.
      The full scale  range may be 0 to 1 ppm, 0 to 10 ppm or
      through several ranges up  to 0 to 10,000  ppm hydrocarbons
      per calibration equivalent.  The range may be varied by ad-
      justing the attenuation and flow rate of  the sample to the
      detector.  Accuracies and  noise  levels are also different
      and are expressed in terms of percent of  the lowest full
      scale  per calibration gas.  Zero drift varies and depends
      on airflow rate,  sample flow rate, fuel flow rate, ambient
      temperature change, detector contamination and electronic
      drift.  Zero drift is usually expressed 1n range percent/24
      hour period.

          Other recording  instruments using an ionization prin-
      ciple  for the determination of trace gases in the parts-per-
      million and in  some cases  1n the parts-per-billion range
      have been developed and are  commercially  available (American
      Industrial, I960).

     This method of gas analysis uses an lonization chamber
to detect an aerosol  which is formed from the contaminant in
a number of ways.  The ionization chamber is composed of a
stainless steel cylinder electrode and an inner electrode of
a heavy wire.  An alpha emitter is placed along the inner
surface of the cylinder.  The two electrodes are given oppo-
site charges which cause  a flow of ion current through the
gas between them.

     With only air present in the chamber, there is a rela-
tively constant flow of ion current.  However, very small
concentrations of particular matter cause a pronounced drop
in this flow of ion current.  The effect of the particulates
on the conductivity of the ion chamber is a measure of the
concentration of the gas or vapor of interest (Mine Safety,
1970).  The formation of particulates for a wide range of
materials is accomplished by: reacting the trace gases with
various reagents to produce particulate matter; direct py-
rolysis of the sample; or a combination of both chemical
reaction and pyrolysis.

     The conductivity is measured by a detector cell  and a
compensator cell placed on two arms of a Wheatstone bridge
circuit.  The presence of particulates causesa drop in ion
current in the cell and results 1n an unbalance of the
bridge which is amplified for driving a meter or a recorder.

     There are portable models that are reported to be
accurate as well as easily operated.  Applicable to the
scope and nature of this survey are the analyses of acid
and alkaline gases, oxides of nitrogen, and amines which
are sensitized through reactions with specific reagents.
For example, ammonia may be used to sensitize such materials
as halide acids, nitric oxide and sulfur dioxide.  A com-
bination of pyrolysis and use of a reagent may be employed
in the determination of organic halides.  On  passing a
chlorinated hydrocarbon through a copper oxide converter
at elevated temperatures, a copper chloride aerosol 1s

     The overall applicability of this  instrument is  limited
as to the variety of gases  and vapors it can  monitor.   Sev-
eral of these gases have been mentioned, but  its strongest
asset 1s not variety but the  sensitivity that the Instrument
can achieve.

     Ultraviolet Spectrometry - Ultraviolet  analysis  measures
the degree of  light absorption  as  it  occurs  1n  a gas  or 0
liquid  sample.  The ultraviolet spectrum  ranges  from  100A to
3800A*,  and those gases  whose  nature  It  is  to  absorb  in this

      ultraviolet region can be measured.   In  general,  ethylene,
      carbonyl  structures and molecules  which  Include  the  benzene
      ring In their structure all  absorb 1n the  ultraviolet  region
      (Holzbock, 1955 and American Industrial,  1960).

           The ultraviolet spectrophotometer will  detect gases
      and vapors at low concentrations  and  will  detect  a great
      array of gases and vapors that absorb in  the ultraviolet
      band.  The Instrument may therefore  be used  either for com-
      plete analyses of samples by running  a complete  spectrogram,
      or for continuous analysis of process streams  by  operating
      at a fixed wavelength critical to  the analysis  being made
      (Eckman, 1950).

           Ultraviolet spectrophotometry may be  applied to the
      detection of trace amounts of benzene, toluene,  and  other
      aromatic hydrocarbons, and chlorinated solvents  1n air 1n
      the ppm range (Univ. of Michigan,  1956),   Polynuclear  aro-
      matics, some of which are believed to be  carcinogenic, have
      been qualitatively detected and quantitatively  analyzed by
      ultraviolet spectrophotometric methods (American  Industrial,
      1960) and as previously mentioned, ethylene  and  carbonyl
      structures absorb In the ultraviolet band  and  are therefore
      detectable with this method.

           Descriptions and Illustrations  of several  types of
      ultraviolet spectrophotometers as  well as  other  spectro-
      photometric Instruments that are commercially available have
      been compiled (Wlllard, 1958).  Special  mechanisms have also
      been described which permit automatic measurement on as many
      as three plant streams by one instrument (Holzbock,  1955).
      Most spectrophotometers are for stationary use  as opposed
      to the more portable photometers.

           Infrared Spectrometry - The infrared spectrum ranges
      from wavelengths of 7800X to 3,000,OOOA.   Those  gases  or
      vapors which will absorb in this heat radiation band can
      therefore be measured.  This applies to  the analysis of
      almost any organic and some inorganic compounds  under  the
      proper circumstances (American Industrial, 1960).

           Of all molecular spectroscopy, the  infrared spectro-
      Dhotometer 1s the most versatile.  Infrared spectrophotom-
      Pters  may be used to analyze an entire  sample or at least
      a number of Us constituents.  However,  there are other
      infrared spectrophotometers that are designed to contin-
      ,,n,ielv analyze for one constituent in a  given sample.
absorption characteristics.   Conversely,  elemental  diatomic
gases like oxygen, hydroqen,  nitrogen,  helium,  chlorine,  etc
do not have absorption characteristics.   Within acceptable
substances the concentration  of any component in a  mixture
can be determined by the specific absorption pattern that is
unique for each component.

     There are three basic types of infrared spectrophotom-
eters:  the dispersion analyzer, the positive nondispersion
analyzer and the negative nondispersion analyzer (Holzbock,
1955).  Common to all three types are:  they have been em-
ployed 1n commercially available Instruments; they  provide
continuous measurement; and they analyze for only one constituent
either alone or in mixture for concentrations as low as ppm.

     The dispersion analyzer concentrates upon a single
wavelength or a narrow wavelength band within the infrared
spectrum.  The radiation of a particular wavelength that
passes through the sample without being absorbed 1s measured.
Generally, light of a second wavelength is simultaneously
applied and the ratio of transmission at two wavelengths is
measured.  The dispersion analyzer may be readily adjusted
for various wavelengths and maximum sensitivity.  However,
the optical system may become relatively complex.

     Unlike the dispersion analyzer, the two nondispersion
analyzers utilize all the energy from an infrared radiation
source and pass it through the  sample, unless  the width of
the wavelength band  1s limited  by filters.   In  addition,
the nondispersive analyzer usually  spHtsthe infrared  radi-
ation into two parallel beams.

     The  positive nondispersive  analyzer operates by passing
one beam  through  the  sample to  be  analyzed  and  from  there
into  a cell which is  filled with the component  to be detec-
ted.  The  other beam  passes directly to a second cell with
the  same  component.   The  gas  in both cells  expandsdue  to
the  heat  absorbed from  the infrared radiation.   The  cell
directly  exposed  to  the  infrared radiation  will expand more
than  the  cell  which  received  Infrared  radiation only after
it had  passed  through the sample.   The relative expansion
of the  two  cells  is  then  usually measured  by means  of  a
variable  capacitor.   Also characteristic of this analyzer
is a  light  chopper  that  interrupts  the radiation of the
infrared  light source periodically, largely eliminating  the
effect  of changing  ambient temperatures.

      In  the  negative nondispersion analyzer, both  beams pass
 through  the  sample,  but one  traverses  an additional cell
which is  filled with the component to  be detected.  The non-

      absorbed  radiation  from  each  beam  1s  then measured  by  a
      thermocouple or  bolometer  and  the  difference measured  as
      an Indication of the  concentration of the component for
      which the sample is analyzed.   This analyzer is  character-
      ized as  being more  sensitive  to ambient  temperature than  the
      positive  type, less sensitive  as an analytical Instrument
      but simpler  and  more  rugged.

           The  prisms  for the  infrared spectrophotometers are
      usually made of  four  materials. The  prism materials include
      lithium  fluoride, calcium  fluoride, sodium chloride and
      potassium bromide but quartz  and cesium  bromide  have also
      been used (Willard, 1958 and  Eckman,  1950).  The  selection
      of the prism is  determined by  the  wavelength range  most
      adaptable to the substance under analysis.

           These Infrared spectrophotometers have been  described
      as being  reliable and easily maintained.  They are  relatively
      rugged and yet versatile;  they  are simple in design, yet
      possess wide construction  tolerance (Conn, 1960).

           Colorimetric - Colorlmetric methods Include  general
      laboratory wet chemical  techniques as well as direct reading
      colorlmetric instrumentation.   The principle 1s  based  on  the
      formation of a specific  color  resulting  from the  reaction of
      the contaminant  with  reagent or number of reagents.  The
      resultant color  1s  then  measured with  a  photometer.  Sensi-
      tive and  specific colorlmetric  reactors  have been developed
      for many  of  the  air pollutants.  Methods of colorlmetric
      analysis  are divided  into  three basic classes:

           1.   Detector tubes
           2.   Laboratory methods
           3.   Automatic  instruments

      The Initial  cost investment and the degree of technical
      skill  required for  application  of  these  methods  are repre-
      sented 1n increasing  order.

           Detector tubes provide a  noninstrumental technique for
      semlquantitatlve analyses  for  many toxic gases and  vapors 1n
      the threshold limit ranges.  The detector tubes  are glass
      tubes  which  contain selective,  reactive  chemicals impregnated
      on sands, alumina gels,  silica  gels or other qels.  The com-
      oosltion  of  the  reactive substrate Is determined  by the gas
      or vapor  to  be tested.   As an  example, for hydrogen sulflde,
      the substrate may consist  of  granular silver cyanide on
      activated alumina.  For  sulfur  dioxide a mixture  of Iodine
      flel  and tetrabase sand may be  used.   A yellow silica gel
      Impregnated  with a  complex slllco-molybdate compound cata-

lyzed by palladium sulfate is used for carbon monoxide.  For
aromatic hydrocarbons, the substrate used may be a mixture
of sulphuric acid gel and paraformaldehyde sand (Univ  of
Michigan, 1956).

     Samples of the contaminated air are drawn through the
tubes by means of aspirators or assorted hand pumps.  Some
hand pumps provide for measured volume samples, while both
aspirators and hand pumps are equipped with variable ori-
fices.  The measured sample is drawn through the substrate
and the gas or vapor reacts with the chemicals in the sub-
strate, producing a stain which is directly proportional  to
the concentration of that 9s or vapor in the suspected
stream.  The length of the stain is measured by an appro-
priate precalibrated scale for the vapor being tested.  The
stain may develop directly with contact of the gas or vapor
or may be developed with the addition of reagents.

     The detector tubes are usually supplied as kits,with
aspirators or hand pumps provided.  A listing of the gases
and vapors for which the Detector tubes can be used are
provided by the manufacturers (Davis Instr, Div., 1963 and
Mine Safety, 1971).  Also Included 1n this listing are those
compounds that interfere with the respective reaction.  The
effect of the Interference should therefore be considered
when measuring for specific compounds.  Necessary reagents
are also provided when applicable.  The detector tubes may
be the ready-to-use type or those that are mixed prior to
use and equipped with special filters and prefilters.  Re-
mote sampling kits are available for sampling inaccessible
areas.  In addition to these accessories, a pyrolyzing unit
may also be provided.  The pyrolyzer separates toxic com-
pounds by the application of heat, whereby specific chemi-
cals are released whose concentration can be related to the
level of the original compound.

     This type of detection is particularly suitable for
the small businessman.  The Initial cost is quite low, the
technical skills required to analyze a sample are practically
nil and small businesses generally emit a single pollutant.

     Laboratory colorimetric methods require a photometer
for measurement of the concentration of a developed color.
Photometers vary 1n complexity and concomitant cost.  Simple
photometers employ a light source, filters, phototube and
readout.  The more complex and expensive photometers use
gratings or prisms to provide light of the desired wave-
length.  The sample which 1s collected must be treated chem-
ically to remove Interfering contaminants, If present,  and
develop the desired colored compound.  This, of course,  re-
quires some technical skill plus associated laboratory

          Automatic colorimetrlc Instrumentation  is  available
     for many contaminants.  The basic instrument consists  of  a
     sampling pump, a scrubber system to remove  the  contaminant
     from the gas stream, automatic liquid metering  of  the  solu-
     tion, automatic reactant feed, associated electronics  and
     controls and direct readout.  The instrument is  versatile
     in that various contaminants can be analyzed by  changing
     the scrubbing solution or reactant.  This type  of  instru-
     ment generally requires a high degree of  technical  skill
     to select the appropriate separation and  reactant  solutions
     and to maintain the instrument in a reliable operating

          Electroconductivity - The electroconductiv1ty  analyzer
     continuously records concentrations of gases or  vapors that
     will ionize in water either directly or after decomposition
     by heat.  The principle of measurement is the differential
     measurement of conductivity before and after a  gas  sample
     ionizes in distilled water.  The conductivity is measured by
     two sets of platinum electrodes  arranged  in  an  alternating
     current bridge circuit.  This type of instrument is  usually
     used to continuously record atmospheric concentrations, but
     has been adapted for process control (Univ.  of  Michigan,
     1956).  Instruments are available which permit  periodic
     check of gases or vapors at 3 to 13 remote  points  and  the
     results are shown on various recorders (Univ. of Michigan,

          A continuous flow of water  scrubs the  gas  sample
     stream.  The mixture is circulated around the platinum
     electrodes and the amount of ionization is  indicated.   The
     water is then passed through a deionizing bed and  recircu-
     lated through the system.  In some instances the sample
     need not pass through the pyrolyzing furnace before  scrub-
     bing with  the water.

          This  type of system makes for a basic  analyzer  and is
     adapted for the measurement of one gas or vapor  in  air in
     the ranges of 1 to 5 ppm or lower if necessary  for  such
     vapors as  chlorine, sulfur dioxide and hydrogen  sulfide
     (American  Industrial, 1960).  Carbon tetrachloride,  and the
     oxides of  nitrogen may also be determined.   The  system may
     also be employed for the determination of organic  chloride
     vapors (Davis Instr. Div., 1963  and American Industrial,

          Halogen Leak Detection - Halogen leak  detectors main-
     tain an ion current by electrically exciting an  emitter.
     The detector sensing element for the halogen leak  detector
     is  a diode consisting of a heater element (positive) and  a
     collector  (negative) arranged coaxlally (General Electric).

A positive ion current flows  at all  times  from the  emitter
to the collector.   When a trace of halogen gas is present
between the collector and emitter the  ion  current increases
markedly.   The increase In ion current 1s  linearly  propor-
tional to  the concentration of halogen from 0  to  1  ppm.
For concentrations from 1 ppm to about 1000 ppm,  the  in-
crease 1s  exponentially proportional.   Above 1000 ppm,  no
further increase of ion current occurs and the element  de-
sensitizes rapidly but cleaning usually restores  original

     These instruments are portable  and are designed  for
continuous operation.  The instruments are nonspecific  in
the determination of halogenated hydrocarbons  or  mixtures
of halogenated hydrocarbons.   Nonhalogenated compounds,
such as hydrocarbons, will not interfere,  however;  and
mixtures of halogenated vapors with  other  vapors  may  be

     Designed for field use,  the normal operating range  is
0 to 500 ppm for most substances.  The lower limit  of de-
tection is about 10 ppm and the accuracy throughout most
of its range is about 10% of the amount being measured.  The
instruments employing this principle are most often used for
the determination of perchloroethylene, trichloroethylene,
carbon tetrachloride, methylene chloride and other  similar
halogenated hydrocarbons (Univ. of Michigan, 1956 and Davis
Instr. Div., 1971).

     Emission Spectrometry - Emission spectrometers are
usually employed in the determination of elements and metal-
lic compounds.  However, this principle has been applied to
the determination of halogenated hydrocarbons in air  (Davis
Instr. Div., 1971).  Briefly, emission spectrometers  employ
an arc spark for producing radiation emissions and  a  grating
for refracting emitted radiation into a spectrum.  The in-
tensity of the radiation lines are measured directly  by means
of phototubes.

      In this particular  type of spectrometer, air containing
halogenated hydrocarbons is passed through a chamber con-
taining an alternating current electric arc between a copper
electrode and a platinum electrode (Univ.  of Michigan, 1956).
A bright  line spectrum of  copper is produced when the air
surrounding the arc contains hallde vapors.  The intensity
of the copper spectrum is  proportional to  the concentration
of hallde vapors  present,  and  the blue lines are measured
with  a blue sensitive phototube fitted with a blue glass
filter.   The change in current output of  the  phototube is
then  measured by  a direct  current vacuum  tube voltmeter.

          This  type  of  Instrument  is expensive and 1s expected to
      have little  application  in  the measurement of absorbent bed
      Inlet  and  outlet concentrations.

          Mass  Spectrometry - The  mass spectrometer separates
      Ions of  various atoms, molecules or their fragments by pass-
      ing  the  ions  of these materials through a magnetic field.
      The  ions are  deflected according to their mass and charge
      and  impinge  upon a  collector.  The current Intensity of the
      collector  1s  proportional to  the number of Ions which Im-
      pinge  on the  collector.

          Mass  spectrometers  are highly specific for individual
      contaminants.   The  disadvantages are that they do not possess
      the  low  level sensitivity of  many other instruments, they are
      expensive  and they  are not  portable.  Sensitivity can be
      increased  by  concentration  of a specific impurity of Interest,
      Mass spectrometers  are amenable to analysis of high concen-
      tration  of contaminants  and the results from this type of
      analysis can  be used to  determine what type of sorbent bed
      should be  used, to  size  the bed,and to predict the life time
      of  the bed.

          Scentometry -  Measurement of odors is, at best, semi-
      quantitative  even when only a single odiferous material 1s
      present.   When  the  test  atmosphere contains a mixture of
      odiferous  materials, even semlquantltative identification
      1s essentially  Impossible.  Fortunately, many processes emit
      only one odor and since  package sorbent devices will be used
      1n many cases for odor control, it will be a straightforward
      task to determine when the  device needs replacement or re-

          A scentometer  1s marketed which provides a means of
      multiple dilution of a sample.  A trained panel smells each
      diluted sample  and, if the  pollutant Is known, an estimation
      can be made of  its  concentration.

          Another method of quantitatively defining odors 1s
      given  in ASTM D 1391-57.  A sample is collected 1n an air
      sampling syringe.   The contents of the air sampling syringe
      1s then  Injected into a  mixing syringe and diluted with
      100 ml of  clean air.  After 15 sec, the syringe 1s placed
      near the nostrils  and the content 1s ejected for 2 to 3 sec.
      Various dilutions  are made  until the odor 1s just percep-
      tible   The  ratio  of the total volume of the diluted sample
      to the volume of the original sample 1n the diluted sample
      < a measure  of the concentration of odor In the original
      sample.  Table  2-IV lists the odor threshold of some air

Recommended Methods for Specific Classes of Impurities


     The hydrocarbons can be measured as a general  class by
flame ionizatlon detection down to the ppm level.   Gas-liquid
chromatoqraphy can be used to separate the various  hydrocar-
bons and,when used 1n conjunction with flame lonlzatlon de-
tection, Individual hydrocarbons in a hydrocarbon mixture can
be measured at the ppm range.  Ultraviolet and Infrared spec-
trophotometry can be used to measure individual  hydrocarbons
1n low ppm ranges, also.  Catalytic combustion provides a
means of detection at low levels, but is generally  applicable
only when a single hydrocarbon (and no other catalytically
combustible material) 1s present.  Thermal conductivity is
applicable down to about the 100 ppm level and can  be used
In conjunction with gas-liquid chromatography for  the detec-
tion of specific hydrocarbons.  Mass spectrometry  is  limited
to measurement at approximately the 0.5* level but  does pro-
vide detection of the individual hydrocarbons.

     Oxygen Containing Organic Compounds

     These compounds exhibit essentially the same  detection
characteristics as the straight chain hydrocarbons.  Again,
flame 1onizat1on is the most sensitive means of detecting
this class of compounds and chromatography can be  used to
separate the various oxygen containing organics.  Sensitivity
colorimetric methods have been developed for measurement of
oxygen containing organics, also.

     Haloqenated Compounds

     The halogenated compounds are particularly sensitive to
electron capture detection.  The electron capture detector
has been used 1n conjunction with gas-liquid chromatography
to separate and detect mixtures of halogenated compounds.
The halogen leak detector 1s a suitable instrument for the
detection and measurement of halogenated compounds, also.
The flame ionlzation detector can be used if a single halo-
genated hydrocarbon is being detected.

     Inorganic Oxides

     Oxides such as CO can be measured by catalytic com-
bustion down to the few ppm level.  Sensitive chemical
methods are available for the detection of other Inorganic
oxides such as S02*, Instrumentation has been developed which
will automatically perform this function.


          They pose the greatest problem 1n terms of quantifica-
     tion.  Fortunately, olfactory detection is the most practi-
     cal means of sensing odors and, since the objective of sorbent
     systems is to remove these odors, an inexpensive means of
     detection exists.  Semiquantitatlve estimations can be made
     with the use of a scentometer.

          Aromatic Hydrocarbons

          These can be detected by the same methods used for
     measurement of straight chain hydrocarbons.   Again, flarte
     ionization detection provides the most sensitive means~of
     detection.  Sensitive colorometric methods have been devel-
     oped for many of the aromatics.  Other methods include ultra-
     violet and infrared spectrophotometry.

          Polynuclear Hydrocarbons

          The polynuclear hydrocarbons are generally of .interest
     in the low ppm to ppb range due to the carcinogenic nature
     of many of these compounds.  Because of this low level re-
     quirement and since other hydrocarbons are generally present
     with the polynuclear hydrocarbons, ultraviolet -or Infrared
     spectrometry are the best means of measuring this class of
     compounds.  Colorimetric methods are also applicable 1n
     certain cases.

A      Glossary
3      Information Sources
C      Tables of Data
D      Methods
E      Sample Calculations
F      Literature Cited
G      Bibliography

                      Appendix A


A      Adsorption potential equal  to [T/Vm] log [pQ/p],
       also a constant
Ac     Area of adsorbent or catalyst bed,  also collision
B      Constant
C      Pollutant vapor concentration, also a constant
Cfc     Pollutant breakpoint effluent concentration

Ce     Pollutant effluent concentration from adsorbent
       or catalyst bed
Cj     Pollutant influent concentration to adsorbent or
       catalyst bed
CQ     Concentration of saturated  vapor
D      Diameter of adsorbent or catalyst bed
Dp     Adsorbent or catalyst particle diameter

Dv     Diffusion coefficient
E      Heat of adsorption on different layers, also
       kinetic energy
F      Total gas flow
AF     Free energy change equal to -RT In  [p0/p]
AH     Heat of adsorption or desorption

K      Reaction rate constants
Kf     Volumetric film heat transfer coefficient
L      Adsorbent or catalyst bed length or depth
Lc     Critical bed length
Ls     Saturated bed length

Lt     Total bed length
Lz     Adsorption zone length
M      Molecular weight
N      Avogadro's number, 6.03 x 102J molecules/mol
P,PA   Total pressure

AP     Pressure drop through adsorbent or catalyst bed
Q      Heat of reaction, total heat
R      Gas constant
S      Surface area
AS     Entropy change

T      Absolute temperature
U      Face velocity to adsorbent or catalyst bed
Um     Intergranular gas velocity In adsorbent bed

     V      Liquid molar volume of pollutant vapor at ambient
     Vm     Liquid molar volume of pollutant vapor at normal
            boiling point

     Vc     Catalyst or adsorbent bed volume
     W      Total weight, adsorbent, catalyst or pollutant
     Wc     Weight of  carbon
     Wr     Total weight of pollutant remaining in carbon after
     Ws     Total weight of pollutant adsorbed in carbon
            saturated  layer

     Wz     Total weight of pollutant adsorbed in carbon ad-
            sorption zone
     a      Superficial area of particles per unit volume
            of  particles
     a,b,c  Component  species
     d      Pore diameter, carbon density
     h      Planck's constant. Wheeler mechanism index

     k      Constant
     n      Number of  adsorbed layers, constant 1n FreundUch
     p      Pollutant  vapor pressure
     Po     Vapor pressure of pure liquid
     r      Pore radius

     t      Time, service time
     tfc     Service time to breakthrough concentration of C^
     tc     Contact time
     v      Volume of  adsorbed vapor
     vm     Volume of  vapor required to form a unlmolecular layer

     a      Polarizability
     ct,3,y  Constants
     Y      Surface tension
     c      Void volume in adsorbent bed
     e-f     Polanyi potential

     n      Viscosity
     6      Fraction of surface area covered
     X      Heat of vaporization of pure liquid
     u      Dipole moment, viscosity of pollutant in air
     v      Characteristic light dispersion frequency

     p      Gas density
     J      Distance between centers of two dlpoles
     T      Multilayer correction factor
     A      Contact angle
     u      Amount pollutant adsorbed per unit quantity of

cor     Mean amount of pollutant adsorbed  per  unit  quantity
       of adsorbent that remains at the end of  regeneration
       Amount pollutant adsorbed in saturated layer  per
       unit quantity of adsorbent
       Mean amount pollutant adsorbed in  adsorption  zone
       per unit quantity of adsorbent

                     Appendix B

        Information Sources on Air Pollution

Trade Associations 1n the Washington, D.C.  Area

     Bureau of Raw Materials for Vegetable  011  & Fats
     Council of Independent Laboratories
     National Drycleaners Institute
     National Paint, Varnish & Lacquer Association
     Manufacturing Chemists Association
     National Fisheries Institute
     Bureau of Fish Oil and Meal Processes
     Research Council of the Graphic Arts Institute

Trade Associations in the Chicago Area

     National Restaurant Association
     Brewers Association
     Institute of Baking
     National Canners - NCA Staff Center
     Candy & Confectioners Institute
     National Renderers Association
     American Meat Institute
     Institute of Paper Chemistry

Pollution Control Agencies

     Allegheny County Dept. of Health
     Pittsburgh, Pennsylvania

     Southwest Air Pollution Control Authority
     Vancouver, Washington

     Los Angeles County Air Pollution Control District
     Los Angeles, California

     District of Columbia Dept. of Public Health
     Washington, D.C.

     Bureau of Air Pollution Control
     Illinois Dept. of  Public Health
     Springfield. Illinois

     Wyoming Dept. of Public Health
     Cheyenne, Wyoming

     Bay Area Air Pollution Control  District
     San Francisco, California

           Environmental  Control  Board
           New  York,  New  York

           Houston  Health Department
           Houston, Texas

           Fulton County  Health Department
           Atlanta, Georgia

           Albuquerque Dept.  of Environmental  Health
           Albuquerque, New Mexico

           Air  and  Water  Pollution Control
           City of  Jacksonville
           Jacksonville,  Florida

           City of  Chicago Dept.  of Air  Pollution  Control
           Chicago, Illinois

           Dept. of Public Health
           Philadelphia,  Pennsylvania

           Denver Air Pollution Control  Agency
           Denver,  Colorado

           Division of Air Pollution Control
           St.  Louis , Missouri

      Plant Trips.  Pollution  Control

           Vic  Manufacturing  Company
           Minneapolis, Minn.

           American Machinery and Foundry
           Minneapolis, Minn.

           Vulcan Cincinnati, Inc.
           Cincinnati , Ohio

           Charleston Rubber Co.
           Charleston, South  Carolina

           Reigel Paper Co.
           Edenburg, Indiana

           Norfolk  Tallow Co.
           Chesapeake, Virginia

                Government Documents

                                                   Wash.  D.C.
Title	    Ident.  No.

