C.f
4974
                                            600R78102
        INTERIM TREATMENT GUIDE FOR CONTROLLING
                  ORGANIC CONTAMINANTS IN
             DRINKING WATER USING GRANULAR
                     ACTIVATED CARBON
                    Water Supply Research Division
                Municipal Environmental Research Laboratory
                   Office of Research and Development
                  U.S. Environmental Protertion Agency
                          Cincinnati, Ohio
                           January, 1978

-------
                 Supporting Documentation

                            for

           U. S. Environmental Protection Agency
        Interim Primary Drinking Water Regulations
"Control of Organic Chemical Contaminants in Drinking Water"

Federal Register. 43, No. 28, 5756-5780 (February 9, 1978)
          INTERIM TREATMENT GUIDE FOR CONTROLLING
                  ORGANIC CONTAMINANTS IN
               DRINKING WATER USING GRANULAR
                     ACTIVATED CARBON
                  Compiled and Edited by

                      James M. Symons
               Water Supply Research Division
        Municipal Environmental Research Laboratory
            Office of Research and Development
                     Cincinnati, Ohio
                       January 1978            Erwbrr
                                               Rev-"' '

-------

-------
                                 TABLE OF CONTENTS
Executive Summary                                                   ES-1
         INTERIM TREATMENT GUIDE FOR CONTROLLING ORGANIC CONTAMINANTS IN
                  DRINKING WATER USING GRANULAR ACTIVATED CARBON
Introduction                                                           1
Review of Organic Removal Unit Processes                               2
     Precipitation                                                     2
     Oxidation                                                         2
     Aeration                                                          5
     Adsorption                                                        5
          Synthetic Resins                                             5
          Activated Carbon                                             8
     Summary                                                          22
Procedure for Collection of Site Specific Design Data                 23
     Collection of Treatability Data                                  23
          Performance Criteria                                        23
          Example Data                                                26
          Calculation of Adsorbent Use Rate                           27
          Summary                                                     35
     Economic Analysis                                                36
          Basic Costs                                                 36
          Influence of Empty Bed Contact Time                         43
          Influence of Type of System Chosen                          45
          Influence of Granular Activated Carbon Cost                 48
          Impact of Inflation                                         48
          Summary                                                     54
Summary                                                               54
Acknowledgments                                                       55
References                                                            55
                                   APPENDIX A
  PERFORMANCE OF GRANULAR ACTIVATED CARBON FOR THE REMOVAL OF ORGANIC COMPOUNDS
Literature Review                                                     Al
     Taste and Odor Control                                           Al
     Removal of Organic Compounds as Measured by General
       Organic Parameters                                             A5
     Removal of Specific Organic Compounds                            A9
     Reactivation of Granular Activated Carbon                       A22

-------
                                       ii
Current Water Supply Research Division Findings                    A34
     Class I Compounds (Taste and Odor Producing)                  A34
     Class II Compounds (Synthetic Source Water Contaminants)      A36
          Naphthalene                                              A36
          Other Polynuclear Aromatic Hydrocarbons                  A3 7
          Carbon Tetrachloride                                     A37
          Trichloroethylene, Tetrachloroethylene and Similar
            Chlorinated Organics                                   A40
          Prediction of Granular Activated Carbon Bed Performance,
            Studies with Humic Acid and Carbon Tetrachloride, Dieldrin,
            PCB, Benzene, and _p_-Dichlorobenzene                    A41
          Mixtures from Actual Waters                              A42
          Unidentified Compounds                                   A52
          Summary                                                  A56
     Class III Compounds (Disinfection By-Product Precursors)      A56
     Class IV Compounds (Disinfection By-Products)                 A63
     General Organic Parameters                                    A66
     Influence of Empty Bed Contact Time                           A69
     Biological Growth and Endotoxin Production                    A76
          Standard Plate Count                                     A76
          Endotoxins                                               A80
Ozone Enhanced Granular Activated Carbon Adsorption (Biological
   Activated Carbon                                                A81
     Literature Review                                             A81
     WSRD Pilot Plant Results                                      A86
Economic Analysis Information                                      A91
     Basic Costs                                                   A91
     Influence of Empty Bed Contact Time                           A92
     Influence of Granular Activated Carbon Cost                   A99
     Influence of Inflation                                        A99
     Cost of Reactivation                                         A108
     Summary                                                      Alll
Acknowledgments                                                   A112
References                                                        A113

-------
                                       iii
                                   APPENDIX B
             ANALYTIC METHODOLOGY FOR MONITORING PILOT COLUMN TESTS

Introduction                                                  Bl
Low Molecular Weight Halogenated Organic Compounds Excluding
    Trihalomethanes - Performance Criterion 1                 Bl
Organic Carbon - Performance Criteria 2 and 3                 B5
     Analytic Method                                          B5
     Sampling                                                 B6
Terminal Summation Trihalomethanes - Maximum Contaminant
    Level Requirement                                         B7
     General                                                  B7
     Effect of Time                                           B8
     Maintenance of Chlorine Residual                         B9
     Effect of Temperature                                   BIO
     Effect of pH                                            BIO
     Loss of Volatile Species                                BIO
     Effect of Bromide or Iodide Contamination               BIO
     Effect of Precursor Contamination                       Bll
     Procedure for Terminal Trihalomethane Determination     Bll
     Summary                                                 B13
Acknowledgments                                              B13
References                                                   B14
                              APPENDIX C
          DESIGN OF PILOT GRANULAR ACTIVATED CARBON COLUMNS
Introduction                                                  Cl
Selection of Materials                                        Cl
Adsorption Columns                                            C2
Establishing Test Conditions                                  C2
Operation                                                     C7
Parts List                                                    C9
Acknowledgments                                              Cll

-------
                                       ES-1





                             EXECUTIVE SUMMARY




     Low concentrations of organic compounds currently occurring in




drinking waters are a hazard to the health of the consumers (see Basis




and Purpose Dcoument) and in many cases cause drinking water to have an




unpleasant taste and odor.  This Guide concerns drinking water treatment




unit processes that are available for the control of the concentration




of organic contaminants. For the purposes of this Guide, organic contaminants




are divided into five general classes:




     Class I, organic compounds that cause taste and odor problems;




     Class II, synthetic organic chemicals that are present in source




waters from upstream discharges or runoff;




     Class III, organic compounds (precursors) that react with disinfectants




to produce "disinfection by-products";




     Class IV, organic chemicals that are the disinfection by-products




themselves; and




     Class V, natural (non-Class III) organic compounds of little direct




toxicological importance.




     The Guide (including Appendices) is a compilation, both from the




literature and from the research being conducted by the Water Supply




Research Division, of treatment techniques for controlling these various




classes of organic contaminants.  This information then leads to a




recommendation of how a water utility requiring additional treatment for




organic contaminant control should proceed towards choosing the proper




treatment design.




     Four general categories of organic treatment unit processes are




currently (Fall 1977) available.  These are: precipitation, oxidation,




aeration, and adsorption.    In the Guide each of these general categories




is reviewed and its advantages and disadvantages discussed.  Although

-------
                                     ES-2







all of these four processes do remove one class or another of organic contaminants




to some degree, adsorption using granular activated carbon is chosen as the




best "broad-spectrum" unit process for controlling concentrations of organic




contaminants. Appendix A of the Guide details what is known about this unit




process.




     To summarize, the literature shows that Class I compounds, those producing




taste and odor, are well removed by granular activated carbon adsorption.




Class II compounds, synthetic source water contaminants, however, are variable




in their adsorbability. Certain types of compounds such as pesticides, polynuclear




aromatic hydrocarbons, polychlorinated biphenyls, and so forth are very




strongly adsorbed, although low molecular weight halogenated compounds such




as carbon tetrachloride are less strongly adsorbed.  Therefore, although




granular activated carbon is effective for the control of this latter type of




compound when fresh, to maintain control, the granular activated carbon must




be thermally reactivated more frequently than when more strongly adsorbed compounds




are present in the water to be treated.  A search of the literature shows




that more than 50 different organic compounds have been investigated as to




their adsorbability on granular activated carbon at concentrations approaching




those that might be found in source  or drinking waters.  Unfortunately, many




of these studies were conducted under conditions so dissimilar to those found




in water treatment practice that they are not directly applicable to the




current problem, but taken in total they do support the concept that granular




activated carbon is a "broad-spectrum" adsorbent.




     Because the problem of disinfection by-products and their precursors is




relatively new, the literature does not contain much information on the




adsorbability of these two classes of compounds.  Some information

-------
                                          ES-3







is, however, available on the treatability of organic compounds in general,




as measured by some general organic parameters, and these results are related




to the removal of Class III, disinfection by-product precursors.  In general,




granular activated carbon beds provide good removal of organic compounds as




measured by these general organic parameters when fresh, but reactivation




would be required more often than for control of taste and odor compounds.




Many Class V compounds, low molecular weight organic acids and alcohols, are




not readily adsorbable by granular activated carbon, but these compounds are




of little toxicological significance.




     The literature shows that currently (Fall 1977) four general types of




reactivation devices are available, multiple-hearth, fluidized-bed, rotary




kiln, and infra-red tunnel furnaces.  Although some change in the reactivated




carbon pore structure and subsequent loss in adsorption capacity during




reactivation is possible, proper reactivation facility design will minimize




this.




     The in-house and extramural research program of the Water Supply Research




Division is evaluating the performance of granular activated carbon adsorption




for the control of the first four classes of organic compounds. This work is




being carried out through bench-, pilot-, and field-scale research projects.




These studies are an important adjunct to those found in the literature




because they are being conducted under conditions of "low concentrations" and




"adsorption in competition with other organics," both conditions that are




found in practice.




     Thus far, the results are supporting the literature and show that,




when fresh,  granular activated carbon can adsorb a wide variety of organic




compounds,  without significant measurable organic compounds initially leaching




off of the adsorbent.   Additionally, granular activated carbon has been shown




to adsorb milligram per liter concentrations of organic carbon, even though




the total concentration of specific organic compounds shown to be

-------
                                       ES-4






adsorbed is only in the microgram per liter range (1000-fold less).   This




further indicates that many other individual organic compounds beyond those




measurable at this time (Fall 1977) are also adsorbed.  On the other hand,




the research data show that capacity of granular activated carbon for the




adsorption of some low molecular weight halogenated organic compounds, disinfection




by-product precursors, and the disinfection by-product, chloroform,  is on the




lower end of the adsorption spectrum as compared to pesticides and taste and




odor causing compounds.  Contact times and reactivation frequencies must be




adjusted to take this adsorption variability into account.  Finally, excessive




bacterial growths or endotoxin production does not occur in granular activated




carbon beds to a significant extent.




     Although research on the use of ozonation prior to granular activated




carbon adsorption, called biological activated carbon, is not complete, initial




results are promising.  In summary, the data available, both from the literature,




and from on-going research, support the recommendation of granular activated




carbon adsorption as the best currently (Fall, 1977) available treatment for




controlling organic contaminants.




     Although granular activated carbon is chosen as the best available




treatment, the design and operation of an adsorption system for a given




utility must be site specific because of the variation in organic contaminants




found from location to location.  The technique a water purveyor should use




to collect the data needed to design and operate a granular activated carbon




facility in a given location is based on the principle of operating two small




granular activated carbon pilot column systems, one with an empty bed contact




time equal to that available in the existing filters at average flow and the




other consisting of three columns in series to evaluate the benefits of




additional contact time (see Appendix C for details) for a few months in the




summer and winter  to determine the necessary reactivation frequency.

-------
                                   - ES-5 -



To assess the performance of the granular activated carbon adsorbent,

three performance criteria, based on weekly effluent sampling have been

selected. These are:

     1)  the concentration of any of the low molecular weight halogenated

organics (excluding the trihalomethanes) that are determinable by a purge and

trap, gas chromatographic, halogen detector type analysis (see Appendix B)

shall not exceed 0.5 yg/£,

     2)  the increase in the total organic carbon concentration (see Appendix

B) shall not exceed 0.5 mg/£ over the total organic carbon concentration in

the granular activated carbon column effluent at the start of the test,  and

     3)  the initial removal of total organic carbon must exceed 50 percent,

(Adsorber Inf. TOC Cone. - Adsorber Eff. TOC Cone/Adsorber Inf. TOG Cone.) x 100 > 50

     In addition, the other organic maximum contaminant levels (MCL) must be

met; therefore, the terminal "summation" trihalomethane concentration* (see

Appendix B) and the pesticide concentrations (rarely involved) in the effluent

shall not exceed the requirements specified in the Interim Primary Drinking

Water Regulations.

     Whichever of these criteria is exceeded first, as measured by a three-

week running average (W. 4- W. , + W. «)/3, value assigned to Week   ) of

weekly effluent analyses during the pilot granular activated carbon column tests

shall determine the design factors for the granular activated carbon treatment

system at that location.

     Because the length of adsorber operating time (reactivation frequency)

before the controlling criterion is exceeded will be related to the adsorber

empty bed contact time in a given situation, the design empty bed contact

time should be chosen such that the reactivation frequency will neither
*Arithmetic sum of the concentration of the individual trihalomethane
species measured.  This parameter is called "Total Trihalomethane" in the
Regulation.

-------
                                          ES-6


be uneconomically short, nor so long as to allow the desorption of previously

adsorbed organic contaminants.  If an initial total organic carbon removal of

50 percent cannot be achieved in the pilot column using the empty bed contact

time available in the existing filters at average flow at that location, even

with the best available granular activated carbon for that situation, then

replacing the media in the existing filter boxes with granular activated

carbon will not be acceptable and post-filter adsorbers with longer empty bed

contact times will have to be used.  The pilot column tests will determine

the most economical method of operation.

     To estimate the cost of installing granular activated carbon adsorption

treatment, calculations were made for both a sand replacement and a post-

filter adsorption type design for two utility sizes based on a set of reasonable

assumptions as detailed in the Guide.   These total unit costs are as follows:

                          10 mgd plant              100 mgd plant
                   Sand       Post-Filter        Sand       Post-Filter
                 Replacement  Adsorption         Replacement Adsorption

cents/1000 gal.    13.2          16.7               6.9          8.7

Because alternative design and operation techniques yield different overall costs

for treatment, examples of how a utility should make an economic analysis to choose

the optimum design and operation configuration for that particular location is

outlined in the Interim Treatment Guide. Specifically, this analysis shows, in

general, that when a longer empty bed contact time adsorber system is considered,

unless the increase in the period between reactivations is proportionally greater

than the increase in empty bed contact time, total unit costs will rise.  The data

collected during the pilot column study outlined above will provide the information

needed to make this analysis in a given location.

-------
                                       ES-7
     In addition, a technique for taking into account future inflationary




trends is discussed, so that a water purveyor could make a choice of a




design that would be cost effective over the life of the treatment facility.




Because the use of post-filter adsorbers is less labor intensive than when




the adsorbent is put in existing filter boxes, post-filter adsorbers will




eventually become less expensive as that system is less influenced by inflation




than are sand replacement systems, although a "present worth" analysis showed




the sand replacement system to be a better investment for the assumptions in




the example.




     In summary, this Guide provides information that demonstrates that




granular activated carbon adsorption is the best available treatment technology,




Fall 1977, for treating water to remove organic contaminants, thereby improving




finished water quality and providing the American consumer with a more healthful




and esthetically pleasing drinking water.

-------
  INTERIM TREATMENT GUIDE FOR




CONTROLLING ORGANIC CONTAMINANTS




IN DRINKING WATER USING GRANULAR




      ACTIVATED CARBON
           Written by




        James M. Symons




              and




        Robert M. Clark
          Reviewed by




      Gordon G. Robeck

-------
                                INTERIM TREATMENT GUIDE




                                   TABLE OF CONTENTS




Introduction                                                          1




Review of Organic Removal Unit Processes                              2




     Precipitation                                                    2




     Oxidation                                                        2




     Aeration                                                         5




     Adsorption                                                       5




          Synthetic Resins                                            5




          Activated Carbon                                            8




     Summary                                                         22




Procedure for Collection of Site Specific Design Data                23




     Collection of Treatability Data                                 23




          Performance Criteria                                       23




          Example Data                                               26




          Calculation of Adsorbent Use Rate                          27




          Summary                                                    35




     Economic Analysis                                               36




          Basic Costs                                                36




          Influence of Empty Bed Contact Time                        43




          Influence of Type of System Chosen                         45




          Influence of Granular Activated Carbon Cost                48




          Impact of Inflation                                        48




          Summary                                                    54




Summary                                                              54




Acknowledgements                                                     55




References                                                           55

-------
   INTERIM TREATMENT GUIDE FOR CONTROLLING ORGANIC CONTAMINANTS




           IN DRINKING WATER USING GRANULAR ACTIVATED CARBON







                              INTRODUCTION




     Historically the problem of organic contamination of drinking water was




related to taste and odor producing compounds.  Although of only indirect




health significance, these compounds are still important as they are the ones




that make drinking water disagreeable and cause some consumers to use poorly




controlled sources of drinking water, bottled water or water treated by home




treatment devices.  In the past five years, through advances in organic




analytic techniques and understanding of organic reactions, three other




classes of organics, although undetected by consumers, have been shown to be




important in drinking water.




     The second of the four classes of concern is contaminants found in




source waters that are of industrial origin.  These are contained in industrial




wastes, municipal wastes, and urban and rural runoff and have a potential




adverse health effect, (see Basis and Purpose).    The third of the four




classes of organic compounds is that with which disinfectants react to produce




disinfection by-products (precursors).  These are largely, but not entirely,




of natural origin, arising from urban and rural runoff and municipal wastes




and have a potential secondary health effect, creating carcinogens and toxicants




when reacting with disinfectants.   The fourth class of organic chemicals is




the disinfection by-products themselves, (trihalomethanes if chlorine is the




disinfectant), either in the source water from upstream disinfection practices




or created in the water treatment process itself.  These too have a direct




potential adverse health effect (see Basis and Purpose).  A fifth class, of




little concern, is the organics of natural origin that are not disinfection




by-product precursors.

-------
                                    - 2 -





Because such a small percentage of the total organic contaminants in drinking




water can be identified and quantified, estimated to be less than 10 percent,




other classes may be identified in the future.




     Therefore, the first four classes of organic compounds should be controlled




to provide consumers with a wholesome and taste-and-odor-free drinking water.




The purpose of this Interim Guide is to summarize what is known about organic




removal unit processes and to recommend a procedure that will allow those




utilities requiring additional organic contaminant control to select an




appropriately sized treatment facility.




                 REVIEW OF ORGANIC REMOVAL UNIT PROCESSES




     The advantages and disadvantages of each of the four general types of




organic removal unit processes will be briefly reviewed.




Precipitation




     The addition of a metal coagulant for the removal of particulates or




raising the pH for removal of calcium and magnesium results in the removal of




some organic matter.  This can be measured both by the reduction in the




organic carbon concentration, see Figure 1, and color. Although partially




effective for the removal of disinfection by-product precursors, see Figure  2




using chloroform formation as an example, this process does little for the




reduction of the other classes of organic compounds that' are in true solution.




Oxidation




     A second possibility for organic removal is to treat the organics with




an oxidizing agent.  If oxidation could be complete, this would be an attractive




approach as the major end-products would be carbon dioxide andXwater.  Although




some organic oxidation has been demonstrated when oxidants such as ozone,




chlorine dioxide, chlorine, and potassium permanganate are added to water,

-------
                             -3-
 O)
 E
CO

 I

CM
 •
CM

II

LLJ
o
z

                                 2
<
o

o


CO


S
2; O


z ^


LJJ < Z


< ° ^
UI LU /-N
^ o: 2
                              3
                              3

                              0
                                                    Z
                                                    O
                                                    <
                                                    u
                                CN
           31V3S 3AI1V13H

-------
                           -4-
\
                                                 00

                                                 •O
                                                 CN
                                                   U
                                                   O
                                                   8
                                                 . 2
                                                 * X
                                                   u
                                                 CN
                                          DC

                                          O
                                          LL

                                          O
                                          CC

                                          O


                                          X

                                          o
                                                   o  t
                                                         _
                                                       cc <
                                          "-2
                                          0°-

                                          ^s
                                          oQ
                                          Zh-
                                          LU <




                                          II
                                          = li.

                                          CM

                                          LU
                                          CC
                                          3
                                          O

                                          li.
      o
      CN
O

O
o
00
o
•o
o
CN
                            WilOdOilOIHD

-------
                                         - 5 -





doses must be uneconomically high to insure nearly complete oxidation, see Figure




3.  If oxidation is incomplete, the organic compounds will not be oxidized to




carbon dioxide and water, but only to some intermediates of unknown toxicologic




hazard.  Thus far, even the addition of ultra-violet radiation to an ozonation




process has not been demonstrated to cause nearly complete oxidation for many




compounds.




Aeration




     Because many of the organic compounds of concern that have recently been shown




to occur in drinking water were determined by a procedure that employs a concentration




step based on purging the compounds from the sample with diffused gas, aeration,




either diffused air or spray, was considered as a possible organic removal process.




The advantage of such a process would be that the organic compounds are truly




removed from the water without the formation of by-products or intermediates.  The




disadvantages are at least two-fold; one, the removed materials are transferred




from the water phase to the air phase, possibly causing another problem, and two,




only organic substances with certain characteristics are removable even with excessive




quantities of air, see Figure 4.  Therefore this approach is self-limiting.




Adsorption




     A fourth possible approach to control of organic contaminants is to remove




them from the water by adsorption onto a surface. Two types of adsorbents are




currently commercially available in the United States. These are:  1) synthetic




resins and 2) activated carbon, either in the powder or granular form.




     Synthetic Resins




     Synthetic resins of the Amberlite XAD type (a class of macroreticular




divinylbenzene-styrene or acrylic copolymers) have a very limited capacity for

-------
                           -6-
   1.6
   1.4
   1.2
   1.0
o>

 *>
 m
o

o
o

O  0.8
O
Q.
z
u  0.6
UJ
   0.4
   0.2
          O
          O
          Q.
        O)
        E
        O
        o
        Q.
                           6" DIA. PYREX CONTACTOR

                           STAINLESS STEEL SPARGER

                           COUNTER FLOW

                           13' WATER  DEPTH

                           CINCINNATI TAP WATER

                           03 GENERATED FROM 02
             o>
             E
             O
             o
             Q.
                      O)
                      E

                    O)  ^
                   O
                   2
                                               0>
                                               E
                                               
-------
                  -7-
l/Bui 'NOIiVHiN3DNOD NO9HVD DINV9HO 31flV39cind-NON
               oo
               d
             •o
             6
   <->
  u o
                         o _
                                         o
                                         m
                                         m
                                         CO
                                         £ 2
                                              CO
                                              <
                                              LLJ
                                              (!)
                                              ce.
                                              D
                                              Q_
                                     O
                                     z
                                     <
                                     LJJ
                                     Z
                                     <
                                     X
                                     I—
                                     LU
                                         cs
                                   LU —
                                   5 *
                                   < t-

                                   ^ Z
                                   o o
                                         S5P
                                              z±
                                              LU 
-------
                                       - 8 -


adsorption, although Ambersorb XE-340 (a carbonaceous adsorbent, chemically

intermediate between the XAD type and activated carbon and manufactured by

Rohm and Haas*),  has a high adsorption capacity for organic compounds such as

chloroform and carbon tetrachloride, but does not have much capacity for the

removal of a wide range of organic compounds as measured by organic carbon

concentrations, see Figure 5.