Atmospheric Emissions from Fuel  Oil  Comb.             AP-2
A Pilot Study of Air Pollution in Jacksonville,
   Fla.                                              AP-3
Air Pollution and the Kraft Pulping  Industry         AP-4
Dynamic Irradiation Chamber Tests of  Automotive
   Exhaust                                           AP-5
Air Pollution in the Coffee Roasting  Industry        AP-9
Community Perception of Air Quality                   AP-10
Reactivity of Organic Substances in  Atmospheric
   Photooxidation                                    AP-14
Selected Methods for the Measurement of Air
   Pollution                                         AP-11
An Air Pollution Management Plan for Nashville       AP-18
Effects of the Ratio of Hydrocarbon  to Oxides of
   Nitrogen, Auto Exhaust                            AP-20
Continuous Air Monitoring Program, Cincinnati        AP-21
Continuous Air Monitoring Program, Washington,DC     AP-23
Atmospheric Emissions from Coal  Comb.                AP-24
Seminar on Human Biometeorology                      AP-25
Workbook of Atmospheric Dispersion Est,              AP-26
Air Pollution Aspect of Tepee Burners                AP-28
Rapid Survey Technigue for Estimating Community
   Air Pollution Emissions                           AP-29
Selection and Training of Judges for Sensory
   Evaluation Diesel Exhaust Fume                    AP-32
Sources of Polynuclear Hydrocarbons  in the
   Atmosphere                                        AP-33
Air Pollution Engineering Manual                     AP-40
Compilation of Air Pollutant Emission Factors        AP-42
A Compilation of Selected Air Pollution Emission
   Control Regulations and Ordinances                AP-43
Handbook of Air Pollution                            AP-44
Interim Guide of Good Practices  for Incineration
   at  Federal Facilities                             AP-46
Air Pollution Translations, Vol.  I                   AP-56
Air Quality Criteria for  Carbon  Monoxide             AP-62
Air Quality Criteria for  Hydrocarbons                AP-64
Air Pollution Translations, Vol.  II                  AP-69
National Survey of  the Odor Problem, Phase  I,
   Jan.  1970, prepared by Copley  Int. Corp.
Nationwide  Inventory of  A1r Pollutant Emissions,
   1968, HEW

      Section  I   General  Information
            Name of  agency: 	
            Address:   City	
            Name of Individual:
            Telephone number:  AC_
            Responsible for pollution control in:   (name)
            Responsible for emission from:  (check)
               Agri cultural

Section II   Emissions  Source  Complaints

The purpose of this  question  is  to  determine  the  need  for
control of the various  small  emissions  sources.

Fill in appropriate  square with  number   of complaints  per
year.  If complainants  did not identify odor  or  other  effect
produced by an air pollutant  emitted by a source,  these  may
be recorded in the last column.   The next to  last column,
ill-effect on health, refers  to  headaches, nose,  lung  or
skin irritations.

                        Number of complaints  per year	


Dry  cleaning

Paint  shop

Print  shop

Electroplating  shop

Gasoline  transfer   |   | j   |


            a a a EH CD c

            CD [Zi iZD IZ) L
 Food  stores,

 Small  chem.  mfg.

  dwelling, school,


Section III   Amount of Air  Pollutant Discharged  From Small Emission Sources

       The  purpose  of  this question  is to determine the problems that might be
encountered  in  the  control of emissions by package sorbent or catalytic com-
bustion devices.   It also supplies information on sizing of the control devices
for  the various  pollutant producing  operations.

       The  following table has  four  column headings under which data may be

       Size of  source  asks for  a  judgment as to the medium size or  size range
for  each  operation  in  terms  used  by  the particular trade.

       Type, odor or chemical formula asks for identification of the pollutant
 in terms  of type of odor, or chemical formula or  class of compounds (halogen,
sulfurated, hydrocarbons) with  some  estimate of molecular weight range.

       Rate of  emission  asks for  data on  amount of pollutant lost to the
 atmosphere  per  unit of time, or per  unit  of material processed or per unit

       Concentration  and/or  rate  of  air exhaust asks for data on pollutant
 concentration with  air exhaust  rate  if data on rate for previous column is  not
 on record.   In  the  case  of  odor,  total air exhaust rate may be the  only
 numerical information  available for  the last  two  columns.  Exhaust  rate should
be identified with  size  of  source.

                                                                    Cone, and/or
                  Size  of       Type,  odor or         Rate of            rate of
 Source            source       chemical  formula       emission         air exhaust
 Dry  cleaning      	      	
 Paint  shop        	      	
 Print  shop
 Food stores,
 Small chemical
 Incinerators ,
  dwel1 ings ,
  schools ,

 Table  C-I  Properties  of Pollutant Vapors Emitted From Small Stationary Sources

2-Ethyl hexyl acetate

Butyl carbitol
Diethyl cyclohexane
Butyl cyclohexane
1 ,8-Octanediol

Hexyl acetate
2-Methyl pentyl acetate
1-Methyl pentyl acetate
CH3 [CH3
Ck Ha
% 9

[(CH,)2CH CH2]2CO





1/\ M
point, LEL,
A ^ ^ ^ ^K
r ppm
_ _ _
421 6,000
A f\ /

345 8,000
302 7,400






er 3>
-^ -a
10 -a
IA to
0 0.
t r>

* Wherever possible the nomenclature approved by the International Union of
  Chemistry was used.

Table C-I  (continued)

2-Ethyl hexanol
Diethyl cyclopentane
Butyl cyclopentane
Butyl benzene
aPinene (turpentine)
Pentyl acetate

Phenyl acetate
Dipropyl sulfide
Difurfuryl ether
Ethyl cyclohexane

Cs H. [CjHs ] .

[(CH,)2C = CH]2CO

3 o 5
[OCH = [CH]2=CCH2]20
CH CO C H [CH3]2
f(CH3 )2CHJ2CO
Table C-I (continued)

Isopropyl benzene
Propyl benzene
Dimethyl cyclohexane
2-Ethoxy-ethyl acetate
1 ,6-Hexanediol
1-Methyl-propyl acetate
Butyl acetate
2-Methyl-propyl acetate
Dllsopropyl ether
Isopropyl propyl ether
2,4-01methy1-l ,3-
2,3-D1methyl-l ,3-penta-

CH,[CH ]3OfCH,]2OH

V ,
ca 246

ca 340
ca 298
ca 156
ca 298


11 ,000




Table C-I (continued)

Dimethyl cyclopentane
Ethyl cyclopentane
P en tachlo roe thane
Ethyl benzene

C13C CHC12




 2-Hydroxy-ethyl propanate

 3-Methyl butanoic acid
 1 ,2-Dichlorobenzene
 Propyl acetate

 Isopropyl acetate



ca 279





__ _



 Table C-I  (continued)
   1,2,2-tn'f luoroethane
 3-Hethyl  butanal
 2-Methyl  butane

 2-Furyl methanol


Butanoic acid
Diethyl ether
1 ,1 ,1-Trichloroethane
Ethyl acetate
[CH3]2CH CO  CH3
[C H 0]CH2OH
C12C3= CC12

[C%H70] CH,

CH3C Cl s



point, LEL,
F ppm










Table C-I (continued)

 Dimethyl  dlsulfide


 [CH3 ],COH
 CH3 [CH2]2CH2OM
 C6 Ms OH

 CH3CH =  CHCH  =  CHj
 [CH3 ]2 S2
 C12C =  C HC1

 [CH3]2CH N02
 [CH3]2CH CH3
      ]20  CHtCH2




ca 109









Table C-I (continued)
Dimethyl sulfoxide
Propanoic acid

Methyl acetate
1 ,3-Butadlene



Carbon dlsulflde

CH2 = CHCH = CH2



_ _ -



Table C-I (continued)
 Di chloromethane
 Ethanoic  acid

 Sulfur dioxide
 Methanoic acid
 Hydrogen  sulfide

 Nitrogen  dioxide
 Hydrogen  chloride
 Carbon monoxide





_ _ _


41 ,000

_ _ _

4,51,55,57,61 ,
1 ,4,5,7,51 ,52,58
1 ,4,7,55,57

                       Appendix  C
Table C-II  Flammabil ity Characteristics of Organic Vapors
in Ai r

Organic vapor
1 ,3-Butadiene
To! uene
Ethyl benzene
Cycl opentane
Dimethyl ether
Diethyl ether

Ethyl propyl ether 88
Heat of
1 ,050
1 ,850
1 ,560
temp. F
_ - _
_ --
- - 

21 ,000
11 ,000
11 ,000
11 ,000

      Table C-II continued
      Organic vapor


      Ethyl acetate      88
      Propyl acetate    102
      Isopropyl acetate 102
  Heat of


temp.  F





                      Appendix D


              Method for Calculating Vm

            Additive Volume Increments  to
            Calculate Vm,  Method  of  LeBas

   Element of factor                    cm'/mol

   Carbon                                14.8
   Hydrogen                               3.7
      Doubly bonded                       '*
      Methyl ester and ethers             9.1
      Ethyl ester and ethers              9.9
      Higher esters and ethers           11.0
      Adds                              12.0
      Joined to S. P. N                   8.3

      Primary amines                     10.5
      Secondary amines                   12.0

   Fluorine                               8.7
   Chlorine                              21.6
   Bromine                               27.0
   Iodine                                ----
   Sulfur                                25.6

      3-membered                        - 6.0
      4-membered                        - 8.5
      5-membered                        -11.5
      6-membered                        -15.0
      Naphthalene                      -30.0
      Anthracene                        -47.5
     To calculate Vm, the number of elements of each type in
the molecule is multiplied by the volume Increment for each
corresponding element.  The totals for each element are added
to get the total molar volume.  If the molecule contains ring
structures, the corresponding volume Increment Is subtracted
from the above total.

                Lower  Explosive  Limits  of Mixtures

           Method  for  calculating  lower explosive limit  (LEL) of
      a  mixture  of combustible organic  compounds, based  on modi-
      fied  Le  Chateliers1  mixture  law.

                                          +  P3/N3

           p-j,  P2,  P3*Pn  "  volume  %  of each flammable vapor.

           NI ,  N2,  N3..Nn    volume  %  of each flammable vapor  1n
                              air  at  the lower explosive limit.

      For the  mixture

                                       LEL of each vapor
                                   P,%     N,%     ppm

           2-Ethoxyethyl  acetate    10.0   1.2    12,000
           Toluene                  36.3   1.4    14,000
           Propanone                53.7   3.0    30,000

           The  calculated LEL  for  the mixture 1s  1.92X  or
      19,200  ppm.

              Van der Waals1  Constant 

     Van der Waals1 a 1s a force of attraction constant
incorporated into the" ideal  gas equation to make it more
generally applicable to nonideal vapors.  The constant b
is a volume factor related to the liquid volume of the

     pv  RT                 Ideal  gas equation

             (v-b)  RT      Van der Waals' equation
                             of state

v is the molar gas volume at pressure p and temperature T.

                         CC14  Test  Method

           The  apparatus  used  to determine  CC14  capacity  1s  shown
      1n  Figure D-l.   The stepwise  procedure  employed  1s  as

           1.   Place  approximately  1/2  gram of sample
               Into  a prewelghed weighing bottle.

           2.   Heat  samples  to 110C  under vacuum,
               <1  mm  Hg  (valve 1  open,  valve  2 closed).

           3.   After  a few hours, remove vacuum, cap
               bottles and place 1n desslcator to cool.

           4.   Weigh  bottles and replace 1n ovens with
               the caps again  ajar.

           5.   Heat to 40C  under vacuum, <1  mm  Hg.

           6.   Freeze CC14 In  reservoir with  cold bath
               of  dry Ice/acetone or  methanol.

           7.   With vacuum still  being  applied,  open
               valve  2 to evacuate the  volume above
               the frozen CC14.

           8.   After  15 minutes,  close  valve  2 and
               remove cold bath,  allowing CC14 to liquefy.

           9.   Apply  wet  1ce bath around CC14 reservoir.

         10.   Close  valve 1 and  disconnect vacuum pump.

         11.   After  0014 temperature equals  0C and
               oven temperature  equals  40C,  open valve
               2 SLOWLY.

         12.   After  3 hours,  close valve 2 and  open valve 1
               to  restore the  oven to atmospheric pressure.

         13.   Recap  bottles and  place  1n desslcator to cool.

         14.   Rewelgh bottles.

         15.   Compute % CC14  activity  as follows:

                 A.    Sample weight - weight  from 4 -
                     weight of  empty bottle

B.  Weight CC14 picked up  weight
    from 14  weight from 4

                     (B) - blank
    % CC14 activity 	rrr	  x TOO

Dewar cold bath
                                                    To vacuum pump
                                                       Heating control
                                                       for vacuum oven
Figure D-l  Static CC14 Test Apparatus

            Activated Carbon Regeneration

     Two computer programs  were written  to  simulate the
desorption of organic compounds from activated  carbon using
heated air and steam as regenerating agents.  These programs
use heat and mass balances  combined  with the  concept of
theoretical  adsorption stages to model  the  bed.   Equation 19
of Chapter 4 is used to calculate the organic compound's
vapor pressure over the carbon bed and  this,  in  turn, used to
calculate the latent heat of desorption.  Equation  19, based
on the Polanye potential  theory of adsorption,  is  reproduced
below with minor variations.

               A  T/[2.3 Vm] In [p0/p]                   (1)

     Correlation of Potential Equation with Amount

     The data of the carbons and three organic  compounds  were
fitted, using least square  error techniques,  to polynomial
expressions  as follows

              In u - a * SA + yA2 + 6A3                   (2)

where u 1s the amount adsorbed, measured 1n cm' 11q/g.

     The values of the constants o, 0, Y and 6  are given  1n
Table D-I for the carbons.

     The predicted value of A, given co, 1s

        A  o + B In u + Y [In "]2 + 6 [In w]s            (3)

     The values of the constants for Equation 3 are given 1n
Table D-II.

     In no case can A be less than zero.  If a  negative value
is obtained, then u> has a larger value than the actual adsorp-
tive capacity of the bed and unbound organic liquid exists.
Under this circumstance the vapor pressure will equal that of
the pure liquid and A is equal to zero.

     Heat of Desorption

     The latent heat of desorption  is equal  to the  heat  of
vaporization of the pure organic  liquid and  its heat  of
wetting on the adsorbent.  The heat  of  wetting can  be related,
using the Clauslus-Clapeyron equation,  to  the vapor pressure

Table D-l  Numerical Values  for the Constants of Equation 2
Table D-
A range
0 0.0125
Values for
the Constants of
-1 .98237E-3
Equation 3
. . -

of the adsorbed organic compound, which in turn can be pre-
dicted using the potential  theory equation (Equation 1).
The additional heat will  be small for small  A values, but
will increase for the larger values.

     The Clausius-Clapeyron equation  has the form

                   In PO -  - jjzf + k                      (4)

where X is the latent heat  of vaporization of the pure
liquid.  When solved for In p, the potential theory Equa-
tion 1 has the form

               In p - In po - 2.3 AVm/T                  (5)

By substitution of Equation 4 into 5, the following equation
1s derived which relates p  to the heat of desorptlon AH,

                           \ 4. O O A V p
                _      .    A T fc J "m1^                  /r\
                In p  k -  	57	                  (v)
                  AH - X + 2.3 AVmR                      (7)
     Since A changes as the amount of organic compound is
desorbed, the heat of desorption also varies.  The amount of
heat absorbed by the system can be found by Integrating the
heat of desorptlon over the amount desorbed.  The equation
for carrying out the integration has the form
                   2.3 VmRWc   ,                           .  .
                   	jj	  / A dw,                     (8)
where   Q      "   Btu  heat  absorbed
        R      -   gas  constant,  3.42  Btu/[lb mol  x  F]
        Wc        ib  carbon  in bed
        u         Ib  organ1c/lb  carbon
        (i>1  U2     Ib  organic/lb  carbon  at  beginning and
                  end  of  stage

      The adsorption  potential A 1s  a function  of to.  For con-
 centrations such  that A  is  large  (small  amount adsorbed) the
 plot  of  log u  versus  A is  a straight line.   For small  values
 of  A,  the  plots  can  be considered  straight  lines on semi-log
 graphs If  the  amount of  adsorbed  organic  compound does not

      vary over a factor of 2.

           The Integrated form of Equation 8  1s as  follows
        /A du  [C8-C9][to)2-ui] + Cg [w2lnu2  - o^lnw]]          (9)
                         C9 ' U

                         C8  Ai-Cg In MI

           A   A2 when u - 012 and A = AI when u  ai].  The  Hm1t  of
      (1  In ui  1s zero when o>i approaches zero.

           The derivation of the Integral form of heat absorption
      equation. Equation 9, Involved the following steps.   Given  a
      straight line potential  plot, the following 1s true

                         A  Cg + Cg In to>                      (10)

      The  Integrals to be evaluated are
                   ^-to)2       x*
                    A  dw   /Cg dw +   /Cg In u> dco
      Integrating the  second expression by parts
            W2        -&2              ~1U)2
           A dto)  Cg to)     + Cg to) In to)     - Cg    / db)
                            -wi 3 * ^9 [to)2lnu2  Wilnu-j]        (12)
          The  constants  Cg and  Cg are evaluated by solving simul-
     taneously  the  following equations

                        Ag  * Cg  + Cg In 2

                        AI   C8  + Cg In MI

     This yields
                         1 n
                    C8 - A-j-Cg In  wj

     Vapor Pressure of Pure Organic Liquid

     The vapor pressure versus temperature  relationships  for
water, trichloroethene, 4-methyl-2-pentanone  and  propanone
were fitted by least mean square techniques to  the  equation
               In p0 - a + 0/T + C In T
The numerical values for these constants are given  1n  Table
Table D-III  Numerical  Values for Constants In Equation 12
  Range C      a        3         C

  25 to 100  50.1243  -6598.49  -4.35917
  25 to 150  10.9567  -3686.53   1.00381
  25 to 150  39.3263  -4774.12  -3.14373
  25 to 125  45.1415  -7059.07  -3.4225
     Logic Diagram for Air-Regeneration

     The programs make use of heat and mass balances and
assume the organic vapors follow the perfect gas laws.  There
1s little information available on heat and mass transfer
coefficients, so a "theoretical plate" type of approach was
used.  The bed 1s divided into 32 theoretical segments and
thermal and mass transfer equilibria are assumed at each
segment.  A given time increment Is chosen, and the gas which
enters during this Interval 1s allowed to come to full equi-
librium with the carbon.  This gas then enters the next seg-

     The logic diagram for air-regeneration is shown  1n
Figure 0-2.  The heat and mass balances are carried out for
each segment with a trial and error search for the amount

                                       For I'th segment
                                       calculate total
                                       moIs of organic
                                       u0> enthaloy of
                                       gas olus carbon.
                                       Set hiah and low
Calculate new bed  temoerature.

Pressure of orqanic vaoor over
carbon  at bed temperature and
                                            Calculate total  organic
Set up gas
                Set  gas
                  Update cumulative
                  suns. Sum remaining
                                                                     of 2

Set low
to ] .
Figure  D-2   Logic  Diagram for  Air-Regeneration

of organic compound remaining adsorbed  on  the bed.   The heat
of desorptlon 1s modeled exactly.

     The following restrictions  and  assumptions  are to be

     1.   The heat capacity of the  organic  vapor  and
         the air are constant with respect to temp-
         erature.  For the temperatures  occurring
         during  stripping, this  1s satisfactory,

     2.   The bed contains  32  equilibrium stages.
         The program can be easily changed to give
         either  fewer or greater stages.   Going  above
         32 stages 1s not  appreciably different; fewer
         stages  will  increase the  amount of regenerant

         The latent heat of the pure compound is con-
         stant with temperature.

     4.  The heat capacity of the organic compound on
         the carbon 1s equal  to the gaseous heat
         capacity.  This is a thermodynamic require-
         ment once assumption No. 3 is made.

     5.  The organic vapor and the air follow the
         perfect gas laws.

     Logic Diagram for Steam-Regeneration

     Figures D-3 and -4 show the logic diagram for the steam
system.  The heat and mass balances are carried out for each
segment, with a trial and error search for water content and
the amount of adsorbed organic compound remaining on the bed.
Again, the heat of desorption of the organic compound is
modeled exactly.

     This program permits several flow regimes to exist simul-

     1.  The segment temperature is below an equilibrium
         temperature.  The entering steam (1f any) con-
         denses on the bed and there is no exit steam.

     2.  The bed Is at an equilibrium temperature.  Steam
         condenses or water evaporates, depending on
         whether the sensible heat of the steam will supply
         the latent heat required by the organic compound.

Set up
bed con-
               Set up gas
Update  time
For I'th  segment
calculate total
nols of organic
uo enthalpy H of
gas and carbon.
                                      Set  gas
                   Update cumulative
                   suns.  Sum remainin
                                        heat  at
                                                  By enthalpy calculate
                                                  new bed  temperature Tj

                                                  Vaoor pressure of
                                                  carbon at bed temperature
                                                  Tj and concentration u^
                                                                 Set  eoulllbrlum  bed
                                                                 temperature to 212F
                                                                               new trial
                                                                          of 2
Figure  D-3    Logic Diagram  for  Steam-Regeneration

                      Trial * error
                      calculation of
                      equllIbrlum bed
                      temperature T;
                                                              to lower
                                 to raise
                                  T,  to
Set high
 to ui
                     all steam
                     enthalpy Ti
                                        thalpy steam
Figure D-4   Logic  Diagram for  Steam-Regeneration,  Figure D-3  continued

           3.   The bed  is  dry and  above the  equilibrium
               temperature.   For this  case,  stripping  pro-
               ceeds  like  the gas  stripping  case.

           The  equilibrium temperature is  the  temperature  at  which
      the  organic  vapor pressure and  the  steam total one atmosphere,

           The  previously  noted  restrictions and  assumptions  still
      apply with  the  following additions:

           1.   The program is written  specifically  for steam
               as  regenerating agent.   Other fluids would
               require  minor  program  changes.

           2.   There  is assumed  to  be  no  interaction between
               the organic compound and the  steam.  In partic-
               ular,  there is no displacement  effect where
               the steam will  "salt out"  the organic.  If
               displacement occurs, the effect would be to
               lower  the steam requirement.

           3.   The latent  heat of adsorbing water on carbon
               1s  small.   The  program  ignores  the latent heat
               of  wetting  and  considers the water vapor pressure
               to  be  equal to  that of  pure fluid.

           4.   The latent  heat of evaporation  of water is
               constant  with  temperature.  This requires the
               heat capacity  of  steam  and water to  be the
               same.  Although  steam has a heat  capacity about
               half of  that of water,  this effect should be
               small  because  the majority of the heat is
               provided  by the latent  heat.

                      Appendix E
                 Sample Calculations

                  Heat of Adsorption
                   (from pape 4-31)

     The following  demonstrates  the procedure  for  calcu-
lating the heat of  adsorption  using the  Clauslus Clayeyron

                     In P  -     + k                       (1)
The data for making the calculations  were  selected  from
Table 4-IX for w  0.30 cm'liq/g  .  When the  pressure  data
for the three temperatures  -9F (250K), 77F (298K)  and
170F (350K) are substituted into  Equation  1,  the  three
following equations are obtained  which  can now  be solved
simultaneously in pairs:

             at -9F, In 2.3  -^  +  k                 (2)

             at 77F, In 38     +  k                 (3)

             at 170F, In 340  -^L- + k                (4)

             0.833  2.000 x 10'J AH  +  k                 (2)

             3.638 - 1.678 x 10'3 AH +  k                 (3)

             5.830 - 1.428 x 10"J AH +  k                 (4)

     By subtracting Equation 2 from 3

      2.805  -0.322 x 10'3 AH; AH  -8,710 cal/mol
                                 k - 18.2

     By subtracting Equation 3 from 4

      2.192 - -0.250 x 10'3 AH; AH - -8,760 cal/mol
                                 k - 18.3

     If a generalized  adsorption curve 1s available for the
carbon  being considered, an estimate of AH can be made with
knowledge of only  the  physical properties of the orqanic
vapor.  See  Potential  Theory Equations, Chapter 4, for dis-

           To  demonstrate  the  procedure  generalized adsorption
      curve of Figure  4-10 (or BPL  II  curve  of  Figure 4-12) was
      selected and  the calculations  carried  out for butane at
      u   0.25 cm3Hq/g.

           From the  curve  at w   0.25  cm'liq/g, A  6.75; or

                      6.75 ' [T/Vm]  log  [p0/p]                (5)

      By substituting  two  values  of  T  and  the corresponding p0 into
      Equation 5, two  p values can  now be  calculated.  The two T
      and  p values  are sufficient data to  calculate AH with Equa-
      tion 1.

           For the  calculations T *  298K  and 350K were selected.
      The  corresponding p0 are 1820  and  7000 mm of Hg.  Vm  96.2*
      cm'/mol.   The  pair of equations  are  as follows

              at 298K,  6.75  [298/96.2]  log [1820/p]        (5)

              at 350K,  6.75  [350/96.2]  log [7000/p]        (7)

                      at 298K p   11.7  mm of Hg

                      at 350K p - 97.7  mm of Hg

           The  above substituted Into  Equation  1 give

                        In  11.7 - -^L_  + k                  (8)
                                 2x298                         '

                        In  97.7  -A2  + k                  (9)
                                 2x350                         '

      The  above  pair of  equations solved simultaneously gives

                        AH   -8,500  cal/mol

                        k  - 16.9 (ave)

               Free Energy of Adsorption

     Calculation of AF for to  0.10 cmsl1q/g  in  Table  4-12
using p0 and f0 in equations below at  100,  150,  200 and
                   AF  -RT In [p0/p]

                   AF  -RT In [f0/p]

Here R  1.98 2 2.0 cal/(C x mol)

     To calculate AF for u  0.1  cm3liq/g in  Table  4-12
the following values for T, p0 and p,  Table E-I  below were
used in Equation 1 above.

Table E-I  Data Used for Calculating AF in Table 4-12
        oF       T
            P0,  atm    mm  of  Hg     atm
     To recalculate AF using fugacity instead of pressure,
only f0 need be considered since at the low vapor pressures
under consideration f  p.  The fugacity f0 for propane
liquid can be obtained approximately by using van der Waals'
Equation of State, Equation 3 below, and Equation 4 based on
Equation 3.
              In f,
In Equation 4, a_ is a function
between molecules and b is the
of the vapor in its compressed
atmospheres and V in liters, a
x atm/mol2
                    of the forces of attraction
                    ultimate van der Waals'  volume
                    state.  When p0 is measured in
                    for propane is 8.66 liter2
and b_ is 0.0844 liters/mol.   These constants are
for many vapors from organic and physical chemistry
 Here R  0.082 liters/(C x mol).

          With appropriate substitutions of p and T from Table
     E-l into Equation 3, the following numerical values were
     calculated for V.

                       JL        V,  1/mol

                       311          1.60
                       339          0.910
                       366          0.450
                       394          0.198

          By substituting the above V and other appropriate
     values for T, a_  and b_ Into  Equation 4 the following numer-
     ical values were calculated for  f0.

                        T         f0, atm

                       311         12.0
                       339         18.8
                       366         26.8
                       394         39.4

          With above  numerical value  of f and other appropriate
     values for T and p substituted Into Equation 2, the AF as
     given in Table E-IIwere calculated.

     Table E-II  AF Based on p0 and f0
                             AF.  cal/mol
                   311       3,760      3,710
                   339       3,970      3,830
                   366       4,250      3,960
                   394       4,480      4,140

                   Adsorption Zone
                   (from page 4-48)

     The adsorption zone length,  L2,  and  the mean adsorptive
capacity uz of the adsorption zone are calculated for the
vapor distribution curve given in Figure  4-15 according to
the following procedure from data derived from the effluent
concentration curve given in Figure 4-14  and additional data
given below.   The adsorption data below and the curve in
Figure 4-14 were obtained from Lockheed (1966) Run Number
119 for Barnebey Cheney AC carbon.  This  carbon is similar
to SXA at A  17.2 and larger.
     Mesh size


Barnebey Cheney AC
28 to 60
43.8C; 317K
5 lb/in2 absolute

0.223 1/min at P  5 lb/1n* abs.
66.8 ppm at P - 5 lb/in2 abs.
0.488 g
0.8 cm
0.502 cm2

2.4 cm
2.98 x 10"2cmjliq/g
     By a summing-up process the amount of butane penetration
can be calculated from the effluent concentration curve in
Figure 4-14.  When the summation is carried from C * 0.01 Ci
to C  Ci, the penetrated amount is equivalent to the amount
adsorbed in the adsorption zone.  The summation as given
below was 1n At increments of 50 min.
          Mean C (C). ppm


                ZAtC - 50 x 245 " 12,200

           To convert  ZAtC  to  butane  weight  penetrated,  1t  1s
      multiplied by
                   ZAtC  x  1.69  x  10"7    0.00206  g

           Weight  of  butane  entering carbon  bed  between t - 600
      and  975  m1n  1s,

             [975-600] x Cf  x 1.69  x 10'7    0.00418 g

           Residual capacity or amount adsorbed  1n time Interval
      t  "  600  to 975  min 1s,

                   0.00418-0.00206 -  0.00212 g

           The total  amount  of  butane  the  0.488  g of carbon can
      adsorb  1s  the amount adsorbed in time  t *  0 to 600.

                600  x Ci  x  1.69  x  TO'7  -   0.00677
                plus residual  capacity     0.00212
                          total  amount     0.00889 g

           The adsorptlve  capacity  of  the  carbon Is,

            0.00889/W  0.0182 g/g; 0.0302  cm3  I1q/g

           The amount adsorbed  1n the  saturated  bed length Ls  1s
      equal to the capacity  of  the  total bed minus the residual
      capacity and amount  adsorbed  1n  the  adsorption zone. I.e.,

            0.00889  - [0.00206 + 0.00212]  0.00471 g

           The saturated bed length 1s then

                          0.00471   L    ,  ,
                     i  . - x -    1.27 cm
                     5    0.0182    W

           The adsorption  zone  length  Is

                     L  .  2.40-1.27   1.13 cm

     The adsorptive capacity of  the  saturated bed 1s

              w  .    . , Q.00738 g/cm1
               5     LAC

     The mean adsorptive capacity of the adsorption zone

                  ~j  :
                             0.00363 g/cnr
     To check the self consistancy of the above numerical
values, they may be substituted Into Equation  4-23.
If correct, t& should be equal to 600 m1n.  Cj substituted
Into Equation 4-23 1s first converted from ppm at reduced
pressure to g/1, as follows:

                    p 97O M I ft  6
     Ci  66.8 ppm                5.06 X 10'5 g/1

               *b '  0.223  5.06 x 10" * ^'l^ ^,,j

               tb   600 min

                Calculation of  Lc  at Different
                          (from  page 4-52)
Using Equation 27 of Chapter 4, given below,  Lc values
                                            L  and the
      were calculated  at different Cb/Ci  using  ws ,  uz,
      effluent concentration  curve in  Figure  4-14.
Ca /<
= 0.
= 0.
- 0.
- 0.
.5 ppm
223 1/min
502 cm2
00738 g/cm
00363 g/cm
13 cm
273 M TO"6
lb/1n2 pres
^ v in-*
      where  P    5  lb/in2,  T  -  317K and  M    58  g/mol.   The  time
      interval  [tb-t-j]  is  measured  on  the  effluent  concentration
      curve  from C   0.01  C^  to  C  Ci.   It  is  related  to Cfc/Ci  as
      given  in  the following  table.




     The  quantity
                                    0.57  cm

The following 1s  a summary of the  calculations  according  to
Equation 4-27.
t [when
WSAC  cn
LC. ci
     To calculate service time, t, when Lc 1s known at any
	-0738 * -502	 [2.40-0.57]
66.5 x 0.223 x 7.6 x 10'7
   600 m1n,
     t [when Cb - Cj]  327 [2.40 + 0.88]  1070 m1n,

             Calculation of L2 and LC  Klotz Equation
                         (from page 4-57}

           The  following Is  a  sample calculation  for 4-methyl-2-
      pentanone,  using  Klotz  Equation 4-28.
              L     ^.' ^  PpUmP]  *    f
              2    T"  L  P  J      [f
 u   -67  ,
 Ji_       log
                   + 2.30  k  Um  log

      The  following conditions are fixed:

           Dp         =   0.35 cm; 4  to  10  mesh
           IT         =   100  ft/min;  Ura is  measured  1n  cm/sec
           p  for  air     1.20 x  10-J  g/cm3at  77F
           y  for  air     1.83 x  10'"  g/(cm  x  sec)  at  77F
                     =   100
           log  C-t/Ck      2
           T           *   77F;  298K
           k           *   0.0035
           e, the  void volume  between carbon  particles, varies  be-
      tween 0.3 to 0.4.  When  e  0.4, a  3.6/Dp.

           U          = Um  [60/30.5] e2/3  1.066 Um

           Un,0'*1      0.97 U0'*1

           In Equation 4-28
              -1'1   0.97 U-%1 x [1.20 x 10-s]9-*1
           yj	[1.83X 10-]'."        " 2-095 U<

          fu I '67
          [j]     ' '284

           log [C^/Cb]  2

       Lz - 0.638 x 2.095 x 0.284 x 2 Dp  Dp -Um  * Dv'0-'1

                           * 0.008 Umx 2

0.80 Dp11Um-*1 (1/DV)-67 + 0.016 U
     When the numerical  values are substituted for DD, Um and
Dv. (Dv - 0.070 cm2/sec)                            H

            L2  7.2 + 1.6  8.8 cm;  3.5 1n.

            Lc  8.8 x 0.77  6.9 cm; 2.7 in.

             Calculations for Adsorbed-Vapor Profile
                   as Presented 1n Figure 4-18
       Conditions for Adsorption of Ternary Solvent Mixture
     Total concentration
        2-Ethoxyethyl acetate

     Solvent properties





            C0 ppm
   2-Ethoxyethyl acetate

Total vapor-air flow, F
Vapor-air velocity, U

Carbon bed area, Ac
Carbon particle size, Dp
Carbon density

Service time, t^
Break concentration, Cj,
Vapor-air temperature, T

  (1)  Molar volume at 25C
                                 - 3,000 ftVmin
                                     110 ft/mi n

                                 - 3,930 in2
                                 - 4 to 10 mesh; 0.35 cm
                                 - 0.50 g/cm3;  0.018 lb/1nj
                                 - 8 hr
                                 - 0.01 C
                                 - 25C;
         Adsorptive capacity at influent concentration


                 A - T/Vm log  [C0/Ci]

                 A  1.83 log  [3400/50]  -  3.36

                 A  2.52 log  [37,000/181]  -  5.81

                 A - 4.03 log  [290,000/268]  12.1
          From  the  BPL  generalized  adsorption  curve  in  Figure  4-19
     the  us  values  given  below  were determined.
                          0.31  cm3liq/g
                          0.23  cm3Hq/g
                          0.074 cms

     To convert above 
           Amount adsorbed In Zones 1 ,  3 and 5
24.2 -
61.2 -
57.0 -
14.8 - 9.4 Ib
9.6 - 51.6 Ib
2.8  54.2 Ib
           Length of Zones  1,  3 and 5

                E     9.4/17.8  0.5 1n    Zone 1
                T    51.6/12.7  4.1 1n    Zone 3
                P    54. 2/  4.1  13.2 1n   Zone 5

           Shape of adsorption zone profile 1s assumed to be Iden-
      tical  to the vapor-phase curve as calculated from log [C-j/C]
      1n  Equation 4*29. In these calculations
                    L    kglog  [u>s/u]
                                   ^w   0.01
                    L    0  when  o>s/u   1.0 and

                    L    L2 when ws/w  100.

                                                             E- 15
       Calculations for Adsorbed-Vapor Profile
             as Presented in Figure 4-19

     Conditions are the same as for adsorbed-vapor profile
presented 1n Figure 4-18 except coadsorptlon 1s taken Into

     Equations for calculating coadsorptlon

                E    CT  x-j CQI [wt]

                T    C2 " X2 CQ2 [ot]

                P    C3  X3 C03
     C-j , Cg and C3 are the equilibrium adsorption vapor

pressures of E, T and P, respectively.  X] , Xj and X3 are
the mole fractions of adsorbed E, T and P.  CQJ 1s the cal-
culated equilibrium vapor pressure of E when the weight of
adsorbed E 1s assumed equal to the sum of the actual adsorbed
weights of E, T and P.  This 1s also true for C02 and CQS.
In these calculations C] , G and C3 are always known for the
saturated carbon zones.  CQI , C02 and CQ3 were estimated
from the generalized adsorption curve for BPL carbon In Fig-
ure 4-12.  By tr1al-and-error, CQ-J , C()2 and CQS are calculated
at different [ot] until XT + X2 + X3 * 1.0.  To simplify cal-
culations, [wt] cm'liq/g was not corrected for differences 1n
liquid densities of E, T and P.  For Zone 1 of Figure 4-19,
the [ut] that gives XT + X2 + X3  1.0 1s 0.32 cm'Hq/g at
A  3.28.

     The following sets of calculations give the steps 1n
determining XT, X2 and X3 and the amount of E, T and P ad-
sorbed per Inch of bed length.

     Adsorption potential equations

                A - [T/VJ log [Co/Of]

            E    A  1.83 log [3,400/C01]

            T    A  2.52 log [37,000/C02]

            P    A  4.03 log [290,000/CQ3]

           C01   C02  and  C03  for  Zone 1
           At A   3.28,  found  by tMal-and-error:
             E     3.28/1.83    log  [3,400/C01]  -  1.79
             T     3.28/2.52    log  [37,000/C02]  1.30
             P     3.28/4.03    log  [290,000/C03]    0.814
                         C01 -      55  ppm
                         Cg2 "   11860  ppm
                         C03 " 44,500  ppm
           XT, X2 and  X3 for Zone  1
                     X1  -  50/55     -  0.910
                     X2   181/1,860   -   0.097
                     X3   268/44,500    0.006
           Pounds of E.  T and  P  adsorbed per  Inch  of bed  length
           1n Zone 1
           Mol fractions converted to  liquid  volume fractions,
                                             L1g vol fraction
           E      0.910 X 163      148              0.925
           T      0.097 x 118       11.4             0.071
           P      0.006 x  74   -     0.44            0.0027
                                 160              0.999
           Liquid volume fraction  converted  to lb/1n.
       CD-   0.018 [I1q vol  fraction] [wtl  [M/Vm] Ac     Ib/ln

 E    0.018 x 0.925 x 0.32 x 0.81  x  3,930      16.9  Ib/in

 T            0.071          0.78           -    1.2  Ib/in

 P            0.0027         0.78           -    0.05 Ib/in

                                      Total    18.1

     Pounds Of T and P adsorbed  per  Inch of bed  length  1n
     Zone 3 at Influent concentrations,  first  approximation
           T    C2    XgCo2  [wt]   *   181  PPm
           P    63    XsCos  [wtl   "   268  PPm
     By tr1al-and-error,  X2 +  X3  -  1*0  when  [wtl  "  0.227  cm*
11q/g at A  5.8.

                   T    XT  -

                   P    X2  -
     Mol fraction converted to liquid volume fraction  and
lb/1n bed length.
     T     0.977 x 118  115

     P     0.025 x  74 -   1.85

 T    us  "  0.018 x 0.983 x 0.227 x 0.78 x 3,930 " 12.3 lb/1n

 P    ws            0.0158          0.78

     Rate of desorption of T and P in Zone 2. second

     The 12.3 Ib/in of T and 0.20 lb/1n of P are partially
desorbed by the adsorption front of E moving at the rate of

                FCi/[us for E], 1n/m1n.
                                      Lio vol  fraction

           3,000  x  1.68  x  10'5/16.9    2.98 x  10"J 1n/m1n

               T    12.3  -  1.2    11.1  Ib/in  desorbed
               P    0.20  -  0.05  0.15  lb/1n desorbed

               T    11.1  x  2.98 x lO'3   3.31 x 10'2 Ib/min
               P    0.15  x  2.98 x TO'3   0.45 x 10"3 Ib/min

           Concentration of  desorbed vapors

               T    3.31  x  10'2/3,000  - 1.10 x 10"5 lb/ft3
               P    0.45  x  10-3/3,000   0.15 x 10'6 lb/fts

               T    [1.10 x 10-5  x  392/92] 10*  47, ppm
               P    [0.15 x 10-  x  392/58] 106 - J[ ppm

           Concentration of  vapors 1n  Zone 3, second approximation

               T    181 + 47   228  ppm
               P    268 +  1   269  ppm

           Pounds T and  P  adsorbed 1n  Zone 3. second approximation
                T    C2  -  X2C02  C"t3   "   228
                P    03  -  X3CQ3  [wt]   "   269  ppm
           By tr1al-and-error,  X2  +  X3 -  1 .0  when  [ut]  0.235 cm*
      I1q/g at A  5.55.
                                       Liq vol fraction

                T    XT    0.975               0.987

                P    X2  -  0.022               0.0138
                          0.997               1.008
       T    os  0.018 x 0.987  x 0.235 x  0.78 x 3,930    12.8  lb/1n
       p    w$ .         0.0138           0.78           0.18  lb/1n

           Rate of desorotlon of T and P  In  Zone  2.  third
       T    2.98 x 10"'  1n/m1n  [12.8 - 1.2]   3.46  x 10'2  Ib/min

  P    2.98 x 10"3  1n/m1n  [0.18  -  0.05]    0.39  x  10*2  Ib/min
     Concentration  of  desorbed vapors
  T    3.46 x 10"2/3,000  1.15  x  10"s  lb/ft3;  49  ppm
  P    0.39 x 10"2/3,000 . 0.13  x  10-'  lb/ft3;  1  ppm
     Ctincentfation  of  vapors  In  Zone  3,  third  approximation
          T    181  + 49   230 ppm
          P    268  +  1   269 ppm
     The third approximation  made  very  little  change  in  vapor
concentration; the  values  for cus on the second  approximation
will be used.
     Pounds of P adsorbed  per inch of bed  length  in  Zone 5
     at influent concentration,  first approximation
     At 3  268 ppm,  A   12.2
        A  - 4.03 log  [290,000/268]  12.2
        s   0.083 or 4.60 lb/1n

           Pounds  of  P  adsorbed  in  Zone  6

           From  Figure  4-17,  Lz  is  2.9  in.  at  U    110  ft/min  and
      V    73  cm'/mol.   From Equation  4-26  and  also Page  4-54,


      ws  is 4.60 Ib/in  and  by the above  equation,  wz is  1.06  Ib/in
      average capacity.   Total amount P  adsorbed 1n Zone 6  is

                       2.9  x  1.06 * 3.10 Ib

           In Zone  2, E  is  being adsorbed while  T  and  P  are desorbed.
      The  changes  taking  place are  as follows;

              E  _  l~ws  increases  from  zero to  16.9 lb/1n
                    j_C-j  decreases  from  50  to  zero  ppm

              T  -  f^s  decreases  from  12.8 to  1.2 lb/1n
                    i C2  increases  from  181 to 230  ppm

              P  _  fws  increases  from  0.18 to  0.20 lb/1n
                     63  increases  from  268 to 269  ppm

          To estimate  the  quantities of E, T  and  P adsorbed  in
      Zone 2,  several simplifying assumptions  were  made.  It  was
      assumed  that  Lz is  determined by the  strongest adsorbed vapor
      in this  case  E, and that Lz is the same  in adsorption of mix-*
      tures as when E is  adsorbed singly.   The second  assumption  1s
      that the factor

                         fill.- -iii-Tl

      from Equation 4-26 applies to vapor E.  The third assumption
      is that  for T and P

          Pounds E adsorbed in Zone 2
          From Figure 4-17, Lz is 3.6 in. at U  110 ft/m1n and
        135 cmVmol.  Using  the first factor above w,  3.9  lb/1n
       erage when us  16.9 lb/1n.  The total amount E adsorbed  1n

Zone 2 is then
                 3.6 x 3.9  14.0 Ib
     Pounds T adsorbed in Zone 2
     Using the second factor above, wz    9.86  lb/1n  average
when b)S  12.0 lb/1n.  The total  amount of  T adsorbed  in
Zone 2 1s then
                 3.6 x 9.86  35.5 Ib
     Pounds P adsorbed In Zone 2
     Using the second factor above, z    0.14  lb/1n  average
when ws  0.18 lb/1n.  The total  amount of  P adsorbed  in
Zone 2 1s then
                 3.6 x 0.14 - 0.50 Ib
     The situation 1s analogous 1n Zone 4.   Lz 1s  determined
by vapor T and is 3.3 in. when U   110  ft/m1n  and  V    106
     Pounds T adsorbed in Zone 4
     wz  2.95 lb/1n average when us    12,8 lb/1n, and total
amount adsorbed 1s
                 3.3 x 2.95 - 9.7 Ib
     Pounds P adsorbed in Zone 4
     wz  3.55 lb/1n average when ws    4.60 lb/1n, and total
amount adsorbed is
                 3.3 x 3.55 - 11.7 Ib
     Profile of adsorbed E
        Total amount adsorbed                       24.2 Ib*
        Amount in Zone 2                            14.0
        Amount in Zone 1                            10.2 Ib
        Length of Zone 1
           [10.2/16.9]         0.60 in
 (1)  See  page E-13

              Length  of  Zone  2       3.6  1n

           Profile  of adsorbed T

              Total amount  adsorbed                       61 9 IK
              Amount  In  Zone  1                                 ID
                 [1.2  x  0.60]
              Amount  1n  Zone  2
              Amount  1n  Zone  4
              Amount  in  Zone  3

              Length  of  Zone  3
                 [15.3/12.8]         1.2 in
              Length of  Zone  4       3.3 1n

          Profile of  adsorbed P

             Total amount adsorbed                       c7   iu
             Amount in Zone  1                            t>/'  lb
                [0.05 x 0.60]                    Ot03
             Amount in Zone 2                    0.50
             Amount in Zone 3
                [0.18 x 1.2]
             Amount in Zone 4
             Amount in Zone 6
             Amount in Zone 5

             Length of Zone 5
                [41.5/4.60]         9.0 in
             Length of Zone 6       2.9 in

         Total  length of carbon  bed

                      Zone            L,  1n



         The  shape of the  adsorption front profile for E 1n Zone
    2, T in Zone 4 and  P in  Zone 6 1s assumed to be identical  tn
    the vapor-phase curve as calculated from log [Ci/C] In

    (1)  See page E-13

Equation 4-28.  In these calculations

                            n b) " 01 s
            L  k log [ws/&>]
                            J w  0.01 u>s,

            L  0 when ws/u  1.0 and

            L - Lz when ws/w  100.

     For the desorptlon profiles of T and P  1n  Zone  2  and  P
1n Zone 4,

                                           -jw   0.99 u>s
L  -k log Cws/(us-w^)] + k log Cws/(ws"w^
                                             U   U>{

Here uj Is the amount adsorbed per Inch of bed  length  In  the
previous zone.  At

                 CD  0.99 ws, L  Lz

              Estimate of Secondary Adsorber Size for Gas
             Regeneration with Secondary Adsorption System

           The following is a sample calculation for estimating
      the size of the secondary adsorber for a carbon-resorb
      system utilizing gas regeneration and recovery of desorbed
      solvent or pollutant by means of a secondary adsorber, page
      6-40.  The bed is to be sized to give a steam-to-solvent
      ratio of 6.0.

      Bases for Calculations

           Primary Adsorber Effluent

306 ft3/min (at 298l
4,000 ppm;  1.02 x 10'
302F; 423K
120 min
           Secondary Adsorber Influent

                Same as  above except  T    298PK
           Secondary  Adsorber

                Mesh  size    
                Density  (d)  
BPL V Type
4 to 10; Dp  0.35

31.2 lb/ft3; 2.6 Ib/in.  per ft2
20 in.
           Bed  length  L  of  20  in.  was  selected  as  the  fixed  bed
      dimension  because  the  preliminary  calculation  showed that
      the  bed diameter would be  near  24  in.   A  20  in.  bed  length
      is a  reasonable  dimension  with  a 24  in. diameter.  Also,
      regeneration  calculations  that  had already been  made apply
      most  accurately  to 20  in.  bed lengths.

           The  equation  used for the  estimate is Equation  4-23,
      reproduced below:

The above equation was modified for this application and
has the form
                   [Ls + 0.23 Lz]
120 m1n, regeneration time
Ls + Lz, 1n.
2.6 lb/1n.  per ft2
bed area, ft2
b)s-h)r, operating capacity, Ib/lb
0.23 Ao>
With the above known values substituted Into the above
equation, It now has the form

             14.1  AcAw[Ls + 0.23 Lz]

To determine Ao for a steam-to-solvent ratio of 6.0, 1t Is
first necessary to determine u>s, using  Equation  4-19
and   Curve V of Figure 4-12.  In Equation 4-19 V  Vm  140
cm'llq/mol, po  9470 ppm and p  4000 ppm.  A  0.8 and from
Curve V, >r that will give a steam-to-solvent ratio of 6.0,
according to the following equation
              Ib steam/lb carbon
At wr  0.17, Ib steam/lb carbon 1s 0.85 which qlves a steam-
to-solvent ratio of 6.0 at Aw  0.14 Ib/lb.  Now

               100 - Ac [Ls + 0.23 L2].

Lz depends on U  F/AC.  From Figure 4-17,Lz can be deter-
mined for various U at Dp  0.35 cm and V  125 cmMlq/mol
for 4-methyl-2-pentanone.  The conditions of the above
equation are satisfied when

                   U   54.5 ft/m1n
                   Lz   2.5 1n.
                   Ac -  5.5 ft2
                   Lc - 17.5 1n.

      The volume of carbon 1s 9.2 ft3 and weight 1s 290 Ib.

           For the 37.5 Ib solvent adsorbed by the 290 Ib, the
      overall operating capacity Is 0.13 Ib/lb compared to 0.14
      Ib/lb where the adsorption 1s more correctly distributed
      over the saturated and adsorption zone bed lengths.

        Calculations  on  Carbon-Resorb  with  Gas
     Regeneration,  Condensation  and  Vapor  Recycle

     These calculations  estimate the number of  cycles  to
attain steady state,  and,  after  the  steady  state  is  attained,
estimate the flow rates  and  concentrations  to  the  on-stream
adsorber and to the condenser and the  weight of vapor  re-
cycled when condenser is operated at 32F.   For purpose  of
calculations refer to page 6-44.

     Fixed Parameters

     Pollutant or solvent vapor  - 4-methyl-2-pentanone
     Carbon type                - BPLV
     Carbon mesh size           - 4  to 10  U.S.  std sieve
     Carbon density             - 31.2 Ib/ft1
     tbi service time            82 hr
     t, regeneration time        120 min
     Ac                          27.2 ft2
     Lt                          21.3 in.

Weight of carbon for each inch of bed depth is 71  lb.   The
regeneration starts on the 80th  hr of service  time and term-
inates on the 82nd.

     First Cycle Adsorption Phase

     F  to adsorber      3,000  ft'/min               .
     Ci to adsorber   -      10  ppm;  2.55  x 10'*  Ib/ft'
     T                      77F;  298K

     First Cycle Regeneration Phase

     F  to  condenser  -    306  ft3/m1n P 77F
     Ce to  condenser     4000  ppm; 1.02 x  10"' Ib/ft

Saturation  vapor concentration  at 32F for 4-methyl-2-
pentanone  is  0.44  x  10"'  Ib/ft1.

      Amount  removed  by  the condenser, at  100*  efficiency,

  Concentration  in  gas stream at 32F      1.11  x  10"3  Ib/ft3
  Saturation concentration at 32F         0.44  x  10*'

  Amount condensed                         0.67  x  10'*  Ib/ft3,

           306 [273/298] x 0.67 x 10'3 - 0.188 Ib/min,
                    120 m1n x 0.188  22.6 Ib.
      Since 37.5 Ib must be condensed, the process has  not come
      to a steady state.
           The total weight recycled is
           306 [273/298] x 120 x 0.44 X TO"3 - 14.7 Ib.
           Second Cycle Adsorption Phase
           The rate of recycle to the Impure  airstream  (during
      regeneration) is
           306 [273/298] x 0.44 x 10'3  0.123 Ib/min.
      The rate of outside vapor input to the impure  airstream
           3000 x 10 x 10'6 [100/392]  0.0076 Ib/min,
      where 392 is the molar volume of air in ft3  at 77F  and  100
      is the molar weight of 4-methyl-2-pentanone.  The  total
      vapor input rate (during regeneration) is
                  0.1230 + 0.0076  0.131  Ib/min,
      The total flow rate to the adsorber  (during  regeneration)
                    3000 + 306 - 3306 ft3/min.
      The Influent concentration (during regeneration)  is
             [0.131/3306] x [392/100] X 10*   155  ppm.
           Second Cycle Regeneration Phase
           It reguires a slightly larger amount  of regenerating
      gas to regenerate the carbon bed on  the second cycle than
      on the first.  The new figure for F  to condenser  is  313  ft3/
      min at 77F.
           The amount desorbed is 37.5 Ib  brought  in from  the  out-
      side and the 14.7 Ib recycled.  The  concentration  Ce to  con-
      denser is

[37.5 + 14.7]  [313  x 120]    1.39  x  10"  lb/ft;  5440  ppm
     Amount removed by the condenser,  at 10035  efficiency,
 Concentration 1n qas stream at 32F      1.52  x  10"s  lb/fts
 Saturation concentration  at 32F        0.44  x  10-J
 Amount condensed                        1.08  x  10"  Ib/ft*,
     313 [273/298]  x 1.08  x  10"s    0.310 Ib/m1n,
              120 m1n x 0.310 - 37.1  1b.
Since 37.5 1b  must be condensed,  the  process has  virtually
come to a steady state.  The total  weight recycled  1s
     313 [273/298]  x 120 x 0.44 x 10"  15.1  lb.
     Third Cycle Adsorption  Phase
     The rate  of recycle to  the  Impure  alrstream (during
regeneration)  1s
     313 (273/298]  x 0.44  x  10"    0.126 Ib/m1n,
The total vapor Input rate (during  regeneration)  1s
            0.126 + 0.0076  0.134  Ib/m1n
The total flow rate to the adsorber (during regeneration)
              3000 + 313 - 3313  ft3/m1n.
The Influent concentration (during  regeneration) 1s
       [0.134/3313] x  [392/100]  x 10* - 158 ppm
     Third Cycle Regeneration Phase
     The required  Increase 1n the amount of regenerating gas
1s nil.  F to condenser  313 ft'/m1n  at 77PF.
     The amount desorbed  1s 37.5 lb brought 1n from  the out-
side and 15.1 lb recycled.  The  concentration Ce to  conden-
ser 1s

      [37.5 + 15.1] [313 x 120] - 1.40 x 10'3 lb/ftj;  5500 ppm.

           Amount removed by the condenser, at 100% efficiency,

       Concentration in gas stream at 32F    1.53 x  10'3 Ib/ft1
       Saturation concentration at 32F     -  0.44 x  10"3

       Amount condensed                       1.09 x  10'3 Ib/ft3

           313 [273/298] x 1.09 x 10'3 - 0.312 Ib/min,

                    120 min x 0.312  37.5 Ib.

           Since 37.5 Ib must be condensed, the  process  is at
      steady state.

           At steady state operation  the following  conditions
      prevail, measured at 77F:

           F  to adsorber      3313  ft3/m1n
           C-|  to adsorber       158  ppm
           F  to condenser  -    313  ftVmln
           Ce to condenser     5500  ppm

           Weight of vapor
             recycled/cycle           15.1  Ib

           Weight of vapor
             condensed/cycle          37.5  Ib

            Calculation of Service Times  for
          Multistage with Displacement of Beds
                    (from page 6-49)

     These calculations are for the T. Melshime type system.
Two organic compounds, 4-methyl-2-pentanone and propanone,
are used to represent the range of pollutants that might be
encountered.  Service times are calculated for  three influ-
ent concentrations, 1,0, 10, and 100 ppm.
     Fixed Parameters

     Carbon type
     Carbon mesh size
     Carbon density
     AC, bed area
     l_t. total bed length
     F, flow
     d, carbon density
     T, adsorption temp.
                              BPL  V
                              4  to 10  U.S.  std  sieve
                              31.2 lb/fts
                              82 ft2
                              16 in.
                              10,000  ft'/min
                              2.6  lb/1n.  per  ft2
     Equation for the Calculations
     A modified form of Equation 4-23 was  used in  the  calcu-
lations.   After the first displacement of  the carbon beds,
                      be given to the amount adsorbed  in the
                      it is always displaced into  the  second
                         has the form
no consideration need
adsorption zone since
stage.  The equation now

or with  fixed  parameter  substituted,

                   t    0.341  [us/Ci]

      Calculated  Results
      The  following  tables  summarize  the  results  of  the
 calculations.   u>s was  determined  for each  Ci  using  mathematical
 term  4-19,  and  the  generalized  adsorption  curve  given  1n  Figure

                              Ci,  Ib/ft'

                   2.55 x 10-7
                   2.55 x 10"*
                   2.55 x l
ws, Ib/lb      t, hr
ws. Ib/lb
t, hr

                      Appendix  F

                   Literature  Cited

Abeles, P.O.  and Gahagen,  H.E.,  "Abscission:  The  Role  of
   Ethylene,  Ethylene Analogues,  Carbon  Monoxide  and
   Oxyqen", Plant Physiol  43,  1253-1258  (1968),

ACGIH, "Threshold Limit Values  of Airborne  Contaminants  and
   Intended Chanqes", American  Conference of  Governmental
   Industrial Hygienists (1970)."

AIHA, "Odor", Air Pollution Manual  Part  I.  American  Indus-
   trial  Hygiene Association,  14125  Prevost,  Detroit,
   Michigan,  Cp. 11, 148,  copyrinht  (1960).

Altshuller.A.P., "An Evaluation of Techniques for the
   Determination of Photochemical Reactivity  of  Orqanic
   Emissions", J. A1r Pollution Control  Assoc.  16,  257

Altshuller.A.P., et.al,, "Products and B1oloq1cal Effects
   from Irradiation of Nitroqen Oxides with Hydrocarbons  or
   Aldehydes Under Dynamic Conditions".  International  Journal
   of Air. Water Pollution, 1_0, 81-98 (1966).