     Activated Carbon

     Activated carbon is a form of carbon that is activated by a carefully

controlled oxidation process to develop a porous carbon structure with a

large surface area.  This large surface area gives the activated carbon a

high capacity to adsorb dissolved organic materials from water. The major raw

materials used in the manufacture of granular activated carbons are petroleum

coke, bituminous coal and lignite.  After preliminary processing, these

materials are heated to a high temperature and reacted with steam to develop

the extensive internal pore structure required for adsorption.  The substance

is then crushed, graded, screened, and packaged.  The raw material and activating

conditions used affect adsorption and physical properties of the carbons

produced.  For additional information the reader is referred to the Journal

of the American Water Works Association,  66, No. 11, 672-681  (1974) for the

AWWA Standard for Granular Activated Carbon, AWWA, B604-74.

     Powdered activated carbon is a good adsorbent with a high surface area

per gram of material, but as it is usually added to the water  to be treated

in a single dose, an equilibrium is established between the organic compounds

and the adsorbent, thereby limiting the extent of removal possible.  Table I

shows that high doses of powdered activated carbon were unable to reduce the

Threshold Odor Number to acceptable levels  (three or less)  in this instance.

Further, Figure 6 shows that high doses of powdered activated  carbon were
  *Mention of commercial firms does not imply endorsement by the U.S. Environmental
Protection Agency.

-------
     -9-
  l/Bui 'NOI1VH1N3ONOO DO1
l/6n 'NOHVaiN30NOO WHIZ
CC

g
LJ.

-------
                                        - 10 -



                                        TABLE I

Powdered Activated Carbon Dose for Nitro, West Virginia Water Plant - 1963
                                    Average Monthly      Finished-Water
                                   Powdered Activated    Threshold Odor No.*
                                   Carbon Dosage
Month
January
February
March
April
May
June
July
August
September
October
mg/£
21
27
21
19
20
20
19
22
22
42
Min.
8
8
6
10
8
8
7
5
8
8
Max.
35
50
26
60
30
19
14
26
26
40
Avg.
21
25
12
17
14
11
9
12
12
13
*At 60°C.

-------
                       - 11 -
|/6rv'iviJ.N3JLOd NOIiVWaOd WUOJOHOIHD
                                                        O
                                                        u.
                                                        O
                                                        X
                                                        u

                                                        z <
                                                        O P



                                                        25
                                                        a> Q.

                                                        < Z
                                                        u £
                                                        < 5
                                                        > *
                                                        i= O
                                                        U "-
                                                        So
                                                        UJ LL.
                                                        o O
                                                        > as

                                                        o2
                                                        o- I
                                                        u. U
                                                        3 O
                                                        "; ^
                                                        z <
                                                        QC U
                                                        3 Z
                                                        O O

-------
                                    - 12 -



required to greatly reduce the concentration of chloroform in tap water, as


well as the concentrations of the disinfection by-product precursors as

                                                        2
measured by the trihalomethane formation potential test.   Note, because of


the problem of interference of residual powdered activated carbon with the


total organic carbon test, evaluating the ability of powdered activated


carbon to remove organic carbon is difficult.


     In contrast, when activated carbon is used in a column, it acts like a


series of powdered activated carbon "doses" being applied to water of continually


improving quality as the water passes through the column.  This "plate effect"


permits an activated carbon column to produce water containing much lower


concentrations of organic matter when contrasted to reasonable additions (10


mg/£ or less) of powdered activated carbon.


     An example of excellent performance of granular activated carbon for


removing Class I type organic compounds, taste and odor producing, is shown


in Figure 7.  In this installation, the granular activated carbon replaced


the sand in existing filter boxes and under these circumstances controlled


tastes and odors for more than 1 year.


     An early example demonstrating the ability of granular activated carbon


beds to adsorb Class II type organic compounds, source water contaminants,


showed that seven compounds, listed below, present in the Kanawha River after


aeration could be reduced to below detectable concentrations by fresh (2-day


old) granular activated carbon beds. .  These compounds were bis-(2-chloroethyl)


ether, 2-ethylhexanol, bis-(2-chloroisopropyl) ether,  ct-tnethylbenzyl alcohol,


acetophenone, isophorone, and tetralin.  A recent review of the literature


(see Table II) has expanded this list considerably.

-------
-13-
                                 O

                                 O
                                 oe.

                                 i— ot.
                                 ^ LU
                                             CO

                                             £*
                                             LU
                                             CO
                                             oc
                                             o
                                             o
                                             O
                                             O

                                             O
                                             X
                                             CO
                                   z
                                   o
                                   CQ
                                   a*
U


£
LU
z
                                      CO
                                      CO
                                      <
                                   o  .
                                 LU
                                      LU
                                   LU
                                 O
              CO
                   es    —
XBBM ei
    NOI

-------
                                     - 14 -
                                   TABLE II
ORGANIC COMPOUNDS THAT HAVE BEEN REPORTED IN THE LITERATURE* TO BE ADSORBABLE
                         ON GRANULAR ACTIVATED CARBON
     acetophenone
     aldrin
     baygon
     a-BHC
     benzocaine
     benzole acid
     butyric acid
     bis-(2-chloroethyl)  ether
     bis-(2-chloroisopropyl)  ether
     dibrom
     dieldrin
     di (n-butyl) phthalate
     di (2-ethylhexyl) phthalate
     diuon
     diquat
     dimethoate
     m-dinitrobenzene
     DDT
     endosulfan
     endrin
     gasoline
     heptachlor
     heptachlor epoxide
     hexachlorobenzene
     isophorone
     juglone
     lindane
                                          linuron
                                          MS-222
                                          malathion
                                          ct-methylbenzyl alcohol
                                          methyoxychlor
                                          nitrobenzene
                                          oil (fuel)
                                          paraquat
                                          parathion
                                          phenylacetic acid
                                          phenols
                                          p-nitrophenol
                                          propionic acid
                                          pyridine
                                          PCB
                                          rotenone
                                          sevin
                                          simazine
                                          strychnine
                                          3-trifluoromethyl-4-nitrophennol (TFM)
                                          2-ethylhexanol
                                          2,4-D
                                          2,4-dinitrophenol
                                          2,4,5-T (ester)
                                          toxaphene
                                          tetrachlorobenzene
                                          telodrin
                                          triazine
                                          tetraline
*References given in Appendix A.   Studies with original concentration 1 mg/£
or less.

-------
                                     - 15 -
     Current research efforts by the Water Supply Research Division continue to




demonstrate the ability of granular activated carbon to adsorb a variety of synthetic




organic contaminants.  Figure 8 is an example of these data, showing the adsorption




of carbon tetrachloride by granular activated carbon.  The data in Table III indicate




the adsorbability of other low molecular weight halogenated organic compounds by




granular activated carbon.  The data in Figure 9 demonstrate that with Ohio River




water, granular activated carbon adsorption treatment can initially produce an




effluent containing only approximately 0.1 mg/£ of organic carbon.  This shows that




under these circumstances, when fresh, the granular activated carbon adsorbed over




one milligram per liter of organic carbon.  Because most of the specific organic




contaminants found in drinking water occur at \ig/H concentrations, 1000-fold less,




this further indicates the ability of granular activated carbon to adsorb a wide




variety of organic contaminants. Finally, the data in Figure 10 further demonstrate




this point, as indicated by the absence in the granular activated carbon bed




effluent of many of the organic compounds that show up as "peaks" on the capillary




gas chromatogram of the influent to the granular activated carbon column.




     Figure 11 is an example of data collected by the Water Supply Research Division




at one of its large-scale field research projects.  These data show the ability of




granular activated carbon to adsorb Class III type organic compounds,  disinfection




by-product precursors.  The data in Figures 12A and 12B show that, when fresh,




granular activated carbon has the ability to adsorb Class IV type organic compounds,




disinfection by-products.  The bromine-containing trihalomethanes are well adsorbed,




chloroform breaking through the granular activated carbon treatment systems first.

-------
                     -16-
|/Brv 'NOIlVdlNBONOO

-------
                                    - 17 -


                                    TABLE III

             USE OF GRANULAR ACTIVATED CARBON BEDS TO REMOVE ORGANIC

                     CONTAMINANTS FROM A GROUND WATER SUPPLY
                                    Influent
                                    Concentration     Average Effluent Cone.,
Contaminant                         Range, yg/£     0-4 weeks  4-8 weeks  8-12 weeks
1,1, 1-trichloroethane
trichloroethylene
tetrachloroethylene
10.9-36.7
1.9-7.7
0.1-16.9
NF NF
NF <0.1
NF <0.1
0.4
<0.1
<0.1
NF = None found

Note:  After six weeks of testing, two new contaminants, 1,1-dichloroethane
and cis-1,2-dichloroethylene began appearing in the influent to the granular
activated carbon column.  No cis-1,2-dichloroethylene has yet been detected
in the treated water, but the average concentration of 1,1-dichloroethane
in the effluent for the 4 to 8 week period after first appearing was 0.7

-------
                     -18-
                                                      LU
                                                      >
                                                      cc
                                                      D
                                                      o

                                                      X
                                                      o
                                                      ID
                                                      O
                                                      cc
   H
   Z
   UJ
                                                    HI
                                                    UJ


                                                    uT
                                                    O

                                                    cc
                                                    UJ
                                                    UJ
l/Buj '(DO1) NOIlVaiN30NOO NOQdVO DINVOdO 1V1O1
LU
CC
CQ
Z
O
CQ
CC
<
O
o
Z
<
o
cc
o
o>

LU
cc
ID
O
LL
LU
CC
h-

cc
UJ
I-
o
                                                         cc
                                                         O

-------
     -  19 -
             UJ
              CO
             — CD i LLJ
             I- CC [Zh- UJ
             O < U. U.
             
             UJQ-W
              4
8*
5£
32
o o
UJ o
                            CC DC
                            O <
                            0<
                              o
                            UJ Z

                            Z <

                            o2
                            z <
                            < DC
                            OH
                            OCX
                            OS
                            &
                            2? 
-------
                                 -20-
  260
  240
  220
  200
   180
   160
O
o
O  140
UJ
O
a.
5
cc
O
LL
UJ
z
5
O
z
120
   100
 80
 60
   40
   20
                                                  ADSORBER
                                                  INFLUENT
                               LOWER MISSISSIPPI TREATMENT PLANT
                               POST-FILTER ADSORBER
                               INITIAL EMPTY BED
                               CONTACT TIME 28 MINUTES
                                                 ADSORBER
                                                 EFFLUENT
                      6     8     10    12
                       TIME IN SERVICE, WEEKS
                                          14
16
18
FIGURE 11. BREAKTHROUGH PATTERN OF CLASS m ORGANIC
          CONTAMINANTS,DISINFECTION BY-PRODUCT PRECURSORS

-------
                            -21A-
                                                            DC
                                                            LU
  UJ
                          UJ
                                       •**     *vi

                                       5    -501

                                       *l5ljj
                                       o    z o» c
                                       DC OQ — CM I-
                                       < oc o .. i-
                                       5 O W O O
                                       ^ CO ~Z <
                                       OQ < p Q P
                                       UJ OC K 9 Z
                                       2  .. X O Q
                                       ~  III l_ — III
                                         uj
                                       UJ

                                                UJ
                                                OQ
CO O Q Q £
tu < uj >- 5
I- O OQ X UJ
                   oc
                   O
                                                    O
                                                    oc
                                                    ffl
                           z

                           CL
                           <



                           ig
                           O CQ
                           _r CC
                           h- <
                           < O

                           * °
                           ± UJ
                           O H
                           S O

                           O <
                                                            UJ
1/Br/ 'SNOI1VU1N30NOO
                                            CM
3NVH13l/VCnVHIdl
                                                            X  CC

                                                            I-  O
                                                               OQ

                                                            ui
                                                            CC
                     CM
                     r-

                     UJ
                     CC
                     3
                     g

                     u.

-------
                                  -21B-
                                                                       CC
                                                                       LU
                           -CO
                           -CM H

                               CC
                               Ul
                               Q.
                               O

                             r- Z

                           E>   uj

                           D   ?
                           I)   "-
O
CO
 I
O
CM
                               V)
                               X
                           -co
                           -CM
                               oc
                               UJ
                               Q.
                               O
                               Ul
o       o
in       o
       'SNOIlVaiN3ONOO
                                        3NVH13WO1VHIU1
                                                              UJ
                                                              QC

                                                              g
                                                              u_

-------
                                      - 22 -
     The disadvantage of any adsorption process is that the adsorbent's




capacity has a finite limit.  This means that the adsorbent either must be




replaced, or processed in a manner that will restore its adsorptive capacity.




In the case of granular activated carbon, as it is contained in a vessel, in




contrast to powdered activated carbon, it can be removed, dried and thermally




reactivated such that the adsorbed organic materials are driven off the




activated carbon and the surfaces reactivated for further use.  Note, proper




treatment of the off-gases is necessary to avoid an air pollution problem.




Summary




     Taking into account all of the advantages and disadvantages of the




various organic removal unit processes briefly discussed above, granular




activated carbon has been selected as the best "broad-spectrum" process for




the removal of organic compounds, without the production of by-products,




currently available at this time (Fall 1977).  Appendix A contains the detailed




analysis of the literature and the Water Supply Research Division research




findings that support this conclusion.  In addition, Appendix A contains a




brief review of a variation of the usual granular activated carbon adsorption




process called biological activated carbon employing ozonation preceding




adsorption.  The next section of this Guide will present the technique a




water purveyor should use to develop the necessary design and operation data




needed to install a granular activated carbon adsorption system.

-------
                                        - 23 -




               PROCEDURE FOR COLLECTION OF SITE SPECIFIC DESIGN DATA




Collection of Treatability Data




     Performance Criteria




     Having chosen granular activated carbon as the best "broad-spectrum"




organic removal unit process currently (Fall 1977) available, several questions




still remain for a given utility required to install this form of treatment.




These are: 1) how should the facility be designed and operated to optimize




performance and 2) which of the currently commercially available granular




activated carbons are best for a given location.  To begin to answer these




questions, the objective of treatment must be defined.  In general terms, the




treatment process should remove, to the degree possible, four of the classes




of organic compounds defined above;  taste and odor producing components,




source water synthetic contaminants, disinfection by-product precursors, and




disinfection by-products.  The fifth class, natural organic compounds that




are not disinfection by-product precursors, are not easily controlled by




granular activated carbon adsorption, but are of little public health concern.




Further, the reactivation cycle should neither be uneconomically short, nor




so long as to allow the desorption of previously adsorbed material,  see Figure




8.




     Within these general constraints, how are the two basic questions to be




answered?  Some performance criteria for the adsorption process must be




chosen so that it can be properly designed and operated.  Experience has




shown that wide differences exist in the adsorbability of various organic




compounds, see Appendix A.    Some toxic chemicals such as pesticides,  and




polynuclear aromatic hydrocarbons are strongly adsorbed as are taste and odor




compounds and should not be a limiting performance factor.  Toward the other




end of the adsorption spectrum are the low molecular weight halogenated




compounds, chloroform,  carbon tetrachloride and so forth.

-------
                                     - 24 -


     Because of the variability of adsorption among organic compounds and of

adsorption competition among different organic compounds, the design and

operation of an adsorption facility must be site specific.  To allow the

evaluation of granular activated carbon as a treatment process in any given

location, performance criteria have been chosen, as detailed below, beyond

the established maximum contaminant levels (MCLs) for trihalomethanes and

pesticides.  In a given location, any one of these performance factors may

be the one that controls design and operation, because to assure maximum

protection for consumers, all three and the MCLs must be met.

     The first performance criterion is:  Although EPA is not now setting

MCL's for other specific organic compounds, reduction of exposure to known

or suspected carcinogens in drinking water to the extent possible is encouraged.

Because evidence is mounting that low molecular weight halogenated compounds

such as carbon tetrachloride, tri- and tetrachloroethylene, vinyl chloride,

and so forth are a health hazard, and are adsorbable by granular activated

carbon, the treatment process should be designed and operated so that the

granular activated carbon is reactivated* when  the concentration of any

organic compound of this type (excluding the trihalomethanes) exceeds 0.5

yg/£ in the effluent.  Additionally, these organic compounds can be measured by

a single analytic procedure, see Appendix B.

     This  criterion leaves a gap in contaminant  control.  Neither known

carcinogens,  such as benzene (an unchlorinated  compound), nor toxicants

that are unable to be analyzed at this time  (Fall 1977) would necessarily be

controlled if only the above criterion were used. Therefore, the second

performance criterion must be related to some group parameter(s). As a group

parameter, organic carbon has merit, as any deterioration in granular
*For this discussion reactivation of the granular activated carbon is assumed
rather  than replacement.

-------
                                        - 25 -
activated carbon bed performance with respect to this parameter would indicate

an increase in the concentration of some unknown organic compound(s) of unknown

hazard in the treated water.  Therefore, to maximize protection for the consumer,

the treatment process should be designed and operated so that the granular

activated carbon is reactivated when the increase in the concentration of total

organic carbon (TOC) exceeds 0.5 mg/£ over the initial effluent TOG concentration

(see Appendix B),  the second performance criterion.

     Finally, to assure that the granular activated carbon is performing adequately

at the start of the test, the third performance criterion is that the initial

TOC removal ((Adsorber Influent TOC Cone. - Adsorber Effluent TOC Cone.)/Adsorber

Influent TOC Cone.)) x 100, must exceed 50 percent.

     In addition to these criteria, to meet the maximum contaminant level (MCL)

portion of the Regulation, the treatment process should be designed and operated

so that the granular activated carbon is reactivated when: 1) the arithmetic

sum* of the terminal trihalomethane concentrations, a technique for estimating

compliance with the Regulations, (see Appendix B), or 2) the pesticide concentrations

(rarely involved) exceed the MCLs (0.10 mg/£ proposed for the trihalomethanes)

specified in the Interim Primary Drinking Water Regulations.

     These performance criteria should be applied as follows.  The simplest

method of retrofitting a large water treatment plant with granular activated

carbon would be to replace the present filter media with granular activated

carbon (sand replacement) and add a reactivation facility. To investigate this

mode of operation in a particular location, a small granular activated carbon

column with an empty bed contact time, volume/flow (EBCT), similar to  that

available in the existing filter boxes during average flow should be fabricated

(see Appendix C) and used to treat settled water. The effluent concentration of:

1) Any low molecular weight halogenated compounds present, 2) Total Organic

Carbon, and 3) Terminal summation trihalomethane (see Appendix B)
*Called "Summation" Trihalomethanes in the Guide and Appendices and "Total"
Trihalomethanes in the Regulations.

-------
                                     -26 -


should be determined weekly to show how long good performance can be maintained

with that particular adsorbent and EBCT.  When Criterion one or two or the MCL

are exceeded, using a three week running average, (W.  + W    + W  «/3, value

assigned to Week.  ) the granular activated carbon would have to be reactivated.
                1+
If Criterion three is exceeded initially, this EBCT is not adequate.

     To investigate the use of a granular activated carbon system with a longer

EBCT, a second system consisting of three pilot granular activated carbon columns

in series each with a 10 minute EBCT should be operated simultaneously with the

one described above on filtered water (see Appendix C).  Sampling after each of

these three additional columns will show the influence of longer contact time on

reactivation frequency as measured by all the Criteria.  Pages A69 to A76 of

Appendix A present an example of making this type of investigation and an analysis

of these data are presented below.  Finally, because experience has shown that

Terminal ETHM concentrations are likely to be greater in the summer and the

concentration of the types of organic compounds related to Criterion one are

often higher in source waters in the winter, collecting these pilot column data

in both seasons would be very helpful.

     Another major question beyond mode of operation, empty bed contact time,

and reactivation frequency is which commercially available granular activated

carbon is most cost effective for a given location.  A utility wishing to evaluate

the performance of competitive commercial products could do so by expanding the

test procedure outlined above to include parallel units containing different

kinds of adsorbents.  Possibly the adsorbent suppliers would help in this evaluation.

     Example Data

     Data collected by the Water Supply Research Division in three locations,

the lower Mississippi River, the central Ohio River, and a southern Florida

ground water will be used to show how these performance criteria would be

-------
                                      - 27 -







applied.  Tables IV and V contain data from a full-scale installation in




the lower Mississippi Valley, rather than the recommended pilot granular




activated carbon columns, but do show that for this installation Criterion




1 and 2 are exceeded after two plus weeks for the sand replacement




system, while the reactivation frequency could be extended to six weeks




through the use of post-filter adsorbers.  The difference in EBCT is too




small to account for this difference.  For the data in Table VI, collected




with a pilot column operating on tap water from the central Ohio Valley,




Criterion 1 controlled after 7 weeks, followed by Criterion 2 on the 8th




week.  The MCL was not exceeded until the 24th week.




     The data in Tables VII, VIII and IX were prepared from pilot granular




activated carbon column studies conducted in southern Florida on a ground




water containing high concentrations of organic carbon and trihalomethane




formation potential,  as well as several low molecular weight halogenated




organic compounds.  These data show that the short EBCT does not provide




good performance for very long and leads to the conclusion that longer EBCTs




are necessary.  Pilot columns with longer EBCTs are now being employed at




this site (see pages A69 to A?6 in Appendix A).  Tables VIII and IX present




data from repeat runs and indicate the general reproducibility of this




approach.




     Calculation of Adsorbent Use Rate




     The data collected by the procedure summarized above can be used to




calculate the use rate (dose) in weight per unit volume of water of granular




activated carbon needed to meet the controlling criterion for any empty




bed contact time.  Adsorbent use rate is calculated by dividing the dry weight




of adsorbent for a given empty bed contact time by the total volume of water




passing through the column(s) until each criterion and the MCL is exceeded.

-------
                                       ~  28 ~
                                    TABLE IV
      APPLICATION  OF PERFORMANCE CRITERIA FROM LOWER MISSISSIPPI PROJECT

                SAND REPLACEMENT - AVERAGE EBCT - 21 min.

Time,
Weeks
Performance
Criteria:

0
1
2+
3
4
5
6
7
8
9
10+
11+
12+
13+
14+
15+
16+
Criterion 1 Criterion 2 Criterion 3
1,2-Di- Tri-
chloro- chloro- TOC
ethane ethylene TOC Removal
Cone. Cone. Cone. Percentage
yg/£ yg/£ mg/Jl %
< 0.5 < 0.5 < A0.5 Initial
over TOC0 > 50%
NF 0.8 77
0.25 NF
0.7* NF 2.0*
1.5 NF 2.0
3.7 NF 1.9
NF
NF
NF
NF
0.03
0.03
0.03
NF
NF
NF
NF
NF
MCL
Term.
ETHM
Cone.
mg/£
< 0.10


< 0.01
0.05
0.07
0.11*
0.13











NF - None found
* - Performance criterion exceeded
Note:  All data are three week running averages, except for TOC

-------
                                        -  29-

                                     TABLE V

APPLICATION OF PERFORMANCE CRITERIA FROM LOWER MISSISSIPPI PROJECT

                 POST-FILTER ADSORBER, AVERAGE EBCT - 24 min.
Time,
Weeks,
  Criterion 1
1,2-Di    Tri-
chloro-   chloro-
ethane    ethylene
Cone.     Cone.
                                    Criterion 2
                                    TOC
                                    Cone.
Criterion 3      MCL

TOC              Term.
Removal          ZTHM
Percentage       Cone.
Perform-
mance
Criteria:
           < 0.5
                     < 0.5
                         mg/£
                         < A0.5
                         over TOC
Initial
> 50%
NF - None Found

* - Performance criterion exceeded

Note:  All data are three week running averages > except for TOC
mg/£
< 0.10
0
1
2+
3
4
5
6
7
8
9
10+
11+
12+
13+
14+
15+
16+
NF
0.2 NF
0.1 NF
0.0 NF
0.1 NF
0.5 NF
0.7* NF
0.9 NF
NF
NF
0.1
0.1
0.1
NF
NF
NF
NF
0.8
0.9
0.8
0.8
0.8
1.0
0.9
1.0
1.1
1.3*
1.3






73
<0.01
0.01
0.02
0.02
0.02
0.03
0.04
0.07
0.08
0.09
0.09
0.09
0.09
0.09
0.10
	

-------
                                -30  -
                              TABLE VI
APPLICATION OF PERFORMANCE CRITERIA FROM OHIO RIVER TAP WATER
  POST-FILTER ADSORBER - UPFLOW MODE - EBCT - 11 minutes
Criterion 1
Carbon Tetra-
Time, chloride
Weeks Cone.
Performance yg/£
Criteria < 0.5

0 ND
1 NF
2 NF
3 NF
4 NF
5 NF
6 NF
7 1.7*
8 3.3
9 5.3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
ND - Not determined
NF - None Found
Note: All data are three week rv
Criterion 2 Criterion 3 MCL
TOC Term.
TOC Removal ETHM
Cone. Percentage Cone.
mg/£ % mg/£
 50%
ND ND ND
0.34 77 NF
0.28 < 0.01
0.31 < 0.01
0.40 < 0.01
0.41 < 0.01
0.57 < 0.01
0.71 0.01
0.85* 0.02
1.21 0.02
0 .03
0.04
0.04
0.05
0.06
0.06
0.07
0.07
0.08
0.08
0.09
0.10
0.10
0.10
0.11*
* - Performance Criterion Exceeded

inning averages, except for TOC_-

-------
                                    - 31 -

                                 TABLE VII
    APPLICATION OF PERFORMANCE CRITERIA FROM SOUTHERN FLORIDA GROUNDWATER
               RAW WATER APPLIED TO COLUMN, EBCT -6.2 minutes

Time,
Weeks
Performance
Criteria
0
1
2
3
4
5
6
7
8
9
10
11
12
13+
14
15
16

A
Cone.
PgM
<0.5
-
NF
NF
<0.01
<0.01
<0.01
<0.01
0.01
0.02
0.04
0.05
0.06
0.27
0.39
0.67*
0.70
_
Criterion 1

B C
Cone. Cone.
yg/£ yg/Jl
<0.5 <0.5
-
NF 0.13
<0.01 1.0*
0.3 2.4
0.04 4.9
0.03 5.8
0.01
0.02
0.02
0.03
0.06
0.05
0.15
0.17
0.30
0.30
_
Criterion 2 Criterion 3 MCL
TOC Term.
D TOC Removal ZTHM
Cone. Cone. Percentage Cone.
yg/£ mg/£ % mg/£
<0.5  50%
3.3 67
NF 4.9* 0.33*
<0.01 6.1 0.51
<0.01 7.6 0.59
<0.01
NF
NF
NF
NF
NF
<0.01
<0.01
0.03
0.05
0.06
0.06
_
NF = None found
* = Performance criterion exceeded
Note:  All data are three week running average, except for TOC .
                                                              o

A = trans -1,2-Dichloroethylene
B = 1,1- Dichloroethane
C = cis-1,2-Dichloroethylene
D = Trichloroethylene
Note:  These data are for illustrative purposes only and are not indicative
of final design and operation based on longer EBCTs, see pages A69 to A76.

-------
                                  -  32  -
                                  TABLE VIII
     APPLICATION OF PERFORMANCE CRITERIA  FROM SOUTHERN FLORIDA GROUNDWATER
             FILTERED WATER APPLIED  TO  COLUMN,  EBCT -  6.2  minutes
Criterion 1

Time,
Weeks
Perfor-
mance

A
Cone.
yg/£
<0.5

B
Cone.
yg/&
<0.5

C
Cone.
yg/&
<0.5

D
Cone.
yg/£
<0.5

E
Cone.
yg/£
<0.5 <
Criteria
0
1
2
3
4
5
6
7
NF
0.13
0.13
0.13
0.04
0.04
0.04
_
NF
NF
0.18
0.18
0.39
0.21
0.21
_
NF
3.57*
4.30
5.25
4.65
6.85
9.19
_
NF
0.20
0.21
0.21
0.01
<0.01
<0.01
_
ND
0.40
0.38
0.32
1.68*
2.65
3.27
_
Criterion 2 Criterion 3 MCL
TOG Term.
TOC Removal ETHM
Cone. Percentage Cone.
mg/£ % mg/A
: A0.5 Initial <0.10
over TOC0 >50%
- -
1.5 76 0.16*
4.3* 0.22
5.2
4.5
4.5
4.8
_
ND = Not Determined
NF = None Found
* = Performance criterion exceeded
Note:  All data are three week running average > except for TOCQ.

A - trans-1,2-Dichloroethylene
B = 1,1-Dichloroethane
C = cis-1,2-Dichloroethylene
D = Trichlorethylene
E = Vinyl chloride
Note:  These data are for illustrative purposes only and are not indicative
of final design and operation based on longer EBCTs, see pages A69 to A76.

-------
                                 - 33 -
                                    TABLE IX
     APPLICATION OF PERFORMANCE CRITERIA FROM SOUTHERN FLORIDA GROUNDWATER
               FILTERED WATER APPLIED TO COLUMN, EBCT - 6.2 minutes
Criterion 1

Time,
Weeks,
Perform-
ance
Criteria
0
1
2
3
4
5
6
7

A
Cone.
yg/&
<0.5

-
NF
NF
NF
NF
NF
NF
0.2


B
Cone.
ygM
<0.5

-
NF
0.
0.
0.
0.
0.
0.



37
43
43
17
37
60*


C
Cone.
ygM
<0.5

-
NF
0.
1.
5.
6.
9.
8.



43
70*
40
80
53
87

D
Cone.
yg/&
<0.5

-
NF
NF
NF
NF
0.07
0.17
0.37


E
Cone.
yg
50%
.4 93
.3*
.5
.9
.6
.0
.4

MCL
Term.
ETHM
Cone.
mg/&
<0.10

-
0.12*
0.23
0.37




NF = None found
* = Performance criterion exceeded
Note:  All data are three week running average, except for TOC

A =  trans-l,2-Dichloroethylene
B = 1,1-DLchloroethane
C = ci^s-1,2-Dichloroethylene
D = Trichloroethylene
E = Vinyl chloride
Note:  These data are for illustrative purposes only and are not indicative
of final design and operation based on longer EBCTs, see pages A69 to A76.

-------
                                      - 34 -
     If the adsorbent use rate decreases with increasing empty bed contact time

then a more than proportional improvement in performance is gained by

increasing the empty bed contact time.   The next sub-section on "Economic

Analysis" will show the importance of knowing whether or not this proportionally

between empty bed contact time and performance occurs in a given location.

Further, the adsorbent use rate can be used to estimate one of the major

components of cost associated with a granular activated carbon adsorption

treatment system, the cost of adsorbent reactivation.  Figures 42 and 43 in

Appendix A, pages A109 and A110, present the unit costs that can be multiplied

by the adsorbent use rate to calculate the reactivation cost per unit volume

of water produced.

     The data in Table X, based on Figures 24 to 27, pages A69 to A76 in

Appendix A, can be used an as example of this type of analysis.

                                    TABLE X

      ADSORBENT USE RATES FOR TREATING A SOUTHERN FLORIDA GROUND WATER

Empty Bed         Criterion 1                 Criterion 2          MCL
Contact Time VC
Minutes Use, mg/1
6.2 560
12.4 370
18.6 170
24.8 230
Cis
Use, mg/1
90
70
NR
NR
TOG
Use, mg/1
1,340
880
530
430
Use, mg/1
860
430
410
410
VC - Vinyl Chloride
Cis = cis-1 ,2-Dichloroethylene
NR = Criterion Not Reached
Note:  Data are from Figures 24 to 27, pages A69 to A76, Appendix A.  These
data are for example purposes only.  Calculated adsorbent use rates at other
locations are much lower.

-------
                                        - 35 -







     These data show that the adsorbent use rate declined for all criteria




as the empty bed contact time increased from 6.2 - 18.6 minutes, (more than




proportional performance improvement),  but that the next increment of empty




bed contact time to 24.8 minutes did not lower the adsorbent use rate for the




MCL requirement.  Therefore, for these data, an empty bed contact time of




about 19 minutes was optimum.




     Summary




     In summary several points are important:




     1)  Granular activated carbon treatment performance criteria have been




selected that will permit collecting the data from pilot granular activated




carbon columns that is necessary for; a) the economical design and operation




of a granular activated carbon adsorption unit process, and b) choosing an




adsorbent in any given location requiring organic contaminant control.




     2)  This scheme is also flexible enough to serve other purposes such as




evaluation of the combination of ozone and granular activated carbon (biological




activated carbon), including an evaluation of the microbiological content of




the effluent, and the performance of reactivated carbon compared to virgin




material, if a sample of reactivated carbon is available for study.




     3)  Because of changing seasons and possible changes in the quality of




source waters, continued monitoring of the full-scale treatment plant will be




necessary for good process control.




     4)   Although the pilot granular activated carbon column tests should




give a good indication of full plant performance, some adjustments in reactivation




frequency should be expected.




     5)  Although the above outlined approach may not be the ultimate solution,




operation of the full-scale granular activated carbon beds to meet the




established performance criteria consistently will provide a large measure of




protection for consumers against the health hazards and esthetic problems of




organic compounds in drinking water.

-------
                                        - 36 -

Economic Analysis
     The final selection of a water treatment plant design based on the
"Collection of Treatability Data", previous sub-section,  may involve an economic
analysis of alternatives.  This sub-section will summarize recommended economic
analysis procedures to aid utilities and consulting engineers in arriving at
a cost-effective design.  More details are included in Appendix A, pages A91 to Alll.
     Basic Costs
     For these analyses, as in the "Interim Treatment Guide for the Control
                                        3 4
of Chloroform and other Trihalomethanes" ' , two types of granular activated
carbon systems will be considered, one using the granular activated carbon in
separate contactors after sand filters (hereafter called post-filter adsorber)
and the other using the granular activated carbon as a replacement for the
media in existing filter beds (hereafter called sand replacement).  Both
systems will be considered with on-site reactivation.  The economics of using
truck transport combined with centralized reactivation furnaces has been
                                                       3 4
explored in the Trihalomethane Interim Treatment Guide. '
     When the assumption is made that granular activated carbon will replace
the media in existing filters, the need to consider the cost of separate granular
activated carbon contactors is eliminated.  For purposes of the sand replacement
analysis, a water treatment plant is assumed to consist of a number of one
mgd filters.  Table XI shows the design parameters that were assumed for the
sand replacement systems.  Table XII contains typical design assumptions for
post-filter adsorber systems of 10 and 100 mgd.
     Note that, for the sand replacement, a granular activated carbon loss
of 10 percent per reactivation cycle is assumed, but a granular activated

-------
                                       - 37 -


                                    TABLE XI

                 DESIGN PARAMETERS FOR GRANULAR ACTIVATED CARBON

                             (Sand Replacement)
     Design Parameters

     Activated Carbon Cost

     Activated Carbon Loss

        per Reactivation Cycle

     Fuel Cost

     Construction Cost Index

     Wholesale Price Index

     Direct Hourly Wage Rate

     Amortization Rate

     Amortization Period

     Volume per Filter

     Loss in Adsorptive Capacity

     Design Capacity

     Empty Bed Contact Time

     Reactivation Frequency
Level

$.45/lb

10 percent



$1.50/mil BTU

273.8 (6/77)

194.6 (6/77)

$5.58/hr (5/77)

7 percent

20 years

856 ft3

0 percent

70 percent

13.4 minutes

Every 1.2 months
Note:  The influence of variations in these assumptions on the final cost is
       discussed in detail in the Interim Trihalomethane Control Guide.

-------
                                       -  38  -

                                     TABLE XII

                      PARAMETERS  FOR GRANULAR ACTIVATED  CARBON

                          (Post-filter  Adsorption)
     Design Parameters

     Activated Carbon Cost

     Activated Carbon Loss

          per Reactivation Cycle

     Fuel Cost

     Construction Cost Index

     Wholesale Price Index

     Direct Hourly Wage Rate

     Amortization Rate

     Amortization Period

     Filter Configuration

          10 mgd plant

               No. of filters

               Diameter of filters

               Vol. per filter

               Empty Bed Contact Time

          100 mgd plant

               No. of filters

               Diameter of filters

               Vol. per filter

               Empty Bed Contact Time

     Loss in adsorptive capacity

     Design capacity

     Reactivation Frequency
Level

$0.45/lb

5 percent



$1.50/mil BTU

273.8 (6/77)

194.6 (6/77)

$5.58/hr (5/77)

7 percent

20 years
12 feet
1470 ft
18 minutes at 5.4 gpm/ft^
28

20 ft.

4396 ft3

19 minutes at 5.5

0 percent

70 percent

Every 2 . 4 months
Note:  The influence of variations in these assumptions on the final cost,is
       discussed in detail in the Interim Trihalomethane Control Guide.

-------
                                   - 39 -
 carbon loss of only 5 percent per cycle is assumed for post-filter adsorbers.




These two assumptions are intended to reflect differences in the operation




of the two systems.  Sand replacement systems are labor intensive and




increase the possibility of granular activated carbon loss because the




activated carbon is changed manually.  In post-filter adsorption systems,




the activated carbon transfer is assumed to be accomplished hydraulically,




leading to fewer possibilities for handling losses.  Representative costs




for both types of systems based on the design parameters in Tables XI and




XII are summarized in Table XIII.




     The costs for the post-filter adsorber system in Table XIII are assumed




as if the system might be added to an existing 100 mgd plant.  In this case,




the post-filter adsorber is assumed to supply an average flow of 70 mgd rather




than for the peak capacity of the plant at 100 mgd.




     Table XIV contains the unit costs for three possible granular activated




carbon system design configurations.  Lines 1 through 3 show the costs




associated with a 100 mgd sand replacement system with an average flow of




70 mgd.  Lines 4 through 6 are associated with a 70 mgd sand replacement




plant treating 70 mgd, and lines 7 through 9 contain unit costs for a




post-filter adsorption system that is designed for and produces an average




flow of 70 mgd.  Therefore, the first sand replacement system has a load factor




of 70 percent, while the other two systems have load factors of 100 percent.




     Each system illustrates a set of conditions that might be met under




field conditions.  Designing a plant for peak capacity, for example 100




mgd, with an average load factor of 70 percent is common.  The second




set of data illustrates a condition in which the peak capacity and




average flows are equal.  Figure 13 shows the cost penalty of excess




capacity. In Figure 13, for given equivalent reactivation frequencies,




the cost in cents per thousand gallons is plotted for the smallest




contact time in each plant design, lines 1 and 4 from Table XIV.  The cost




penalty decreases as the time between reactivations increases.

-------
                                                    - 40 -
X!
W


•K
*
£_i
cu a
4-1 0
rH -rl
•rl 4-1
>4H CX
1 rl
4-1 O
cn cn
O 13
PH <
13
60

0
O 4J
rH C
0)
B
CU
13 O
G cd
Cfl rH
CO CX
cy
pi





*
•K
rl
CU C
4-1 0
rH -rt
•H 4-1
MH CX
1 rl
4-1 O
cn cn
0 13
13 PH <3
60
6

o
rH
4J
C
cu
0
T3 CU
C 0
CS cfl
C/> rH
CX
cu
Pn1























g
4J
CO
>-,
to






0 O
o o
o o
rt •»
VO CO
rH r-~
O CO
A n
CM I---





O
o
o
n
o
0 1
CO
n
CM








0 O
0 O
0 O
rt r«
CO rH
o in
r- -3-
rt
rH






O
o
o
1
-vf
CO
OO








^~v

**^s
^^
•oo- cn
-— ' !H
cn o
4-1 CU 4-1
cn o o
O cfl cfl
U C 4J
Cn rl C
4-> 4-1 ^ O
cn o En u
O 0)
u •!->
o
i-H M
cfl PH
4-J
•H
ex
CO
O






0
o
o
•t
o
rH
CM
A
CO





0
0
0
rt
o
0
CO
M
CM








o
0
o
n
vO
rH
CO








O
O
0
r.
0
C CO
O CM
,0
rl
Cfl
o

13
cu
4-1
Cfl /-N
> 
•H **-^
4J
O CU


o
r> _
rH



^-\
•Cfl-
N.^

4-J
cn
O
0

4-J
u
cu
•i~)
o
M
PH

rH
Cfl
4-1
O
H









O
o
o
rt
r-.
oo
i —
*«
CO
CM




0
0
0
n
m
00
^D
•\
oo








o
o
o
r\
CO
vO
*£)
n

0
o
*
CM




T3
a
cfl

4-1
O
cu
•r-)
O
rl -K
PH 4-1
•— ' cn
o
rH O
Cfl
4-1 s~^
•H 4-1
cx cn
cd cu
O rl
cu
rH 4J
Cfl C
4J H
5


m rH
VO O
• •
d- CO
ON CM
• •
CO ON









•P
cn
o
CJ>

rH
cd
4-1 ^ ^^
•H rH 4-1 rH
CX Cd CO Cfl
Cd 60 O 60
o u
0 O
4-1 O 60 O
•iH O CO
C i-H -H rH
fj -^- 4J -^
cn cfl cn
13 4-1 (-1 4J
cu C cu C
N CU CX CU
•HO O O
4-1 <^ ^~s
M 4-1
0 -H
a c
< P





vO
vO
•
oo







•
^>
•vl-
ON
o
oo
00 O
•
VO >4H
o

rl
o
4-1
a
cd
U-l
>,
!-J
cu

o
a
CU
CO iH
r-~
rH
vo cd
rH 4-1
•H
CX
cd
a

cfl
Cfi
13
rH
cu
•H
>.
r-~
rH M
M
co cd
rH CU
>-,

^N 0
CN
rH
cd *
60 S-S
r^.
O
O 4-1
0 cfl
rH
\ 4-1
co cn
4-> 0
d u
0)
O rH
*~s cd
4-1
4-1 -H
cn cx
o cd
u o

4-1 13
•H 0)
C N
P -H
4J
I-H rl
cfl O
4-" I
0 <
H *
4-1
C
cd
rH
CX

4-1
C
cu

4-J
cd
cu
M
4-1 •
>->
60 4-1
C iH
•H CJ
4J Cfl
cn cx
•H cd
X 0
Q)
4-1
Cf— (
c
cd cd
i — i
O CX
4-J
T~H
C cd
Of4
C
*H *H
4-1 00
. i _ i
•rl 'rl
13 SH
T3 O
cd
cu
rH ^
Cd 4J
4J
C •+-!
CU O
E
OJ 4—1
I r~*
t-t n
o cu
C o
•H J-l
cu
C CX
cd
0
cn r~-
cd
a)
i3 ^Q
CU
6 o
S 4-J
CO
CO '"O
C^ QJ
g
CO 3
•H CO
CO
g cd
cu
4-* CO
CO Cfl
>-. 15
cn
>>
M 4-J
QJ «H
,£} U
s-i cd
O £X
co cd
T3 O
* e
M 3
OJ 6
4-> 'H
	 | LJ
I— 1 ?S
•H CTj
,, . f-j
M-l B
1
4-J CO
CO 4-»
O 'H
CX
13
g g
£ rt
H
*
•K

-------
                   -41-












































£>
H
W
|_J
Pg
H


























































C/3
*rj
w
H
CO

CO

^
0

s?
o

Q
W
H
5;
H
H
U
^

H
CO
o
o
















o
0
VO
0
e
s~^
c
Q)
Q)
4->
M

CO

C
O
s


>> rH
o cfl
C M
0)
2 0
cr o
CU O
M rH
[T | •» — _
tn
C 4-1
o c
•rl 0)
4-1 O
cd
> C
•H -H
o en
Cfl 4J
cu co
Pi O
c_>










in

Q
e
VJ-

g
"
CO

p-J
^
CM

i
K
I— 1

*»
p
^
if
^^
•
c



o


•



0
a

CM

0

o
a
r_|

o
e
Tj -H ^
CU H CO
PQ 0)
4-1 4J
>~, O 3
4-> cd C
fi 4-> -H
6 C sB
O

>-.
•H ^-N
^ O -O
cd cd oc
cu ex g
CM cfl ^^
a
a
•rH *"CJ
W ^ M
DOB
Q rH £<





a
0) cu
4-1 pt
>i H

rH -
CM -j- in

vD ^0 CN]
• • •
CN| <± vD

ON O O
CM iO r^*

m 
CO rH
O -H


CM
OX

m
•
Ox

tH
O
, — 1

Ox
•
0

m
•
CM
rH
ax
•
vO
tH



^O
•
OX
rH





OX
^
CM




oo
•
ax
CO


O
CN
00








O
CO





0





o
r^.




M
01
•8
o
Cfl

-------
                    - 42 -
                                            CO
                                            OC
                                            o
                                   S3
                                 co 5  Q-
                                   *0
O
(O
O
if)
o
CO
o
CM
                                   O _,
                                   E CO
                                   £ UJ
                                   z O
                                   o X
                                   f= UJ
                                 "2u.

                                   §H
                                   s§
                                      o
                                      UJ
                                            CO

                                            UJ
                                            QC
                                            z>
                                            g
                                            u.
                                               CO
                                               2
                                               UJ

                                               CO

                                               CO

                                               H

                                               UJ
                                               5
                                               LJJ
                                               O
                                               Q.
                                               UJ
                                               OC
                                             •H
                                             J-l
                                             a
                                             rt
                                             OJ
                                             0)
                                             1?
                                             jj
                                             a)
                                             pa
                                             •H
                                             H
         (|B6 OOCH/D) 1SOO W31SAS

-------
                                       -43 -

     Influence of Empty Bed Contact Time



     The tradeoff between bed depth (empty bed contact time) and reactivation



frequency for a given quality of water is an important relationship.   To



make this comparison the ratio between reactivation frequency in months and



the EBCT in minutes was calculated.  For example, for one of the systems in



Table XIV with an EBCT of 10 minutes and a reactivation frequency of 0.5



months, the ratio, "R", is as follows:
     R- °'5        = 0.05                (1)
         10 minutes


The R values were calculated for the sand replacement systems in Table XIV



XIII with contact times of 10 and 20 minutes (lines 4 and 5) .   The system



with 20 minutes contact time has twice the bed depth of the system with 10



minutes contact time.



     The cost in cents per thousand gallons is plotted versus R in Figure 14



for these two systems.  These data show that the longer contact time bed is



always more expensive for the same value of R.  When the R value remains the



same, this means that a direct proportionally exists between empty bed



contact time and reactivation frequency, which may or may not be true in a



given situation.  The data collected with the pilot columns described in the



previous sub-section will provide the necessary information for this analysis.



     For a given value of R and a given EBCT (P^ the straight line drawn



horizontally to the curve representing the longer contact time bed (?„) ,



represents the increase in the period between reactivations that is required



for the longer contact time to be economically equivalent to the shallower



bed. In Figure 14 at P ,  R = 0.01,  the total unit cost for the 10 minute



empty bed contact time bed is 45.1 cents/1000 gal.  Drawing a horizontal



line to the 20 minute contact time curve yields an R at P~ of 0.0124.  The



necessary reactivation frequency for cost equivalancy can be calculated as



follows:


            P2   =   0.0124                  (2)
            20



            P
             2 = 0.25 months

-------
            -44-
(|B6 000170) 1SOO IN31SAS
g
LL

-------
                                       - 45 -





Therefore, the period between reactivations would have to increase by 150




percent to be able to use a iQO percent longer EBCT system at the same cost.