American Document Institute, "ADI Document  2936", 1719 N.
   St., N.W., Washington 6, D.C.

American Industrial Hyqiene Association, Air Pollution
   Manual> Part  I, "Evaluation", Am.  Ind. Hyg, ASSOC.  (1960).

American Industrial Hygiene Association, "Hygienic Guide
   Series",  Am.  Ind. Hyq.  Assoc. J.t  January (1971).

Anderson,  R.D.,  et.al., "Catalytic Oxidation of Methane",
   Ind. &  Enq.  Chem.,  53,  (10), 809 (1961).

Arrhenius, S.,  Theories of Solution.  Yale University Press,
   68  (1912).

ASHRAE, Handbook of  Fundamentals, "Odors", Cp. 12, Am. Soc.
   Heat.  Refrlg. A1r-Cond. Engrs., New  York, 175  (1967).

ASHRAE,  "Odor  Control", Guide  and Data  Book, Applications.
   Cp. 80, 962  (1966  and  1967).

ASTM,  STP-310,  "Properties of  Commerlcally Available  De-
   greasing  Solvents", Am. Soc.  Testing Mater.  - Handbook
   of Vapor  Decreasing (1962).

     Avery,  D.A.  and Boiston,  D.A.,  "The  Recovery  of  Solvents
        from Gaseous Effluents",  The Chemical  Engineer.  January/
        February  (1969).

     Bailor, W.C.,  A1r Pollution  Engineering  Manual FS2.300,
        AP-40,  Washington,  D.C.,  393 (1967).

     Balandln,  A. A., "The  Theory  of  Heterogeneous  Catalytic
        Reactions,   The Multiplet Hypothesis.   A Model for De-
        hydration  Catalysis",  Zeit Phys.  Chem.  132t 289  (1929).

     Barnebey,  H.L., "Removal  of  Exhaust  Odors  from Solvent
        Extraction  Operation  of Activated Charcoal Adsorption",
        J.  Air  Pollution Control  Assoc..  Jj5_ (4), 422  (1965).

     Barr,  A.,  "From Waste  Materials:  Tallow  and High  Protein
            ".  Chem. Eng..  June (1966).
     Barrett,  E.P.,  Joyner,  L.G.  and  Halenda,  P.H.,  "The Deter-
        mination  of  Pore Volume and Area  Distributions  in  Porous
        Substances,  I.  Computations from  Nitrogen  Isotherms",
      ^J.  Am.  Chem. Soc.. 3_.  373  (1951).

     Barth, D.S.,  Romanovsky,  J.C., Knelson, J.J., Altshuller, A. P.
        and Horton,  R.J.M.,  "Discussion", J. Air Pollution Control
        Assoc..  21_ (9),  535  (1971).

     Bay Area  APCD,  "Source  Inventory of  Air Pollutant  Emissions,
        San Francisco Bay Area",  Bay  Area Pollution  Control  Dis-
        trict,  San Francisco,  California  (1969).

     Bay Area  APCD,  "Regulations  for  the  Bay Area",  Air Pollution
        Control  District, Bay  Area  (1971).

     Beckman Instruments, Incorporated, "Hydrocarbon Analyzer",
        Bulletin  P70565-1269-106F (1969).

     Beeck, 0., Advances in  Catalysis.  2, Academic Press,  151

     Beeck, 0.,  "Hydrogenation  Catalysts",  Discussions  of  Fara-
        day Soc.  8.  118  (1950).                               "

     Bendix Process  Instruments  Division. "Total Hydrocarbon
        Analyzer", Bulletin  A530SB-1070 (1970).

     Berg,  C.  and  Bradley, W.E.,  "Hypersorption-New  Fractionating
        Process",  Petroleum  Eng., 18. 115 (1947).

     Bera.  C..  "Hypersorption-A  Process for Separation  of  Light
        Gises", Gas  23  (1),  32  (1947).

     Berq.  D.B. and  Lewandowski,  R.,  "Trichlorofume  Control  Saves
        $780 per  Week",  Chem.  Process., January  (1968).

Berzelius, "On the Progress of Physical  Sciences", report
   to Swedish Academy of Sciences.  Mar.  31  (1835).

Brodski, A.M.. et.al., "Quantitative Characterization of the
   Reliability of Chemical  Reaction Mechanisms". Dold. Akad.
   Nauk SSSR. Proceedings of the Academy of Science of the
   USSR, 174 (4). 865 (1967).

Brunauer, S., Emmett. P.M.  and Teller, E.,  "Adsorption of
   Gases in Multimolecular  Layers", J. Am.  Chem. Soc.. 60.
   309 (1938).                      	"  

Brunauer, S., Emmett, P.H.  and Teller, E.,  "On the Theory
   of Van der Waals Adsorption of Gases", J.  Am. Chem. Soc.
   62, 1723 (1940).                       	~	
Bruno, M.H., "Printing Processes", in Kirk-Othmer Encvclo-
    tedia of Chemical Tecl
    ork, J, 494 (1968).
  v p  * p    (*ii*^j  iw*^*^w^ p  i   ^   ^ _. _v^ _w_  _  *	* * * j 
pedia of Chemical Technology. Interscience Pub., New
Burg, S.P. and Berg, E.A., "Molecular Requirements for the
   Biological Activity of Ethylene", Plant Physio! 42. 144-
   152(1967).                       - --

California, "Emission Inventories", California Air Resources
   Bojrrd, the Resources Agency, November (
Carter, J.W., "Adsorption Processes", Chem. and Process Eng.
   37,, August (1966).                              "

C.E.P., "Fluid Flow Through Packed Columns", Chem. Eng.
   Progr. 48 (2), 89, February (1952).

Committee on Hazardous Materials, "Evaluation of the Hazard
   of Bulk Water Transportation of Industrial Chemicals",
   National Academy of Sciences. Washington, D.C. (1970).

Conn, G.K. and Avery, D.G., Infrared Methods. Academic
   Press (1960).            -

Conrad, E., Interview at Haynie Products, 5010 York Road,
   Baltimore, Maryland (1971).

Cooper, B.J., "Platinum-Carbon Catalysts with Molecular
   Sieve Properties", Platinum Metals Review 14 (4), 133,
   October (1970).

Cooper, J.C. and Cunniff, F.T., "Control of Solvent Emissions'
   Metropo'li tan Engineers Council on Air Resources. New  York,
   N . Y . 11967).

Copley Report, "National Survey of the Odor Problem",
   National Air Pollution Control Administration.  Coolev
   International Corp., Contract CPA Ul>-b9-70  (1970).

     Corrigan, I.E.,  "Catalysis and Absorption",  Chem.  Eng.  61,
        236 (1954).                                          """

     Cowan, J.C., "Soybeans",  1n K1rk-0thmer Encyclopedia  of
        Chemical 'Technology.  Intersclence PUD., hew  Vork,  18,
        599 11969).

     Crocker, W., Hltchock, A.E. and Zimmerman, P.M.,  "Similari-
        ties 1n the  Effects of Ethylene and  the Plant  Auxins",
        Contribution  from the  Boyce Thompson Institute.  7  (3),
     Davis Instrument Division,  "Toxldty Limits  and  Detector
        Methods",  Bulletin 7020763 (1963).

     Davis Instrument Division,  "Flame Ion1zat1on Meter",  Bulletin
        11-65-2 (1965).                                    -

     Davis Instrument Division,  "Davis Hallde  Meter",  Bulletin
        P/N 11-9000 (1971).                           -

     Davlson Chemical Division,  W.R.  Grace  and Co., "Davlson
        Molecular  Sieves", Sales Catalog. 101  North Charles St.,
        Baltimore, Md.

     Debye, P., "Van der Waals  Cohesion Forces",  Physlk. 2.. 21,
        178 (1920).                                -  

     Debye, P., "Molecular Forces and Their Electrical  Interpre-
        tation",  Physlk. Z. (Journal  of Physics), 22,,  302  (1921).

     D'Imperlo, J., "Rubber-Compounding Equipment", A1r  Pollution
        Engineering Manual . Danlelson, J.A., ed., U.s.  uept. of
        Health, Education and Welfare, 999-AP-40, 375  (1967).

     Dmuchovsky,  B., Freerks, M.C. and Zrenty, F., "Metal  Oxide
        Activities In the Oxidation of Ethylene", J.  Catalysis
        , 577  (1965).

     Dobler, L.A., "Solvent Emissions from  Stationary  Sources
        1n Denver", Correspondence from Denver Air Pollution
        Control,  Denver, Colorado (1971).

     Dowden, D.A.  and Reynolds,  P.W., "Heterogeneous  Catalysis:
        Theoretical and  Hydrogenatlon by Binary Alloys", J. Am.
        Chem. Soc.. 242-271 (1950).                           "

     Dowden, D.A.  and Wells, D., Proceedings of the Second Inter.
        national Congress on Catalysis, vol. II.  Paris.  Prance


Dubinin,  M.M,  and  Sayerina,  E.t  "The  Porosity  and  Sorptive
   Properties  of Active  Carbons",  Acta.  Physicochim.  USSR,
   , 647 (1936).                  - -

Dubinin,  M.M.  and  Timofeev,  D.P.,  "Adsorbabi 1 i ty  and  Physico-
   chemical  Properties  of Vapors,  I.  Adsorption of Vapors by
   Active Carbons",  J.  Phys.  Chem.  (U.S.S.R.)  21.  1213
   (1947).                "

Dubinin,  M.M., Zaverina, E.D.  and  Radushkevlch, L.V.,  "Sorp-
   tion and  Structure of Active  Carbons,  I.  Adsorption  of
   Organic  Vapors",  J.  Phys.  Chem.  (U.S.S.R.)  21.  1351  (1947).

Duhl, R.W.,  "Factors Affecting Recovery  of Rotogravure  Sol-
   vents",  Vulcan-Cincinnati,  Incorporated (1970).

Easton, E.R.,  "Food  Products", cp.  6  in  U.S.  Industrial
   Outlook,  1970,  U.S.  Dept.  of  Commerce  (1969).

Eckman, D.P.,  Industrial Instrumentation.  John Wiley  and
   Sons,  Inc.  (                       "
Emmett, P.M. and Brunauer, S., "The Use of Low Temperature
   Van der Waals1 Adsorption Isotherms in Determining  the
   Surface Area of Iron Synthetic Ammonia Catalysts",  J. Am.
   Chem. Soc.. 59_, 1553 (1937).

Environmental Protection Agency, "Federal Register",  Part
   II, 36_ (21), January 30 (1971).

Environmental Protection Agency, "National Primary and
   Secondary Ambient Air Quality Standards", Federal  Regis-
   ter, Environmental Protection Agency. 36 (84), April  30
   (1971T                               "~

Erdco Engineering Corporation, "Tox-Ex", Bulletin 7160 (1970).

Eyring, H., Colburn, C.B. and Zwol1nsk1, B.J., "The Activated
   Complex  in Cheml sorption and Catalysis", Discussions  of
   Faraday  Soc. 8. 39 (1950).

Falckenhagen, H., "Cohesion and Equation of Condition in
   Dipolar  Gases". Physlk. Z. 23. 87  (1922).

Freundlich, H. and Hase, E.,  "The Velocity of Absorption
   Retrogression", Physik. Z. 89, 417  (1915).

Freundlich, H.,  Kapil larchemie, Akademische Verlagsgesel 1-
   schaft m.b.h., Leipzig, 232  (1933).

Gadomski, R.R.,  David,  M.P. and Blahut,  G.A.,  "Evaluation  of
   Emissions  and  Control Technologies  in  the  Graphic  Arts
   Industry", Graphic Arts Technical  Foundation,  U.S. Dept.
   of  Health. Education and Welfare.  No.  CPA-22-69-72 (1970).

     General Electric Company,  "Type H-2,  H-3,  H-4 and  H-5P  Leak
        Detectors and Control  Units", Service Manual.

     General Monitors, Incorporated, "Combustible  Gas Detectors",
        Bulletins on Models  510  and  410  (1970).

     Grant, R.J., Manes,  M.  and  Smith,  S.6.,  "Adsorption  of  Normal
        Paraffins and Sulfur Compounds  on  Activated  Carbons",  J.
        Am. Inst. Chem.  Enq. 8  (3).  403-406  (1962).            

     Grant, R.J.  and Manes,  M.,  "Adsorption  of  Binary Hydrocarbons
        Gas Mixture  on Activated Carbon",  Ind.  Enq.  Chem.  Fund. 5
        490-8 (1966).	~	 -*

     Gregg, S.J.  and Sing,  K.S.W., Adsorption,  Surface  Area, and
        Porosity. Academic  Press, New York (1967).

     Gustafson, P. and Smith,  Jr.,  S.H.,  "Removal  of Organic Con-
        taminants from Air  by Type  13X  Molecular Sieve",  NRL
        Report 5560, U.S. Naval  Research  Laboratory, WashTJig'ton.
        U.C., Aug. (1961).

     Heck, W.W. and  P1res,  E.G., "Growth  of  Plants Fumigated with
        Saturated and Unsaturated Hydrocarbon Gases  and Their
        Derivatives", Texas  Agr1. Exp.  Station  Pub.  603 (1962).

     Heln, G.M.,  "Odor Control  by Catalytic  and High Temperature
        Oxidation",  Annals  of  the New York Academy of Science.
        116. 656  (196TF:	

     Hlndawl, I.J.,  "Air  Pollution  Injury  to  Vegetation",  National
        Air Pollution Control Administration, AP-71  (1971)1	"

     Hlnshelwood, C.N.,  "Homogeneous Catalysis", Trans. Faraday
        Soc. 24.  552 (1928).

     Hlavacek,  V., "Aspects  in Design of  Packed Catalytic  Re-
        actors",  Ind. &  Eng. Chem.,  ^2  (7),  8 (1970).

     Holzbock,  W.G., Instruments for Measurement and Control.
        Relnhold  Publishing  Corporation,  New  York  (1955).

     Horesch, A.J.,  "The  Role of Odors and Vapors  In Allergic
        Disease", Journal of Asthma  Research  4  (2),  (1966).

     Hougen, O.A. and Watson, K.M.,  "Solid Catalysts and  Reaction
        Rates",  Ind.  & Enq.  Chem. 35, 529  (1943).

     Hwang,  S.T.  and Parravano,  G.J., "Transient Response  of
        Chemically Interacting Solid Gas  Systems",  J. Electro-
        chem Soc.. 114, 478  (1967).

IITRI, "Identification of Chemical  Constituents  in  Rendering
   Industry Odor Emissions",  Illinois  Institute  of  Technology
   (Research Institute),  Final  Report,  IITRI  Project  No.
   C8172. by T.  Burgwald  to  Fats  and Proteins  Research  Foun-
   dation, Incorporated,  Des  Plaines,  111.,  D.M.  Doty,  Di-
   rector (1971).

Jacobs, M.B., The Analytical  Chemistry  of  Industrial  Poisons;
   Hazards and Solvents,  Interscience  (1949);  The Analytical
   Toxicology of Industrial  Inorganic  Poisons  (1967);  The"
   Chemical Analysis of Air  Pollutants, Interscience  (1960).

Jaffe, H.H. and Orchln, M.,  Theory and  Application  of Ultra-
   violet Spectroscopy. John Wiley and  bons,  inc.
Jaquet, E,, "The Theory of the Adsorption of Gases",  Physik
   U. Physik. Chem. (Chemical Physics and Physical  Chemistry)
   181 , 437 U9Z5) .

Johnson-Williams Products, Division of American Bosch Arma
   Corporation, "Super-Sensitive Indicator", Bulletin SS-P.
   May (1970).

Jost, W., Diffusion In Solids. Liuids. Bases, Academic Press,
   Inc., New York, N.Y.
Joyner, L.C., Barrett, E.P. and Skold, R., "The Determination
   of Pore Volume and Area Distribution 1n Porous Substances,
   II. Comparison between Nitrogen Isotherm and Mercury
   Porosimeter Methods", J. Am. Chem. Soc. 73, 3155 (1951).

Juhola, A.J. and W11g, E.O., "Pore Structure in Activated
   Charcoal, I. Determination of Micropore Size Distribu-
   tions", J. Am. Chem. Soc. 71. 2069 (1949).

Juhola, A.J., Palumbo, A.J. and Smith, S.B., "A Comparison
   of Pore Size Distributions of Activated Carbons Calculated
   from Nitrogen and  Water Desorption Isotherms", J. Am.
   Chem.  Soc. '74, 61  (1952).

Kearney,  T.J. and Kircher, C.E., "How to  Get the Most  from
   Solvent-Vapor Degreasing, Parts I and  II, Metal Progress.
   93,  May  (1960).

Keesom, W.H.,  "Van  der  Waals Attractive  Force",  Physik.  Z.
   2,  129  (1921).

KHmisch, R.L.,  "Oxidation of  CO and  Hydrocarbons  Over Sup-
    ported Transition  Metal Oxide Catalysts",  First National
    Symposium on  Heterogeneous  Catalysis  for  Control  of Air
   Pollution,  National  Air Pollution  Control  Administration,
    175-197  (1968).

     Klotz, I.M., "The Adsorption Wave",  Chapter  8  of  Summary
        Technical Report of Division  10,  NDRC  (1946).

     Kovats, E., Keulemans. A.I.M.,  "The  Kovats Retention  Index
        System". Analytical Chemistry 16. (8), July (1964).

     Lacasse, N.L., Moroz,  W.J., Handbook of Effects Assessment.
        Vegetation Damage.  National  Air Pollution Control  Admin-
        istration U969),

     Langmuir, I., "A Chemically Active Modification of  Hydrogen",
        J. Am. Chem. Soc..  34_, 1310  (1912).

     Langmuir, I., "The Constitution  and  Fundamental Properties
        of Solids and Liquids", J. Am.  Chem. Soc.t  3, 2221  (1916).

     Langmuir, I., "The Adsorption of Gases  on Plane Surfaces  of
        Glass. M1ca and Platinum", J. Am. Chem. Soc..  40,  1361
        (1918).                                       ~~

     LeBas, G., "The Molecular Volumes  of Liquid  Chemical  Com-
        pounds", Longmans Green, London (1915).

     LeChatelier, H., "Estimation of  Fire Damp and  Inflammability
        Limits", Ann. Mines. J, 380  (1891).

     Leonardos, G., Kendall, D. and  Barnard, N.,  "Odor Threshold
        Determination of 53 Odorant  Chemicals", J.  Air Pollution
        Control Assoc.. J9. (2), 91 (1969).                   "

     Levy, A. and Miller, S.E., "Final  Technical  Report  on the
        Role of Solvents in Photochemical Smog  Formation",
        Scientific Circular No. 79,  National Paint. Varnish  and
        Lacquer Association, Washington,  U.C.  (19/0).        """"""

     Lewis, U.K., GlllHand, E.R., Chertow,  B. and  Cadogen,  W.P.,
        "Pure Gas Adsorption", J. Ind.  and Eng. Chem.  42.  13-19

     Lewis, W.K., Gilliland, F.R., Chertow,  B.  and  Cadogen,  W.P.,
        "Adsorption Equilibria-Hydrocarbon Gas  Mixtures  -  Pure
        Gas Isotherms". Ind. & Eng.  Chem. 42.  1314-1332, July

     Lin, K.H., "Applied Reaction Kinetics", Ind. & Eng. Chem^.
        61 (3), 42 (1969).

     Linde Company, Division of Union Carbide Corporation,
        "Physical Properties of Types 3A, 4A,  5A, 10X  and  13X",
        Sales Catalog F-9947-D. P.O.  Box 44, Tonawand, N.Y.

Lockheed, "Trace  Contaminant  Removal  System  Design",  MSAR
   Subcontract No.  28-5145  with  Lockheed  Missies  and  Space
   Co.  (1970).

London, F., "Theory of Systematics  of Molecular  Forces.  A
   General  Survey of Intermolecular Forces  from  the Stand-
   point of Quantum Mechanics",  Physik." Z.  BIT.  222 (1930).

London, F., "The  General  Theory  of  Molecular Forces",  Trans.
   Faraday  Soc. 33, 8 (1937).

Lorenz, R.  and Lande. H., "Adsorption and Corresponding
   States", Z. anorg. u.  allgem.  chem.  (Journal  of  Inorganic
   and General Chemistry) 125,  47 (1922).

Los Angeles County APCD,  "Rules  and Regulations". Air Pollu-
   tion Control District. County of Los Angeles  (ly/U).

Lunche. R.G., Stein, A.,  Seymour, C.J.  and Weimer,  R.L.,
   "Emission from Organic Solvent Usage in Los  Angeles
   County", J. Air Pollution  Control Assoc.  I,  (4), 275
   (1957).                             "~

MacGraw, M.J., "Air Pollution Emission Factors", U.S.  Dept.
   of Health. Education and Welfare, draft copy (1970).

Mango, F.D. and Schacktschneider, J.H., "Molecular  Orbital
   Symetry Conservation in Transition Metal Catalyzed Trans-
   formations", J. Am. Chem.  Soc. 89. 2484  (1967).

Maryland,  "Air Quality StandardsNational Capitol Inter-
   state Air Quality Control  Region", Maryland State  Depart-
   ment of Health and Mental  Hygiene. Baltimore, Maryland
   October 20  (1970).

Mattia, M.M.,  "Process for Removing Organic Contaminants
   from Air",  U.S. patent 3,455,089  (1969).

Mattia, M.M.,  "Process for Solvent Pollution Control", Chem.
   Eng. Progr. 66, (12), 74 (1970).

MCA, "Laboratory Waste Disposal Manual", Manufacturing Chemists
   Assoc.  (1970).

McGraw,  M.,  Bokhe, K., Fensterstock, J.  and Duggan, G.,
   "St.  Louis  Air  Pollutant Emission Inventory", U.S. Depart-
   ment  of Health, Education and Welfare, Durham, North
   Carolina,  August  (1968).

McGraw,  M. and Duprey, R.L., "Air  Pollutant Emission  Factors",
   Environmental  Protection Agency.  Preliminary  Document, 58

     Mecklenburg,  N.,  "Layer Filtration,  A  Contribution  to  the
        Theory of  Gas  Masks"; Z.  Elektrochemlc  31  (Journal  of
        Electrochemistry),  488 (
     Melnhold,  T.F.  and  Jones,  D.L.,  "Objectionable Odors Snuffed
        Out by  Flameless  Oxidation".  Chem.  Process..  (76), Novem-
        ber 18  (1963).                -___

     Miller,  M.R.  and  Wllhoyte,  H.J.,  "A  Study of Catalyst
        Support Systems  for  Fume-Abatement  of Hydrocarbon Sol-
        vents", J. Air Pollution  Control  Assoc.  17 (121  (1967).

     Miller,  M.R.  and  Sowards,  D.M.,  "Solvent Fume Abatement by
        Ceramic Honeycomb  Catalyst  Systems", Proceedings. First
        National Symposium on Heterogeneous Catalysis for Control
        of. Air  Pollution.  National  Air Pollution Control Arim
National Institute of Dry Cleaning, "Estimation of Solvent
   Vapor Emission from Petroleum Dry Cleaning Plants". Feb.
   v 1 * 7 1 y ,

NFPA, "Hazardess Chemicals Data", National  Fire Protection
   Association. No.  49 (1969).     --

Otake, T.,  Kunigita, E.  and Naleao, K.,  "Analysis of an
   Overall  Gas-Liquid Reaction  Rate Based on Two-Film
   Theory11, Kagaku Kogaku 31  (7), 691  (1967).

Paasche Airbrush Company, "Air-finishing Booths", Catalog
   AFB-865. 1909 Diversey Parkway, Chicago,  II 1 inois (197l ) .

Parker, D.H., Principles of Surface Coating  Technology.
   Interscience Publishers, John Wiley & Sons,  New York,

Perry's Chemical Engineers Handbook. Perry,  R.H., Chilton,
   C.H., and Kirkpatrick, S.D., "Humidification and Drying",
   McGraw-Hill, New  York, Section 15-25  (1963).

Phillip, T., Communications on  assembled costs  of after-
   burners, Shell Development,  Emery,  California (1971).

Polanyi, M., "Adsorption and  Capillarity from Standpoint
   of 2nd Law of Thermodynamics", Verhandl.  deut. physik.
   Ges.t Ji6_, 1012 (1914).         -

Ray, G.C. and Box, Jr.,  E.O., Joint Symposium on Adsorption,
   Am. Chem. Soc. Meeting. Atlantic City, N.J., Sept. 18-23
Ray, G.C. and Box, Jr., E.O., "Adsorption of Gases on Acti-
   vated Charcoal", Ind. & Eng.  Chem. 42. 1332, July (1949).

Ritter, H.L. and Drake, L.C., "Pore Size Distribution in
   Porous Materials, Pressure Porosimeter and Determination
   and Complete Macropore Size Distributions", Ind.  & Enq.
   Chem.. Anal. Ed. 17. 782 (1945).

Robell, A.J., Ballow, E.V. and Borgardt, F.G., "Basic
   Studies of Gas-Solid Interactions III", Lockheed  Missies
   and Space Co.. Report No.  6-75-65-22. April U9b5).

Robell, A.J. and Merrill, R.P.,  "Gaseous Contaminant Removal
   by Adsorption:  II. Adsorption Dynamics in Fixed Beds",
   Chemical Engineering Progress Symposium 65 (96).  100 (1969).

Rogers, T.H., "Natural Rubber",  in Kirk-Othmer Encyclopedia
   9f-Chei!lJ9a^ Technology. Interscience Pub., New York. 17
   660 U968).           "                               

     Romeo,  P.L.  and Warsh,  A,,  Catalytic  Incineration  Design
        Parameters  and  Operating Practices.  Public  Health  gh
        Training  Manual,  Feb.  (1967).

     Rosen,  A. A., Peter,  J.B.  and  Mlddleton,  P.M.,  "Odor Threshold*
        of Mixed  Organic  Chemicals". J.  Water Pollution Control
        Federation.  3  (1),  7  (1962).            ~ -- - - L

     Ruch, W.E.,  Quantitative  Analysis of  Gaseous Pollutants
        Ann  Arbor-Humphrey,  Science Publishers, London  (1970)*.

     Sabatier,  P.,  LaCatalyse  en Chemle  Organlgue.  C. Beranger
        Paris,  France (
     Scott Aviation Daves Products Company, "Vapotesters" . Bulle
        tin Nos.  17031.  17025. November (1969).            '   '

     Sheehey, J.P. and Henderson, J.J., "A Pilot Study of A1r
        t on In Jacksonville, Florida", U.S. Dept. of Health
        tlon and  Welfare. U.S. Dept. of Commerce, Pb IbUUWL
        U963J.                                            *

    Smith, M.F., "A1r Pollutant Emission  1n Greater Cincinnati"
        Division  of A1r Pollution Control, Department of Sewers '
        Cincinnati, Ohio (1970).                         wers,

    Spray Engineering Company,  "Spraco Spray  Nozzles and Indus-
        trial  Finishing Equipment".  Catalog 70. 100 Cambridge St
        Burlington, Massachusetts (1971).   -                   '

    Stein,  K.C.,  et.al., "The Oxidation of Hydrocarbons on  S1mrl
       Oxide  Catalysts", J.  Air Pollution Control  Assoc.  10
       275-281  (1960).    - -- - ^

    Stern,  A.C.,  Air  Pollution.  Vol.  II.  Chapter 23, Academic
       Press,  New York,  325  U965T

    Straus, W.,  "The  Development of  a  Condenser  for  Odor Control
       from Dry  Rendering Plants",  Journal  of  the  Air  Pollution
       Control  Association.  ]  (10), 4U4,  October  114)64)" - ~

    Sullivan,  R.J., "Preliminary Air Pollution Survey  of Odorous
       Compounds  - A  Literature  Review",  National  A1r  Pollution
       Control  Administration.  No.  APTD 6fc'-42  (IM). -     -

    Sundaresan,  B.B.,  Harding,  C.I., May,  F.P. and Hendrickson
       E.P.,  "Adsorption of  Nitrogen Oxides from Waste  Gas",
       Environ.  Sci.  and Techno!. 1  (2),  151,  February  (1967).

    Suter, H.R.,  "Range  of Applicability  of Catalytic  Fume
       Burners".  J. Air  Pollution Control Assoc. 5 (3), 173
       (1955).                                  -

Taylor. H.S. and Halsey, G.f "Contact Catalysis between
   Two World Wars", J. Chem. Phys. 15. 627 (1947).
Temkin, M.I., Problems in Kinetics and Catalysis. Moscow
   \ iy*ty).
Thiele, E.W., "Relation Between Catalytic Activity and Size
   of Particle". Ind. & Eng. Chem. 31. 916 (1938).
Thomson, W.T., "On the Equilibrium of Vapor at a Curved
   Surface of Liquids", Phil. Mag. 158.  448 (1871).
Trappe, W., "Separation of Biological Fats from Natural
   Mixtures by Means of Adsorption Columns",  Biochem.  Z.
   (Journal of Biochemistry) 305. 150 (1940).         '
Turk, A., "Catalytic Reactivation of Activated Carbon  In
   Air Purification Systems", Ind. & Eng. Chem. 47 (1)
   966 (1955).	
United States Public Health Service, Selected  Methods  for
   the Measurement of Air Pollutants, Publication No.  999-
   AP-11  (1965).
University of Michigan, Encyclopedia of  Instrumentation for
   Industrial Hygiene. Institute of Industrial Hygiene (1956).
UOP,  Universal Oil  Products Company, Darien,  Connecticut,
   Handbook on catalytic and thermal incinerators.
U.S.  Department of Commerce. "Graphic Arts",  cp. 5,  U.S.
   Industrial Outlook, 1970 (1969).
U.S.  Department of Commerce, "Food Products",  cp. 6, U.S.
   Industrial Outlook, 1970 (1969).
Vulcan-Cincinnati,  Inc., "Solvent Recovery Installations11,
   Sales literature. Vulcan-Cincinnati,  Inc.,  Cincinnati,
   Ohio (1971).
Walsh, R.T., "Solvent Degreasers", in Chapter 7, Air Pollu-
   tion Engineering Manual. Danielson, J.A.,  ed., Public
   Health Service,  999-AP-40, 383 (1967).
Webb, A.N., "Hydrofluoric Acid  and Acidity of Alumina",
   Ind. & Eng. Chem. 4 (2), 261 (1957),
Weiss, S.M., A1r Pollution Engineering Manual. FS2.300,
   AP-40, Washington, 387 (1967).

     Weisz, P.B. and Prater,  C.D.,  Advances  in  Catalysis.  Vol.
        V^, Academic Press,  New York (1954).

     Wheeler, A., Advances  in Catalysis.  Vol.  III.  Academic
        Press, New York (1951).

     Wheeler, A., Catalysis.  Vol.  II,  Ch.  2,  Relnhold  Pub. Co.,
        New York (1955).

     Wiig,  E.G.  and Juhola,  A.J.,  "The Adsorption of Water Vapor
        on  Activated Charcoal", J.  Am. Chem.  Soc. 71.  561  (1949).

     Wilhelm, R.H., Johnson,  W.C.  and  Acton,  F.S.,  "Conduction,
        Convection and  Heat  Release in Catalytic Converters",
        I rid.  & Eng. Chem.  35  (5),  562  (1943).

     Willard, H.H., Merritt,  L.L.  and  Dean, J.A., Instrumental
        Methods  of Analyses,  D. Van Norstrand Co.,  Inc., PHnce-
        ton,  New Jersey,  3rd  Edition (1958).

     Wonlers, H.C.  and  Jackson, W.E.,  "Air Pollution Emissions
        in  the Delaware Valley  for  1965", The Environmental
        Engineering and Science Program. Drexel Institute of
        Technology, Philadelphia, Pennsylvania  (1968).

     Yang,  K.H.  and Hougen, O.A., "Determination of the Mechanism
        of  Catalyzed Gaseous  Reactions", Chem.  Eng. Proqr. 46.
        146  (1950).	

     Young,  D.M.  and Crowell, H.O.,  Physical Adsorption of Gases.
        Butterworths, London, 371-391  (1962).

     Zimmerman,  P.W., "Anesthetic Properties of Carbon Monoxide
        and Other Gases  in Relation  to  Plants,  Insects and
        Centipedes",  Contribution from  the Boyce Thompson Insti-
        tute.  7  (2),  147-155  (1935)."

                       Appendix G

            Introduction to the Bibliography

     Information sources considered pertinent to the Package
Sorption Device Study are included in the subsequent bibli-
ography,,  Important sources include technical literature,
related government contract or in-house studies, trade pub-
lications and data obtained by letter or interviews.,

     The bibliography has been divided into subsections with
references in each subsection placed in alphabetical order by
author,,  This was done in order to make it easier to locate
cited references and additional collateral information ,  The
sections are titled and self-explanatory

     In the text of the report, the author or principal
investigator is named first followed by the year the article
was publishedo  In the bibliography the major author's name
appears first in the same way it appeared in the text.  This
is followed by additional information needed to clarify and
enable the reader to locate the article.  The year of pub-
lication is enclosed by parenthesis at the end  Often the
author is unknown, an article may be published by a govern-
ment agency, professional or trade society, or in some other
manner,,  In these cases the key word and date of publication
are used in the text,,  Abbreviations are sometimes used,
again followed by the year of publication.  Key words or
abbreviations appear the same way in the bibliography followed
by additional identifying information spelling out the abbrevi-

     The titles of periodicals, serials and non-serial publi-
cations are abbreviated in the bibliography according to the
"American Standard for Periodical Title Abbreviations", and
can be located in Chemical Abstracts,,


                           Appendix G


     Air  Pollutant Emission Sources

     Alkire,  H. L., "Air Pollution in Caroline County Maryland",
        Division of Air Quality Control, Jan. (1960).

     Altshuller, A. P., and Bufalini, J. J., "Photochemical Aspects
        of Air Pollution:  A Review", Environ. Sci. Techno!. 5
        (4)  (1971).                                          ~

     Annual  Survey of Manufacturers, "Paper Coatings", U. S. Dent.
        of Commerce,  (1969).

     Anon, "Know Your Solvents", Screen Process Magazine. 4 (4),
        28 (1967).

     Anon, "Purifiers Remove Formaldehyde Fumes", Air Eng., June

     API  Bulletin  2513, "Evaporation Loss in The Petroleum Indus-
        try  -  Causes  and Control", American Petroleum Institute
        Bulletin 2513.  (1959).

     API  2516, "Evaporation Loss from Low-Pressure Tanks", Ameri-
        can  Petroleum Institute Bulletin 2516, (1962).

     API  Bulletin  2517, "Evaporation Loss from Floating-Roof
        Tanks", American Petroleum Institute Bulletin 2517. (1962).

     ASHRAE,  cp. 31,  "Odor Control", in Guide and Data Book
        Systems. Am.  Soc. Heat. Refrig, Air Cond. Engrs., New York,
          459  (1970).

     ASHRAE,  cp. 12,  "Odors",  in Handbook of Fundamentals. Am.
        Soc.  Heat. Refrig. Air-Cond. Engrs., New York,  173

     ASTM, STP, 310,  "Properties of Commercially Available De-
        greasing Solvents", Am. Soc. Testing Mater.; - Handbook
        of Vapor   Degreasingf  (1962).

     Bailor,  W. C. , "Dry Cleaning Equipment", in cp. 7, Air Pol-
        lution Engineering Manual , Danielson, J. A. ed.,  PuFTTcT"
        Health Service, 999-AP-4U.   393  (1967).

     Barnebey.H. L.,  "Removal  of Exhaust Odors from Solvent Ex-
        traction Operation by  Activated Charcoal Adsorption", J^
        Air  Pollution Control  Assoc., 15  (4) , 422  (1965).

Air Pollutant Emission  Sources  (continued)
Barnhart, R.  F.,  "Rubber Compounding",  in  Kirk-Othmer  Ency-
   clopedia of Chemical
   York, 12,  543  (1968).
clopedia of Chemical  Technology,  Interscience  Pub.,  Mew
  -k, 17,
Barr, A., "From Waste Materials:  Tallow  and  High  Protein
   Meal", Chem. Eng..(1966).

Bell, K.  E., "Leather and Fur Processing",  in  McGraw-Hil1
   Encyclopedia of Science and Technology,  New York,  7_t
   445 (1960).

Benforado, D. M. , Rotella, W. J.  and Norton, D.  L. ,  "Develop-
   ment of an Odor Panel for Evaluation  of  Odor Control  Equip-
   ment", J. Air Pollution Control  Assoc.,  1_9_ (2) ,  101  (1969).

Benjamin, M.', Douglass, I. B. , et al , "A General  Description
   of Commercial Wood Pulping and Bleaching  Process", J.  Air
   Pollution Control Assoc., 1J9 (3) ,  155 (1969).

Berg, D.  B. , and Lewandowski, R. , "Trichloro-Fume Control
   Saves  $780 per Week", Chem. Process,  Jan. (1968).

Blank, D. H., "Rubber and Plastics  Products",  cp. 16, in  U.S.
   Industrial Outlook, 1970, U.S. Dept.  of  Commerce,  (1969).

Brewer, G. L.,  "Odor Control for Kettle  Cooking", J.  Air
   Pollution Control Assoc., ]3_ (4) (1963).

Bruno, M. H., "Printing Processes", in Kirk-Othmer Encyclo-
   pedia of  Chemical Technology. Interscience Pub., New  York,
   16 ,  494  (1968).

Bureau of Census, "Chocolate and Cocoa", U.S.  Dept. of Com-
   merce.  (1971).

Bureau of Census, "Coffee Roasting", U.S. Dept. of Commerce,

Bureau of  Census, "Tobacco Stemming and Redrying", U.S.  Dent.
   of Commerce,  (1967).

Byrd, J.  F., and  Phelps,  A.  H., cp. 23,  "Odor  and its Measure-
   ment",  in Air  Pollution.  Vol. II, Stern, A. C., ed., p  305,
   Academic  Press,  N.  Y.  (1968).

Byrnes,  M.  L.,  "Leather  and  Leather Products", 1n  cp. 11,
   U.S.  Industrial  Outlook.  1970,  U.S.  Dept.  of  Commerce,

      A1r Pollutant  Emission  Sources  (continued)
      California  (A1r  Resources  Board),  "Emission Inventories
         The  Resources  Agency,  (1969).
      Census  of  Manufacturers,  "Adheslves".  U.S. Dept. of Commerce.

      Census  of  Manufacturers,  "Graphic Arts", U.S. Dept. of Com-
         merce.  (1967).

      Census  of  Manufacturers,  "Vegetable 011 Extraction", U.S.
         Dept. of  Commerce.  (1963).

      Chass,  R.L.,  Kanter, C.V.,  and  Elliott, J.H., "Contribu-
         tion of Solvents  to A1r  Pollution and Methods for Control-
         ling their Emissions", J. A1r Pollution Control Assoc..
         13  (2), 64 (1963).                                    '

      Chass,  R.L.,  Holmes, R.G.,  and  Burlln, R.M., "Emissions
         from Underground  Gasoline Storage Tanks", J. A1r Pollution
         Control Assoc.. JI3  (11)  (1963).                      "	

      Chatfleld, H.E.,  "Pipe-Coating  Equipment", cp. 7, In A1r
         Pollution  Engineering  Manual. Danlelson, J.A. ed., U.S.
         Dept. of  Health,  Education and Welfare, 999-AP-40, 390

      Collins, J.W., Webb, A.A.,  et al., "Components of Wood
         Pulp Bleach",  Pulp  Manufacturers Research League Inc..
         Appleton,  Wise. 11968).

      Cooper, J.C., and Cunniff,  F.T., "Control of Solvent Emis-
         sions", Metropolitan Engineers Council on A1r Resources.
         New  York,  N.Y. (1967).

      Copley  Report "National Survey  of the  Odor Problem", National
         A1r  Pollution  Control  Administration, Copley International
         Corp.,  Contract CPA 22-69-50, (1970).

      Cowan,  J.C.,  "Soybeans",  1n Klrk-Othmer Encyclopedia of
         Chemical  Technology. Intersdence Pub., New York, 1^8,
         599  (1969).

      Coward, H.F., and Jones,  G.W.,  "Limits of Flammablllty of
         Gases and  Vapors",  Bulletin  503. Bureau Of Mines. (1952).

Air Pollutant Emission  Sources  (continued)

Crouse, W.  R.,  and Flynn,  N.  E.,  "Summary of  Oraanic  Emissions
   from the Dry Cleaning  Industry",  Bay  Area  Air  Pollution
   Control  District. San  Francisco  3,  Calif.  (1963).

Current, Fishery Statistics,  "Fisheries  of  the  U.S.", No.
   5600, U.S. Dept of Commerce.  (1970).

Dammakoehler, P. E., "Inventory  of  Emission for The  City of
   Chicago", J. Air Pollution Control  Assoc.. Ij5  (3)  (1966).

Denver, "Meeting the Challenge", Denver  Air Pollution Control,
   Feb. (1971).	

Dickinson,  J.,  Chass, R.  L.  and  Hamming, W. J., "Air Contami-
   nants",  cp.  2. Air Pollution  Engineering Manual,  Daniel-
   son, J.  A. ed., Public Health Service,  999-AP-40,   12

D'Imperio,  0.,  "Rubber-Compounding  Equipment",  CD.  7, In
   Air Pollution Engineering Manual. Danielson, J.  A., ed.,
   Kubllc Health Service, 999-AP-40,  375  (1967).

Dobler, L.  A.,  "Solvent Emissions from Stationary Sources
   in Denver",  Correspondence from  Denver  Air Pollution Con-
   trol . Denver, Colorado (1971).

Drury, H. F., "Establishments 1n Commercial Printing", BDSA
   Quarterly Industry Report; Printing and  Publishing, 9_ (2)
   , 15 (1968).

Duhl  , R. W., "Factors Affecting Recovery of Rotogravure
   Solvents", Vulcan-Cincinnati, Inc., (1970).

Dunn, H., et al., "Inks" 1n Kirk-Othmer  Encyclopedia of
   Chemical Technology. Interscience Pub.,  New  YorK,  TT,
   611  (1966).

Easton, E.  R. ,  "Food Products", cp. 6, in  U.S.  Industrial
   Outlook,  1970, U.S. Dept. of Commerce,  (1969).

ESSC,  "Solvent  Emission Control  Laws and the Coatings and
   Solvent Industry", Environmental Science Services Corp.,
   24  Danbury  Rd.,  Wilton, Conn.

Feuss, J. W. and Flowers, F. B., "State of the Art:  Design
   of  Apartment House, Incinerators", J. Air Pollution Con-
   trol Assoc., 19,  (3) (1969).       	

     A1r Pollutant Emission Sources  (continued)

     Gadomski, R. R., David M. P. and Blahut, G. M. , "Evaluation
        of  Emissions and Control Technologies In the Graphic Arts
        Industries", Graphic Arts Technical Foundation, U.S. Dept.
        of  Health. Education and Welfare. No. CPA22-69-7i!, (1970).

     Glllett, T. P., "Chemicals and Allied Products", 1n cp. 15,
        1n  U.S.  Industrial Outlook, 1970, U.S. Dept.  of Commerce.
        (1969).	-

     Glaser, J.  R. and Ledbetter, J. 0., "Air Pollution from
        Sewage Treatment", Paper No. 69-44. Univ. of Texas at
        Austin.  (1968).

     Kama,  G. M., "Corrosion of Combustion Equipment by Chlori-
        nated Hydrocarbon Vapors", A1r Eng.. April  (1965).

     Hama,  G., "How  to Ventilate Coin Operated Dry  Cleaning Equip-
        ment", Air Eno., Dec.,  34 (1961).

     Hangebrauck, R. P., von Lehmden, D. J., Meeher,  J. E.,
        "Sources of  Polynuclear Hydrocarbons in the Atmosphere",
        U.S. Dept. of Health, Education, and Welfare. AP-33, (1967)

     Hardlson, L. C., "Controlling Combustible Emissions", Paint
        and Varnish  Production, July (1967).                    '

     Hardisty, J. T., "Selected Services", cp. 39,  in U.S. Indus-
        trial Outlook, 1970, U.S. Dept.  of Commerce (1969).

     Harris, T.  L. ,  "Chocolate and Cocoa", 1n Kirk-Othmer Ency-
        clopedia of  Chemical Technology, Interscience Pub., New
        York 5_ ,  363 (1969).

     HendMckson, E. R. , Roberson, J. E. and Koogler, J. B.,
        "Control  of Atmospheric Emissions in the Wood Pulping
        Industry", U.S. Dept of Health,  Education and Welfare.
        Final Report^Contract No.  CPA  22-69-18, Vol.  1 of (3)
        Vol's.  (1970).

     Herbert, E.  L., "Beverages", cp. 7  in U.S. Industrial Outlook,
        1970, U.S. Dept. of Commerce. (1969).

     Herbert, E.  L., "Food Products", cp. 6, in U.S.  Industrial
        Outlook,  1970,  U.S.  Dept. of Commerce, (1969).

     Heuser, A.  R. ,  "Trends  in Air Conservation 1n  the  Graohic
        Arts Industry". Gravure. Feb. (1969).

Air Pollutant Emission  Sources  (continued)

lanetti, M.,  "Solvents  for  Letter  Press  and L1thoqraph1c
   Inks", American Ink  Maker,   51,  Oct.  (1969).

IITRI, "Identification  of Chemical  Constituents  in  Rendering
   Industry Odor Emissions",  Illinois  Institute  of  Technology
   [Research  Institute],  Final  Report,  IITRI  Project  No.
   C8172, by  A.  Burgwald  to Fats  and  Proteins  Research
   Foundation, Inc., Des  Plaines,  111.,  D.  M.  Doty, Director,

Oenson, L. B. , Microbiology of Meats, Gerrard Press,  Champaign,
   111.,  (1945).

Jones, G. W.  , "Inflammability of Mixed Gases", Technical
   Paper  450, Bureau of Mines , (1929).

Juve, A., "Rubber Products", in McGraw-Hill Encyclppedia  of
   Science and Technology. New York, 11  . 664 U96QJ.

Kearney,  T. J., and Kircher, C. E. , "How to Get the Most
   from  Solvent-Vapor Decreasing", Parts I and II  Metal
   Progress, April,  87 and May,   93  (1960).

Kempf,  N. W., "Cocoa Powder and Chocolate", in McGraw-Hill
   Encyclopedia of  Science and Technology. New York,  ,
   252  (1960).	

Kennedy,  M.  D. , "Food Products", cp. 6,  in U.S. Industrial
   Outlook,  1970, U.S. Pent, of Commerce,  (1969).

Kerka,  W. E. , and Kaiser,  E. R.,  "An  Evaluation of Environ-
   mental Odors", J. Air Pollution Control Assoc.. 7_ (4;,
   297  (1958) .

Knape,  S. D. ,  "Tobacco",  cp. 8, in U.S.  Industrial Outlook,
    1970, U.S.  Dept. of Commerce,  (1969).

 Kovats, E. and  Keulemans,  A.  I.,  "The  Kovats  Retention Index
    System",  J.  Anal. Chem.  36  (8)  ,  (1964).

 Kuratsune,  M.,  "Polycyclic Aromatic  Hydrocarbons  in  Coffee
    Soots",  Proceedings of  the  American  Association for
    Cancer Research. Los  Angeles,    1-15 U948).

 Lange, N. A., and Forker,  G.  M.,  Handbook  of Chemistry,  10th
    ed., McGraw-Hill,  New York (19677;

     Air  Pollutant  Emission Sources  (continued)

     LeBas, G.,  "The Molecular Volumes of Liquid Chemical Com-
        pounds", Longmans Green, London  (1915).

     LeChatelier, H.,  "Estimation of F1re Damp and Inflammability
        Limits", Ann.  Mines,  1_9,  388  (1891),

     Levy, A., and  Miller, S.E., "Final  Technical Report on the
        Rule  of  Solvents  in Photochemical Smog Formation", Scien-
        tific Circular No. 79. National  Paint. Varnish and Lacquer
        Assoc..  Wash., D.C.,  (1970).	

     Lockhart, E.E., "Coffee  Processing", in McGraw-Hill Encyclo-
          ?edia of Science  and  Technology, Interscience Pub., NevT~
          ork, 3,  , 265  U960).

     Los  Angeles APCD, "Rules  and Regulations", Air Pollution
        Control  District. County of  Los  Angeles. (1970).

     Lowry, H.H., Chemistry of Coal  Utilization. Supp. Vol.,
        Lowry. H.H. ed.,  John  Wiley  and  Sons, Inc., New York

     Lunche,  R.G.,  Stein, A.,  Seymour, C,J. and Weimer, R.L.,1
        "Emission from Organic Solvent Usage in Los Angeles County
        J. A1r Pollution  Control Assoc., 7_ (4)  , 275  (1957),

     MacKnlght,  R.L. and  Williamson, J.E., "General Refuse In-
        cinerators" cp. 8, in Air Pollution Engineering Manual.
        Danielson,  J.A. ed.,  U.S. Oept.  of Health, Education and
        Welfare, 999-AP-40,   419 (1967).

     Maryland, "Air Quality Standards---National Capitol Inter-
        state Air Quality Control Region", Maryland State Dent, of
        Health and  Mental Hygiene, Division of Air Quality Control.
        Baltimore,  Md. (1970).

     McGraw,  M.J. and  Duprey, R.L.,  "Air Pollutant Emission
        Factors", Environmental Protection Agency, Preliminary
        Document,   58, (1971).

     McGraw,  M.J. and  Duprey, R.L.,  "Evaporative Loss Sources",
        in Air Pollutant  Emission Factors, U.S. Environmental
        Protection  Agency,35 (1971).

Air Pollutant Emission Sources  (continued)
McGraw, M.J.  and Duprey,  R.L.,  "Fermentation",  in Air
   Pollutant  Emission Factors,  Preliminary  DocumenTT" U.S.
   Environmental Protection Agency,    74  (1971).
McGraw, M. ,  Bohke, K., Fensterstack,  J.  and Duggan,  G.
   "St. Louis Air Pollutant Emission  Inventory",  U_._S_. Dept.
   of Health, Education and Welfare,  Durham, N.C. (1968).
Mellan, I.,  Handbook of Solvents,  Reinhold  Pub.  Corp., New
   York, Vol. ]_, (1957, 1959).
Mellan, I.,  Sourcebook of Industrial  Solvents.  Reinhold
   Pub. Corp., New York,  Vol's  ,  3. (1957).
Mills, J.L.,  Danielson, J.A. and Smith,  L.K., Control of
   Odors from Inedible Rendering and  Fish Meal  Reduction  in
   Los Angeles County, Presented at  the  60th Annual  Meeting
   of Air Pollution Control Assoc.,  Cleveland,  Ohio  (1967).
Mills, J.L.,  Walsh, R.T., et.al.,  "Quantitative Odor Measure-
   ment", J.  Air Pollution Control  Assoc..  TJ. (10) (1963).
Moores, R.G., "Instant Coffee", in Kirk-Othmer Encyclopedia
   of Chemical Technology, Interscience  Pub., New York,  5_
    757 (1964).
NAPCA 1970,  "Nationwide Inventory of Air Pollutant Emissions'
   National  Air Pollution Control  Administration, AP-73,
Nashkovtsea,  Tabak, 16 (1), 46  (1955).
National Institute of Drycleaning, "Estimation of Solvent
   Vapor Emission from Petroleum Drycleaning Plants", Feb.,
National Dry Cleaners Association, personal  communication,
NFPA,  "Standard for Drycleaning Plants", National Fire Pro-
   tection Assoc., NFPA No. 32, (1970).
Paasche Airbrush  Co., "Airfinlshing Booths", Catalogue AFB,
   Diversey Parkway,  Chicago,  111.,   865,  1909,  (1971).

     Air Pollutant Emission Sources (continued)

     Parker, D.H., "Industrial Finishes", cp. 42, in Principles
        of Surface Coating Technology, Interscience Pub., New
        York (1965).

     Partee, F., "Air Pollution in the Coffee Roasting Industry",
        U.S. Dept. of Health Education and Welfare. 999-AP-9,

     Polglase,  W.L., Dey, H.F. and Walsh, R.T., "Deep  Fat Frying",
        cp. 11, in Air Pollution Engineering Manual. Danielson,
        J.A. ed.. Public Health Service,  999-AP-40,   755 (1967).

     Polglase,  W.L., Dey, H.F. and Walsh, R.T., "Food  Processing
        Equipment", cp.  11, in Air Pollution Engineering  Manual.
        Danielson, J.A.  ed., Public Health Service, 999-AP-40.
        , 750 (1967).                                        *

     Randolph,  A.F., ed., "Cellular Plastics",  cp.  12,  in Plastics
        Engineering Handbook,  Reinhold Pub.  Corp.,  New  York  (i960)".

     Randolph,  V.A., ed., "Vinyl  Dispersions",  cp.  16,  in Plastics
        Engineering Handbook,  Reinhold Pub.  Corp.,  New  York  (I960),

     Rogers,  T.H., "Natural  Rubber", in Kirk-Othner Encyclopedia
        of  Chemical  Technology. Interscience Pub.,  New  York,  17
        ,  660 (1968).                                        

     Rosse, D.G.,  "Chemicals and  Allied Products",  cp.  15, in U.S
        Industrial  Outlook,  1970,  U.S.  Dept.  of Commerce.  (1969).*

     Rowe,  N.R.,  "Controlling  of  Gaseous  Contaminants Through
        Adsorption".  Bulletin  T-101, J. Amer. Assoc. Cont. Contr
        March (1964).	'	~*

     Sallee,  E.D.,  "Four Types  of  Air  Pollution", Metal Decoratinn
        Sept. (1968).                            	*

     San Francisco  Bay Area, "Source Inventory  of A1r Pollutant
        Emissions,  San Francisco  Bay Area",  Bay Area A1r  Pollution
        Control District,  San  Francisco,  Calif.  (l969).

     Scheflen,  J.  and Jacobs,  M.P., The Handbook of Solvents. D.
        Van Nostrand  Co.,  Inc., New York,  (1953).

Air Pollutant Emission Sources  (continued)

Scott Research Lab, Inc.. "Investigation  of  Passenger  Car*
   Refueling Losses",  National  Air Pollution Control Admini-
   stration. Contract  CPA-69-68,  (1970).

Sheeny, J.P, and Henderson, J.J.,  "A Pilot  Study  of  Air
   Pollution in Jacksonville,  Florida",  U.S. Dept. of  Health.
   Education and Welfare, U.S.  Dept. of  Commerce, PB 168888,
   April, (1963).

Shreve, R.N., The  Chemical  Process Industries.  McGraw-Hill,
   New York, (1945).

Slnitsyna, E.L., "Investigations  into Certain Aspects  of the
   Health of People Working in  the Main  Shops of  Tanneries",
   Hygiene Sanitarian. 3 (6)  , 336 (1965).

Skeist, I., ed., Handbook of Adheslves.  Reinhold  Pub.  Corp.,
   New York (1962T1

Skolholt, 0., "Bread Production",  1n McGraw-Hill  Encyclopedia
   of Science and  Technology.  New  York,  5_,    412  (1960).

Smith, M.F., "Air  Pollutant Emissions 1n Greater  Cincinnati",
   Division of Air Pollution Control. Dept.  of  Sewers. C1n-
   cinnati, Ohio (1970).

Solomon, G. and Petrone,  J.P.,  "A  Compilation of  Solvents  for
   Flexographic and Gravure Inks", American  Ink Maker.   51

Spray Engineering  Co., "Spraco  Spray Nozzles and  Industrial
   Finishing Equipment",  Catalogue 70. 100  Cambridge St.,
   Burlington, Mass. (1971).

Stansby, M.E., "Fish and  Shellfish", in  Kirk-Othmer  Encyclo-
   pedia of Chemical Technology.  Interscience Pub.,  New York,
   |f^   560 (1966).

Strauss, W., "The  Development of  a Condenser for  Odor  Control
   from Dry Rendering  Plants",  J.  Air Pollution Control Assoc..
   ] (10)  , 424 (1964).

Strauss, W., "Odor Control  for the Process  Industries", Chem.
   & Proc. Eng., (London),  March  (1965).

      Air  Pollutant  Emission Sources  (continued)

      Sullivan,  J.L.,  Kafka, F.L.  and  Ferrari, L.M., "An Evaluation
         of  Catalytic  and  Direct  Fired Afterburners for Coffee and
         Chicory Roasting  Odors",  J.  Air Pollution Control Assoc.,
         ]  (2), 583 (1965).

      Sullivan,  R.J.,  "Preliminary Air Pollution Survey of Odorous
         Compounds - A Literature  Review", National Air Pollution
         Control Administration.  No.  APTD 66-42 (1969).

      Summer,  W., Methods  of A1r  Deodorization. Elsevier, Amster-
         dam (1963).

      Tl-3 Petroleum Committee,  "Control of Atmospheric Emissions
         from  Petroleum Storage  Tanks", J. Air Pollution Control
         Assoc.. 21  (5),  260 (1971).

      Tl-9 Pulp  and  Paper  Committee,  "A General Description of
         Commercial  Wood  Pulping  and  Bleaching Process", J. Air
         Pollution Control Assoc.. ] (3), 155 (1969).

      Taylor,  R.L.,  "Chemicals  and Allied Products", cp. 15, in
         U.S.  Industrial  Outlook,  1970, U.S.  Dept. of  Commerce.