This non-proportional relationship between EBCT and reactivation frequency




is caused by the increased activated carbon inventory that must be maintained




for deeper beds.




     At larger values for R (longer periods between reactivation), however,




the difference between systems becomes small so that little economic penalty




occurs when choosing a system with a longer EBCT and a longer period between




reactivations.  Performance data from the recommended pilot column test (see




previous sub-section), to determine whether or not a more than proportional




lengthening in reactivation frequency will occur with an increase in contact




time, can be used to make this type of analysis in a given location.




     Influence of Type of System Chosen




     In the following analysis when post-filter adsorbers and sand replacement




systems are compared, the comparison will be made between the last two systems




in Table XIV because the design empty bed contact times are equal.  Figure 15




shows the cost in cents per thousand gallons for a sand replacement system and




the post-filter adsorber having equal empty bed contact times (10 minutes).  For




very short reactivation periods (less than 2 weeks) post-filter adsorbers are




always less expensive than sand replacement. When the reactivation period is




greater than 0.5 months, however,  sand replacement becomes less expensive because




of less capital expense.




     Figure 16 shows the relationship between a sand replacement system with a




10 minute EBCT and a post-filter adsorber with a 20 minute EBCT.  As might be




expected from the previous analysis, with smaller R values the longer EBCT post-




filter adsorber is less costly than the shorter EBCT -sand replacement system,




but as the period between reactivations increases, the sand-replacement system




becomes relatively less expensive.  Therefore, under this set of

-------
              -46-
                                           UJ
                                      N    m
                                           o co
                                           < 2
                                           _l UJ
                                           Q. H
                                           LU CO
                                          DC >•
                                         f Q CO

                                      «  o Z Z
                                      ^  s < O
                                         z CO d
                                        = Q t
                                         2 O
                                         £H
                                         Q. co

in
6
                                               3z=!
                                               S O  "r
                                                  co  H
                                                  E  co
                                                  <  O
                                                  OL  Q-
o
U)
o
co
(|B6 OOOI./0) 1SOO W31SAS
                                               Q
                                           o z
                                           o <
                                           in
     UJ
     oc

     O
     u.
                                                      w
                                                      c
                                                      o
                                                      •H
                                                      4J
                                                      Cfl
                 U
                 cfl
                 0)
                Pi

                 c
                 OJ
                                                      0)
                                                      M


                                                      01


                                                      •H
                                                      H

-------
                         -47-
                                                            g
                                                            H

                                                            > Q
                                                            H Z
                                                            O <
                                                      m
                                                      q
                                                      d
                         LJJ
                         CC
      Z
      UJ
uj  Q 2
i  Z LU
H  < O
                                                         o
                                                         <
                      o
                      o
                                                               Q.
                                                               LU
                                                               cc
o
(O
o
CM
              (|B6 0001/3)) 1SOO IN31SAS
                                                         >  < z
                                                         o  H <
                                                         £  ZCO
                                                         =3  O M-
Q                                                               ii  ^^
                                                                  H

                                                         2  DO (/) tt
                                                         <  > O O
                                                      g  >  e o to
                                                         UJ
                                                         tr
                                                         u.
                                                         O

                                                         O
   U. LU CO
   U. DC O
   LU U. Q.

   CO

   HI
   CC
   D
   g
   LL.

-------
                                     - 48 -
assumptions, the increased cost of building post-filter adsorbers cannot be




overcome by obtaining longer periods between reactivations through the use




of longer EBCTs, if the reactivation frequency is greater than 0.4 months.




Because the cost differential is not great, however, other considerations




may dictate final designs.  Note: the cost comparisons of other combinations




than those presented in Figures 13-16 can be made using the data in Table




XIV.




     Influence of Granular Activated Carbon Cost




     To minimize costs, a water purveyor might consider the use of the least




expensive granular activated carbon available.  A lower cost activated




carbon, however, may also require shorter periods between reactivation, if




its performance is reduced.  For example, using the data shown in Figure 17,




for a 10 mgd post-filter adsorption plant, if a pound of activated carbon




costs $0.70 and the reactivation frequency is three months, then the system




cost would be 17.5 cents/1000 gal.  With a less expensive activated carbon,




perhaps one costing $0.30/pound, the reactivation frequency would have to be




2 months or greater to achieve a favorable economic tradeoff under these




assumed conditions.




     Influence of Inflation




     Table XIII shows that under static economic conditions, sand replacement




systems are slightly less expensive than post-filter adsorption systems.




Because post-filter adsorption is less labor intensive than sand replacement,




it might become less expensive some time in the future because of inflation.




Figure 18 illustrates the impact of inflation on the two 100 mgd systems




(one of each type) assuming an inflation rate of 5 percent per year and




shows that sand replacement system does become more expensive than post-




filter adsorption in year 18.

-------
                  - 49 _
              .0.0.0
                                       o
                                       0
                                       t-
                                    CO
                                    CM  m

                                       Q
                                       O
                                       QC
                                       UJ
                                       Q.
q
in
CJ
q
d
q
in
q
d
q
in
q
6
           (|B6 00017$)
        IN31SAS dO 1SOO
                                   H o
                                   O (0
                                   < o
                                          O LLI
                                          CO H
                                          O CO
                                          8o.
                                >  cc  =


                                i  s§
                                5  t "•
                                UJ  < .
                                £  >fc
                                          UJ CO
                                            O
                                   UJ
                                   oc
                                   D
                                   O
                                   LL

-------
       - 50 -
                                     Q. CO
                              O
                              CM
                              in
   q
   in
q
6
q
in
q
6
   (|e6 00(H/3>)
IAI31SAS dO 1SOD
                                      o>
                                        LU
                                 z
                                 g
                                 i-
                                 o
                                 D
                                 CC

                                 (/)
                                 Z
                                 o
                                 o
                                 £  z
                                 t  2
                                 DC
                                 <
                                 LU
                          £2
                          °H
                          h- 0-
                          s§
                          og
                          z <
                          UJ Z
                          I- UJ

                          25
                          CC ui
                                     u_
00


LiJ
CC
3
O

-------
                                      -51 -





     Figure 19 shows the impact of inflation on the two granular activated



carbon configurations, over a 20-year period assuming inflation at 7 percent



per year.  Post-filter adsorption becomes less expensive than sand replacement



in year 12 under this set of assumptions for a 100 mgd plant.



     Figures 18 and 19 illustrate that over the life of the two types of



systems, because of the labor intensive nature of the sand replacement type,



it eventually becomes more expensive than a post-filter adsorber on a yearly



expenditure basis.  This phenomenon occurs, of course, because the capital



expenditure remains fixed over the life of the investment, while operating



costs, particularly labor costs, are subject to inflation.



     Figures 18 and 19 also illustrate that the total expenditure over time



is less for the sand replacement system than for post-filter adsorbers.  To



account for total expenditures, a "present value" analysis was made for the



systems listed in Table XIII.   Two discount rates (6% and 8%) and three



inflation rates (5%, 7% and 9%) were used in the analyses.  The results are



summarized in Table XV.   As can be seen from the Table, for the larger



plant, at the highest inflation rate, for both discount rates, the difference



in present value for the two systems is small.  In no case, however, is the



present value of the expenditure for post-filter adsorbers less than those



for sand replacement.



     The unit costs in Table XIII show that small treatment systems in



general are more expensive,on a per unit of product basis, than larger



systems.  These costs can be reduced significantly, however, by the use of



truck transport and regional reactivation systems.  This effect is illustrated


                                              3 4
in the Trihalomethane Interim Treatment Guide. '

-------
            -52 -
                                        CC
                                        LU
                                        I-
                                        o
                                        o_
                                        •o
                                        O)
                                        E
                                        o i
                                        o
                                          UJ

                                          CO
                                        oz
                                        og
                                        z H
                                        uj o
                                        I-CO
                                        
-------
                            - 53 -








                       TABLE XV




           SUMMARY OF PRESENT VALUE ANALYSIS




FOR POST-FILTER ADSORPTION AND SAND REPLACEMENT SYSTEMS
System
Sand Replacement
Inflation at
Inflation at
Inflation at
10 mgd 100 mgd
Discount Discount Discount Discount
at 6% at 8% at 6% at 8%

5% 225.67 190.96 119.43
7% 262.37 219.50 140.86
9% 305.30 252.58 166.12

101.
117.
137.

67
64
13
Post-Filter Adsorber
Inflation at
Inflation at
Inflation at
5% 259.96 221.20 134.69
7% 289.80 244.72 150.63
9% 323.35 270.50 169.12
114.
127.
141.
76
15
40

-------
                                     - 54 -
     Summary




     This sub-section has discussed some of the important factors that a




water purveyor must consider when making decisions regarding the use of




granular activated carbon systems for removal of organic contaminants.




Because the economics of choosing and designing a granular carbon system




are complicated, individual utilities or their consultants, or both may need




assistance in developing cost and economic design criteria.  The Water




Supply Research Division is prepared to assist those utilites affected




by treatment regulations.  Computer programs currently being utilized by




WSRD and instructions in the use of these programs will be made available




to interested parties.  The Water Supply Research Division is prepared




to provide a limited economic analysis for individual utilities affected




by the treatment regulations.  This assistance will aid the utility in




making general decisions regarding overall implementation strategies,




but will not be sufficient for specific designs.




                                      SUMMARY




     In summary, the Interim Treatment Guide provides information that demonstrates




that granular activated carbon adsorption is the best available treatment technology,




Fall 1977, for treating water to remove organic contaminants, thereby improving




finished water quality and providing the American consumer with a more healthful




and esthetically pleasing drinking water.  For a more detailed summary the reader




is referred tc the Executive Summary at the front of this document.  Three




Appendices with more detailed information on various aspects of this subject




follow this Guide.

-------
                                    - 55 -

                                ACKNOWLEDGMENTS

     The Authors wish to extend their special appreciation to Ms.  Maura

M. Lilly without whose dedication, secretarial skill, diligence and

patience, this Guide could not have been completed in a timely manner.

They also wish to thank the Organic Contaminants Research Staff, J.K.

Carswell, J. DeMarco, P. Dorsey, W.C. Elbert, W.A. Feige, D.L. Guttman,

D. D. Hinderberger, K.L. Kropp, O.T. Love, Jr., B.W.  Lykins, L.A.  Moore,

D.R. Seeger, C.J. Slocum, B.L. Smith, A. A. Stevens and R. Stevie, all

of whom contributed to this document.

                                   REFERENCES

1.   Dostal, K.A., Pierson, R.C., Hager, D.G. and Robeck, G.G., "Carbon
     Bed Design Criteria Study at Nitro, West Virginia," JAWWA, 57, No. 5,
     663-674 (May 1965).

2.   Stevens, A.A. and Symons, J.M^, "Measurement of Trihalomethane and
     Precursor Concentration Changes,"  JAWWA, 69, No.  10, 546-554
     (October 1977).

3.   Symons, J.M., et al., "Interim Treatment Guide for the Control of
     Chloroform and Other Trihalomethanes," Water Supply Research Division,
     U.S. Environmental Protection Agency, Cincinnati,  Ohio, June 1976,
     mimeo, 48 pp. plus 4 Appendices, unpublished.

4.   Clark, R.M., Guttman, D.L., Crawford, J.L. and Machisko, J.A.,
     "The Cost of Removing Chloroform and other Trihalomethanes from
     Drinking Water Supplies," Municipal Environmental  Research Laboratory,
     Office of Research and Development, U.S. Environmental Protection
     Agency, Cincinnati, Ohio 45268, EPA-600/1-77-008,  March 1977.

     See Also:

     "Activated Carbon in Water Treatment" Proceedings  of a Water Research
     Association Conference at the University of Reading, April 3-5, 1973,
     Water Research Centre, Medmenham, United Kingdom,  Available from the
     Water Research Centre, Henley Road, P.O. Box 16, Medmenham, Marlow,
     United Kingdom, SL7 2HD.

     Translation of Reports on Special Problems of Water Technology -
     Volume 9 - Adsorption" Proceedings of a Conference in Karlsruhe,
     Federal Republic of Germany, 1975, EPA-600/9-76-030, December 1976.

-------
          APPENDIX A

   PERFORMANCE OF GRANULAR

  ACTIVATED CARBON FOR THE


REMOVAL OF ORGANIC COMPOUNDS
         Written By:
         J.K. Carswell
         R.M. Clark
         J.  DeMarco
         P.  Dorsey
         W.  A. Feige
         D.L. Guttman
         O.T. Love, Jr.
         B.  W. Lykins
         A.  A. Stevens
         R.  Stevie
         J.M. Symons
         Reviewed By

      Gordon G. Robeck

-------
                                   APPENDIX A
                                TABLE OF CONTENTS
Literature Review                                                     Al
     Taste and Odor Control                                           Al
     Removal of Organic Compounds as Measured by General
        Organic Parameters                                            A5
     Removal of Specific Organic Compounds                            A9
     Reactivation of Granular Activated Carbon                       A22
Current Water Supply Research Division Findings                      A34
     Class I Compounds (Taste and Odor Producing)                    A34
     Class II Compounds (Synthetic Source Water Contaminants)        A36
          Naphthalene                                                A36
          Other Polynuclear Aromatic Hydrocarbons                    A37
          Carbon Tetrachloride                                       A37
          Trichloroethylene, Tetrachloroethylene and Similar
             Chlorinated Organics                                    A40
          Prediction of Granular Activated Carbon Bed Performance,
              Studies with Humic Acid and Carbon Tetrachloride,
              Dieldrin, PCB, Benzene, and _p_-Dichlorobenzene          A41
          Mixtures from Actual Waters                                A42
          Unidentified Compounds                                     A52
          Summary                                                    A56
     Class III Compounds (Disinfection By-Product Precursors)        A56
     Class IV Compounds (Disinfection By-Products)                   A63
     General Organic Parameters                                      A66
     Influence of Empty Bed Contact Time                             A69
     Biological Growth and Endotoxin Production                      A76
          Standard Plate Count                                       A76
          Endotoxins                                                 A80
Ozone Enhanced Granular Activated Carbon Adsorption (Biological
    Activated Carbon)                                                A81
     Literature Review                                               A81
     WSRD Pilot Plant Results                                        A86
Economic Analysis Information                                        A91
     Basic Costs                                                     A91
     Influence of Empty Bed Contact Time                             A92
     Influence of Granular Activated Carbon Cost                     A99
     Influence of Inflation                                          A99
     Cost of Reactivation                                           A108
     Summary                                                        Alll
Acknowledgments                                                     A112
References                                                          A113

-------
       PERFORMANCE OF GRANULAR ACTIVATED CARBON FOR THE REMOVAL OF


                             ORGANIC COMPOUNDS


     The purpose of Appendix A is to summarize the current state of


knowledge concerning the performance of granular activated carbon


adsorption as a treatment unit process.  Appendix A will be in two major


parts, one, a review of the literature, and two, a summary of the current


status of the Water Supply Research Division's projects on this subject.


The literature review will not be exhaustive, but will be a summary of


the subject.  Because many of the research projects discussed in the


second portion of Appendix A are on-going, the research findings to date


(Fall 1977)* will be presented as progress reports.  Detailed papers on


these studies will be published in the technical literature as they are


completed.


Literature Review


     Taste and Odor Control


     Many water treatment plants in the United States are practicing taste


and odor control by the addition of powdered activated carbon, which also


removes some organic matter.  Currently (Fall 1977) however, about 35 plants


are using granular activated carbon, either alone or on top of some sand, as


both a filter media for particulate control and an adsorption media for


organic contaminant control.  This type of system is hereafter called a sand


replacement system and is used primarily for taste and odor control.  A few


water purveyors installed these granular activated carbon beds because of


the organic pollution in their raw water as well as taste and odor problems.


     In their 1964 paper, Woodward,  Dostal, and Robeck  reported on five


installations of granular activated carbon beds during the 1930's.  A

                  2
more recent survey  of the known granular activated carbon installations in


this country, see Table I,  shows that approximately 80 percent of the water

*Some data collected beyond that contained in the November 1977 draft of

Appendix A has been included in the report.

-------
                         - A 2 -







                     TABLE I




DATE GEANULAR ACTIVATED CARBON FIRST INSTALLED2




                Experimental Use (9 Plants) Routine Use (38 Plants)
1935
1961
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
0
0
0
0
0
1
3
1
1
1
1
0
1
1
1
2
1
4
6
7
7
5
3
1

-------
                                       -A3  -


treatment plants routinely using granular activated carbon beds have


been installed since 1970.


     A discussion of some of the early installations is of interest.  In


the fall of 1963 the management of the Nitro, West Virginia, water utility


installed two beds of granular activated carbon to investigate treatment


for removal of taste and odor.  During the 3-month testing period the


Threshold Odor Number (TON) of the raw water varied between 500 and


1000. During this same time period the TON of the influent water to the


experimental granular activated carbon beds varied between 100 and 200,


with one excursion to 400.  These beds, with empty bed contact times (EBCT)


of about eight minutes, were able to produce an odor-free water for as

                3
long as 26 days.   The success of these tests encouraged the water purveyor


to convert the entire plant to granular activated carbon beds in 1965.


Another interesting feature of this installation was the construction of an


on-site 10,000 Ib/day multiple-hearth reactivation furnace.  From that time,


until the plant was closed several years later, because the municipality


obtained a different source of water, the granular activated carbon was


reactivated approximately every six months.


     The success of granular activated carbon beds for controlling taste and


odor demonstrated at Nitro, West Virginia, has been duplicated at many other


water utilities.  For example, at Piqua, Ohio; Mt. Clemens, Michigan; Lawrence,

             4
Massachusetts ; and Davenport, Iowa, the water purveyors were having difficulty


providing their consumers with an acceptable drinking water in spite of the


use of large doses of powdered activated carbon. After conversion to granular


activated carbon beds, these purveyors reported successful control of their


taste and odor problems.  Figure 1 shows that at Lawrence, Massachusetts,


during a time period when the TON in the settled water was in the 6-10


range, the granular activated carbon beds controlled the effluent TON to


acceptable levels for over a year.

-------
                         -A4-
                           (CI1OHS3}IHiMO139)
                                                       O
                                                       =>
                                                       O
LU


00

fid
LU
00
                                                          Z
                                                          O
                                                          CO
0<
Q>
Ofi
_ u



  LU CO
  Z<


  O  ..
  U_ LU

-z z
LU fi£ UJ
fid UJ Qd

o<5
U.O. _i
                                                       to
                                                       LU
                                                       a:
                                                       X
aOI«3d >I33M  CL iiOd  NO1 39V33AV

-------
                                      -A5-
     When conversion was first contemplated, the Massachusetts State Health


Department insisted that some sand be left in the bottom of the filters to


avoid any problems with particulate (turbidity)  breakthrough.   A six-month


comparison of the turbidity in the effluent of a conventional sand filter


and an all granular activated carbon system, see Figure 2, showed that their


performance was equal.  Note, pilot plant scale comparisons by the Water


Supply Research Division Laboratory has confirmed this finding.  When the


entire plant was converted to granular activated carbon, no sand was left in


the filter boxes.

                              2
     In the survey cited above  the water purveyors were asked how long


their current charge of granular activated carbon had been in service and


whether or not the beds were still effective for taste and odor removal. The


shortest effective life reported was 23 months and some beds had been in


service for four years and were still effective.  Reports from the United


Kingdom  substantiate these findings from United States practice.


     Removal of Organic Compounds as Measured by General Organic Parameters


     The previous sub-section detailed the performance of granular activated


carbon beds for the removal of taste and odor causing compounds, but investigators


have also been interested in the performance of the granular activated


carbon adsorption process with respect to the removal of analytic parameters


that would be reflective of the "total" organic content of water.  As pointed


out by Stevens and Symons,  no analytic test currently available measures


the "total concentration of organic compounds" in water. Several tests have


been proposed to approximate or be proportional to this parameter.  A few of


these are:  1) the organics-carbon adsorbable test producing a carbon chloroform


extract (CCE), 2) chemical oxygen demand (COD),  and 3) total organic carbon (TOG).

-------
                        -A6-
u
z
LU
ae
z
z
LU
s
oc
LU
Q_
X
u
z
LU
ot
>-

o
LU
X
—
z
<
o
                                                         Z
                                                         O
                                                         co
V
                                                      V
                                                      <
                                                      s
                                                      Q£
                                                      Q.

                                                      <
                                                      z
                                                      <
                                                         a:
                                                         O
      O
      Z
      <
      CO
      u.
                                                         si
   CO  Q_
   "  S^
      89
       . OQ
      a: »-
      =>c*
      OO
                   simn

-------
                                  -A7-
These tests have been used to monitor various operating granular activated

carbon adsorption beds.
                                               o
     In 1965, Dostal, Pierson, Hager and Robeck  reported that the percent

removal of chemical oxygen demand declined from 72 percent at the start of

the test to 56 percent after 31 days at the Nitro, West Virginia water

treatment plant, mentioned previously, while 97.5 percent of the odor was

still being removed after 31 days.  More recently, monitoring of operating

granular activated carbon beds has been carried out at five field locations:

Nitro, West Virginia; Piqua, Ohio; Mt. Clemens, Michigan; Lawrence, Massachusetts;

and Davenport, Iowa. This monitoring began in September 1968 and spanned

several years. Because analytic capabilities have changed and improved

gradually, the same monitoring techniques were not used at all of these

installations. In general, however, the purpose of the monitoring was to

compare the time of breakthrough (first detectable increase) of organic

compounds as measured by a general organic parameter with the time of odor

breakthrough.

     In every case, the breakthrough of odor occurred much later than the

breakthrough of the parameter measuring the removal of general organic

compounds.  Except for the Nitro, West Virginia situation, the control of odor

was successful for years, whereas control of organic compounds as measured by

a general organic parameter was successful only for weeks.  The following

are some typical data:  Nitro, West Virginia, CCE-hf* removal** lasted 5 weeks;

Piqua, Ohio, CCE-lf removal lasted 10 weeks; Mt. Clemens, Michigan, CCE-m

removal lasted less than 28 weeks; Lawrence, Massachusetts, CCE-m removal

lasted 16 weeks, and Davenport, Iowa, CCE-m removal lasted 6 weeks. Figure 3

from Lawrence, Massachuetts, is a typical data plot of the breakthrough

curve for a general organic parameter.
*The lower case letters refer to the method of operation of the activated
carbon adsorption collection and the chloroform extraction system and are
defined in Reference 8.
**Effluent concentration approximates influent concentration.

-------
                         -A8-
                            Q31VAI10V M3N'
                                                         CO
                                                         CN
                                                         «  2
                                                            UJ
                                                            UJ
                                                         o  z
                                                            z

                                                            UJ
                                                                Z
                                                                UJ

                                                                S
                                                                UJ
                                                         a. ^
                                                         UJ
                                                         DC z

                                                         o 2
                                                         < o
                                                         2 ffl
                                                            CC

                                                         ui <
                                                         o °
                                                         Z Q
                                                         UJ UJ
                                                         DC |-

                                                         ^ 5
                                                         < >
                                                         -I H
                                                                -
                                                                53
                                                                o 2
                                                                UJ
es

CO
        cs
                  00

                  o
•o
6
CN

O
UJ

O
O


CO

UJ
CC
                                                                  O

                                                                  UJ
I/Bui ' (uj-300) 1QVU1X3 IflaOJOUOIHO N08UVO
                                                                O

-------
                                    -A9-

                       6 *
     Figure 4 from Ford    indicates that experiences in the United Kingdom

are similar to those in the United States.  The experience in Germany is
                                                         9*
similar to that in the United Kingdom as shown by Heymann   of Duisburg who

investigated the breakthrough pattern of the parameter dissolved organic

carbon (DOC) through a granular activated carbon bed with sampling points at

various depths.  Figure 5 shows that the maximum depth, corresponding to an

empty bed contact time of 4 minutes, showed some increase in DOC after about

4 to 5 days.   These reports lead to the generalized conclusion that if

controlling the general organic content of drinking water is desirable, the

useful life of a given charge of granular activated carbon in an adsorption

bed will be much shorter than might be anticipated from performance based on

taste and odor compound control.

     Removal of Specific Organic Compounds

     Early demonstrations of granular activated carbon's ability to remove

specific organic compounds from water came from its use as an analytic

procedure in which organics were adsorbed on granular activated carbon and

desorbed (extracted) with a solvent, chloroform or ethyl alcohol.  Analysis

of these extracts revealed what organic compounds that were in the original

sample were able to be adsorbed onto granular activated carbon under prescribed

conditions, and then be desorbed with a solvent.  For .example, in 1956,

Middleton and Rosen   found the following organic compounds or classes of

compounds in a carbon chloroform extract (CCE) from a surface water:
*Note, the entire document "Activated Carbon in Water Treatment," a Water
Research Association conference held at the University of Reading, April 3-
5, 1973 and available from the Water Research Centre, P.O. Box 16, Henley Rd.,
Medmenham, U.K., SL7 2HD, is an outstanding volume well worth reviewing.  A
second document "Translation of Reports on Special Problems of Water Technology
Volume 9 - Adsorption" a conference held in Karlsruhe, Federal Republic of
Germany, 1975, EPA-600/9-76-030, December 1976 is an excellent companion
document to the one just previously cited and summarizes the experiences of
water purveyors in continental Western Europe with activated carbon.

-------
                        -A10-
o- 6
e* ?«/T
< C LU
UZ*^
IP
*z £ 4
^^ ^J ill
\J H-

-------
                             -All-
                                              o
                                              CM
                                                        in
                                                        O
                                                        Z <

                                                        1«
                                                      .. < o
                                                      o u-
                                                      Q S O
                                                      O
                                                  CO

                                                  Z
                       < s



                       S i
                                                      LU
                                                      O
  cs
  cs
         oo
•O
6
cs
o
|/6ui 'NOIiVUiN3DNOD NO8HVD DINVOHO QHAlOSSia

-------
                               -A12-





     Substituted benzene compounds         Kerosene



     Polycyclic hydrocarbons               Phenyl ether



     Acrylonitrile                         Alcohols



     Aldehydes                             Ketones



     Organic acids                         Esters



     DDT



     In the same year, Rosen, Middleton and Taylor   isolated alkyl



benzene sulfonate from a carbon-alcohol (ethyl)  extract (CAE).   In


                              12
1963 Rosen, Skeel and Ettinger   isolated the following organic compounds



from a river water CCE.



     Naphthalene                             Tetralin



     Styrene                                 Acetophenone



     Ethyl benzene                           Bis- (2-chloroisopropyl) ether



     2-Ethylhexanol                          Bis- (2-chloroethyl) ether



     Di-isobutyl carbinol                    Phenyl methyl carbinol



     2-Methyl-5-ethylpyridine


                                         13
     In 1970, Rosen, Mashni and Safferman   reported on finding both geosmin



and 2-methyl-isoborneol in a CCE from an Ohio Lake.  In 1972 a report was


         14
published   concerning studies on organic contamination of drinking



water conducted in the lower Mississippi River area.  Organic compounds



were collected on granular activated carbon and desorbed by both heat



and solvents.  One sample contained the following compounds:



     acetylene dichloride                     ethyl benzene



     benzene                                  methyl chloride



     carbon tetrachloride                     propyl benzene



     chloroform                               toluene



     1,2-dichloroethane                       vinyl benzene



     dimethyl sulfoxide

-------
     A second series of samples

     acetone

     acetophenone

     benzene

     b r omob en z ene

     bromochlorobenzene

     bromoform

     bromophenyl phenyl ether
       (positional isomer?)

     butyl benzene


     a-camphanone

     chlorobenzene

     chloroethyl ether

     chloromethyl ether*

     chloroform

     chloronitrobenzene

     chloropyridine

     dibromobenzene

     dichlorobenzene
        (positional isomer?)

     1,2-dichloroethane
   -A13-


contained the following organic materials:

         dichloroethyl ether

         dimethoxy benzene

         2,6-dinitrotoluene

         endo-2-camphanol

         ethyl benzene

         exo-2-camphanol

         hexachlorobenzene


         l-isobromobenyl-4-isopropyl
             benzene (1,2 isomer)

         isocyanic acid

         methyl biphenyl

         methyl chloride

         nitrobenzene

         o-methoxyphenol

         p-menth-en-l-8-ol

         tetrachloroethylene

         toluene

         1,1,2-trichloroethane


        vinyl benzene
*The identity of this compound is very questionable.

-------
                                       -A14-


Also in 1972 Kleopfer and Fairless   found the following compounds in CCE's

from drinking water taken from the Ohio River at Evansville,  Indiana,

using gas chromatographic - mass spectrometric techniques.

     Bromodichloromethane                   Toluene

     Dibromochloromethane                   Tetrachloroethylene

     Bromoform                              Xylene

     Ethylbenzene                           Bis-(2-chloroisopropyl) ether

     Styrene                                Bis-(2-chloroethyl) ether

     Hexachloroethane                       Hexachlorobenzene

     Chlorohydroxybenzophenone

     Finally, Symons and Stevens  reported on organic compound identification

from an ethyl alcohol extract of granular activated carbon (CAE) that included

the following:

     Trichlorobiphenyl              Di (2-ethylhexyl) phthalate**

     Tetrachlorobiphenyl            Methyl Ester of Benzoic Acid*

     Pentachlorobiphenyl            Methyl Ester of Benzene Sulfonic Acid

     Hexadecane                     Methyl Ester of Palmitic Acid*

     Octadecane                     Methyl Ester of Stearic Acid*

     Eicosane                       Methyl Ester of Lignoceric Acid*

     Docosane

     Diethylphthalate**
     * - Probably esterified during extraction.
     ** - Plasticizers

     These reports show some of the types of organic compounds that can

be adsorbed from water by granular activated carbon as evaluated by

this technique.

-------
                                           - A15-
     Below are typical references reporting on the removal of specific



organic compounds in conjunction with other organics by granular activated



carbon beds in use at operating water treatment plants.  In 1965, Robeck,



Dostal, Cohen and Kreissl   demonstrated that coal-base granular activated



carbon, partially exhausted for CCE-hf removal, could reduce the concentration



of endrin, dieldrin, lindane, 2,4,5-T ester, DDT, and parathion dosed into


                                                                 3
river water.  In the same year, Dostal, Pierson, Hager and Robeck  showed that



the seven compounds listed below that were present in the Kanawha River water



after aeration could be reduced to below detectable concentrations by fresh (2-



day old) granular activated carbon beds. These compounds were bis-(2-chloroethyl)



ether, 2-ethylhexanol, bis-(2-chloroisopropyl) ether, a-methylbenzyl alcohol,



acetophenone, isophorone and tetralin.



     Forty days later, however, all of these compounds with the exception of



acetophenone were detected at a bed depth equal to an empty bed contact



time of about 8 minutes.  Providing an additional 2 minutes of empty bed contact



time did remove these seven compounds at this time (40 days), although another



organic compound, ethyl benzene, was detected at a bed depth equal to 15 minutes



of empty bed contact time.



     Stieglitz, et al.   reported on removals by 2 month-old granular activated



carbon beds of 61 organic compounds amenable to analyses by the Grob closed loop


                                            18
stripping and gas chromatographic procedures  .  The granular activated carbon



appeared to have lost its effectiveness for adsorbing compounds eluting from the



chromatograph early, such as chloroform and trichloroethylene, but was still



quite effective for later eluting chlorinated aromatics.  Some questions remain



as to documentation of analytical recovery efficiencies and the operation of the



adsorption beds, however.  The authors claim that some higher concentrations of

-------
                                          -A16-






a few aliphatic and aromatic hydrocarbons in the adsorber effluent can be explained




by chromatographic or biological effects on the adsorber. Dissolved organic



chlorine (DOC1) was still being reduced from 85 to 2 yg/£ through the adsorber at



this time.  The sum, as organic chlorine, of the concentration of chloroform,



carbon tetrachloride, trichloroethylene, and tetrachloroethylene in the treated



water accounted for all the DOC1.   To summarize, after two months of operation,



all of the typical low molecular weight chlorinated compounds were breaking



through the bed although DOC1 was significantly removed from a chlorinated influent.


                                             19
     In 1965 Kolle, Sontheimer, and Stieglitz   reported on studies of pilot



granular activated carbon adsorbers receiving ozonated pre-filtered Rhine River



water.  Two or three adsorbers in series were used, each one meter in depth.



Each meter of depth represented four minutes of empty bed contact time.  After



six months of operation, granular activated carbon samples were taken from the



top of each section and from the bottom of the last section.  These activated



carbon samples were extracted with dioxane and the extracts analyzed for specific



organics.  The following organics were found in various concentrations:  chloroform,



1,2-dichloroethane, 1,2-dichloropropane, tetrachloroethylene, trichloroethylene,



bis-(2-chloroisopropyl) ether, o-dichlorobenzene, hexachlorobutadiene, hexa-



chlorocyclohexane, and tris-(2-chloroethyl)-phosphate.



     According to the authors, the substances identified in these extracts can



be classified into three groups, the aromatic chlorohydrocarbons, (o-dichlorobenzene)



that were completely adsorbed by the uppermost layer of the adsorber, the aliphatic



chlorohydrocarbons (hexachlorobutadiene and hexachlorocyclohexane) that were much



less strongly adsorbed, but still are adsorbed well enough so that the lowest



filter layers do not contain these substances, and the oxygen-containing organic

-------
                                        -A17-
 chlorine compounds, bis-(2-chloroisopropyl) ether and tris-(2-chloroethyl)



phosphate, that were breaking through the activated carbon beds and were identified



in the bottom layer of the filter.



     In addition to the above information, chloroform, 1,2-dichloroethane, and



1,2-dichloropropane were present at various levels throughout the adsorbers,



although tri- and tetrachloroethylene were confined to the upper and middle



layers of the three-layered adsorber.  The presence of these first three compounds



throughout the depth of the adsorber would tend to weaken the authors' gross



classification of the adsorption of aliphatic hydrocarbons (see above).


                           20
     In 1977 Suffet, et al.   reported on the performance of granular activated



carbon and some adsorbent resins for the removal of trace organics from Philadelphia



drinking water.  Suffet used computer-reconstructed gas chromatographic profiles



plus mass spectrometric identifications to assess the ability of the adsorbents



to remove twenty-seven identified organic compounds.   For the activated carbon



column (Calgon F-400*,  9.7 minute empty bed contact time) in one experiment, the



adsorbent was shown to be quite effective for removal of most of the compounds



identified although exhaustion was noted for the organic compounds with lower



boiling points by the 18th week of the run.  Gas chromatographic profile analysis



of the F-400 column effluent indicates chloroform and trichloroacetone first



broke through after 3 weeks and dibromochloromethane and tetrachloroethane first



broke through after 4 weeks.
* Mention of commercial products does not imply endorsement by the U.S. Environmental


Protection Agency.

-------
                                     -A18-






     In a second experiment, breakthrough patterns were presented for 1,2,7,9 and




15 weeks of the run plus the respective organic profiles for the influent water.




Twenty-nine compounds were identified from the gas chromatographic profiles for




this experiment.  Empty bed contact time in this experiment was 7.3 - 7.5 minutes.




Again, detection in the effluent of organic species with lower boiling points occurred




sooner than organic compounds with higher boiling points.  Suffet   cautions the




reader, however, that the data are largely qualitative and interpretation of




results is complicated by the highly variable nature of the organic content of




the influent to the adsorbent column.



                            21
     In 1977 McCarty, et al.   reported on the performance of "Water Factory 21"




for removal of organic materials.  Water Factory 21 is an advanced waste treatment




facility designed to reclaim wastewater to provide injection water needed for a




sea water barrier system to protect ground waters in Orange County, California.




Part of the treatment train includes packed-bed, upflow pressure adsorbers filled




with Calgon Filtrasorb 300.  The empty bed contact time is 30 minutes.




     During a period when the plant was operating on a continuous basis, single




activated carbon adsorber influent and effluent samples were taken and subjected


                                                               1 ft
to a rather rigorous organic analysis by closed loop stripping.    Relative




influent and effluent concentrations were reported for sixteen compounds and




absolute concentrations for twelve of these. The general trend was toward removal




of these compounds to widely varying degrees.  An examination of the gas chromatograms




verifies this trend for a large number of unidentified compounds (see section on




"Unidentified Compounds" , page A52, for results of a similar comparison).  The




meaning of the results presented by McCarty, et al., however, are difficult to




interpret in the context of predicting activated carbon adsorber life for




removal of the specific compounds identified, because few data are presented on

-------
                                            -A19-
the condition of the activated carbon itself (time in-place, reactivation frequency,



and so forth).   Measurement of influent/effluent organic concentration profiles



vs. time is part of the planned future Water Factory 21 work, however.



     Most of the information on reducing various concentrations of trace organics



has been gathered through laboratory studies and pilot-scale experiments.  The


                                                   22
National Interim Primary Drinking Water Regulations   established maximum contaminant



levels for six organic chemicals:  endrin, lindane, methoxychlor, toxaphene, 2,4-



D and 2,4,5-TP (Silvex).  These six specific organic contaminants can be grouped



under the general term "pesticides."  The "Manual of Treatment Techniques for


                                                       23
Meeting the Interim Primary Drinking Water Regulations"   reported that adsorption



on granular activated carbon is the most effective treatment process for reducing



the concentrations of these contaminants.



     The U.S. EPA library in Cincinnati made a computer search of the literature



on the subject of adsorption of organic contaminants on granular activated carbon.



Listed below in Table II are 50 organic compounds in addition to those reported



in the text that have been reported to be reduced in concentration through granular



activated carbon treatment.  Only those studies where the concentration of the



specific organics before treatment were below the one mg/Jl level were included,



thus eliminating studies on industrial wastes where the concentrations are usually



much higher.  Even these concentrations are higher than usually found in source



waters, making direct extrapolation or prediction of adsorption behavior speculative



when low concentrations are present. These data are based on both isotherm and



column type studies.

-------
                                       -A20-
                                    TABLE II
ADDITIONAL ORGANIC COMPOUNDS THAT HAVE BEEN REPORTED IN THE LITERATURE TO BE
                    ADSORBABLE ON GRANULAR ACTIVATED CARBON
     acetophenone
           39
          i
           36
                 29
 ,, .  39,43,49
aldrin
     baygorT
     a-BHC
          43,49,50,54
     benzocaine
               32
                 29
     benzole acid
                 29
     butyric acid
     dibrom
     ,.  ,, . 25,39,43,49
     dieldrin
     di  (n-butyl) phthalate
                      48
     di (2-ethylhexyl)  phthalate
                                48
     diuon
     diquat
          50
      30,51
     dimenthoate
                49,50
     m-dinitrobenzene
       P27,39,43,46,49
                     38
     DDT
     endosulfan
            ),5
             26
               49,52
  ,  .  50,53
endrin
     gasoline
     heptachlor
          49
     heptachlor epoxide
                      49
     hexachlorobenzene
            32
                       49
     juglone~
     lindane
       27,31,49,50
                                           49
linuron
MS-22232
malathion
                                             43,49
                                                 49,95
                                    methyoxychlor
                                         ,       38
                                    nitrobenzene
                                    oil (fuel)26>35,37,45
                                            30,51,52
                                    paraquat"
                                    parathion
         50
                                                    ,29
phenylacetic acid
      ,33,34,40,42,45
                                    phenols"
                                    p-nitrophenol
             42
propionic acid
        29
pyridine
  .39^41,49
                                                  29
                                    PCB"
                                    rotenone
                                       .  36
                                    sevin
                                    simazine
        24,32
        28,31,50
          53
strychnine
3-trif luoromethyl-4-nitrophenol (TFM)
I.A-D4"'50
2 , 4-dinitrophenol
2,4,5-T (ester)43'50
     ,    24,50
toxaphene
                                                                              32
                                    tetrachlorobenzene
                                            49
                                    telodrin

                                         triazine
                                                 28

-------
                                      -A21-







     A report by Giusti, et al.   includes a list of 12 alcohols, 8 aldehydes, 11




amines, 4 pyridines (and morpholines), 8 aromatics (benzene derivatives),




11 esters, 3 ethers, 14 glycols and glycol ethers, 2 halogenated hydrocarbon




solvents, 10 ketones, 8 acids, and 2 oxides that were studied.  Single dose




studies (isotherms not determined) were carried out on each of the compounds




with 100 ml of a 1000 mg/£ solution being dosed with 0.5 g of activated




carbon (equivalent dose 5000 mg/£).   Concentrations of solute dosed were




less where solubilities so dictated.  These data were used to draw conclusions




and test hypotheses about effects of pH, polarity, functional groups,




molecular weight, and other differences in physical and chemical characteristics,




in-so-far as amenability to adsorption is concerned.  These data might




relate to relative adsorbability of the respective compounds at these very




high concentrations, but do not evaluate competitive effects of mixtures or




the relative effects of 5 to 6 order of magnitude lower concentrations of




organic matter and adsorbent that are experienced in the drinking water




treatment situation.




     Isotherm tests were run for only 5 compounds at varying pH and although




linear, the isotherms were determined only at high concentrations. Parallel




column studies and multi-solute isotherm studies were conducted on these




compounds and the authors claim a fairly high level of predictability of




the column capacities from the isotherm data.  Whether this translates well




to lower concentrations (where isotherms may become non-linear) or to real




systems where the solute species number in the hundreds, cannot be determined




from these studies.




     In summary, although the individual literature citations are often




vague on critical details of the study, and data have been collected under




a variety of circumstances, many atypical, in total they do demonstrate




that granular activated carbon is correctly described as a "broad-spectrum"




adsorbent.

-------
                                    -A22-





     Reactivation of Granular Activated Carbon




     Although the internal pore structure providing a large surface area




per unit weight is developed during initial manufacture (thermal activation)




of granular activated carbon, the surface area is finite and eventually




becomes covered with adsorbate, and adsorption ceases.  To continue with




effective adsorption the granular activated carbon is processed to remove




these adsorbed materials (reactivation).  The most common technique for




processing granular activated carbon to renew the adsorption capacity of




its surfaces is to drive off and oxidize the adsorbed organic compounds in




environment containing steam and little oxygen at high temperature (approaching




1000°C). The problem is to design a facility and choose reactivation conditions,




such that the maximum amount of adsorbed materials is removed with a minimum




of change in the properties of the granular activated carbon.




     The four basic types of furnaces currently (Fall 1977) in use for the




reactivation of granular activated carbon are:  the multiple-hearth, the




rotary-kiln, the infra-red-tunnel, and the fluidized-bed furnace.  Of the




five currently operating on-site reactivation facilities in Europe, two are




multiple-hearth furnaces and three are various designs of  the fluidized bed




furnace.  Figure 6 is a schematic diagram of a multiple-hearth furnace,




Figure 7 is a cross-section of a one-bed fluidized bed furnace, Figure 8 is




a diagram of the infra-red-tunnel furnace, and Figure 9 shows a rotary-kiln




furnace.  The Water Supply Research Division currently has underway three




projects for the evaluation of reactivation factilites.




     Proper design and operation of a thermal reactivation facility is




necessary to avoid any unwanted change in the properties of granular activated




carbon.  For example, Juhola   reported  a change in pore size distribution




upon several cycles of reactivation as shown in Table III.

-------
                        -A23-
                FEED IN
                 GAS OUT
BURNER, STEAM
ENTRY OR AIR
INJECTION    •
      GAS FLOW
       PRODUCT OUT
                       BRICK HEARTHS
                       SUPPORTED AT
                       WALL ONLY

                       RABBLE ARM (NOT
                       SHOWN ON OTHER
                       HEARTHS)
                                                GRANULAR
                                                ACTIVATED
                                                CARBON
CENTER SHAFT
  (ROTATES)
                                          SHOWING SOLIDS IN
                                          PERSPECTIVE (TYPICAL
                                          FOR ALL HEARTHS)
         FIGURE 6  CROSS SECTION MULTIHEARTH FURNACE

-------
                                -A24-
GAS INLET
                 GRANULAR
                 ACTIVATED
                  CARBON
                    INLET
     GAS
     OUTLET
                           OUTLET
                        GAS CHAMBER
 INCOMING
AIR CHAMBER
                                                    REACTIVATED
                                                     GRANULAR
                                                     ACTIVATED
                                                      CARBON
     FIGURE 7  CROSS SECTION OF FLUIDIZED BED FURNACE

     NOTE: THIS DIAGRAM DEPICTS A ONE-BED FURNACE AS AN EXAMPLE.
          OTHER DESIGNS OF FLUIDIZED BED FURNACES ARE COMMERCIALLY
          AVAILABLE.

-------
                             -A25-
UJ
OC
       OC
       UJ
       N
   l_      5
UJ < -I   ^
o oc o   z
< UJ OC   UJ
z o. j-   o
OC 2 Z   >•
3 UJ O     .
u. h- O   O
                                                            UJ
                                                            o
                                                            UJ
                                                            o
                                                            UJ
                                                            cc
                                                            o
                                                            z
                                                            o

                                                            o
                                                            UJ
                                                           CO
                                                           (0
                                                           o
                                                           oc
                                                           o

                                                           00
                                                           UJ
                                                           oc

                                                           o

-------
-A26-
                         LU
                         O
                         QC
                         D
                         U.
                         O
                         cc
                         U.
                         O
                         o
                         LU
                         CO
                         (/)
                         O
                         CC
                         O
                          O)
                          LU
                          CC
                          D
                          g
                          u_

-------
                                      -A27-
                                     TABLE III

CHANGES IN GRANULAR ACTIVATED CARBON PROPERTIES ON SUCCESSIVE REACTIONS
Cycle
Ash Content - %
Iodine Number*
Measure of
Pores Between
1 nm and 2.5 nm
Molasses Number*
Measure of
Pores Greater
than  2.8 nm   Bulk Density
Initial
1
2
3
5.7
7.6
8.6
9.5
1090
1040
935
940
250
310
290
350
0.469
0.468
0.469
0.473
*These numbers are based on the adsorption of the respective material under
a standard set of conditions.

     This change in pore structure was then reflected in a decrease in

performance as measured by a decrease in initial percent chemical oxygen demand

reduction from 55 percent to 40 percent over eight reactivation cycles.

     Smith   agreed pointing out that based on Figure 10, practically no change

occurred in pore size distribution in the 1 to 1.9 nm diameter range, although a

marked change occurred in the pores between 1.9 and 2.5 nm in diameter.  He also

noted that from 2.5 nm diameter and above, again, very little change occurred in

                                                 58
the pore structure.  Finally, DeJohn and Hutchins   also reported that the

effect of thermal reactivation may be to reduce the surface area in the

micropores of granular activated carbon.  They further state that the properties

of granular activated carbon derived from lignite are not changed as much as

the properties of granular activated carbon derived from another source,

bituminous-coal.  Their data showed a decline in specific surface area of
                                                  2
bituminous-coal based activated carbon from 1100 m /gm for virgin material to 700
 2
m /gm (36 percent) after four reactivations, although lignite-based activated
                               i-\
carbon only declined from 650 m^/gm for virgin material (initially lower) to 500
 2
m /gm (23 percent) after five reactivations.