      Taylor,  O.C.,  "Injury Symptoms  Produced by Oxidant Air
         Pollutants",  Chapter  IV,  in  Handbook of Effects Assess-
         ment. Lacasse, N.L.,  and  Moroz, W.J., ed., Penn State
         Univ.,  (1969).

      Teller,  A.J.,  "Odor  Abatement  in the Rendering and Allied
         Industries",  J.  Air Pollution Control Assoc.,  13^  (4),
         148 (1963).

      Thomas,  G., "Mechanical  Equipment", cp. 7, in Air Pollution
         Engineering Manual . Danlelson, J.A., ed., U.S. Dept. of
         Health, Education and  Welfare, 999-AP-40. 325 (1967).

      Tso, T.C., "Tobacco", in  Kirk-Othmer Encyclopedia of
         Chemical Technology.  Interscience Pub., New York,  20.,
         504 (1969).

      U.S. Department of Commerce. "Graphic  Arts", cp. 5,  U.S.
         Industrial  Outlook,  1970 (1969).

Air Pollutant Emission  Sources  (continued)

Viessman, W., "Gaseous  Air-Pollution-Its  Sources  and  Control",
   Air Eng. ,  July (1968).

Vulcan-C1nc1nnati,  Inc.,  "Solvent  Recovery  Installations",
   Sales Literature,  Vulcan-Cincinnati,  Inc..  Cincinnati,
   Ohio (1971).

Walsh, R.T.,  Leudtke, K.D.,  and Smith,  L.K.,  "Fish  Canneries
   and Fish  Reduction Plants",  cp. 11,  in Air  Pollution
   Engineering Manual ,  Danielson,  J.A.  ed.,  Public  Health
   Service,  999-AP-40,    760 (1967).

Walsh, R.T.,  "Solvent Degreasers", 1n Chapter  7,  Air  Pollution
   Engineering Manual .  Danielson,  J.A., ed.,  Public Health
   Service,  999-AP-40,    383 (1967).

Weast, R.C.  and Selby,  S.M., "Handbook of Chemistry and
   Physics",  48th ed.,  The Chemical Rubber  Co., Cleveland

Weiss, S.M.,  "Surface Coating Operations",  en. 7, Air
   Pollution  Engineering Manual, Danielson, J.A.  ed., Public
   Health Service, 999-AP-40,  : 387 (1967).

Wohlers, H.C. and Jackson, W.E., "A1r Pollutant Emissions
   in the Delaware Valley for  1965", The Environmental
   Engineering and Science Program, Drexel  Institute  of
   Technology. Philadelphia, Pa.  (1968).

Yocum, J.E.  and  Duffee, R.A..  "Controlling Industrial Odors",
   Chen. Eng., June  15 (1970).

Zegel, W., "What's Going Out the  Stack?", Ind. Finish,
   (London)  Dec. (1970).

Zimmerman, O.T.  and  Lavine,  I., "Handbook of Material Trade
   Names", 1953  ed., Suppl.  1,  (1956), Suppl.  2, (1957),
   Suppl. 4,  (1965), Industrial Research Services, Dover,
   New  Hampshire.

     Air Pollutant Effects on Environment
     Abeles, F.D. and Gahagan, H.E., "Abscission: The Role of
        Ethylene, Ethylene Analogues, Carbon Monoxide and
        Oxygen", Plant Physio!.. 43,,  1253 (1968).

     ACGIH, "Threshold Limit Values of Airborne Contaminants and
        Intended Changes", American Conference of Governmental
        Industrial Hygienists. (1970).

     AIHA, "Air Pollution Manual, Part 1 Evaluation", Am. Ind.
        Hyg. Assoc.. Detroit, Mich. (1960).

     AIHA, "Air Pollution Manual, Part II Control Equipment",
        Am. Ind. Hyg. Assoc., Detroit, Mich. (1968).

     AIHA, "Aldehydes", Am. Ind. Hyg. Assoc. J.. 29 (5) .  505
        (1968).                     	  ~

     AIHA, "Phenol and Cresol", Am. Ind. Hyg. Assoc. J.. 30
         ,425  (1969).                       	  

     AIHA, "Sulfur Compounds", Am. Ind.  Hyg. Assoc. J.. 31 (2)
         ,253  (1969).                      	  

     Altshuller, A.P., "An Evaluation of Techniques for the
        Determination of Photochemical Reactivity of Organic
        Emissions", J. Air Pollution Control Assoc.. 16  , 257

     Altshuller, A.P., and Bufalini, J.J., "Photochemical Aspects
        of Air Pollution: A Review", Environ. Sci. Techno!. 5
        (4), (1971).                 	

     Altshuller, A.P., et.al., "Products and Bioloqical Effects
        from Irradiation of Nitrogen Oxides with Hydrocarbons
        or Aldehydes under Dynamic Conditions", International
        Journal of Air. Water Pollution. 1_0  , 81-98 (1966).

     Altshuller, A.P., "Reactivity of Organic Substances in
        Atmospheric Photooxidation Reactions", Public Health
        Service. AP-14. July (1965).           """

     ASHRAE, cp. 31, "Odor Control", in  Guide and Data Book
        Systems, Am. Soc. Heat. Refrig.  Air-Cond. Engrs., New
        York,   459 (1970).

Air Pollutant Effects  on  Environment  (continued)

ASHRAE, cp. 12, "Odors",  in Handbook  of  Fundamentals,  Am.
   Soc. Heat. Refriq.  A1r-Cond.  Engrs.,  New  York,  173  (1967).

Barth, D.S., Romanovsky,  J.C.,  Knelson,  J.J.,  Altshuller,
   A. P. and Horton, R.J.M., "Discussion",  J. Air  Pollution
   Control Assoc.. 2J_ (9),  535  (1971).

Barynin, J., "Measuring Odour Pollution",  New  Scientist.
   Oct. 15 (1970).

Bay Area APCD, "Requlations for the Bay  Area", Air Pollution
   Control District. Bay Area,  (1971).

Benforado, D.M., Rotella, W.J.  and Horton, D.L. ."Develop-
   ment of an Odor Panel  for Evaluation  of Odor  Control
   Equipment", J. Air Pollution Control  Assoc.,  19. (2),  101
  atnik, J.A., "History of Federal Pollution Control
   Leqislation", cp. 2, in industrial Pollution Control
   Handbook, Lund. H.F. ed., McGraw-Hill, New Vork imi).
Brado, I.M., "Transplant Experiments with Corticulous
   Lichens Usinq a New Technique", Ecology,  (4),
Brandt, C.S. and Heck, W.W.. cp. 12, "Effects of Air Pollu-
   tants on Vegetation", in Air Pollution, Vol. 1, Stern,
   A.C. ed., 401, Academic Press, New York (1968).

Brennen, E.G., Leone, I. A. and Dalnes, R.H., "Atmospheric
   Aldehydes Related to Petunia Leaf Damage", Science, 143
   818-819  (1964).

Brown,  E.A. and Colombo, N.J., "The Asthmoqenic Effect of
   Odors, Smells and Fumes", Ann, of Allergy. TJ?  (1954).

Bufalinl, J.J. and  Alfshuller, A. P., "Synergistic Effects
    in  the Photooxidation  of Mixed Hydrocarbons",  Environ.
    Set. Techno!.,  1  (2),  133  (1967).              -

 Burg,  S.P,  and Berg,  E.A.,  "Molecular  Requirements  for  the
    Biological  Activity  of  Ethylene",  Plant  Physio! . ,  42,
    144-152  (1967).                    _  

 Byrd,  J,F.  and  Phelps,  A.M.,  cp.23, "Odor and  Its Measure-
    ment",  in  Air  Pollution Vol .  II, Stern,  A.C. ed.,  305,
    Academic Press,  N.  Y.  (1968).

     Air Pollutant Effects on Environment (continued)

     Byrd, J.F., Mills, H.A., Schellhaser. C.H. and Stokes, H.E.,
        "Solving a Major Odor Problem In a Chemical Process", J.
        A1r Pollution Control Assoc. , ^4 (12) , (1964).        

     California A1r Resources Board, "Emission Inventories", The
        Resources Agency, (1969),                            """""

     Committee on Hazardous Materials, "Evaluation of the Hazard
        of Bulk Water Transportation of Industrial Chemicals",
        Nat. Acad. Scl . . Wash.. D.C. (1970).

     Copley Report, "National Survey of the Odor Problem", National
        Air Pollution Control Administration. Copley International
        Corp., Contract CPA-22-69-50, (1970).

     Crass, F.L. and Ross, R.W., "Single-Pass Dryer Reduces A1r
        and Water Pollution", Food Engineering. (1969).

     Crocker, W., Hitchcock, A.E. and Zimmerman, P.W., "Similar-
        ities in the Effects of Ethylene and the Plant Auxins",
        Contributions from the Boyce Thompson Institute. 7 (3)
        , 231-248                                        ~"
     Dickinson, J., Chass, R.L. and Hamming, W.T., "Air Contamin-
        ants", cp. 2, 1n Air Pollution Engineering Manual. Daniel-
        son, J.A. ed., Public Health Service, 999-AP-40,   12

     EPA, "The Clean A1r Act", Environmental Protection Agency.
      .  Wash., D.C. (1970).

     EPA, "National Primary and Secondary Ambient Air Quality
        Standards", Federal Register, Environmental Protection
        Agency, 36. (84) April 30 (197lT                     ""

     EPA, "Requirements for Preparation and Submittal of Imple-
        mentation Plans", Federal 'Register, Environmental Pro-
        tection Agency, 3^6 (158), Aug. 14 (1971).

     EPA, "Standards of Performance for New Stationary Sources",
        Federal Register, Environmental Protection Agency. 36
        (159), Aug. 17 (1971).

Air Pollutant Effects on Environment (continued)
ESSC, "Solvent Emission Control  Laws  and  the Coatings  and
   Solvents Industry", Environmental  Science Services  Corp.,
   24, Danbury Rd., Wilton, Conn.

Fowler, P.B.S., "Printer's Asthma".  The Lancet.  2,  Oct.

GCA, "Control Techniques for Polycyclic Organic  Matter
   Emissions", GCA Technology Division, Air Pollution  Control
   Administration, (1970).

Gillette, D.G. and Waddell, I.E.,  "Estimate of Air  Pollution
   Damage-Guidelines for Research  and Control  ,  presented  at
   Annual Meeting of Air Pollution Control  Assoc.,  U970).

A Symposium of the British Ecologi
       -  -
Goodman. G.T., et.al., "Ecology and the Industrial  Society"
               of the British Ecological Society. April
Guderian, R.H., et.al., "Problems of the Recognition and
   Evaluation of the Effects of Gaseous Air Impurities on
   Vegetation", Z. Pflanzerk lankh. Pf 1 anzenschutz, 67,
   257-264, U.S. Public Health Service Report A 61-37.

Haagen-Smit. A.J. and Wayne, L.G.. "Atmospheric Reactions
   and Scavenging Processes", cp. 6, in Air Pollution, Vol.
   1, Stern, A.C. ed . , Academic Press, New York (1968) .

Haagen-Smit, A.J., et.al., "Investigations on Injury to
   Plants from  Air Pollution in the Los Angeles Area",
   Plant  Physio!.. 27_  , 18-34 (1952).

Hama, G.M., "Corrosion of Combustion Equipment by  Chlori-
   nated  Hydrocarbon Vapors", Air Eng., April (1965).

Hangebrauck, R.P., von Lehmden, D.J.,  Meeker, J.E.,  "Sources
   of Polynuclear Hydrocarbons  in the  Atmosphere",  U.S.  Dept.
   of Health,  Education and  Welfare, AP-33  (1967).

Hardison, L.C., "Controlling Combustible  Emissions", Paint
    and  Varnish  Production,  July (1967).

Heck, W.W.  and  Pires,  E.G.,  "Growth  of Plants  Fumigated with
    Saturated  and Unsaturated Hydrocarbon  Gases  and their
    Derivatives". Texas Agri. Exp.  Station.  Pub.  603, (1962).

      A1r  Pollutant  Effects  on  Environment  (continued)

      Heck,  W.H.i  "Measurement  of  Photochemical Air  Pollution
        with  a  Sensitive  Monitoring  Plant",  J. Air  Pollution
        Control Assoc.. 2  (2),  (1970).

      Hegnestad, H.E.,  "Diseases  of Crops and Ornamental Plants
        Incited by  Air Pollutants",  Phytopathology. 58, 1089-
        1096  (1968).                                "~

      Hemeon,  W.C.L.,  "Technique  and  Apparatus for Quantitative
        Measurement of Odor Emissions", Air  Pollution Control
        Assoc.. 1_8  (3), 166 (1968).

      Heuss, J.M.  and  Glasson,  U.A.,  "Hydrocarbon Reactivity and
        Eye Irritation",  Environ. Sci. Techno!.. 2  (12), 1109
        (1968).                           	  -

      Hill,  A.C.,  "A Sneclal  Purpose  Plant  Environmental Chamber
        for Air Pollution Studies",  J. Air Pollution Control
        Assoc.. V7  (11),  (1967).     	

      H111 ,  A.C.,  "Inhibition of Apparent Photosynthesis by
        Nitrogen  Oxides", Atmospheric Environment.  4_ (1970).

      Hill,  A.C.,  et.al.,  "Ozone,  Effect on Apparent Photosyn-
        thesis, Rate  of Transpiration and  Stomatal  Closure in
        Plants",  Center for Environmental  Biology,  Univ. of Utah.
        (1969).   	   	'

      Hlndawf, I.J.,  "Air  Pollution Injury  to Vegetation", National
        Air Pollution  Control  Administration, AP-71 , (1971).   	

      Hlndawl, I.J.  and  Altshuller, A.P., "Plant Damaqe Caused by
        Irradiation  of  Aldehydes", Science,  146, 540-552 (1964)..

      Horesch, A.J.,  "The  Role  of  Odors and Vapors in Allergic
        Disease", Journal of Asthma  Research, 4_ (2), (1966).

      Jacobs,  M.B.,  Analytical  Chemistry of Industrial Poisons.
        Hazards arid  Solvents,  Vol. 1, Interscience  Pub., N.Y,

      Jaffe, L.S,, "Effects  of  Photochemical  A1r Pollution on
        Vegetation  with Relation  to  the Air  Quality Requirements",
        J. Air Pollution  Control  Assoc., T7_, 38-42  (1967).

      Jaffe, L.S., "Photochemical  Air Pollutants and Their Effect
        II Adverse  Effects", Archtv. Environ. Health, 1_6, 241


Air Pollutant Effects on Environment (continued)

Kerka, W.E. and Kaiser, F.R., "An Evaluation of Environmental
   Odors". J. Air Pollution Control  Assoc..  21 (4) 297 (1958).

Kniebes, D.V., Chisholm, J.A. and Stubbs,  R.A., "Direct Odor
   Level Measurement Instrumentation",  Gas Age. Aug.  (1969).

Knight, L.I. and Crocker, W., "Toxicity of Smoke", Botan.
   Gaz., 55, 337-371 (1913).                       	

Kuratsune, M., "Polycycllc Aromatic  Hydrocarbons  in Coffee
   Soots", Proceeding of the American Association for Cancer
   Research, Los Angeles, 1-15 (1948).

Lacasse, H.L., et.al., "A Cooperative Extension-Based System
   of Detecting and Evaluating A1r Pollution Damage to
   Vegetation", presented at the 63rd Annual Meeting  of the
   Air Pollution Control Assoc., Paper  No. 70-107 (1970).

Lamphear, F.D., "A1r Pollution Injury to Plants 1n St. Louis",
   63rd Meeting Air Pollution Control Assoc.. Paper No. 70-133,

Leonardos, G., Kendall, D. and Barnard, N..  "Odor Threshold
   Determination of 53 Odorant Chemicals", J. Air Pollution
   Control Assoc.. 1 (2), 91 (1969).

Los Angeles County APCD, "Rules and  Regulations", Air Pollu-
   tion Control District. County of  Los Angeles  (1970).

Levy, A. and Miller, S.E., "Final Technical  Report on the
   Role of Solvents in Photochemical Smog  Formation",
   Scientific Circular No. 79, National Paint. Varnish and
   Lacquer Assoc.. Wash.. D.C. (1970).

MacKniqht, R.L. and Williamson, J.E., "General-Refuse Incin-
   erators", cp. 8, in Air Pollution Engineering  Manual.
   Danielson, J.A. ed., U.S. Dept. of Health, Education and
   Welfare, 999-AP-40, 419 (1967).

Martin, D. and Tikvart, J.A., "A General Atmospheric Diffusion
   Model for Estimating the Effects  of One or More Sources on
   Air Duality", Preprint National Air Pollution Control  Ad-
   ministration (1968).

MCA, "Laboratory Waste Disposal Manual", Manufacturing
   Chemists Association  (1970).          -^.

Mills, J.L., Walsh,  R.T., Leudtke,  K.D. and  Smith, L.T.,
   "Quantitative Odor Measurement",  presented  at  56th  Annual
   Meeting of  the  A1r Pollution  Control Assoc..  June  (1963).


      Air Pollutant  Effects  on  Environment  (continued)

      Moncrlef,  R.W.,  Odor Preferences t  Leonard Hill, London  (1968).

      NAPCA,  "Air Quality Criteria  for Hydrocarbons", National Air
         Pollution Control Administration,  AP-64,(1970^

      NAPCA,  "Air Quality Criteria  for Photochemical Oxidants",
         U.S.  Dept.  of Health.  Education and  Welfare, AP-63 (1970).

      NFPA,  "Hazardous Chemicals  Data".  National  Fire Protection
         Association,  No. 49 (1969).      ---_

      Ottoson, D.G.R., "How  We  Recognize Odours", New Scientist.
         Oct.  15 (1970).

      Patty,  F.A. ed.. Industrial Hygiene and Toxicology. 2nd ed.,
         Vol.  II, Toxicology,  Intersclence  Pub.,  Mew York (1965).

      Peckham, B.W., "Some Aspects  of A1r Pollution: Odors, Visi-
         bility  and  Art", National  Air Pollution  Control Admini-
         stration, (1969).

      Pendray  and Co., "Opinion Survey on Odors and Fumes as  A1r
         Pollution Problems",  New York City,  March  (1955).

      Polglase,  W.L.,  Dey, H.F. and Walsh,  R.T.,  "Deep  Fat  Frying",
         cp.  11, 1n  Air Pollution Engineering Manual. Danielson,
         J.A.  ed., Public Health  Service, 999-AP-40, 755 (1967).

      Philadelphia Dept. of  Public  Health,  "Air Management  Regu-
         lation  V -  Control  of  Organic Substances from  Stationary
         Sources", Air Pollution  Control  Board  (1971).

      Rosen, A.A., Peter, J.B.  and  Middleton, P.M., "Odor Thres-
         holds of Mixed Organic Chemicals", J. Water Pollution
         Control  Federation. 3,4 (1), 7 (1962T!

      Sallee,  E.D.,  "Four Types of  Air Pollution",  Metal Decor-
         ating.  Sept.  (1968).

      Sax,  N.I., Dunn, M.S., et.al., Dangerous Properties of
         Industrial  Materials,  3rd  ed.,  Relnhold  Pub. Co.,  N.Y.


      Schonbeck, H., "A Method  for  Determining  the  Biological
         Effects of  Air Pollution by Transplanted Lichens", Staub-
         Relnholt. Luft, 2S>, No.  1  (1969).

      Seldman, G., et.al.,  "Environmental Conditions Affecting the
         Use  of  Plants as  Indicators of  Air Pollution", J.  A1r
         Pollution Control Assoc.,  J5,  (4),  (1965).

Air Pollutant Effects on Environment (continued)

Shepherd, M., Rock, S.M., Howard,  R.,  and Stormes,  J.,
   "Isolation, Identification and  Estimation of Gaseous
   Pollutants of Air", J. Anal.  Chem.. 23 (10)  ,  1431  (1951).

Sim, V.M., "Effect of Possible Smog Irritants on  Human
   Subjects", J. Am. Med. Assoc.,  Dec. (1957).

Sinitsyna, E.L., "Investigations into  Certain Aspects  of the
   Health of People Working in the Main Shops of  Tanneries",
   Hyg. Sanitation. 30 (6) ,  336 (1965).

Snow, G.F., et.al., "Procedures  Used to Estimate  the Economic
   Impact of Air Pollution on Vegetation  in California",
   Rent. 70-06 California Dept.  of Agriculture, presented at
   63rd Annual Meeting of the Air  Pollution Control  Assoc..

Stephens, E.R. and Burlson, F.R.,  "Analysis of the  Atmosphere
   for Light Hydrocarbons". J. Air Pollution Control Assoc..
   !_ , 147-153, (1967).

Stephens, E.R., Darley, E.F.  and Burlson, F.R., "Sources and
   Reactivity of Light Hydrocarbons in Ambient Air", Proc.
   Div. Refining. American Petroleum Institute, 47.  466-483

Strauss, W. , "Odor Control for the Process Industries",
   Chem. & Process Eng. (London),  March (1965).

Strauss, W., "The Development of a Condenser for Odor Control
   from Dry Rendering Plants", J.  Air Pollution Control  Assoc
   1_4_ (10) , 424 (1964).       	"~^~

Sullivan, R.J., "Preliminary Air Pollution Survey of Odorous
   Compounds-A Literature Review", National Air Pollution
   Control Administration, No. APTD 66-42, (1969).

Taylor, O.C., "Effects of Oxidant Air Pollutants",  J,
   Occupational Med., 10  (9), (1968).

Taylor, O.C., "Importance of  Peroxy-acetyl Nitrate  (PAN) as
   a Phytotoxic Air  Pollutant", J. Air Pollution Control
   Assoc.. 19 , 347-351  (1969).

      A1r  Pollutant  Effects on Environment  (continued)

      Taylor,  O.C.,  "Injury Symptoms Produced by Oxldant A1r
        Pollutants",  cp.  IV, 1n Handbook of Effects Assessment,
        Lacasse,  N.L.,  and Noroz, W.J., ed., Penn State Univ.

      Teller,  A.J.,  "Odor  Abatement 1n  the  Rendering and Allied
        Industries",  J. A1r Pollution  Control Assoc., 13 (4),
        148  (1963).

      Thompson,  C.R.,  "Effects of A1r Pollutants on Growth, Leaf
        Drop,  Fruit Drop, and Yield of Citrus Trees", Environ.
        Sc1.  Techno!..  3_  (1969).

      Thompson,  C.R.,  "Effects of A1r Pollutants on Lemons and
        Navel  Oranges", California Agriculture, 22^ (9), (1968).

      Thompson,  O.C.,  et.al., "Effects  of Continuous Exposure of
        Navel  Oranges to  Nitrogen Dioxide", Atmospheric Environ-
        ment.   (1970).

      TR-1134,  "Hazard Nuisance Pollution Problems Due to Odors-
        Symposium IM, Japan Society for A1r Pollution. 8th Annual
        Meeting.    (1), 11-59 (1967).

      Tremalne,  B.K.,  "All About Odors  - Detection, Analysis,
        Control", A1r Eng.. Oct. (1959).

      Treshow,  M., "Evaluation of Vegetation Injury as an Air
        Pollution Criterion", J. Air Pollution Control Assoc..
        J_5 (6),  (1965).

      Turk, A.,  "Industrial Odor Control",  Chemical Engineering
        Deskbook  Issue. April 27 (1970).

      Van  Sandt,  W.A., Graul, R.J. and  Roberts, W.J.,  "The
        Determination of  Acroleln in the Presence of  Other
        Aldehydes", presented at the 15th  Annual Meeting of  the
        American  Industrial Hygiene Association. Chicago, 111.

Air Pollutant Effects  on Environment  (continued)

Vlessman, VI., "Gaseous Air Pollution  -  Its  Sources  and
   Control". Air Eng., July (1968),

Walsh, R.T., Leudtke,  K.D. and  Smith, L.K.,  "Fish  Canneries
   and Fish Reduction  Plants",  cp.  11,  in  Air  Pollution
   Engineering Manual, Danielson,  J.A.  ed.,  Public  Health
   Service, 999-AP-40,   760 (1967).

Weidensaul, T.C. and Lacasse, N.L.,  "Results of  the State-
   wide Survey of Air  Pollution Damage  to  Vegetation",
   63rd Annual Meeting of Air Pollution Control  Assoc..
i ng
   Paper No. 70-108, (1970).

Wilby, F.V., "Variation in Recognition Odor Threshold of
   a Panel", J. Air Pollution Control  Assoc., 19 (2)  ,  96
   (1969).   	  ~~~

Wohlers, H.C., et.al., "Recommended Procedures for Measuring
   Odorous Contaminants in the Field", J.  Air Pollution Con-
   trol Assoc.. V7 (9) ,  609 (1967).   	'

Yocum, J.E. and McCaldin, R.O., cp. 15, 'Effects of Air
   Pollution on Materials and the Economy", 1n Air Pollution,
   Vol. 1, Stern, A.C., ed.,   617, Academic Press, New
   York T1968).

Zimmerman, P.W., "Anaesthetic Properties of Carbon Monoxide
   and Other Gases in Relation to Plants,  Insects  and Centi-
   pedes", Contributions  from the Boyce Thompson Institute,
   7_ (2),   147-155 (1935).


     AIHA and ACGIH, "Table of Sorbents for Contaminants Listed
         in ACGIH 1970 Threshold Limit Values". American Indus-
         trial Hygiene Association. June (1971).

     ALCOA Product Data, "Activated and Catalytic Aluminas",
         Aluminum Company of America, July 14 (1969).

     American Document Institute, "ADI Document 2936", 1719 N.
         Street, N.W., Mash. 6, D.C.

     Antonson, C.R.  and Dranoff, J.S., "Nonlinear Equilibrium
         and Particle Shape Effects 1n Intrapartlcle Diffusion
         Controlled Adsorption", Chemical Engineering Progress
         Symposium, 65. (96) ,  20 (1969).

     API Bulletin 2513, "Evaporative Loss in the Petroleum
         Industry-Causes and Control", American Petroleum Insti-
         tute. Bulletin 2513 (1959).

     Arrhenius, S.,  Theories of Solution. Yale Univ. Press (1912).

     ASHRAE, "Guide  and Databook Systems", Chapter 31, in
         American Society of Heating, Refrigerating and Air-
         Conditioning Engineers. Inc.. New Vork U970).

     ASHRAE, "Handbook of Fundamentals", Chapter 12, 1n American
         Society of Heating. Refrigerating and Air-Conditioninq
         Engineers. Inc., New York (1967).

     Avery, D.A. and Boiston, D.A., "The Recovery of Solvents
         from Gaseous Effluents", The Chemical Engineer. Jan/Feb

     Bailor, W.C., "Dry Cleaning Equipment", in Chapter 7, Air
         Pollution Engineering Manual, Danlelson, J.A. ed.,
         Public Health Service, 999-AP-40 ,  393 (1967).

     Barnebey, H.L., "Activated Carbon for Air Purification",
         Heating, Piping and Air Conditioning, March  (1958).

     Barnebey Cheney, T-108, "Air Purification and Recovery with
         Activated Charcoal", Bulletin T-108, Sept.  (1963).

     Barnebey Cheney, T-196, "A1r Recovery by Charcoal Adsorption",
         Barnebey Cheney Technical Bulletin T-196 (1958).

Sorptlon (continued)

Barnebey,  H.L.,  "Removal  of Exhaust  Odors  from Solvent
   Extraction Operation by Activated Charcoal  Adsorption",
   J.  Air  Pollution Control Assoc..  1^ (4)  ,  422  (1965).

Barrer, R.M., McKenzie, N. and  Reay, J.S.,  "Capillary Con-
   densation in  Single Pores",  J.  Colloid  Sc1..  H. ,  479

Barrett, E.P. Joyner, L.G. and  Halenda,  P.M.,  "The Deter-
   mination of Pore Volume and  Area  Distributions  in  Porous
   Substances",  J. Am. Ghent. Soc., 73,,  373 (1951).

Beeck, 0., "Hydrogenation Catalysts", Discussions  Faraday
   Soc..  , 118 (1950).

Bohart, G.S. and Adams, E.Q., "Some  Aspects of the Behavior
   of Charcoal with Respect to  Chlorine",  J.  Am.  Chem. Soc.
   4  , 523 (1920).

Brooman, D.L. and Edgerly, E.,  "Concentration and  Recovery
   of Atmospheric Pollutants Using Activated Carbon", J.  Air
   Pollution Control  Assoc.t 1_6 (1)  , 25 (1966).

Brown, R.A. and  Foster, Jr.,'A.G., "The Sorption of Amines
   by Silica Gels". J. Phys. Chem..  5_6 , 733 (1952).