-------
                           -A28-
O
O
O

uT
O

UJ
cc
O
Q.
O
OC
O
UJ
O
.22

.20

.18

.16

.14

.12

.10

.08

.06

.04

.02
             ORIGINAL
	_TJT^--3rd REACT.
Z-~-	1st REACT.
                              	5th REACT.
        L 5TH
                  ORIGINAL
       1.0    1.5  2.0   3.0 4.0 5.0 6.0  8.0 10.0  15.0  20.0
                MICROPORE DIAMETER , NANOMETERS

FIGURE 10. CUMULATIVE MICROPORE VOLUMES AS A FUNCTION
           OF MICROPORE DIAMETER
           REFERENCE 57

-------
                                   -A29-
                                          59
     In contrast to these reports, Juntgen   noted that, based on Figure 11, an

appropriate time-temperature relationship (800°C - 20 min.) during reactivation

can be developed that will not alter activated carbon significantly from its

virgin state.  In support of this, Weissenhorn   collected data on the decline of

granular activated carbon performance as measured by ultraviolet absorbance of

the effluent with relation to the volume of water treated per unit volume of

granular activated carbon and showed, according to Figure 12, that virgin activated

and reactivated carbon were nearly equivalent. Finally, the Shirco Company of

Dallas, Texas, claims to have collected data as shown in Figure 13 and Table IV.

(unpublished promotional literature).


                                  TABLE IV
   TYPICAL RESULTS OF CARBON REACTIVATION IN THE SHIRCO ELECTRIC INFRARED FURNACE
                                         Apparent
       Activated                         Density      Iodine     Molasses
       Carbon                            (g/cc)       Number       Number
ICI HYDRODARCO 3000
Virgin
Spent
Reactivated
CALGON FILTRASORB 400
Virgin
Spent
Reactivated
WESTVACO NUCHAR WV-L
Virgin
Spent
Reactivated
CARBORUNDUM
Virgin
Spent
Reactivated

0.352
0.458
0.397

0.48
0.62
0.48

0.594
0.678
0.593

0.525
0.667
0.523

550
413
596

1167
480
1122

834
588
868

950
320
1132

333
302
380

355
250
376

371
178
380

220
137
230

-------
                  -A30-
  Z 30
  UJ
  o
  oc
  ui 20
  UJ
  2-10
01
z g
5 u
     12
< H -20
Q UJ
"g
gS-40
                             t=20 min_
                             t=15 min
          2   4    6   8   10   12
             REACTIVATION CYCLE
                                    14
T=800 °C
      t=15 (min)
                             t=20 (min)
                            fe30
                                  in)
           2   4   6   8   10   12   14
              REACTIVATION CYCLE
     FIGURE 11. EFFECT OF RESIDENCE TIME
               ON ACTIVITY AND CHANGE OF
               WEIGHT OF ACTIVATED CARBON
               REFERENCE 59

-------
                       -A31-
     100
     50
   a
   cc
   o
   (A
   O
   3 20
                     ACTIVATED

                     CARBON "F"
                  REACTIVATED


                     VIRGIN
                I
     10
       024   6  8  10 12

     THROUGHPUT, CUBIC METERS /LITER

          OF ACTIVATED CARBON
    100

      A
   3? 50
   a.
   cc
   O
   (A
   O
   3 20
                     ACTIVATED

                     CARBON "L"
VIRGIN


 REACTIVATED
     10
       024   6  8  10  12

    THROUGHPUT, CUBIC METERS/LITER

         OF ACTIVATED CARBON
FIGURE 12  INFLUENCE OF REACTIVATION ON

          GRANULAR ACTIVATED CARBON

          PERFORMANCE

          REFERENCE 60

-------
                     -A3 2-
DC
UJ
  100 -I
±  70^
O
O
      EFFECT OF REPEATED REACTIVATION BY
              THE SHIRCO PROCESS
        VIRGIN ACTIVATED CARBON (MINIMUM)
       S: SPENT ACTIVATED CARBON O
       R: REACTIVATED CARBON •
       VIRGIN S-1 R-1  S-2 R-2  S-3  R-3
   FIGURE 13
INFLUENCE OF REACTIVATION ON
IODINE NUMBER
FROM SHIRCO, INC. PROMOTIONAL
LITERATURE

-------
                                       -A3 3-






     In summary, some changes may occur in granular activated carbon properties




during thermal reactivation, and any reactivation system should be carefully




designed to minimize the changes.  Because some granular activated carbon losses




occur through burn-off and mechanical attrition, fresh granular activated carbon




will be added during each operating cycle. The addition of this fresh material




will help overcome any losses in performance of the reactivated granular carbon,




as compared to virgin material.




     A final design consideration for any reactivation system must include proper




consideration of handling the off-gases.  Both dust collecting devices and gas




after-burners may have to be considered in certain circumstances. Reactivation




systems, however, can be designed in order to avoid any possible problems of air




pollution during the reactivation process.




     Because, as discused above, most water utilities using granular activated




carbon as part of their water treatment process are doing so to control tastes




and odors, and because tastes and odors are removed by granular activated carbon




adsorption beds for periods of several years, reactivation of granular activated




carbon for drinking water purposes is not widespread.  In most cases, when a




water purveyor changes the granular activated carbon charge in its treatment




plant,  the material that is replaced in the beds is virgin rather than




reactivated granular carbon.  The only known on-site reactivation facility




at a water treatment plant in the Unted States was at Nitro,  West Virginia,




a plant that has now been closed for several years because the utility




built a new plant on a cleaner source of water.   At this writing (Fall




1977) the five known on-site reactivation facilities at water treatment




plants  are all in Europe, one in Switzerland, two in the

-------
                                       -A3 4-
Federal Republic of Germany, one in Sweden, and one in the United Kingdom.


     If  granular activated carbon adsorption systems are to be used to control


other organic contaminants beyond taste and odor producing compounds, as noted in


the literature review above and as will be discussed in the sub-sections below,


the period between reactivations will be much shorter than currently practiced.


Current Water Supply Research Division Research Findings


     In an attempt to provide additional information not currently in the literature


on the performance of granular activated carbon as an adsorption medium, the


Water Supply Research Division has an active in-house and extramural research


program on this subject  The results of these experiments, many still on-going,


are summarized below.  Detailed papers on many of these projects will eventually


be published in the technical press.


     Class I Compounds (Taste and Odor Producing)


     Snoeyink at the University of Illinois*working under EPA sponsorship found


that the odorous compounds 2-methylisoborneol (MIB) and geosmin are both strongly

                             f\~\ f\ 9
adsorbed by activated carbon.   '   When present, humic substances significantly


reduce the capacity of activated carbon for adsorption of these compounds, more


so before equilbrium is achieved than at equilibrium.  Commmercial humic acid


(HA) and the humic substances from well water each had differing competitive


effects on MIB.  The capacity of activated carbon for geosmin adsorption was


reduced by commercial HA to a greater extent than was observed for MIB.  The


performance of laboratory columns was consistent with the isotherm results.


Application of distilled water  to a partially saturated activated carbon bed


resulted in almost no elution of MIB indicating  that it was strongly adsorbed.

-------
                                           -A3 5-
     Using the data collected in this study and assuming, 1) complete saturation



of the activated carbon, 2) no desorption and 3) no biological activity, Snoeyink



predicted the bed life for the reduction of MIB or geosmin from 10 yg/& to its



threshold odor level of 0.1 yg/£ in a 7-8 minute empty bed contact time bed to



be much greater (several months to years) than the predicted life for the reduction



of humic substances from 5 to 1 mg/Jl (1 to 2 months).  When both MIB, or geosmin,



and humic substances must be removed, humic substance removal will control the



life of the bed.



     Chlorophenols are adsorbed very strongly by activated carbon at the yg/£



level, which is near the threshold odor limit for these compounds. The extent of



adsorption of 2,4-dichlorophenol (DCP) and 2,4,6-trichlorophenol  (TCP) is a



function of pH.  The neutral species of these compounds predominate at pHs below



the pK  (pH at which the concentrations of the free acid and the acid anion are
      3.


equal) values (7.85 and 6.00, respectively, at 25°C) and are adsorbed more strongly



than the anionic species.  As the number of chlorine atoms substituted on the



phenol increases, the solubility of the neutral species decreases and the adsorbability



increases.  As substitution increases, the pK  of the species is lowered, however.
                                             a


     When water containing phenol is chlorinated with low levels of chlorine, a



mixture of chlorophenols will form and thus the extent of adsorption of one



chlorophenol in the presence of another chlorophenol is an important consideration.



Significant reductions in adsorption capacity (up to 50 percent) of one chlorophenol



was caused by the presence of a second chlorophenol. Evaluation of the competitive



effects of commercial HA, soil fulvic acid (FA) and leaf FA showed that the



presence of these materials decreased the capacity of activated carbon for chlorophenol



adsorption and that each of the materials competed somewhat differently.

-------
                                       -A3 6-





Even in the presence of humic substances and another chlorophenol species,




however, the adsorption capacity is even greater for chlorophenol than it is for




MIB and that bed life for chlorophenol adsorption will be greater than for MIB




and much greater than for humic substances.




     In another study, a joint effort between the AWWA Research Foundation, 14




water utilities, the EPA Water Supply Research Division,  and the Universities of




Illinois, Iowa State and Missouri-Columbia,  is underway to determine the removal




of trace organics (particularly taste and odor compounds  and haloforms) on granular




activated carbon and polymeric adsorbents.




     In this study, pilot scale columns containing seven  different adsorbents are




located at the Kansas City Water Treatment Plant, Kansas  City, Missouri.  The




adsorbents include granular activated carbon made from bituminous coal, lignite,




petroleum, and peat, and carbonaceous anion exchange resins.  The adsorbers are




presently arranged so that both a coal-base-and a lignite-base granular activated




carbon system can provide up to 33 minutes empty bed contact time (EBCT).   The




applied water is unstable because of precipitative lime softening and a lack of




recarbonation, resulting in calcium carbonate deposition  on the adsorbents.  The




effect of this will be evaluated.   Work will continue on the pilot column studies




until the spring of 1978 and the final report will be available by early summer.




Taste and odor removal results are not available at this  time (Fall 1977).




     Class II Compounds (Synthetic Source Water Contaminants)




          Naphthalene




     A long-term (started July 7, 1974) experiment comparing the adsorption of




naphthalene with that of background organic content of Cincinnati tap water on




granular activated carbon beds was terminated during June 1975.

-------
                                       -A3 7-
330-day span of the experiment, 18,500 liters of Cincinnati tap water spiked with

an average concentration of 30 yg/£ of naphthalene were passed through the granular

activated carbon column (16.9 min EBCT).  Although the non-purgeable organic

carbon (NPOC) 50 percent removal front had penetrated the entire 27-inch length

of the column by May 1975, the maximum penetration of the naphthalene 50 percent

removal front was only about 6.5 inches (this was quite variable throughout the

experiment). Therefore granular activated carbon columns can be expected to

remove compounds of low polarity and solubility such as naphthalene for a much

longer period of time than they can remove the diverse organic group represented

by NPOC.

     The variability of the depth of the naphthalene penetration into the adsorbent

(1-6.5 inches) is something of a mystery.  Suspected causes are (1) variable

influent concentrations,  (2) variable constitution of NPOC (competitive adsorption),

(3) biological activity,  (4) variable flow,  (5) variable temperature, or (6)

influence of backwashing.

          Other Polynuclear Aromatic Hydrocarbons
                                                      r n
     Snoeyink, at the University of Illinois, reported   that limited

experimentation with the  polynuclear aromatic hydrocarbon  (PAH) anthracene  led

to the conclusion that no significant association between  its adsorption and that

of humic substances occurred.  Thus the possibility of PAH passage through

granular activated carbon beds because of its association with the less adsorbable

humic substances is not a cause for concern.

          Carbon Tetrachloride

     The adsorption of this contaminant on granular activated carbon beds was

observed in  the U.S. EPA  Cincinnati Laboratory during the  fall and winter of

1976-77 when carbon tetrachloride was  in  the tap water.  At that time,  granular

activated carbon columns  that had been in service for two  months were being

monitored for  trihalomethane reduction.   Table V shows the monthly average  carbon

-------
                                        -A38-
tetrachloride concentrations in the tap water and corresponding monthly average

concentrations in the effluent.  In spite of being in service for two months, the

granular activated carbon was effective for 3 to 5 months, probably because this

contaminant was "new" to the adsorbent when it occurred during the third month.

Low level desorption of the carbon tetrachloride continued for several months

after the contaminant disappeared from the influent.


                                  TABLE V

    ADSORPTION OF CARBON TETRACHLORIDE ON GRANULAR ACTIVATED CARBON
                                  Time in Service, months
                          2     3     4     56789    10    11
Monthly Average
Influent,
Monthly Average
Effluent, yg/£
NF
NF
NF    5
13    15   44   NF   NF   NF    NF
NF    NF    NF   NF    <6*  14   16   13
Filtrasorb 400 granular
activated carbon
10 min. Empty bed contact time
NF = none found
* - Single Value during month
     During the first week of January 1977 a 6-inch diameter by 30-inch

depth bed (11 minutes EBCT) of virgin Calgon Filtrasorb 400 granular

activated carbon was placed in upflow operation treating Cincinnati tap water,

Coincidently with the start of operation, the concentration of carbon

tetrachloride in the tap water began to increase.

     Figure 14 shows the performance of the adsorbent bed during the first

43 weeks of 1977.  During the first 7 weeks, the granular activated carbon

effectively removed carbon tetrachloride during the period of extremely high

influent concentrations.  As the influent concentration returned to the

limit of detection (0.1 yg/£), the adsorbent, acting under the influence

of the adsorption-desorption equilibrium phenomenon, began to desorb carbon

-------
                        -A39-
O)
oc

LU
O
O
o
oc
O
oc
LU
O
m
oc
<
o
   60
   50
   40
   30
   20
   10
                               UPFLOW OPERATION
                                11 MINUTES EBCT
             INFLUENT (CINCINNATI TAP WATER)
^ GRANULAR ACTIVATED
CARBON COLUMN EFFLUENT
     0    5   10   15   20   25  30  35  40   45   50
                  TIME IN SERVICE, WEEKS

     FIGURE 14 PERFORMANCE OF GRANULAR ACTIVATED
              CARBON BED FOR CARBON TETRACHLORIDE
              REMOVAL

-------
                                         -A40-
tetrachloride.  This desorption, which was monitored approximately every 6 weeks,

has continued for about 9 months.

     Figure 14 is important because it illustrates the protection a granular

activated carbon adsorption barrier can afford the consumer during periods of

sudden contamination of the raw water supply.  These data also demonstrate that

granular activated carbon reactivation will probably be necessary fairly frequently

to afford continuing consumer protection and guard against desorption of unwanted

contaminants.

          Trichloroethylene, Tetrachloroethylene and Similar Chlorinated
          Organics

     Because several incidences of tri- and tetrachlorethylene contamination of

drinking water supplies have been reported, the Water Supply Research Division is

currently (Fall 1977) studying the effectiveness of granular activated carbon

beds for removing these pollutants.  In most instances the water source is ground

water and the contamination can be linked to some industrial activity involving

the present or past use of cleaning solvents in the aquifer recharge area.

Columns containing granular activated carbon (10 minute empty bed contact time)

have been installed in three locations in New England and the first few weeks of

data indicate that tri- and tetrachloroethylene are very well removed when the

adsorbent is fresh, see Table VI.  As indicated in the footnote to Table VI, the

contaminant, 1,1-dichloroethane was not as well removed.  At another location

(not shown in Table VI) the tetrachloroethylene contamination has exceeded 2500

Pg/&> Yet the granular activated carbon bed produced an effluent with less than

0.5 yg/£ tetrachloroethylene, for 15 weeks.  These tests are continuing.

-------
                                        -A41-

                                     TABLE VI

USE OF GRANULAR ACTIVATED CARBON BEDS TO REMOVE ORGANIC CONTAMINANTS FROM

                                 A GROUND WATER
Contaminant
Influent
Concentration
Range, yg/£
                                            Average Effluent Con., yg/£
                                            0-4 weeks    4-8 weeks   8-12 weeks
1,1,1-trichloroethane   10.9-36.7

trichloroethylene        1.9-7.7

tetrachloroethylene      0.1-16.9

NF = None Found
NF

NF

NF
                                                            NF
                                                  0.4
Note:  After six weeks of testing, two new contaminants, 1,1-dichloroethane
and cis-l,2-dichloroethylene began appearing in the influent to the granular
activated carbon column.  No cis-1 , 2-dichloroethylene has yet been detected
in the treated water, but the average concentration of 1,1-dichloroethane
in the effluent for the 4 to 8 week period after first appearing was 0.7
          Prediction of Granular Activated Carbon Bed Performance, Studies with
          Humic Acid and Carbon Tetrachloride, Dieldrin, PCB, Benzene, and
          p-Dichlorobenzene

     Weber, at the University of Michigan, is studying the adsorption of

several specific organic compounds on granular activated carbon.  The main thrust of

this WSRD grant activity is to investigate the possibility of predicting activated

carbon column adsorber performance on the basis of mathematical parameters

obtained from simple batch kinetic and equilibrium (isotherm) studies.  Although

the studies include only the use of model systems (humic acids as the

competitive species) , certain general conclusions about relative activated

carbon adsorber operating life can be drawn on the basis of some of the data

obtained to date from the modeling studies.

     The results of the work with carbon tetrachloride support the in-house work

reported above in that an adsorber life of several weeks to months can be expected.

The work with dieldrin indicates that this compound is very strongly adsorbed,

supporting evidence to that effect previously reported in the literature. ^3

PCB mixtures resemble pesticides in properties and are also strongly adsorbed.

-------
                                     -A42-



Adsorber operating life for PCB's are likely to be on the order of months to one


year.  Adsorber operating periods for treatment of benzene and jD-dichlorobenzene


are likely to range somewhere between carbon tetrachloride and the PCS mixtures.


          Mixtures from Actual Waters


     In an attempt to further evaluate the capability of granular activated


carbon beds to remove actual mixtures of raw water organic contaminants in competition


with other organics, several studies are being conducted on actual waters at


locations other than Cincinnati, Ohio.  One of these research projects, using


bench scale pilot granular activated carbon columns is being performed with a


ground water in Southern Florida. The granular activated carbon columns are 1-


inch in diameter and contain 30-inches of activated carbon.  The bed was loaded

           2
at 3 gpm/ft , which resulted in a 6.2 minute empty bed contact time.  The influent


to the granular activated carbon column was untreated ground water for the first


phase of the study and finished water for the second and third phase. The source


has a color of approximately 50 color units, a TOC concentration of about 10


mg/£, and a pH of about 7.1.


     The project includes routine analysis for the concentration of 19 organic


compounds, total organic carbon, and trihalomethane formation potential, both


before and after the granular activated carbon column. The average influent


concentration to the granular activated carbon bed for the duration of the study


is shown in Tables VII, VIII, and IX. Tables X, XI, and XII show the corresponding


effluent concentrations for specific substances with time over the duration of


each study.  These data show that at these short  contact times, some organic


compounds did break through fairly early  in the test, but that the chlorinated


aromatic compounds were well removed.  Longer empty bed contact time tests are


now underway, see pages A69 to A78.

-------
                                   -A43-




                                TABLE VII




  AVERAGE CONCENTRATIONS OF TENTATIVELY IDENTIFIED ORGANICS INFLUENT TO




                      GRANULAR ACTIVATED CARBON BED




                       (Jan. 18 - May 20, 1977)









     Compound                                     yg/£




Vinyl chloride                                    0.8




Methylene chloride                                0.08




Methyl iodide                                      nil




trans-1,2-Dichloroethylene                         1.3




1,1-Dichloroethane                                 0.3




cis-1,2-Dichloroethylene                          29.0




Chloroform           (Tr. to 2.1)                  0.16




1,1,1-Trichloroethane




1,2-Dichloroethane	^(Sum)                     0.11




Carbon tetrachloride




Trichloroethylene                                  0.13




Bromodichloromethane    (Tr to 0.9)                0.11




Tetrachloroethylene                                0.06




Dibromochloromethane    (Tr. to 0.4)               0.04




Chlorobenzene                                      0.19




Bromoform               (Tr. to 0.19)              0.02




p-Chlorotoluene                                    0.11




o,m,p-Dichlorobenzene (sum)                        1.1

-------
                                   -A44-

                              TABLE VIII
AVERAGE CONCENTRATION OF TENTATIVELY IDENTIFIED ORGANIC COMPOUNDS
IN INFLUENT TO GRANULAR ACTIVATED CARBON  BED (Aug.  26 to Oct.  18, 1977)
                         EBCT =6.2 minutes
      Chemical Name
   Vinyl chloride                                5.4
   Methylene chloride                         not determined
   trans-1,2-Dichloroethylene                    1.0
   1,1-Dichloroethane                            0.3
   cis-1,2-Dichloroethylene                     19
   Chloroform                                   57
   1,1,1-Trichloroethane
   1,2-Dichloroethane	-(Sum)                  5.3
   Carbon tetrachloride
   Trichloroethylene                            0.1
   Bromodichloromethane                        39
   Tetrachloroethylene                          nil
   Dibromochloromethane                        27
   Chlorobenzene                                0.8
   Bromoform                                    2.5
   £-Chlorotoluene                              0.2
   m-Dichlorobenzene
   £-Dichlorobenzene—— (Sum)                   0.3
   o-Dichlorobenzene

-------
                                        -A45-







                                       TABLE IX




AVERAGE CONCENTRATIONS OF TENTATIVELY IDENTIFIED ORGANIC COMPOUNDS IN INFLUENT




                 TO GRANULAR ACTIVATED CARBON BED, EBCT =6.2 min.




                      (Run No. 2, Nov. I, 1977 - Jan.  3, 1978)









             Compound                             yg/£




     Vinyl chloride                               7.4




     Methylene chloride




     trans-1,2-Dichloroethylene                   0.63




     1,1-Dichloroethane                           0.11




     cis-l,2-Dichloroethylene                    24.5




     Chloroform                                  76.6




     1,1,1-Trichloroethane




     1,2-Dichloroe thane	^(sum)                7.9




     Carbon tetrachloride




     Trichloroethylene                            0.52




     Bromodichloromethane                        48.4




     Tetrachloroethylene                         nil




     Dibromochloromethane                        29.0




     Chlorobenzene                                0.56




     Bromoform                                    1.95




     £-Chlorotoluene                              nil




     m-Dichlorobenzene                            nil




     p-Dichlorobenzene                            0.19




     o-Dichlorobenzene                            0.09

-------
                                       -A46-

                                       TABLE X

       CONCENTRATIONS OF TENTATIVELY IDENTIFIED ORGANICS PRESENT IN EFFFLUENT
              FROM ADSORBER,  EBCT = 6.2 min. (Jan.  18  - May 20,  1977)

                     All data in yg/£ - Compare to  Table VII
Time,
Weeks
0
1
2
3
4
5
6
7
8
9
10
11
12
13+
14
15
16
17


0

0

0
0
0
0
0
0
0
0
0
0
0
A
N
N
N
.002
N
.002
N
.002
.04
.02
.05
.07
.07
.66
.45
.90
.76
.60
B
N
N
N
0.02
0.08
0.007
0.013
0.015
0.02
0.03
0.04
0.1
0.02
0.32
0.17
0.38
0.34
0.23
C
N
0.18
0.2
2.7
4.2
7.8
5.3
12.0
18.0
17.3
21.6
10.2
18.7
25.7
30.4
23.3
21.9
21.3
D
N
N
N
0.002
N
N
N
N
N
N
N
0.01
0.006
0.06
0.07
0.05
0.06
0.028
E
0.007
0.006
0.004
0.001
0.002
N
N
N
N
N
N
N
N
N
N
N
N
N
F
N
N
N
0.01
N
N
N
N
0.001
N
N
N
N
N
N
N
N
N
G
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
H
0.02
N
0.02
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
I J
0.
0.
N
0.
N
0.
N
0.
0.
0.
0.
N
0.
0.79 0.
0.70 0.
0.81 0.
0.75 0.
0.78

07
003

003

006

002
003
003
002

03
01
10
06
05
N
- = Not determined
N =
A =
B =
C =
None found
trans-1 , 2-Dichloroethylene
1 , 1-Dichloroethane
cis-1,
2-Dichloroethylene
D = Trichloroethylene
E = Tetrachloroethylene
F = Chlorobenzene
G = p-Chlorotoluene
H = m,p,o-dichlorobenzene
I = Vinyl chloride
J = fl,l,l-Trichloroethane
   |l,2-Dichloroethane
   ^Carbon tetrachloride

-------
                                        -A47-

                                   TABLE XI



CONCENTRATIONS OF TENTATIVELY IDENTIFIED ORGANICS PRESENT IN EFFLUENT FROM ADSORBER

           EBCT = 6.2 min., (Run No. 1, Aug. 26 - October 18, 1977)

                  All Data in \ig/i - Compare to Table VIII
Time
Weeks A
0
1
2
3
4
5
6
7
N =
A =
B =
C =
N
N
0.
N
N
0.
N
N
B C D E
N N N N
N N N N
40 N 10.7 0.59 N
0.55 2.19 0.03 N
N 2.87 N N
12 0.62 8.9 0.003 N
N 8.77 N N
N 9.91 N N
F G H I J
N N N - N
N N N 0.165 N
N N N 0.64 N
N N N 0.33 N
N N N N 0.31
N N N 4.7 0.34
N N N 3.25 0.09
N N N 1.85 N
None found
trans-1,
2-Dichloroethylene

1 , 1-Dichloroethane
cis-1 , 2-Dichloroethylene
D = Trichloroethylene
E = Tetrachloroethylene
F = Chlorobenzene
G = _p_-Chlorotoluene
H = m,p,o-Dichlorobenzene
I = Vinyl chloride
J =p.,l,l-Trichloroethane
   <1,2-Dichloroethane
   (Carbon tetrachloride

-------
                                     -A48-

                                  TABLE XII

CONCENTRATIONS OF TENTATIVELY IDENTIFIED ORGANICS IN EFFLUENT FROM ADSORBER

          EBCT = 6.2 min. (Run No. 2, Nov. 1, 1977 - Jan. 3, 1978)

                    All data in yg/£, Compare to Table IX
Time
Weeks A
0
1
2
3
4
5
6
7
8
9
N =
A =
B =
C =
N
N
N
N
N
N
N
N
0.58
0.97
None found
B C
N N
N N
N N
1.05 1.28
0.21 3.8
N 11.1
0.30 5.5
0.8 12.0
0.66 9.1
N 9.1

D
N
N
N
N
N
N
0.22
0.29
0.57
0.04

E
N
N
N
N
N
N
N
N
N
N

F
N
N
N
N
N
N
N
N
N
N

G
N
N
N
N
N
N
N
N
N
N

H
N
N
N
N
N
N
N
N
N
N

I
0.30
N
N
N
N
N
N
N
N
N

J
N
N
N
N
N
N
N
N
N
N

K L
N N
3.2 1.2*
8.2 N
N N
5.2 N
N N
1.7 l.f
1.8 4.9
2.5 4.C
4.4 4.7

trans-1 , 2-Dichloroethylene
1 , 1-Dichloroethane
cis-1, 2-Dichloroethylene
D = Trichloroethylene
E = Tetrachloroethylene
F = Chlorobenzene
G = £-Chlorotoluene ?•
H = m-dichlorobenzene
I = _p_-Dichlorobenzene
J = jo-Dichlorobenzene
K = Vinyl chloride
L = fl,l,l-Trichloroethane
   t1,2-Dichloroethane
    (Carbon tetrachloride

-------
                                     -A49-





     Another actual plant site research project is being performed using full-




 scale granular activated carbon beds located at a water treatment plant in




 the  lower Mississippi Valley.  In this project one bed is used as a post-filter




 adsorber and another as a sand replacement system in parallel operation.  The




 post-filter adsorber is an existing conventional rapid sand filter that had




 the  30 inches of sand replaced by 30 inches of 12x40 mesh granular activated




 carbon.  The post-filter adsorber is in series after a rapid sand filter,




 so that the granular activated carbon received coagulated, settled, softened, and




 filtered water.  The empty bed contact time was 27 minutes at the start of




 the  test and 20 minutes at the end because of granular activated carbon




 loss, caused by inadvertent excessive backwashing.




     The sand replacement system is an existing rapid sand filter that had




 the  top 24 inches of the 30 inches of sand replaced by 24 inches of 12x40 granular




 activated carbon.  This unit receives coagulated, settled and softened water




 directly from the precipitator of the full-scale plant.  The empty bed contact




 time was 24 minutes at the start of the test and 18 minutes at the end because




 of granular activated carbon loss caused by inadvertent excessive backwashing.




     The project includes routine analysis for the concentration of at least




 35 organic substances and total organic carbon and trihalomethane formation




 potential both before and after the granular activated carbon beds.  Table XIII




shows the effluent concentration for each substance found, as contrasted to




 the  influent concentrations listed at the bottom of the Table for the duration




 of the study, 25+ weeks.  For this report only the post-filter adsorber data




 will be shown as an example.  Note, for this project, too, all data is




 preliminary in nature as the gas chromatograph-mass spectrometry confirmations




 have not yet been evaluated for consistency of identification (Fall 1977).

-------
                                         -A50-

                                    TABLE XIII

            REMOVALS  OF  TENTATIVELY  IDENTIFIED ORGANICS BY POST-FILTER
                        GRANULAR ACTIVATED CARBON ADSORBER

                  Empty  Bed  Contact  Time, 27 min. Start,  20 min.  End
Time,
Weeks
0
1
2+
3
4
5
6
7
8
9
10+
11+
12+
13+
15+
20+
25+
Benzene
Effl. Cone.
ygM
ND
0.1
ND
ND
ND
0.1
ND
ND
ND
ND
ND
0.1
ND
ND
ND
ND
ND
1 , 2-Dichloroethane
Effl. Cone.
ygM
0.4
0.2
ND
0.1
ND
0.3
1.1
0.7
0.8
1.3
1.4
1.0
2.0
3.5
4.5
10.1
9.4
Trichloroethylene
Effl. Cone.
vg/A
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.4
ND
ND
ND
ND
ND
ND
ND
Toluene
Effl. Cone.
yg/£
ND
0.1
ND
ND
ND
0.1
ND
ND
ND
ND
ND
0.2
ND
ND
ND
ND
ND
Avg.  Inf.
Cone. yg/£  0.3         8.0

Range Inf.
Cone. yg/£  ND-2.1    1.2-23.7
                                         0.3
                                      ND-0.9
 0.2
ND-0.9
ND
Not detected

-------
                                    -A51-







     Within the precautions stated, Table XIII can be used to show the




trends for removal by the granular activated carbon bed.  The adsorber




removed the ambient concentrations of toluene, benzene and trichloroethylene




more efficiently than other substances consistently present during the




first 10+ weeks of operation.  The 1,2-dichloroethane was removed at an




average of greater than 80 percent for about 8 weeks, but the effluent




concentration consistently exceeded 0.5 yg/£ after six weeks.  Desorption




occurred from about the 16th week to the end of the test.  Preliminary




data show that trace concentrations (ng/£) of chlorinated hydrocarbon insecticides




were generally reduced to below detectable concentrations throughout the 25+




week study.  Finally, the sand replacement system performed similarly to the




post-filter adsorber, but the removals were not as long-lasting.




     A third study on the performance of granular activated carbon beds is




being conducted in full scale at a water treatment plant in the upper Ohio




Valley.  As this project is just starting, the data are preliminary, mainly




because the identification of specific organics thus far is based only on




gas chromatographic retention times without mass spectrometry confirmation.




     The following generalizations are made after 52 days of virgin granular




activated carbon being on-line.  Compounds for which granular activated




carbon adsorption may serve as a control mechanism and their influent concentration




ranges are: 1,1,1-trichloroethane or carbon tetrachloride or both (0.17 -




1.14 yg/£, not resolved); 1,2-dichloropropane or trans-1,3-dichloropropylene




or both (0.10-0.64 yg/£) (not resolved); trichloroethylene (0.16-0.96 yg/£);




cis-1,3-dichloropropylene or 1,1,2-trichloroethane or both (0.15-24.0 yg/£)(not




resolved); dichloroiodomethane (0.10-1.22 yg/£) and chlorobenzene (0.24-3.36




yg/£).  Although present occasionally, these compounds were not consistently




present in the influent to the adsorber, therefore removals could not be calculated.

-------
                                      -A52-

          Unidentifled Compounds


     After concentration, gas chromatography can be used to separate many organic

compounds, producing a gas chromatogram in which the separated organic compounds


are represented by "peaks" on a chart.  Although unidentified, the absence of


certain peaks after a given type of treatment gives an indication of the success
                                                                           Q
of the treatment. In 1972, influent and effluent carbon chloroform extracts

from an operating granular activated carbon sand replacement system in Lawrence,

Massachusetts were compared gas chromatographically.  Figure 15 shows the


reduction of many of the organic peaks when the granular activated carbon was

fresh, but after 16 weeks of operation the influent and effluent gas chromatograms

were similar, Figure 16.  Note, Figure 3, page A8, shows the breakthrough pattern

of the general organic parameter CCE-m from this same treatment unit.

     Recently, in the WSRD laboratories this approach was repeated using

improved analytic procedures. Weekly one gallon samples of influent and effluent

from a virgin granular activated carbon bed (about 9 minute EBCT), receiving


coagulated, settled, and dual media filtered Ohio River water, were collected

and extracted with one 250 ml and two 100 ml portions of redistilled methylene

chloride.  After concentration of the extract 2 yg of anthracene, in methylene

chloride was added as an internal standard.

     In-house analyses are being conducted on a gas chromatograph employing a

30 meter SP-2100 wall coated glass capillary column.  Injections are made at

20°C, and after 5 minutes the oven temperature is programmed at 2°/min to

240°C.  Detection limits for anthracene (internal standard) using a flame ionization


detector is approximately 0.4 ng or approximately 20 ng/£ for similarly responding

compounds from the original 4£ sample.  Preliminary results, Figure 17, indicate


that after one week almost all of the compounds present in the influent were

not detectable in the effluent.  Further, no major new peaks occurred, indicating

that organics detectable by this procedure were not leaching off the granular

-------
-A53-

-------
-A54-
    *  I


    <  O
    LU  _

    Q.  LU
u

10  >

-------
                       -A55-
  UJ
  >

  O
  CO
   o
  fiUj  UJ
  Si 
si
U2
o o
iu o
ii
              
-------
                                          -A56-



activated carbon.   Analysis of later samples showed;  1)  that the influent


quality was very variable and 2)  that although some "peaks" began occurring


in the effluent it always contained fewer "peaks" than the influent,  even


after 3 months.  These studies are continuing.  Figure 18 shows the increase


in effluent NPOC concentration during the time of the collection of the

                                                          9
above samples. Similar data have been collected by Heymann  in Duisburg,


Federal Republic of Germany.


     Note, because the influent to the pilot plant (Ohio River water) is


stored in open tanks for several days prior to treatment, relatively volatile


chlorinated organics were not present in the influent to the granular activated


carbon bed.


          Summary


     Taken in total, these results, which were collected under conditions


representative of those found in drinking water, support the concept of


granular activated carbon adsorption being a "broad-spectrum" organic removal


unit process.  These data do, however, indicate the variation adsorbability


of different organic compounds and show the importance of site specific


evaluations of granular activated carbon performance as outlined in the


Interim Treatment Guide (see pages 23 to 26).


     Class III Compounds (Disinfection By-Product Precursors)


     Because no direct measurement exists for Class III organics (for example,


trihalomethane precursors), the degree of precursor removal can be judged by


comparing trihalomethane concentrations upon chlorination of an untreated

                                                                /TO
control water (called the terminal trihalomethane concentration)   to similar


data collected on a treated water after similar chlorination and storage.


For example, if the effluent from a sand filter that was chlorinated and


stored for two days yielded 50 yg/£ chloroform and the same effluent passed


through an adsorbent, then chlorinated and stored under similar conditions


produced 25 yg/£ chloroform, the adsorbent would be 50 percent effective in


removing chloroform formation potential.  This example assumes that no

-------
                       -A57-
                                                     o

                                                     z
                                                     <
                                                     o
                                                     DC
                                                     o

                                                     LU
                                                     _J
                                                     OQ
                                                     <
                                                     LU
                                                     O
                                                     cc
                                                     O
                                                     Z
                                                     u_
                                                     o

                                                     z
                                                     o

                                                     o
                                                     DC
                                                     < CD
                                                     HI DC

                                                     CC <
                                                     00


                                                     111
                                                     DC


                                                     O
I/Bui 'NOIlVdlN30NOO NOSUV3 3INV9UO 319VOand-NON

-------
                                     -A58-





trihalomethanes were present in the filter effluents,  that is the terminal



trihalomethane concentration and trihalomethane formation potential are



equal.


                                                  64
     Using Ohio River water as source, Love et al.    observed that the



relative effectiveness of granular activated carbon adsorbers to prevent the



formation of trihalomethanes was highest for chloroform and lowest for



dibromochlormethane.  A granular activated carbon bed receiving coagulated,



settled, but undisinfected water was initially effective for preventing



trihalomethane formation upon subsequent chlorination (see Figure 19).



     Similar pilot plant studies have been carried out by Sylvia at the



Lawrence Experiment station in Lawrence, Massachusetts under a WSRD research



Contract.  The Merrimac River receives considerable industrial contamination



upstream from the Lawrence study site, however, relative to the Ohio River,



the Merrimac has a very low turbidity (2-4 NTU vs 10-75+ NTU, for the Ohio



River at Cincinnati), and experimentation has shown that trihalomethane



precursors in the Merrimac are only slightly reduced in concentration through



coagulation and settling, yet are removed for long periods by granular



activated carbon adsorption.   For example, three different types of granular



activated carbons exposed to treated yet undisinfected Merrimac River water



were found to reduce the trihalomethane formation potential (2 day) of 40-60



yg/£ to less than 1 yg/& after 6 months of operation.  Details on this will



be contained in the final project report, which is due in early 1978.

                     f O

     Work by Snoeyink   at the University of Illinois on the adsorption of



humic and fulvic acids, major trihalomethane precursors, relates to this phase



of the overall problem of organic contamination.  He found that activated carbon



adsorbed humic substances in all cases  that were studied, but the adsorption



properties of the substances from different sources varied widely as did the extent

-------
                          -A59-
  100
- 75
O)
=5.
w 50
oc
LU
  25
  100
o> 75
3.
  50
cc
LJ
H

  25
       4 DAY

       25°C
          INFLUENT
                  SAND REPLACEMENT ADSORBER EFFLUENT

                                 (F-200 EBCT=10  MIN.)
                   10      15       20

                TIME IN OPERATION, WEEKS
                                   25
                                                    30
  _
o-a~—
                      / SAND REPLACEMENT

                        ADSORBER EFFLUENT

                              (HD 10x30 EBCT=20 MIN.)
                            i _ i _ i _
                                           25
                                           30
      0       5      10       15       20

                  TIME  IN OPERATION, WEEKS


FIGURE 19  USE OF GRANULAR ACTIVATED ACTIVATED CARBON

           FOR REMOVING TRIHALOMETHANE FORMATION

           POTENTIAL. REFERENCE 64

-------
                                           -A60-






to which they competed with selected trace organics for adsorption sites on




activated carbon.




     Humic substances from leaf- and soil extract,  a well water, and a commercial




source were examined in detail.  Extent of adsorption depended upon solubility,




with the less soluble humic acid (HA) fraction being more adsorbable than the




fulvic acid (FA) fraction from the same source. - The lower molecular weight




species from a given FA or HA fraction are more adsorbable than the high molecular




weight species, presumably because more surface area is accessible to them.




     The adsorption characteristics of the humic substances are also dependent




on the method of analysis used to quantify them.  The species that fluoresce the




most were found to be the lower molecular weight species and these adsorb best.




Ultra-violet absorbing species did not adsorb as well as those that fluoresce.




Solution pH and phosphate concentration also had a marked effect on adsorbability




of the humic materials, with adsorption generally improving with decreasing pH




and increasing phosphate concentration.  The trihalomethane formation potential




of the humic substances varied widely from source to source, with only one




exception, but no dependence on molecular weight was found for fractions of FA




or HA.  This work reemphasizes the need for on-site pilot studies to determine




adsorbability for that particular location.




     The removal of disinfection (chlorination) by-product precursors was also




studied at the three projects operating at the actual water treatment plants




cited above.  The data in Table XIV and Figure 20 show the same variability in




the treatability of trihalomethane precursors as noted in the previous studies.




Note:  The data in Table XIV show that although breakthrough did occur, complete




exhaustion for the removal of trihalomethane formation potential did not, at least




during the time of the study.  In summary, removal of trihalomethane precursors




by adsorption on granular activated carbon beds is variable and site specific.

-------
                                       -A61-
                                TABLE XIV

     REMOVAL OF SUMMATION** TRIHALOMETHANE FORMATION POTENTIAL (THMFP)
                    BY GRANULAR ACTIVATED CARBON BEDS
Southern Florida Ground Water
6.2 min. EBCT
       Inf.   Eff.     Percent
Time   THMFP  THMFP    Reduction
Weeks   yg/&   yg/&
  Lower Mississippi River Water
EBCT-28 min. start, 20 min. end
      Inf.       Eff.     Percent
Time  THMFP*     THMFP*   Reduction
Weeks  yg/£       yg/£
0
1
2
3
4
5
6
7
8
9
10
11
12
17+


384
485
758
878
807
733
710
936
617
639
616
348
655
575


11
335
558
560
579
557
524
437
342
424
445
397
452
394


97
34
28
37
30
27
29
53
45
34
39
-17
33
31


0
1
2+
3
4
5
6
7
8
9
10+
11+
12+
15+
20+
25+
59
56
70
113
104
203
239
151
159
217
184
240
236
205
297
218
2.3
1.3
6.7
22
18
25
44
49
92
86
73
100
89
93
148
157
96
98
91
80
82
87
81
68
42
60
60
58
62
54
50
28
*Note:  These data are shown in Figure 11, page 20, of the Interim Treatment Guide.
**The arithmetic sum of the individual trihalomethane species determined.
This parameter is called "Total" Trihalomethane Formation Potential in the
Regulations.

-------
                         -A62-
o>
5.
<

CC
LU
o
z
o
o
HI
h-
o
Q.
cc
o
LL.

UJ
UJ

2

O
Z

O
CO
   260
   240
   220
   200
180
    160
    140
    120
    100
     80
     60
     40
     20
                                           I-
                                           A
                WVW 14x40 GRANULAR

                ACTIVATED CARBON


               Upper Ohio  Valley

                 Water Treatment

                 Plant
         SAND REPLACEMENT

         ADSORBER

         INFLUENT
                    SAND REPLACEMENT

                    ADSORBER EFFLUENT
       -"  APPROX. EMPTY BED CONTACT TIME = 6.5 MINUTES
          i  i  i  i
                                       i  i  i  i
        11 13 \S 18 20 22 25 27 29 1  3 5 8 10 12 15 17 19 22 24 26 29 31

              JULY               AUGUST


  FIGURE 20 REMOVAL OF TRIHALOMETHANE PRECURSORS BY

            GRANULAR ACTIVATED CARBON BEDS

-------
                                       -A63-
     Class IV Compounds (Disinfection By-Products)




     To determine the effectiveness of granular activated carbon to remove




trihalomethanes, Cincinnati tap water, which contains these compounds, was




exposed to granular activated carbon columns at various hydraulic loadings




(different contact times).   Chloroform was reduced 90 percent or more for about




three weeks, (10 min. EBCT),  then breakthrough was steady until the adsorber




was exhausted at about the ninth or tenth week.  The trihalomethanes containing




bromine were effectively reduced by the granular activated carbon for 30 or




more weeks.  Figure 21 shows the difference in effectiveness of six different




types of granular activated carbon for removing all of the trihalomethanes




summed together.   In earlier field studies   where the applied water and the




effluent from granular activated carbon beds were sampled, the findings regarding




trihalomethane reductions were very similar to the laboratory results.




     The second phase of the Florida study, which studied adsorption of organics




from chlorinated water,also showed that chloroform was the trihalomethane least




effectively adsorbed, while bromoform was adsorbed the best of the trihalomethanes.




At the lower Mississippi River water treatment plant,ammonia is added after the




addition of chlorine so the concentrations of trihalomethanes reaching the




granular activated carbon bed were low, chloroform (1.8-46 yg/£), bromodichloromethane




(NF-6.2 yg/£), dibromochloromethane (NF-12.8 yg/£), bromoform (NF).  Under




these conditions the bromine-containing trihalomethanes were well removed,




although chloroform began to appear consistently in the adsorber effluent at




low concentrations after the fifth week.  Adsorption-desorption cycles




began after 13 weeks.




     Figure 22, data from the water treatment plant in the upper Ohio Valley,




shows a breakthrough pattern for the trihalomethanes very similar to the pilot




plant data described above and shown in Figure 21. Here too, the bromine-




contadning trihalomethanes were removed better than chloroform.

-------
-A64-

-------
                                       -A65-
       Upper Ohio Valley Water
         Treatment Plant
                                                    DICHLOROBROMOMETHANE
                                                     ADSORBER INFLUENT
          CHLOROFORM
            ADSORBER INFLUENT
                                               SAND REPLACEMENT
                                               ADSORBER EFFLUENT
                     SAND REPLACEMENT
                       EFFLUENT
                                            SUMMATION TRIHALOMETHANES
             DIBROMOCHLORO METHANE
                                                     ADSORBER INFLUENT
                  ADSORBER INFLUENT
                                             SAND REPLACEMENT
                                             ADSORBER
                                             EFFLUENT
           SAND REPLACEMENT
           ADSORBER EFFLUENT
FIGURE 22.  REMOVAL OF TRIHALOMETHANES BY GRANULAR ACTIVATED CARBON BEDS-
           6.5 MINUTE EMPTY BED CONTACT TIME-VIRGIN WVW 14x40 GRANULAR
           ACTIVATED CARBON

-------
                                         -A66-







     In summary, if the aromatics, taste and odor compounds, and certain pesticides




can be categorized as strongly adsorbed onto granular activated carbon,  then




chloroform is located near the other end of the adsorption spectrum.




     General Organic Parameters




     Although not directly related to any single class of organic compounds discussed




above, the use of an easily measured general parameter to monitor the performance




of an adsorption treatment system is appealing from the standpoint of convenience




and cost.  The Water Supply Research Division pilot plant studies routinely




include the following general organic measurements:  non-purgeable organic




carbon (NPOC)  , ultra-violet absorbance (UV) at 254 nm  , and fluorescence —




both the emission scan (EMS)"  and the fixed wavelength procedure described by




Silvia as the rapid fluorometric method (RFM)  .