Brunauer, S., Emmett, P.A. and  Teller, E.t "Adsorption of
   Gases in Multimolecular Layers",  J. Am. Chem.  Soc. , 60
   , 309 (1938).

Brunauer, S., The Adsorption .of Gases and  Vapors,  Princeton
   Univ. PressT27  (1945).

Brunauer, S., "New Approaches to Pore Structure Analysis",
   Chemical Engineering Progress Symposium. 6_5_ (96) , 1

Brunauer, S., Deminq,  U,  Deminq, W. and Teller,  E.,  "On
    the  Theory of  van  der  Waals Adsorption  of  Gases",  J.  Am.
    Chem.  Soc.,  62_ ,1723  (1940).

Brunauer,  S., Emmett,  P.H.,  "The  Use  of Low Temperature
    van der  Waals  Adsorption  Isotherms  in  Determininq  the
    Surface  Areas  of  Various  Adsorbents",  J.  Am. Chem. Soc.,
    59   , 2682  (1937).

      Sorption  (continued)

      Buchwald, H., "Activated Silica Gel as an Adsorbent for
        Atmospheric Contaminants", presented at 26th Meeting of
        American  Industrial Hygiene Assoc.. May (1965).

      Burnett,  R.W. and Turnbull , W.T., "Industrial Applications
        of Molecular Sieves", British Chemical Engineering. 11
        (4), (1966).                                        ~~

      Carter, J.W., "Adsorption  Processes", Chem. and Process Enq.
        . 37,  Aug. (1966).                 '	a-

      C.E.P., "Fluid Flow Through Packed Columns", Chem. Enq.
        Progr., 48. (2) , 89, Feb. (1952).

      Chass, R.L., Kanter, C.V.  and Elliott, J. H., "Contributions
        of Solvents to Air Pollution and Methods for Controlling
        Their  Emissions", J. Air Pollution Control Assoc., 13
        (2) ,  64  (1963).      ~                     ~"~  ~"

      Cohan, L.H., "Sorption Hysteresis and the Vapor Pressure of
        Concave Surfaces", J. Am. Chem. Soc.. 60 . 433 (1938).

      Collins,  J.J., "The LUB/Equi1ibrium Section Concept for
        Fixed  Bed Adsorption",  59th Annual  Meeting of American
        Institute of Chemical EFglneers, Detroit, Dec* U966).

      Cooper, B.J., "Platinum-Carbon Catalysts with Molecular
        Sieve  Properties", Platinum Metals  Review, 14 (4) , 133

      Cooper, D.E., Griswold, H.E., Lewis, R.M. and Stokeld, R.W.,
        "Improved Desorption Route to Normal  Paraffins", Chem.
        Eng. Progr.. 62_ (4) ,  69 (1966).                 	

     Cooper, J.C. and Cunniff, F.C., "Control of Solvent
        Emissions", Metropolitan Engineers  Council on Air Re-
        sources,  New York (1967).

     CoughUn,  R.W. and Ezra, F.S., "Role of Surface Acidity 1n
        the Adsorption of Organic Pollutants on the Surface of
        Carbon",  Environ. Sci. Techno!.. 2_ (4), (1968).

     Coward, H.F. and Jones, G.W., "Limits  of FlammablHty of
        Gases and Vapors", Bulletin 503, Bureau of Mines. U.S.
        GOV't.  Printing Office  (1952).

Sorptlon (continued)

Davison Chemical  Division,  W.R.  Grace &  Co.,  "Davison Mole-
   cular Sieves", Sales  Catalogs.  101 North  Charles  St.,
   Baltimore,  Md.

de Boer, J.H.  and Custers,  J.F.H.. "Regardinq the Nature  of
   Adsorption", Z. Phvsik.  Chem. (Journal  of  Physical Chem-
   istry) B 25 ,  225  (1934).

Debye, P., "Molecular Forces  and their Electrical Inter-
   pretation", Physik.  Z. (Journal of Physics),  2_2_ ,  302

Debye, P.. "van der Waals1  Cohesion Forces",  Physik  Z.,  21_
   , 178 (1920).

Dubinin, M.M.  and Timofeev, D. P., "Adsorbability and
   Physicochemical Properties of Vapors I. Adsorption of
   Vapors by Active Carbons", J. Phys. Chem.  (USSR), 21_
   , 1213 (1947).	

Dubinin, M.M., Vishnyakova, M.M., Zhukovskaya, E.G.,
   Leont'ev, E.A., Luk'yanovich, V.M. and Sarakhov,  A.I.,
   "Investigation of the Porous Structure of Solids  by
   Sorption Methods", Russ. J. Phys. Chem. . 3 , 959 (1960).

Dubinin, M.M., "Theory of the Bulk Saturation of Micro-
   porous Activated Charcoals During Adsorption of Gases
   and Vapours", Russ. J. Phy. Chem. . 39^ (6)  , 697 (1965).

Dubinin, M.M.  and Saverina, E., "The Porosity and Sorptive
   Properties  of Active Carbons", Acta. Physicochim. USSR,
     , 647 (1936).	

Dubinin, M.M., Zaverina, E.D. and Radushkevich,  L.V.,
   "Sorptlon and Structure of Active Carbons, I. Adsorption
   of Organic Vapors", J. Phys. Chem., USSR.  21  , 1351

Emmett,  P.H. and  Brunauer, S.,  "The  Use of Low Temperature
   van  der Waals  Adsorption  Isotherms in  Determining  the
   Surface Area  of Iron  Synthetic  Ammonia Catalysts", J.  Am.
   Chem.  Soc., 59^ ,  1553 (1937).

Falckenhagen,  H,,  "Cohesion  and Equation  of  Condition 1n
   Dipolar Gases", Physik. Z.,  23. ,  87  (1922).

     Sorption  (continued)

     Foster, A.G.,  "Sorption Hysteresis, Some Factors Determining
         the Size of Hysteresis Loop", J. Chem. Soc. (London).
         1806 (1952),

     Fraust, C.L. and Hermann, E.R., "The Adsorption of Aliphatic
         Acetate Vapors onto Activated Carbon", Am. I rid. Hyg.
         Assoc. J..  Sept-Oct (1969).

     Freundlich, H., Kapil larchetnie, Akademische, Verlagsgesell-
         schaft, m.b.h. Leipzig,   232,  (1933).

     Grace, W.R. and Co., "Davison Molecular Sieves", ADS-96-670.
         Davisori Chemical Division.

     Grant, R.J. and Manes, M., "Adsorption of Binary Hydrocarbon
         Gas Mixtures on Activated Carbon", Ind. and Enq. Chem.
         Fund..  ,  490-498 (1966).

     Grant, R.J., Manes, M. and Smith,  S.B., "Adsorption of Normal
         Paraffins and Sulfur Compounds  on Activated Carbon", Am.
         Inst.  Chem. Engrs. J.. 8_  (3), (1962).

     Grant, R.J., "Basic Concepts of Adsorption on Activated
         Carbon", Pittsburgh Chemical Co., Activated Carbon
         Technical Information.

     Gregg, S.J. and Sing, K.S.W., Adsorption. Surface Area and
         Porosity, Academic Press, N.Y.  U967J.

     Gurvitsch, L., "Theory of Heterogeneous Catalysis", J. Phys.
         Chem.  Soc.  Russ., 4_7  , 805 (1915).

     Gustafson, P.  and Smith, Jr., S.H., "Removal of Contaminants
         from Air by Type 13X Molecular  Sieve", U.S. Naval Research
         Lab..  NRL Report 5560, Aug.  (1961).

     Hasz, J.W. and Barrere, Jr., C.A., "Prediction of Equilibrium
         Adsorption  Capacity and Heats of Adsorption by the Polanyl
         Potential Theory", Chemical  Engineering Progress Symposium
         6 (96)  , 48 (1969).

     Hodgson,  A.S., "Portable Smoke  and Gas Removal Unit for
         Personnel Shelters-Feasibility  Study", Naval Civil
         Engineering Laboratory, (1968).

     Jaquet, E. "The Theory of the  Adsorption of Gases", Physik.
       V Phvsik.  Chern  (Chemical Physics & Physical Chemistry)
         TST,437 l

Sorption (continued)

Jones, G.W., "Inflammability of Mixed  Gases",  U.S.  Dept.
   of Commerce. (1928).                               

Jost, W., D1ffus i o n i n So 1 i d s ,  Liquids,  Ba s e s .  Academic
   Press, New York (I960).

Joyner, L.C., Barrett, E.P,  and Skold, R.,  "The Determination
   of Pore Volume  and  Area  Distribution  in  Porous  Substances,
   II Comparison between Nitrogen  Isotherm  and  Mercury
   Porosimeter Methods", J.  Am. Chem.  Soc.t  7^ ,  3155  (1951).

Juhola, A.J., Palumbo, A.J.  and Smith, S.B.,  "A Comparison
   of Pore Size Distributions of Activated  Carbons  Calcu-
   lated from Nitrogen and  Water Desorption  Isotherms",  J^
   Am. Chem. Soc.. 74. ,  61  (1952).

Juhola, A.J. and Wiig, E.G., "Pore  Structure  in Activated
   Charcoal, 1. Determination of Micro Pore  Size  Distri-
   bution". J. Am. Chem. Soc..  7 ,  2069 (1949).

Juhola, A.J., "Report  40059", Pittsburgh Carbon Company.
   Nov. (1951).

Keesom, W.H., "van der Waals Attractive  Force", Phvsik.  Z.,
   22_ , 129 (1921).

Kiselev. A.V., The Structure and Properties  of  Porous  Solids
     195, Butterworth, London (1958).

Klotz, I.M., "The  Adsorption Wave",  Chem.  Rev.. 39  ,  241

Kraemer, E.O., in  A Treatise on Physical Chemistry ed.  by
   Taylor, H.S.,  , 1661, Macmillan,  N.Y. (1931).

Langmuir, I., "The Adsorption of Gases on Plane Surfaces
   of Glass, Mica and Platinum", J.  Am.  Chem.  Soc.t 40
   , 1361 (1918).                                    

Langmuir, I., "A  Chemically Active Modification of Hydrogen",
   J. Am. Chem. Soc.. 3 , 1310 (1912).

Langmuir, I., "The Constitution and Fundamental Properties
   of Solids  and  Liquids", J.  Am. Chem. Soc.   38  , 2221
   (1916).                                     "~

     Sorptlon (continued)

     Ledoux, "Avoiding Destructive Velocity Through Adsorbent
        Beds", Chem. Eng.. March (1948).

     Leonard-Jones, J.E., "Processes of Adsorption and Fusion on
        Solid Surfaces", Trans. Faraday Soc,, 28 , 333 (1932).
     Lewis, W.K., G1ll1land, E.R., Chertow, B. and Cadogen, W.P.,
        "Adsorption Equilibria-Hydrocarbon Gas Mixtures" - "Pure
        Gas Isotherms". Ind. & Eng. Chen. 42,   1314-1332, July

     L1nde Company, Division of Union Carbide Corp., "Physical
        Properties of Types 3A, 4A, 5A, 10X and 12X", Sales
        Catalog F-9947-D. P.O. Box 41, Tonawand, N.Y. TT9T7).

     Lippens, B.C., Linsen, B.G. and DeBoer, J.H.. "Pore Systems
        1n Catalysts", J. Catalysis, 3, , 32 (1964).

     Lockheed,  "Trace Contaminant Removal System Design", MSAR
        Subcontract No. 28-5145, with Lockheed Missies and Space
        Co.,  (1970).

     London,  F.,  "The General Theory of Molecular Forces", Trans.
        Faraday Soc. 33 . 8 (1937).

     London,  F.,  "Theory  and Systematics of Molecular Forces.
        A General Survey  of Intermolecular Forces from the Stand-
        point of  Quantum  Mechanics", Z. Physik., B 11 ,   222

     Lorenz,  R. and Lande,  H.,  "Adsorption and the Corresponding
        States",  Z. Anorg.  U. Allgem. Chern.  (Journal of  Inorganic
        and General Chemistry), 125, 47TT922).

     Loven, A.W., et.al.,  in J. Makowskl,  et.al. "Collective
        Protection Against  CB Agents",  Eighth  Progress Report  on
        Contract  No. DA  18-035-AMC-279  (A), AiResearch Mfg.  Co.
        Report  Ho. CB-1008, Feb.  (1966).

     Maher, P.K.,  "Silica  (Amorphous)",  1n Kirk-Othmer  Encyclo-
        pedia of  Chemical  Technology.  Intersdence  Pub., New  York,
        18, ,  61  (1969).

     Mattla,  M.M., "Process for Solvent Pollution  Control",  Chenu.
        Eng.  Progr., 66. (12)  ,  74  (1970).

Sorption (continued)

McBain, J.W., "An Explanation of Hysteresis in the Hydra-
   tion and Dehydration of Gels", J.  Am.  Chem. Soc..  57
   , 699 (1935).	'  

McKee, D.W., "The Sorption of Hydrocarbon Vapors  bv Silica
   Gel", J. Phys. Chem.. 3 , 1256 (1959).

Mecklenburg, W., "Layer Fi1tration--Theory  of the Gas Mask"
   Z. Elektrochem. (Journal of Electrochemistry)  31  , 488

Myers, A.L. and Prausnitz, "Thermodynamics  of Mixed-Gas
   Adsorption", Am.  Inst. Chem.  Engrs.  J.,  11  ,  121-127
   (1965).	  ~"

Othmer, D.F. and Sawyer, F.G., "Correlating Adsorption Data",
   Ind. & Enq.  Chem..  35.. 1269  (1943).

Othmer, D.F. and Josefowitz, S., "Correlating Adsorption
   Data", Ind.  & Eng.  Chem.. 4 (4) ,  723 (1948).

Papee, D. and Tertian, R., "Aluminum  Compounds"!  in  Kirk-
   Othmer Encyclopedia of Chemical Technology, Interscience
   Pub., New York, 2_  , 42 (1963).

Pierce, C.  and  Smith,  R.N., "Adsorption in  Capillaries",
   J. Phys. Chem.. 57_  ,64 (1953).

Pierce, C.  and  Smith,  R.N., "Adsorption-Desorption Hysteresis
   in Relation  to Capillarity of Adsorbents", J.  Phys. Chem.,
   4  , 784 (1950).

Pierce, C., Smith, R.N., Wiley,  J.W.  and  Cordes,  H.,  "Adsorp-
   tion of  Water by  Carbon", J.  Am.  Chem. Soc.,  73 ,4551
   (1951).                    	  "~

Pierce, C., Wiley, J.W. and Smith, R.N.,  "Capillarity and
   Surface  Area of Charcoal", J. Phys.  Chem.. 53   ,669
   (1949).                     	*	  

Pierce, C.. "Heats of  Adsorption", J.  Phys. Chem.t 57, 149

Polanyl, M., "Adsorption and Capillarity from Standpoint of
   2nd Law  of Thermodynamics", Verhandl.  deut. physik. Ges.
   16  , 1012 (1914).           	"	

     Sorption (continued)

     Ray, G.C. and Box, Jr., E.O., "Adsorption of Gases on
        Activated Charcoal", Ind. & Eng. Chem.. 42  ,1332,
        July (1949).                            ~~

     Rltter, H.L. and Drake, L.C., "Pore Size Distribution in
        Porous Materials, Pressure Porosimeter and Determination
        and Complete Macropore Size Distributions", Ind. & Eng.
        Chem.t Anal. Ed.. ]_  , 782 (1945).          ~~

     Robell, A.J., Ballow, E.V. and Borgardt, F.G., "Basic Studies
        of Gas-Solid Interactions III", Lockheed Missies and Space
        Co.. Report No. 6-75-65-22, Apri I  11965).

     Robell, A.J. and Merrill, R.P., "Gaseous Contaminant Removal
        by Adsorption: II. Adsorption Dynamics in Fixed Beds",
        Chemical Engineering Progress Symposium. 65 (96)   100

     Robinson, C.S., "Commercial Activated Carbon Systems", cp.  13
        in The Recovery of Vapors, with Reference to Volatile
        Solvents, Relnnold Pub., N.Y. (194ZJ.

     Rowe, N.R., "Controlling of Gaseous Contaminants Throuqh
        Adsorption". Bulletin T-101, J. Am. Assoc. Cont. Contr..
        March (1964).                	

     Rowson, H.M., "Design Considerations  in Solvent Recovery",
        Proceedings MECAR Symposium. New Developments in Air
        Pollution Control, Metropolitan Engineers Council on
        Air Resources, New York (1967).

     Sheehy, J.P., Achinger, W.C. and Simon, R.A., Handbook of
        Air Pollution. National Center for Air Pollution Control.
        Durham, N.C., 999-AP-44 (1970).

     Sing, K.S.W. and Madeley, J.D., "The  Surface Properties of
        Silica Gels. I. Importance of pH in the Preparation from
        Sodium Silicate and Sulfuric Acid", J. Appl. Chem.
        (London) 3_ (1953).

     S1ng, K.W.W. and Madeley, O.D., "Surface Properties of Silica
        Gels. II. Adsorption of Water Vapor", J. ADD!. Chem.
        (London),  , 365 (1954).

     Strauss, W., "Odor Control for the Process Industries",
        Chem. & Process Eng.. March (1965).

Sorption (continued)
Sundaresan, B.B., et.al., "Adsorption of Nitrogen Oxides
   from Waste Gas", Environ. Sci. Techno!.. J_ (2), (1967).

Teller, A.J., "Odor Abatement in the Renderinq and Allied
   Industries", J. Air Pollution Control Assoc..  13 (4)
   , 148 (1963).	  ~

Thomson, W.T., "On the Equilibrium of Vapor at a  Curved
   Surface of Liquids", Phil. Mag. 158 : 448 (1871).

Trappe, W., "Separation of Biological Fats from Natural
   Mixtures by Means of Adsorption Columns", Biochem.  Z..
   (Journal of Biochemistry) 305 , 150 (1940).

Turk, A., "Industrial  Odor Control", Chemical  Engineering/
   Deskbook Issue. April (1970).

Turk, A., Sleik, H. and Messer, P.O., "Determination  of
   Gaseous Air Pollution by Carbon Adsorption", Industrial
   Hygiene Quarterly,  1J. (1) ,  23 (1952).

Turk, A.. "Source Control by Gas-Solid Adsorption", Chapter
   47, in Air Pollution, Vol. Ill, Stern, A.C. ed., 497,
   Academic Press, N.Y. (1968).

Viessman, W., "Gaseous Air-Pollution-Its Sources  and  Control"
   Air Enq. . July (1968).

Walsh, R.T., "Solvent  Deqreasers", 1n Chapter 7,  AjTr
   Pollution Engineering Manual, Danielson, J.A.  ed.,
   Public Health Service, 999-AP-40,   383 (1967).

Webb, A.N., "Hydrofluoric Acid and Acidity of Alumina",
   Ind. & Enq. Chem. 49 (2) ,  261 (1957).

Weiss, S.M., "Surface  Coating Operations", in Chapter 7,
   Air Pollution Engineering Manual, Danielson, J.A.  ed.,
   Public Health Service, 999-AP-40,   387 (1967).

Wiig, E.O. and Juhola, A.J., "The Adsorption of Water Vapor
   on Activated Charcoal", J. Am. Chem. Soc.. 71    561
   (1949).                 	  "

      Sorntlon  (continued)

      Yocum,  J.E.  and  Duffee,  R.A..  "Controlling  Industrial Odors",
         Chem.  Enq., June  15  (1970).

      Young,  D.M.  and  Crowell,  A.O.,  Physical Adsorption of Gases,
         Butterworths,  London,  371-391  (1962).

 Sorbent  Systems

 API Bulletin 2518, "Evaporation Loss from Fixed-Roof Tanks"
   American Petroleum Institute Bulletin 2518. (1962).

 ASHRAE,  Handbook of Fundamentals. "Odors", cp. 12, Am. Soc.
   Heat. Refrig. Air-Cond. Engrs., New York, 175 (1967).

 Aunln, W.J. and Kolarl, C.E., "Meat and Meat Products", 1n
   Kirk-Qthmer Encyclopedia of Chemical Technology. Inter-
   science Pub,, New York, 13, 176 (1967).

 Bailor,  W.C., "Dry Cleaning Equipment", 1n cp. 7, Air Pollu-
   tion  Engineering Manual, Danlelson, J.A.  ed.. Public Health
   Service, 999-AP-40, 393 (1967).

 Barnebey, H.L., "Activated Carbon for Air Purification",
   Heating. Piping and Air Conditioning. March (1958).

 Barnebey, H.L.  and Davis, W.L., "Costs of Solvent Recovery
   Systems", Chem. Eng.t Dec. 29 (1959),

 Barnebey, H,L., "Removal of Exhaust Odors from Solvent
   Extraction Operation by Activated Charcoal  Adsorption",
   J.  Air Pollution Control Assoc., ] (4),  422 (1965).

 Berg,  C. and Bradley, W.E., "Hypersorption-New Fractionatinq
   Process", Petroleum Enn. .  JJB, 115 (1947).

 Berq,  C,, "Hypersorntion-A Process for Separation of Light
   Gases", Gas, 23, (1), 32 (1947).

 Berg,  D.B. and  Lewandowski, R., "Trichloro-Fume Control Saves
   $780 per week", Chem. Process., Jan. (1968).

 Bloomfield, B.D.,  "Control of Gaseous Pollutants", Heating.
   Piping and Air  Conditioning. Jan. (1968).

Burchsted, C.A.  and Fuller, A.B., "Hiqh-Efficiency A1r
   Filtration Systems for Nuclear Application", ORNL-NSIC-65,
   Oak Ridge National  Laboratory, Jan.  (1970),

Burnett, R.W.  and  Turnbull, W.T., "Industrial  Applications
   of  Molecular Sieves", British Chem.  Eng.. Vl^ (4), (1966).

Calgon Corp.,  Activated Carbon Adsorption Handbook, Pitts-
   burgh Activated Carbon Div.                      	

      Sorbent Systems  (continued)

      Chass, R.L,,  Kanter,  C.V.  and  Elliott.  J.H.,  "Contribution
         of Solvents  to  Air Pollution  and  Methods for  Controlling
         Their Emissions".  J.  Air  Pollution Control Assoc..  13  (2),
         64 (1963).

      Cooper. J.C.  and Cunniff,  F.T.,  "Control  of Solvent  Emissions",
         New Developments  in Air Pollution. Metropolitan Engineers
         Council  on  Air  Resources.  (MECAR), New York  (1967).

      Duhl, R.W.,  "Factors  Affectinq Economic Recovery of  Roto-
         qravure  Solvents", Vulcan-Cincinnatit  Inc.

      Glaser, J.R.  and Ledbetter,  J.O.,  "Air  Pollution from  Sewaqe
         Treatment",  Paper  No. 69-44.  Univ. of  Texas,  at Austin.

      Gravure, "Exhaust  Air Purification and  Solvent Recovery in
         Gravure  Printing11, Gravure. March  (1969).

      Hardison,  L.C.,  "Controllinq  Combustible  Emissions". Paint
         and Varnish  Production. July  (1967).              "

      Harrinqton,  R.E.,  "Gas-Solid  Sorption Study", National A1r
         Pollution  Control  Administration.  Nov.  12  (1968).

      Lauber, J.D.,  "The Control of  Solvent Vapor Emissions", New
         York State  Dept.  of Health. Paper  No.  69-42.  Jan. (1969T.

      Mattia, M.M.,  "Process for Solvent Pollution  Control", Chem.
         Eng. Progr..  66. (12), 74  (1970).                    	

      McGraw, M.J.  and Duprey, R.L., "Air  Pollutant Emission
         Factors",  Environmental Protection Agency. Preliminary
         Document,  58  (1971).

      Moncrieff,  R.W., "Industrial  Odors,  Part  2 -  General Methods
         of Treatment",  Industrial  Water and  Wastes.  Sept-Oct (1961).

      Perry's Chemical Engineers Handbook.  Perry, R.H., Chllton,
         C.H, and  Klrkpatrick, S.D., "Humidlfication  and Drying",
         McGraw  H111,  New  York,  section  15-25  (1963).

      Richardson,  N.A. and  Middleton,  W.C., "Evaluation of Filters
         for Removing  Irritants  from Polluted A1r", American
         Society  of  Heating* Air Conditioning; Engineers, Journal
         Section,  Nov. (1958).

Sorbent Systems  (continued)

Robinson, C.S.,  "Commercial  Activated  Carbon  Systems",  cp.
   13 in The Recovery of Vapors  with  Reference  to  Volatile
   Solvents. Relnhold Pub.,  New  York  (1942).

Rowe, N.R., "Controlling of  Gaseous  Contaminants Through
   Adsorption",  Bulletin T-101,  The  Journal  of  the American
   Association for Contamination Control .  March U964).

Rowe, N.R., "The Economics of Activated  Charcoal", Barnebey
   Cheney, Air Enq.. May (1963).

Rowson, H.M., "Desiqn Considerations  in  Solvent Recovery",
   Proceedings MECAR Symposium,  New  Developments  in Air
   Pollution Control, Metropolitan Enqlneers Council  on A1r
   Resources, New York  (1967).

Santry, Jr., I.W.. "Hydroqen Sulflde Odor Control  Measures".
   Hater Pollution Control Federation, March
Stenburq, R.L., "ControlUnn Atmospheric E"1 J',     1tll
   Paint and Varnish Operations, Part 1", Paint and Varnish
   Production. Sept. (1959).

Strauss, W., "Odor Control for the Process  Industries", Chem.
   &  Process Enq., March  (1965).
 Teller,  A.J.,  "Odor Abatement 1n the Renderlnq "d Allied
    Industries", J. Air  Pollution Control Assoc.. jj, (4),
    148  (1963).

 Turk, A.,  "Source  Control  by Gas-Sol1d Adsorption", 1n
    cp.  47, Air Pollution,  Vol.  Ill, Stern, A.C. ed., 497
    Academic Press. New  York  (1968).

 Vulcan-Cincinnati, Inc.,  "Adsorption", Chem.  Enq.. Oct.

 Westvaco,  "Sulfur  Recovery Process  Under  Development",  Puljj.
    and  Paper.  April  (1971).

 Yocum,  J.E. and Duffee. R.A..  "Controlling  Industrial  Odors",
    Chem. Enq.. June  15  (1970).


      Accomazzo,  M.A.  and  Mobe, K., "Catalytic Combustion of Ci to
         C3  Hydrocarbons",  Ind. & EngV Chem., Process Design De-
         velop..  4_ (4),  425  (1965).

      Acres, G.J.K.,  "Platinum Catalysts for the Control of Air
         Pollution",  Research Labs. Johson Matthey & Co.. Ltd.

      Acres, G.J.K.,  "The  Elimination of Organic Fume by Catalytic
         Combustion",  Platinum Metals Review, 14 (1), 2 (1970).

      Anderson, R.B.,  et.al., "Catalytic Oxidation of Methane",
         Ind.  & Eng.  Chem.,  53, (10), 809 (1961).

      Arrhenlus,  S.,  Theories of Solution, Yale Univ. Press, 68

      Balandln, A.A.,  "The  Theory of Heterogeneous Catalytic
         Reactions.   The Multlplet Hypothesis.  A Model for De-
         hydration  Catalysis", Ze1t Phys. Chem.. 132. 289 (1929).

      Beeck, 0.,  "Hydrogenatlon Catalysts", Discussions of Fara-
         day Soc.,  8.,  118  (1950).           	

      Berzellus,  "On  the Progress of Physical Sciences", report
         to  Swedish Academy  of Sciences. Mar. 31 (1835).

      Bloomfleld,  B.D.,  "Control of Gaseous Pollutants", Heating.
         Piping and A1r  Conditioning. Jan. (1968).

      Brewer,  G.L., "Fume  Incineration", Chem. Eng., Oct. 14

      Brewer,  G.L., "Odor  Control for Kettle Cooklno", J. A1r
         Pollution  Control  Assoc.. 1_3_ (4) (1963).

      Brodskl, A.M.,  et.al.,  "Ouant1tat1ve Characterization of the
         Reliability  of  Chemical Reaction Mechanisms", Dold, Akad.
         Nauk  SSSR, (Proceedings of the Academy of Science of  the
         USSR). 174 (4), 865  (1967).

      Brunauer, S., Emmett,  P.H, and Teller, E., "Adsorption of
         Gases 1n  Multlmolecular Layers", J. Am. Chem. Soc., 60,
         309 (1938).

      Christian,  J.6.  and  Johnson, J.E., "Catalytic Combustion of
         Aerosols", Ind. A  Eng. Chem., Prod. Res. 4 Develop.,  
         (3)  (1963).


Catalysis (continued)
Christian, J.G.  and  Johnson.  J.E.,  "Catalytic  Combustion  of
   Atmospheric Contaminants Over  Hopcallte",  International
   Journal of A1r Water Pollution.  Perqamon  Press.  9.  1-10

Cooper, B.J., "Platinum-Carbon  Catalysts  with  Molecular
   Sieve Properties".  Platinum  Metals  Review.  14  (4),  133

Crouse, L.F., "Efficient Design of  After  Burners  for  Incin-
   eration of Many Industrial  Fumes",  Air Eng.. Aug.  (1967).

Corrigan, I.E.,  "Catalysis  and  Absorption",  Chem. Eng^. 61,
   236 (1954).

Dmuchovsky, B.,  Freerks, M.C.  and Zrenty, F.t  "Metal  Oxide
   Activities 1n the Oxidation  of Ethylene",  J. Catalysis
   1, 577 (1965).

Dowden, D.A.  and Reynolds,  P.W., "Heterogeneous  Catalysis:
   Theoretical and Hydrogenatlon  by Binary Alloys", J. Am.
   Chem. Soc.. 242-271 (1950).

Dowden, D.A. and Wells, D.,  Vol.  II, Proceedinos  of the
   Second International Congress  on Catalvsls, Paris,  France

Edwards, F.R., "Catalytic Combustion of Fumes  and Odors",
   presented at Air and Water Pollution Control,  5th  Annual
   Seminar (1969TI~

Emmett, P.H. and Brunauer,  S.,  "The Use of Low Temperature
   van der Waals' Adsorption  Isotherms 1n Determining the
   Surface Area of Iron Synthetic Ammonia Catalysts", J.  Am.
   Chem. Soc.. 59., 1553 (1937).

Eyring, H., Colburn, C.B. and Zwolinskl,  B.J., "The Activated
   Complex in Chem1sorpt1on and Catalysis", Discussions  Fara-
   day Soc.. 8., 39 (1950).

Freundlich, H. and Hase, E.,  "The Velocity of Absorption
   Retrogression", Zelt Physik. Chem.  89. 417 (1915).

Hardlson, L.C., "Controlling Combustible Emissions", Paint
   and Varnish Production. July (1967).

      Catalysis  (continued)

      Harkness,  A.C., Murray.  F.E. and Glrard, L., "Catalytic
         Oxidation  of Sulfurous Air Pollutants", Atmospheric
         Environment* 2., 303,  Grt. Brltlan  (1968).

      He1n,  R.M.,  "Odor Control by Catalytic and Hlqh Temperature
         Oxidation", Ann. N.Y. Acad. Scl.,  116. 656 (1964).

      Hlnshelwood,  C.N., "Homoqeneous Catalysis", Trans. Faraday
         Soc.  24_,  552 (1928).

      Hlavacek,  V.,  "Aspects  1n Deslqn of Packed Catalytic Reac-
         tors",  Ind. & Eng. Chem., 1_2_ (7),  8 (1970).

      Houqen,  O.A.  and Watson, K.M., "Solid Catalysts and Reaction
         Rates", Ind. & Enq.  Chem., 3,  529 (1943).

      Hwano, S.T.  and Parravano,  G.J., "Transient Response of
         Chemically Interacting Solid Gas Svstems", J.  Electro-
         chem  Soc..  114, 478  (1967).                     	

      Johns, P.H.,  "Kinetics  and  Catalyst Selection 1n  the Com-
         bustion of AtmosnheMc Trace Contaminants", Chemical
         Engineering Progress  Symposium. 6_2_ (63), 81 (1966).

      KHmlsch,  R.L., "Oxidation  of CO and  Hydrocarbons  Over
         Supported  Transition  Metal Oxide Catalysts", First
         National  Symposium on Heterogeneous Catalysis  for Con-
         trol  of A1r Pollution, National A1r Pollution  Control
         Administration, 175-197  (1968).

      Krenz, W.B.,  Dlckenson,  J.  and Chass, R.L., "An Appraisal of
         Rule  66 of the Los Angeles Count'/  A1r Pollution Control
         District",  J. A1r Pollution Control Assoc., 18, 743
         (1968).                                    "~

      Lamb,  A.B.,  "The Preferential Catalytic Combustion of
         Carbon  Monoxide 1n Hydrogen", J. Am. Chem. Soc., 44,
         738-57  (1952).

      Lanqmulr,  I.,  "The Adsorntlon of Rases on Plane Surfaces
         of  Glass,  Mica and Platinum", J. Am. Chem. Soc.. 40.
         1361  (1918).

      Lin,  K.H., "Applied Reaction Kinetics", Ind. & Enq. Chen,.
         61  (3), 42 (1969).

Catalysis (continued)

Mango, F.D.  and Schacktschneider,  J.H.,  "Molecular Orbital
   Symetry Conservation In Transition Metal  Catalyzed
   Translormations"   J. Am. Chem.  Soc..  89,  2484 (1967),

Miller, f'.R. and Wilhoyte, H.J., "A Study of Catalyst Support
   Systems for Fume Abatement of Hydrocarbon Solvents", J.
   Air Pdllution Control Assoc.. H (12)  (1967).

Mi-lls, G.A., "Catalysis", in Kirk-Othirer Encyclopedia of
   Chemical  Technology, Interscience Fub., New York, 4_  , 534

Ctake, T., Kunigita,  E. and Naleao, K.,  "Analysis  of an
   Overall Gas-Liquid Reaction Rate Based on Twp-Film
   Theory". Kagaku Kogaku 31 (7)  ,691 (1967).

Paulus, H.J., "Nuisance Abatement by Combustion",  cp. 48 1n
   Air Pollution, Vol.  Ill, Stern, A.C.  ed., 521,  Academic
   Press, N.Y.  (1968).

Pitzer, E.C. and Frazer, J.C.W.,  "The Physical Chemistry of
   HopcalUe Catalysts", J. Phys. Chem., 5 (1941).

Romeo. P.L. and Warsh,  H., Catalytic  Itineration  Design
   Parameters and Operating Practices,  Public Health Service
   Training Manual, Feb.  (1962).

Sabatler, P., LaCatalvse  en Chemie Organique, C. Beranger
   Paris, France (1920).

Selwood,  P.W.,  "Magnetism  and  the  Structure of Catalytically
   Active Solids", in  Advances  in  Catalysis,  3_  . 27  (195U-

Sherwood, P.W.,  "Get  Rid  of That  Smell",  Food Manufacture,
   July  (1961).

Stein, K.C.,  et.al.,  "The  Oxidation  of  Hydrocarbons  on
   Simple Oxide Catalysts",  J.  Air Pollution  Control  Assoc..
   l_p_ ,  275-81  (1960).

Strauss,  W.,  "Odour  Control  for the  Process Industries",
   Chem.  &  Process Eng.,  March (1965).

Suter, H.R.,  "Range  of Applicability of Catalytic Fume
   Burners",  J. Air  Pollution Control Assoc., 5_ (3), (1955).

     Catalysis  (continued)
     Taylor, H.S. and Halsey, G., "Contact Catalysis between
        Two World Wars", J. Chem. Phys. 1.5 , 627 (1947).
     Temkln, M.I.. Problems in Kinetics and Catalysis, Moscow
        (1949).    	""
     Thlele, E.W., "Relation Between Catalytic Activity and Size
        of Particle", Ind. & Eng. Chem., 3J_ , 916 (1938).
     Thomaldes, L., "Why Catalytic Incineration?", Pollution
        Engineering, May/June (1971).
     Viessman,  W., "Gaseous A1r-Pol1ution-Its Sources and Con-
        trol",  A1r Eng., July (1968).
     Welsz, P.B.  and Prater, C.D., Advances in Catalysis. Vol. VI,
        Academic  Press, New York  (l9b4j.
     Wheeler, A., Catalysis, Vol. II, cp. 2, Relnhold Publ. Co.,
        New York  (1955).
     Wheeler, A., Advances in Catalysis. Vol. Ill, Academic
        Press,  New York (1951).
     Wllhelm, R.H. Hohnson, W.C.  and Acton, F.S., "Conduction,
        Convection and Heat Release  in  Catalytic Converters",
        Ind. &  Eng. Chem., 35,  (5) ,  562  (1943).
     Yang, K.H. and Hougen, O.A., "Determination of  the Mechanism
        of Catalyzed Gaseous Reactions", Chem. Eng.  Progr., 46,
        146  (1950).
     Yocum, J.E.  and Duffee, R.A..  "Controlling  Industrial Odors",
        Chem. Eng.. June  15  (1970).

Catalytic Incineration

Acres, G.J.K., "The Elimination of Organic  Fume  by  Cataly-
   tic Combustion", Platinum Metals Review.  14  (1)  .  2

Bloomfield, B.D., "Control  of Gaseous  Pollutants",  Heating.
   Piping and Air Conditioning, Jan.  (1968).

Brewer, G.L., "Odor Control  for Kettle Cooking",  J. Air
   Pollution Control Assoc., V3 (4),  (1963).

Brewer, G.L., "Fume Incineration", Chem.  Eng.,  Oct. 14

Grouse, L.F., "Efficient Design of Afterburners  for
   Incineration of Many Industrial Fumes",  Air  Eng..  Aug.

Cooper, J.C. and Cunnlff. F.T., "Control  of Solvent Emissions",
   Metropolitan Engineers Council  on Air  Resources. New  York,
   N.Y. (1967)."~

Dmuchovsky, B., Freeks, M.C. and Zrenty,  F., "Metal Oxide
   Activities 1n the Oxidation of  Ethylene", Journal  of
   Catalysis 4  . 577 (1965).

Edwards, F.R., "Catalytic Combustion of Fumes and Odors",
   presented at Air and Water Pollution Control,  5th  Annual
   Seminar  (1969).

Hardison, L.C., "Controlling Combustible  Emissions",  Paint
   and Varnish Production, July (1967).

He1n, G.M., "Odor Control by Catalytic and High Temperature
   Oxidation", Ann. M.Y. Acad. Sci.. 116  , 658  (1964).

Klimisch, R.L., "Oxidation of CO and Hydrocarbons Over
   Supported Transition Metal Oxide Catalysts", First
   National Symposium on Heterogeneous Catalysis for  Con-
   trol of Air Pollution. National Air Pollution control
   Administration,   175-197 (1968).

Lin, K.H., "Applied Reaction Kinetics", Ind. ft Eng. Chem..
   61  (3)  , 42  (1969).

Mattia, M.M., "Process for Solvent Pollution Control",  Chem.
   Eng. Progr. 66  (12)  , 74  (1970).                     	

     Catalytic Incineration (continued)
     McGraw, M.J. and Duprey, R.L., "Evaporative Loss Sources",
        In Air Pollutant Emission Factorst   35 (1971).

     Mills, J.L., Walsh, R.T., et.al., "Quantitative Odor
        Measurement", J. Air Pollution Control  Assoc., 13 (10)
        (1963).       	  ~

     Miller, M.R. and WUhoyte, H.J., "A Study of Catalyst Support
        Systems for Fume-Abatement of Hydrocarbon Solvents", J.
        A1r Pollution Control Assoc.. 17 (12)  (1967).       

     Partee, F., "A1r Pollution 1n the Coffee Roasting Industry",
        U.S. Dept. of Health, Education and Welfare. 999-AP-9,

     Paulus, H.O., "Nuisance Abatement by Combustion", cp. 48 1n
        'A If Pollution, Vol. Ill, Stern, A.C. ed.,   521, Academic
        Press, New York (1968).

     Romeo, P.L. and Warsh, H.t Catalytic Incineration Design
        Parameters and Operating Practices. Public Health Service
        Training Manual, Feb. (1962).

     Strauss, W., "Odor Control for the Process Industries", Chem.
        & Process Eng.t March (1965).

     Sullivan, J.L., Kafka, F.L. and Ferrari, L.M., "An Evalu-
        ation of Catalytic and Direct Fired Afterburners for
        Coffee and Chicory Roasting Odors", J.  Air Pollution
        Control Assoc.. 1_5 (2) , 583 (1965).

     Suter, H.R., "Range of Applicability of Catalytic Fume
        Burner", J. Air Pollution Control Assoc., 5 (3) , 173
        (1955).                                   "

     Taylor, H.S. and Halsey, G., "Contact Catalysis between Two
        World Wars", J. Chem. Phys.. ^5 , 627 (1947).

     Wilhelm, R.H., Hohnson, W.C. and Acton, F.S., "Conduction,
        Convection and Heat Release in Catalytic Converters", Ind,
        & Eng. Chem., 35. (5) , 562 (1943).

     Yocum, J.E. and Duffee, R.A., "Controlling Industrial
        Odors", Chem. Eng., June 15 (1970).

Thermal  Incineration
Anon., "Direct Flame Method  of Incineration  for  Combustible
   Solvents", Air Eng..  April  (1968).

Benforado, D.M. and Waitkus, J.,  "Exploring  the  Applicability
   of Direct-Flame Incineration to  Wire-Enamel Ing  Fume  Con-
   trol". Wire and Wire  Products. Nov.  (1967).

Bloomfield, B.D., "Control  of Gaseous  Pollutants",  Heating,
   Piping and A1r Conditioning, Jan.  (1968).

Brewer, 6.L., "Fume Incineration".  Chert.  Eng.. Oct.  14

Chass, R.L., Kanter, C.V.  and Elliott,  J.H.,  "Contribution
   of Solvents to Air Pollution and Methods  for  Controlling
   Their Emissions", J.  Air Pollution  Control  Assoc.. ^ (2)
   . 64 (1963).

Cooper, J.C. and Cunniff,  F.T., "Control  of  Solvent Emissions",
   Metropolitan Engineers  Council  on Air  Resources,  New York,
   N.Y. (1967).

Coward, H.F. and Jones,  G.W., "Limits  of  Flammability of
   Gases and Vapors", Bulletin 503. Bureau  of Mines,  (1952).

Grouse, L.F., "Efficient Design of  Afterburners  for Incin-
   eration of Many Industrial Fumes",  Air Eng.,  Aug.  (1967).

GCA, "Control Technigues for Polycyclic Organic  Matter
   Emissions", GCA Technical Division.  Air  Pollution Control
   Administration. (1970).

Hardison, L.C., "Controlling Combustible  Emissions", Paint
   and Varnish Production,  July (1967).

He1n, G.M., "Odor Control  by Catalytic and  High  Temperature
   Oxidation", Ann. N.Y. Acad. Sci.. 116  ,  656 (1964).

McGraw, M.J. and Duprey, R.L., "Air Pollutant Emission
   Factors", Environmental  Protection Agency. Preliminary
   Document,   58 (1971).

McGraw, M.J. and Duprey, R.L., "Evaporative Loss Sources",
   in Air Pollutant Emission Factors.   35 (1971).

     Thermal Incineration (continued)

     Mills, J.L., Walsh, R.T., et.al., "Quantitative Odor
        Measurement", J. Air Pollution Control Assoc., 13 (10)

     Paulus, H.J., "Nuisance Abatement by Combustion", cp. 48,
        1n Air Pollution. Vol. Ill, Stern, A.C. ed.,   521
        Academic Press, N.Y. (1968).

     Polglase, W.L., Dey, H.F. and Walsh, R.T., "Deep Fat Frying",
        cp. 11, 1n A1r Pollution Engineering Manualt Danlelson,
        J.H. ed., Public Health Service, 999-AP-40,   755 (1967).

     Polglase, W.L., Dey, H.F. and Walsh, R.T., "Food Processing
        Equipment", cp. 11, In Air Pollution Engineering Manual.
        Danlelson, J.A. ed., Public Health Service, 999-AP-40,
        746 (1967).

     Sullivan, J.L., Kafka, F.L. and Ferrari, L.M., "An Evaluation
        of Catalytic and Direct Fired Afterburners for Coffee and
        Chicory Roasting Odors", J. Air Pollution Control Assoc..
        ! (2) , 583 (1965).     	

     Thomaldes, L., "Why Catalytic Incineration?", Pollution
        Engineering, May/June (1971).

     Vlessman, W., "Gaseous Air Pollution-Its Sources and
        Control", Air Eng.. July (1968).

     Walsh, R.T., Leudtke, K.D. and Smith, L.K., "Fish Canneries
        and Fish Reduction Plants", cp. 11, In Air Pollution
        Engineering Manualt Danielson, J.A. ed., Public Health
        Service, 999-AP-40,   760  (1967).

     Yocum, J.E. and Duffee, R.A.. "Controlling Industrial Odors",
        Chem. Eng., June 15 (1970).


Liquid Absorption

ASHRAE, "Odor Control", cp. 31, in Guide and Data Book
   Systems, Am. Soc. Heat., Refriq., Air-Cond. Engrs.. New
   York,   459 (1970).

Barnebey, H.L., "Removal of Exhaust Odors from Solvent
   Extraction Operation by Activated Charcoal Adsorption",
   J. Air Pollution Control Assoc.. 1_5 (4)  4?2 (1965).

Bloomfield, B.D., "Control of Gaseous Pollutants", Heating.
   Piping and Air Conditioning. Jan. (1968).

Calvert, S., "Source Control by Liquid Scrubbing", in Air
   Pollution, chapter 46, Vol. Ill, Stern, A.C. ed.,  "T57
   Academic Press, New York (1968).

Chass, R.L., Kanter, C.V. and Elliott, J.H., "Contribution
   of Solvents to A1r Pollution and Methods for Controlling
   Their Emissions", J. Air Pollution Control Assoc.. 13
   (2) , 64 (1963).       "	  ~

Chatfield, H.E.,  "Pipe-Coating Equipment", chapter 7, in
   Air Pollution  Engineering Manual. Danielson, J.A. ed.,
   U.S. Dept. of  Health, Education and Welfare, 999-AP-40,
   390 (1967).

Cooper, J.C. and  Cunniff, F.T., "Control  of Solvent Emis-
   sions", Metropolitan Engineers  Council on Air Resources.
   New York, N.Y. (1967).

Mattia, M.M., "Process for Solvent Pollution Control", Chem.
   Eng. Proqr.. 6 (12) , 74 (1970).

Mills, J.L., Walsh, R.T., et.al.,  "Quantitative Odor
   Measurement",  J. Air Pollution  Control Assoc.. 13 (10)

Partee, F., "A1r  Pollution in the  Coffee Roasting Industry",
   U.S. Dept. of  Health, Education and Welfare. 999-AP-9,

Strauss, U., "The Development of a Condenser for Odor Con-
   trol from Dry  Rendering Plants", J. Air Pollution Control
   Assoc.. V4 (10) , 424 (1964).

Strauss, W., "Odor Control for the Process Industries",
   Chem. & Process Enq. . March (1965).

      Liquid  Absorption  (continued)

      Teller,  A.J.,  "Odor Abatement In the Renderlno and Allied
         Industries", J. A1r Pollution Control Assoc., 13 (4),
         148  (1963).

      Walsh,  R.T., Leudtke, K.D. and Smith, L.K., "F1sh Canneries
         and  F1sh Reduction Plants", cp. 11, 1n A1r Pollution
         Enolneerlnq Manual, Oanlelson, J.A. ed.f 999-AP-40,
         760  (1967).

      Yocum,  J.E. and Duffee, R.A., "ControlUnq Industrial
         Odors", Chem. Eng. June 15 (1970).

Absorption Systems

Bloomfield, B.D., "Control  of Gaseous  Pollutants1*.  Heating,
   Piping and Air Conditioning,  Jan.  (1968).

Calvert, S., "Source Control  by  Liquid Scrubbing",  cp.  46,
   In A1r Pollution, Vol. Ill, Stern,  A.C.  ed.,  457,
   Academic Press,  N.Y. (1968).

Cooper, J.C. and Cunnlff, F.T.,  "Control  of Solvent
   Emissions", Metropolitan Engineers  Council  on A1r  Re-
   sources, New York, N.Y,  (1967^

McGraw, M.J. and Duprey, R.L., "Air Pollutant  Emission
   Factors", Environmental  Protection  Agency,  Preliminary
   Document, 58  (1971).
Strauss, W., "The Development of a Condenser ^ dr Con-
   trol from Dry Rendering Plants", -i -  Air Pollution Control
   Assoc.. 14  (10), 424 (1964).

Strauss, W., "Odor Control for the Process Industries", Chenr
   & Process Eng.. March  (1965).

Teller, A.J.,  "Odor Abatement 1n the Rendering and A]]{ed
   Industries", J. A1r Pollution Control Assoc., H V^/
   148  (1963).

Walsh,  R.T., Leudtke, K.D. and Smith, L.K., "F1sh Canneries
   and  F1sh Reduction Plants", cp. 11, 1n A1r Pollution
   Engineering Manual . Danlelson, J.A. ed.. Public Health
   Service, 999-AP-40, 760 (1967).

     A1r Pollutant Detection
     American  Industrial Hygiene Association, Air Pollution Manual.
        Part I, "Evaluation", Am. Ind. Hyg. Assoc. J. (1960).
     American  Industrial Hyqlene Association, "Hyq1en1c Guide
        Series", Am. Ind. Hyg. Assoc. J., January (1971).
     Barvnln,  J.. "Measuring Odor Pollution". New Scientist.
        Oct. 15 (1970).
     Beckman Instruments, Incorporated, "Hydrocarbon Analyzer",
        Bulletin P70565-1269-106F (1969).
     Bendlx Process Instruments Division, "Total Hydrocarbon
        Analyzer", Bulletin A530SB-1Q70 (1970).
     Conn, G.K. and Avery, D.G., Infrared Methods. Academic
        Press  (I960).
     Davis Instrument Division, "Davis Hallde Meter", Bulletin
        P/N 11-9000 (1971).                           	
     Davis Instrument Division, "Flame Ion1zat1on Meter", Bulletin
        11-65-2 (1965).                                   
     Davis Instrument Division, "Toxldty Limits and Detector
        Methods", Bulletin 7020763 (1963).
     Eckman, D.P., Industrial Instrumentation. Wiley, New York
     Eckman, D.P., Industrial Instrumentation. John Wiley and
        Sons,  Inc. (1950).
     Environmental Protection Agency. "Federal Register", Part
        II, 36. (21), January 30 (1971).
     Erdco Engineering  Corporation,  "Tox-Ex", Bulletin 7160
     General Electric Company, "Type  H-2, H-3, H-4 and H-5P  Leak
        Detectors and Control Units". Service Manual.
     General Monitors,  Incorporated,  "Combustible Gas Detectors",
        Bulletins on Models 510 and  410 (1970),

A1r Pollutant Detection (continued)

Hemeon, W.C.L., "Technique  and  Apparatus  for  Quantitative
   Measurement of Odor Emissions",  J.  Air Pollution  Control
   Assoc., 18 (3), 166 (1968).

Hochhelser, S., Burman, F.J.  and  Morgan,  G.B.,  "Atmospheric
   Surveillance, The Current  State  of  the Art",  Environ.
   Scl. Techno!., 5. (8), 678  (1971).

Holzbock, W.G., Instruments for Measurement  and  Control.
   Relnhold Publishing Corporation, New York  (1955).

Jacobs, M.B., The Analytical  Chemistry of Industrial  Poisons;
   Hazards and Solvent's, Intersdence  (1949); The Analytical
   Toxicology of Industrial Inorganic  Poisons (1967);  The.
   Chemical Analysis of A1r Pollutants. Intersdence  (1960).

Jaffe. H.H. and Orchln, M., Theory and Application of  Ultra-
   violet Spectroscony. John WHey and Sons, Inc. (1962).

Johnson-Williams Products, D1v. of American Bosch Jrma
   Corporation, "Super-Sensitive  Indicator", Bulletin  SS-P,
   May (1970).

Kay, K.,  "Analytical Methods Used 1n A1r Pollution Study",
   Ind.  and Eng. Chem.,., 44, 1383  (1952).

Knlebes,  D.V.,  Chlsholm, J.A. and Stubbs, R.C., Direct
   Odor  Level  Measurement  Instrumentation", Gas Age, w-ti,
   August  (1969).

Mills, J.L.,  Walsh, R.T.,  Leudtke. K.D.  and Smith. L.K..
   "Quantitative  Odor  Measurement", presented at the 56th
   Annual Meeting of  the A1r Pollution Control Association.
   June  (1963).

Mine Safety  Appliances  Company.  "B11Hon-a1re, Trace Gas
   Analyzer",  Bulletin  0706-2  (1970),

Mine Safety  Appliances  Company,  "Combustible Gas  Detection",
   "Combustible Gas Analyzer", Bulletins 0703-18 and  0703-4

Mine  Safety  Appliances Company,  "Combustible Gas Indicators",
   Bulletin  Nos.  0806-10.  0804-11  (1971).

      A1r  Pollutant  Detection  (continued)

      Mine Safety  Appliances Company, "Universal Tester*. Bulletin
        0815-5  (1971).

      Prince,  R.G.H.  and  Inc1, J.H., "The Measurement of Intensity
        of Odors",  J. Appl. Chem.. 8;. 314 (1958).

      Ruch, W,E.,  Quantitative Analysis of Gaseous Pollutants, Ann
        arbor-humphrey,  Science  Publishers, London (1970).

      Sax, N.I., Dunn, M.S., et.al. eds.. Dangerous Properties of
         Industrial  Materials, 3rd ed., Relnhold Pub. Co., New
        York  (1968).

      Scott Aviation Oaves  Products Company, "Vapotesters",
        Bulletin  Nos. 17031,  17025, November  (1969).

      Stern, A.C., ed.,  A1r Pollution. , 2nd  ed.  (1968).

      Turk, A.,  Slelk, H.  and  Messer, P.J., "Determination of
        Gaseous A1r Pollution by Carbon Adsorption", Ind. Hyq.
        Quarterly,  JJ. (1), 23 (1952).                	

      Turk, A.,  "Measurements  of  Odorous Vapors 1n Test Chambers,
        Theoretical", J.  Am.  Soc. Heating. Refrlg. & A1r Cond.
        Engrs..  (10),  55 (1963).

      United States  Public  Health Service, Selected Methods for
         the Measurement  of A1r  Pollutants, Publication No. 999-
        AP-11 (1965).

      University of  M1ch1nan,  Encyclopedia of  Instrumentation for
         Industrial  Hygiene,  Institute of Industrial Hygiene

      WHlard, H.H., Merrltt,  Jr., L.L. and Dean,  J. A., Instru-
        mental  Methods  of  Analyses, D. Van Norstrand Co., Inc.,
        Princeton,  New  Jersey,  3rd Edition (1958).

      Wohlers, H.C., "Recommended Procedures for Measuring Odorous
        Contaminants  1n  the  Field", J. Air Pollution Control
        Assoc.,  V7  (9),  609  (1967).

                   1- Report No.
3. Recipient's Accession No.
Package Sorption Device System Study
                                                               5. Report Date
                                                                 April 1973
7. Author(s)
       Juho1af  Program  Manager
                                                               8' Performing Organization Kept.
9. Performing Organization Name and Address
MSA Research Corporation
Division of Mine Safety Appliances Company
Evans City. Pennsylvania 16033
                                                               10. Project/Task/Work Unit No.
                                                               11. Contract/Grant No.

                                                                EHSD 71-2
12. Sponsoring Organization Name and Address
EPA, Office of Research and Monitoring
NERC/RTP, Control S ystems Laboratory
Research Triangle Park, North Carolina 27711
                                                               IX Type of Report & Period

15. Supplementary Notes
16. Abstract,The report. Q) identifies and characterizes the numerous small, but
objectionable,  sources of gaseous pollutants; (2) ranks sources according to relative
importance; (3) assesses  the equipment and technology available for controlling these
sources, in handbook form; (4) investigates the potential and need for development of
technology and new sorbents; and (5) details research and development recommend-
ations to improve existing devices or to develop new control methods.  Emission
sources amenable  to control by package sorption devices contribute 15% of the total
organic emissions from all  sources and can be divided into two broad categories:
(1) solvent-user industries that emit solvents essentially unchanged, and (2) process
industries that generate pollutants by chemical, biological, or thermal reactions.
Regenerative activated carbon adsorbent systems show promise for the control of
pollutants from these sources, particularly at concentrations below 700 ppm. At
                                                    higher  concentrations,  catalytic
17. Key W'otds and Document Analysis. 17o. Descriptors
Air Pollution
Activated Carbon
17b. Identifiers/Open-Ended Terms
Air Pollution Control
Stationary Sources
*Packaged Control Devices
Organic Emissions
                                                    and thermal incinerators are
                                                    more economical except where
                                                    the would-be pollutant is a
                                                    valuable solvent recoverable by
                                                    the adsorbent system.
17e. COSATI Field/Croup
18. Availability Statement
                                                    19. Security Class (This
                                                    UK ?  "^-^-'"-'irir-iJ
                                                    20. Security Class (This
                                                                        21. No. of Pages
                                                                        22. Pt
                                                                        USCOMM-OC I4J2-MI