     Figure 23 shows the pattern of each general organic parameter for the




first 15 weeks after a fresh granular activated carbon adsorber was put into




service.  In an attempt to develop a simple test to predict the organic carbon




content of effluents, UV, RFM and MS data were correlated to NPOC.  Table XV




summarizes the regression analysis.  The most promising relationship thusfar is




between RFM and NPOC.  A clearer understanding of these relationships and




others, such as between some general parameter and trihalomethane  (Class IV)




or by-product precursor  (Class III) concentrations is being sought through




additional statistical analyses and should be available in the near future.




                                TABLE XV




      RELATIONSHIP BETWEEN NPOC AND OTHER GENERAL ORGANIC PARAMETERS




             FOR MONITORING A GRANULAR ACTIVATED CARBON ADSORBER




                                     UV           RFM          EMS
r\
(R ) Coefficient of Determination
(R) Correlation Coefficient
Significant at 95% by F and t test
95% Confidence, interval, mg/£
0.22
0.47
Yes
±0.73
0.67
0.82
Yes
±0 . 48'"
0.52
0.72
Yes
+0.58

-------
                         -A67-
          '30NV9dosav A n
                              S1INH NtfOS NOISSIW3 3DN30SdOmd
                                                      .
                                                   
-------
                                        -A68-



     Thusfar, the best predictor of NPOC, RFM, has such a high 95 percent



confidence interval (+0.48 mg/Jl) that its potential use is discouraging.


      64
Sylvia  ,  however, obtained better results correlating RFM measurements with



granular activated carbon bed performance as measured by CCE-m, so some



possibilities for this approach may exist in certain waters .



     As noted above, in addition to correlating a more easily measured general



organic parameter against NPOC, the use of a general parameter to predict



performance of a granular activated activated carbon bed with respect to



breakthrough of a specific organic compound or group of specific organic compounds



is also an attractive concept. In the lower Mississippi River water treatment



plant study, four general parameters have been correlated to total trihalomethanes,



trihalomethane formation potential, and the sum of all of the non-trihalomethane



chlorinated organic compounds in the adsorber effluent.  Table XVI shows that,



except for UV adsorbance, the correlation coefficients have good statistical



significance.  The confidence limits on these data are, however, not available



at this time (Fall 1977).  This effort will continue at this field site, as



well as in the Water Supply Research Division pilot plant, and at other water



treatment plant project locations.

-------
                                        -A69-
                                     TABLE XVI


RELATIONSHIP BETWEEN VARIOUS GENERAL ORGANIC PARAMETERS AND THREE CLASSES
                        OF ORGANIC CONTAMINANTS
NPOC
(50
a
r
OBS)
Signif.
EMS
(42 OBS)
r Signif.
RFM
(47 OBS)
r Signif.
UV
(50 OBS)
r Signif
Class II
Total Non-
Trihalo-
me thane
Organics  0.585  .00001    0.622  .00001    0.740 .00001   0.282   .047

Class III
XTHMFP    0.723  .00001    0.658  .0001     0.770 .00001   0.366   .009

Class IV
ETHM      0.631  .00001    0.745  .00001    0.747 .00001   0.306   .031
a - "r" is the correlation coefficient

b - "Signif" is the significance of the correlation coefficient, r (the
   smaller the value for "Signif" the greater the significance.)

OBS - Observations

     Influence of Empty Bed Contact Time

     One phase of the project studying the use of adsorbents to remove

organics from a southern Florida groundwater includes the effect of empty

bed contact time on organic removal using the techniques similar to those

outlined in the "Procedure for Collection of Site Specific Design Data" in

the Interim Treatment Guide (pages 23 to 26) and Appendix C.  Finished water

from the treatment plant was diverted to four pilot granular activated

carbon columns connected in series.  Each column contained 30 inches of

12x40 mesh activated carbon and was operated at a hydraulic loading of 3
      2
gpm/ft .   The nominal empty bed contact time for each column was 6.2 min.

and thus the contact times were 6.2 min., 12.4 min., 18.6 min., and 24.8

min., respectively for the four columns.  Twenty purgeable halogenated

organic substances were monitored at the inlet and outlet of each column.

Also TOC and Terminal THM concentration data were collected.

-------
                                      -A70-
     The ability to achieve a given quality of effluent for longer time periods




(larger water volumes)  is related to longer empty bed contact times (larger




activated carbon volumes), Figures 24,  25,  26 and 27.  Figure 26 for vinyl




chloride is an exception as the relationship is not as definite for the data




collected to date.   Prior experience also showed that vinyl chloride removal by




granular activated carbon adsorption was the most sporadic of all substances




tested at this site.  So far this has been the only substance to show such an




erratic pattern.  Figures 24, 25, and 27 show a "broad-wave" front for the




substances tested,  thus providing a more gradual approach toward equilibrium




concentration than if a more "narrow-wave" front had occurred.




     In order to assess the period of time between activated carbon reactivations




for a given empty bed contact time, the three performance criteria discussed in




the Interim Guide were applied to the data in Figures 24-27.  Whenever a




performance criterion was consistently exceeded, the first time at which the




criterion was exceeded (see arrow on Figures) was used and the corresponding




operating time and water volume treated calculated. The concentrations between




data points were taken as linear.




     Table XVII presenting the maximum duration of time in days of operation




and liters treated when complying with each criterion shows that the TOG




criterion is the most limiting of the three at all contact times for the




conditions tested.  TOG would have limited the operating time to 1.5 days, 4.5




days, 11.2 days and 18.6 days for the respective empty bed contact times.




     Table XVIII is a different display of the prior data in which the activated




carbon volumes used and water volumes treated are recalculated using the 6.2




min EBCT data as unity and presenting all data as ratios. Thus for the most




stringent criterion, TOG, if two times the activated carbon volume is used,

-------
                      -A71-
                                                    o
                                                    LU
                                                    O
                                                    o
                                                    o

                                                    o
                                                    HI
                                                    CD
                                                    O.

                                                    5
                                                    LU

                                                    LL
                                                    O

                                                    LU
                                                    O -J
                                                    z <

                                                    m £
                                                    D O

                                                    l
                                                 £  ± cc
                                                    LU z
                                                    I O
                                                    i- m

                                                    LU
                                                    <
                                                    X ^
                                                    LU O
                                                    CM

                                                    LU
                                                    DC
                                                    D

                                                    O
I/Bui 'NOIlVaiN30NOD NO8dVD OINVOUO 1V1O1

-------
         -A72-
                                                                at

                                                                ui
                                                                Z
                                                                iu
                                                                ^
                                                                >-
                                                                O
                                                                DC

                                                                O
                                                                z
                                                                o
                                                                u
                                                                <
                                                                I—
                                                                z

                                                                8
                                                                u
                                                                z
                                                                O

                                                                LU
                                                                _J
                                                                Q_
                                                                10
                                                                CM
                                                                O
|/Bn 'NOIlVaiN3DNOD

-------
-A73-

-------
                            -A74-
0.80
                            15       20      25
                         TIME IN SERVICE, DAYS
30
35
    FIGURE 27  EXAMPLE OF THE INFLUENCE  OF EMPTY BED CONTACT
              ON TERMINAL SUMMATION TRIHALOMETHANE REMOVA

-------
Empty Bed
Contact Time
Minutes
6.2

12.4

18.6

24.8
                     -A75-


                     TABLE XVII

     DURATION OF RUN ALLOWED BY EACH CRITERION

Criterion No. 1     Criterion No. 1    Criterion No, 2
(Vinyl Chloride)  (cis-l,2-Dichloro-      (TOG)
                     ethylene)
                   Days    Liters
                                                          MCL
Days  Liters

3.5    313.8

10.8   966.4

35.7  3196.1

35.3  3096.5
   _      	      Days   Liters  Days  Liters

17.5     1573.3      1.5    132.5   2.3   204.8

60.5     5437.4      4.5    402.3   9.3   832.7

                    11.2    999.6  14.5  1297.1

                    18.6   1653.7  19.5  1733.2
- = Not yet exceeded criterion.



                                    TABLE XVIII

    RATIO OF VOLUME OF WATER TREATED IN COMPLIANCE WITH EACH CRITERION
EBCT
Min.
6.2
12.4
18.6
24.8
V.C. =
cis =
Ratio of Act.
Carbon
Required
1
2
3
4
Vinyl chloride
Criterion No. 1
Specific Organics
(V.C.) (cis)
1 1
3.1 3.5
10.2
9.9
Criterion No. 2
TOG
1
3.0
7.5
12.5
MCL
1
4.1
6.3
1.5
cis-1, 2-Dichloroethylene
- = Not yet exceeded criterion

-------
                                       -A76-
three times the amount of water can be treated before the criterion is exceeded.




Further, again doubling the empty bed contact time increases the length of time




of operation over six more times, a favorable improvement.   Similar beneficial




results occurred relative to criterion one and the MCL requirements with a




doubling of the "base" empty bed contact time, but redoubling produced only a




proportional increase in operating time according to the MCL requirement.




     This evaluation of empty bed contact time shows that increased quantities




of activated carbon, at fixed flow conditions, resulted in the ability to treat




more water before exceeding a given criterion.  In each case the first doubling




of empty bed contact time resulted in adsorber operating times that were proportionately




greater than the increase increse in empty bed contact times for all criteria.




This has an important impact on the economics of treatment, see pages 43-45 in




the Interim Guide.




     Two in-house studies on the influence of empty bed contact time on adsorber




operating times, Figures 28 and 29, showed that for the two organic compounds,




chloroform and carbon tetrachloride, the time to reach a given breakthrough




point was approximately proportional to the empty bed contact time.  Because of




these different results in two different locations, this type of information




should be collected in the location under study, as indicated on pages 23-26




of the Interim Treatment Guide.




     Biological Growth and Endotoxin Production




          Standard Plate Count




     Controlling bacterial populations (and particularly killing or inactivating




pathogenic microorganisms) is a primary goal of water treatment.  Some concern,




therefore, has been expressed about the possibility of bacteria proliferating




within granular activated carbon beds.  To investigate this, the bacterial




quality of the untreated and treated water from the WSRD pilot plant was

-------
                           -A77-
 100
   AVERAGE CHLOROFORM CONCENTRATION
IN APPLIED CINCINNATI, OHIO TAP WATER=46A»g/L
              GAC DEPTH: 90 cm (36 INCH)
              GAC TYPE: FILTRASORB 400
                                         50% EFFECTIVE
                                         CHLOROFORM = 23 ftg/L
                  234567
                      TIME IN  SERVICE WEEKS
FIGURE 28.  EFFECT OF EMPTY BED CONTACT TIME ON CHLOROFORM
           REMOVAL FROM TAP WATER.
                                       8

-------
                   -A78-
                                                              I-  >
                                                              *-  o
                                                                   .
                                                          o   Z
                                                          "   8e
                                                                  at
                                                              Q  O
                                                              >•  %
                                                              f  <
                                                              £  *
                                                            «/> 5  i—
                                                            ^ iz  "j
                                                            UJ UJ  |_
                                                              Z  O
                                                              CM

                                                              UJ
                                                              fid
                                                              3
                                                              O
o
10
O
CO
O
CN
  \/Brf 'NOIlVillN3DNOD

-------
                                      -A79-
routinely monitored over a four-month period by the staff of the WSRD
Microbiological Treatment Branch using the Standard Plate Count (SPG) test.
Samples were also collected intermittently and analyzed for total and fecal
coliforms, however, these indicator organisms seldom survived the coagulation
and settling processes and were never detected in the filter or adsorber effluents.
The monthly average SPC (expressed as the geometric mean) for the pilot plant
studies (see Table XIX), in general, show a two log (99 percent) reduction in
the bacterial count through the treatment plant.  Note that no disinfectant was
added anywhere in the treatment process.  Although an increase in bacterial
populations was expected a. priori, the SPC of the effluent from the granular
activated carbon adsorber was slightly lower than the SPC in the companion
sample taken from the dual media filter effluent.  No attempt was made to
measure attached growths.  An attempt was made, however, to isolate and identify
the predominant populations in the pilot plant.  From the adsorber effluent,
five or six different types of colonies could be recognized and two genera,
Flavobacterium and Xanthomonas were identified from smear plates.

                               TABLE XIX
                 MONTHLY MEAN (G ) STANDARD PLATE COUNT
                                m
               (Pilot Plant Studies - Ohio River Water)
                   all bacterial counts are No./I ml
Time in
operation,
Months ,
1
2
3
4


Raw
19,600
12,000
7,170
6,680


Settled
1650
1000
790
700
Effluent from
Dual Media
Filter
137
270
80
100
Effluent from
Granular Activated
Carbon Adsorber
63
72
29
37

-------
                                      -A80-

          Endotoxins

     Endotoxins are lipopolysaccharide-protein complexes produced in the cell

walls of Gram-negative bacteria.  The lipopolysaccharide portion of the complex
                                                                       r Q
is pyrogenic and as little as 1 yg can produce fever in a 700 kg horse.

Concern, therefore, has been expressed regarding the possible formation of

endotoxins in granular activated carbon adsorbers because of bacteriological

activity.  For a 6-month period in 1977, the U.S. EPA Health Effects Research

Laboratory (HERL) monitored bacterial endotoxin concentrations in untreated and

treated water from the WSRD organics removal pilot plant.  These were companion

samples with those collected for Standard Plate Count analyses.

     Using the Limulus lysate bioassay  , HERL scientists observed a marked

reduction in pyrogenic activity as a result of chemical coagulation (and settling)

and a slight additional decrease through filtration by either dual media or

granular activated carbon.  The encouraging finding was that no increase in

pyrogenic activity occurred in the effluent from the granular activated carbon

bed.  These data are shown in Table XX.  An extramural project entitled "Pyrogenic

Activity of Carbon-Filtered Waters" (EPA Grant No. R-804420010) is underway at

Texas A&M University. Samples for endotoxin concentrations are being collected

from full-scale water treatment plants utilizing granular activated carbon

adsorption and the results from this definitive study should provide further evidence

as to whether or not a problem exists.

                               TABLE XX

       MONTHLY MEAN ENDOTOXIN CONCENTRATIONS IN OHIO RIVER WATER
                          (Pilot Plant Studies)

                               Endotoxin Concentration, yg/£

Time                                            Dual   Granular Activated
in Operation,                                   Media  Carbon Adsorbent
Months         Raw       Coagulated/Settled     Effl.    Effluent
2
3
4
5
6
7
158
236
205
500
45
35
16
63
36
66
20
11
16
7
41
16
5
11
9
6
11
15
4
11

-------
                                        -A81-


 Ozone Enhanced Granular Activated Carbon Adsorption (Biological Activated Carbon)

     One organic removal unit process being used in some locations has not yet

been discussed in this report.  It is currently (Fall 1977) receiving much attention,

therefore, although much still needs to be known about the process, a discussion

of a variation of granular activated carbon adsorption for organic contaminant

control in which ozonation  precedes the granular activated carbon treatment processes

is included here.  These two processes in combination are frequently called biological

activated carbon.

     Literature Review

     As part of a research grant with the WSRD, Dr. Rip Rice, in collaboration

with Public Technology, Inc., Washington, D. C. prepared a chapter of the final

report on the "Status of Ozonation and Chlorine Dioxide Technologies for Treatment

of Municipal Water Supplies" project summarizing what is known about biological

activated carbon.  The first few pages of that chapter, with some editing, are

presented below.

                 A REVIEW OF THE STATUS OF OZONATION PRIOR TO GRANULAR
          ACTIVATED CARBON FOR REMOVAL OF DISSOLVED ORGANICS AND AMMONIA
                              FROM WATER AND WASTEWATER

1.  Introduction

     In a recent article that discusses the use of granular activated carbon in

water treatment, McCreary & Snoeyink   state that "beds of granular activated

carbon (GAG) are a convenient place for microorganisms to grow because bacteria

attach themselves to the irregular external surfaces of the activated carbon

particles and are very difficult to dislodge via backwashing procedures."  In the

presence of soluble carbonaceous matter, which serves as food for these organisms,

and in the absence of oxygen, anaerobic bacteria can develop.  There are numerous

instances in which sulfidic odors have been reported emanating from granular

activated carbon columns used for the removal of dissolved organic materials

contained in sewage treatment plant effluents'     and drinking water supplies.^

-------
                                    -A82-





On the other hand, with sufficient dissolved oxygen and carbonaceous matter,  the



bacteria that develop in activated carbon beds will be aerobic.  These do not



produce sulfidic odors.



     Many of the advantages of biological granular activated carbon (BAG) were



first recognized by German water treatment scientists in the 1960's in drinking



water plants along the Rhine River in the Dusseldorf area.   Subsequently, BAG



processes also have been installed in Swiss and French drinking water treatment



plants, and are subjects of active pilot studies in Holland and Belgium.  In the



United States, the U.S. Environmental Protection Agency's Water Supply Research



Division in Cincinnati, Ohio has been testing a pilot BAG column since late in



1976 (see below.)



2.  Fundamental Principles



     Granular activated carbon is made biologically active by the deliberate



introduction of sufficient dissolved oxygen (DO) to aqueous streams just before



they are passed through GAG columns.  As long as the water contains sufficient DO



to maintain aerobicity of the bacteria and sufficient dissolved carbon to provide


                                                                      74
food, the aerobic bacteria will thrive in this environment.  Eberhardt   has



likened bacterial activity in such an ideal environment to a "herd of cows grazing



in a luscious meadow."  Pre-ozonation can convert larger, less biodegradable



organic molecules into smaller, more biodegradable organics, for example, into



acetic and oxalic acids.  Sontheimer   has summarized the German findings to date



which have led to the current theories of operation of BAG.

-------
                                      - A83-
     Although aerobic bacteria are necessary to obtain the benefits from BAG, so




also is the adsorptive capacity of the GAG for the dissolved organic materials




that will serve as food for these bacteria.  This means that the surface area and




pore volume of the activated carbon should be high.  Stated another way, the




organic materials present in solution should be adsorbable onto the activated




carbon column, because the contact times of solutions with the activated carbon




particles in the columns or beds are normally short (15-30 minutes).  This does




not necessarily give the bacteria sufficient time to degrade larger organic




carbon molecules, ideally to carbon dioxide and water.  Therefore, retaining the




dissolved organic molecules in the activated carbon columns so that the bacteria




then will have sufficient time to degrade them is important, even though the




actual contact times involved are rather short.




     Many organic materials are readily adsorbed onto GAG, but many others are




not.  For example, high molecular weight natural humic acids, so prevalent in




drinking water supplies, are not readily adsorbed by activated carbon.    If




solutions of these non-sorbable organic materials are ozonized before passage




through the GAG columns, they are converted to more readily biodegradable organic




materials.  '  '    At the same time, ozonation introduces a large quantity of




oxygen into the water, which promotes aerobic bacterial growth.

-------
                                        -A84-


3.  Advantages of Biological Activated Carbon

     In European pilot studies and in drinking water treatment plants by many

workers  '   '       have shown that ozonation followed by granular activated

carbon adsorption results in:

     - More effective removal of dissolved organics from solution by the BAG,

     - Increased operating life of the activated carbon columns before
       having to be reactivated especially if the GAG can be kept free of halogenated
       organics.

     - Biological conversion of ammonia in the GAG columns,

     - Use of less ozone for removing a given amount of organics than using
       ozonation alone.  (BAG is more cost-effective over ozonation in
       removing Dissolved Organic Carbon - DOC),

     - Filtrates from BAG columns in drinking water plants can be treated
       with small quantities (0.1-0.5 mg/£) of chlorine or chlorine dioxide,
       which produces drinking water of acceptable bacterial quality
       (zero coliforms) and provides a residual disinfectant for distribution
       systems.

     Independent studies on physical-chemical treated sewage at the Cleveland
                        7 ft              Rf*
Regional Sewer District,   and in Israel   have confirmed these advantages with

respect to removing organic materials.

4.  European Background

     Introduction of granular activated carbon into European drinking water treatment

practices occurred after World War II.  Its initial application was for dechlorination,
                                 Q 7
then for tastes and odor removal.   Many surface waters containing ammonia undergo

break-point chlorination at the beginning of the treatment process.  This technique

effectively removes ammonia, but produces considerable amounts of residual chlorine
                                      Q C
and chlorinated products in the water.    German water treatment objectives are

to process surface waters to the same quality as that of natural groundwater

(which does not have to be treated in many cases).  Therefore, waters

-------
                                         -A85-





treated by break-point chlorination have to be dechlorinated before they are


                               Q Q            QQ

treated further or distributed.    Schalekamp   points out that in Switzerland a




residual chlorine concentration of only 0.05 mg/H is permitted in finished drinking




water.  Therefore, the raw water chlorine dose is removed by granular activated




carbon prior to the addition of 0.05 mg/£ of chlorine dioxide for final disinfection.




     Combinations of ozone and granular activated carbon first were installed in



                                       87
the Dusseldorf area in the early 1960's   but nearly ten years passed until the




biological activity in the activated carbon columns was recognized as being beneficial.




By the early 1960's, the lower Rhine River water quality had declined, and advantage




was taken of filtration of organic materials from the river water through the sand




banks of the Rhine.  Wells were dug into the river banks and water is drawn from




these wells as the treatment plant raw water.




     In the intervening years since the introduction of ozone/activated carbon at




Dusseldorf, the beneficial effects of biological activity in the activated carbon




columns have been recognized, characterized and optimized.  After ozonation, the




water is allowed to stand for 20-30 minutes to allow the more refractory organic




compounds sufficient time to react with residual ozone.  This retention time also




allows residual ozone to be utilized, rather than simply to be destroyed when




passed through the activated carbon column.




     On the other hand, the Rhine River also contains considerable amounts of




chlorinated organic materials that are not removed during river bank filtration.




These halogenated materials also are more resistant to oxidation by ozone than are




non-halogenated organics and thus are less likely to be converted into easily




biodegradable materials.  In addition, halogenated organics are variably



                               1 9 89
adsorbed onto activated carbon.  '

-------
                                     -A86-




     Combining the variable adsorptivity of halogenated organics compounds on



granular activated carbon with their lesser reactivity upon ozonation and



their lower biodegradability means that breakthrough of halogenated organic



compounds occurs sooner than does breakthrough of non-halogenated organic



compounds, even if the granular activated carbon columns are biologically



activated.  Thus German water works along the Rhine in the Dusseldorf area



monitor their activated carbon column capacities for Total Organic Chlorine


      90—92                  93
(TOC1)      as well as by DOC   and/or UV absorption.  Activated carbon



columns at three Dusseldorf plants along the Rhein (Flehe, Am Staad, Holthausen)



are backwashed every 4-6 weeks and reactivated every 6 months.



     When Dusseldorf activated carbons are thermally reactived, however,



only some 80 percent of the activated carbon charge is taken out of the



columns.  This leaves a portion of biologically active activated carbon in



the column so that the level of bioactivity will not drop significantly when



fresh or reactivated activated carbon is added.  With fresh activated carbon



columns, about 15 days of operation usually are required for biological



activity to build up to an effective "steady state", particularly for ammonia


      i 94
removal.



     WSRD Pilot Plant Results



     In late 1976, WSRD began a pilot plant study on the use of ozone to



extend  the life of a granular activated carbon adsorber used as a sand



replacement system as measured by Total Organic Carbon - and Trihalomethane



Formation Potential (2 day, 25°C) concentrations.  This study was prompted



by the  successful results from pilot- and  full-scale BAG tests  conducted  in



Europe  (principally West Germany) cited above.

-------
                                       -A87-
     In this test a granular activated carbon adsorber receiving settled ozonated




water and a granular activated carbon adsorber receiving settled unozonated water




were operated for a 10 month study period. Both units had about a 9 minute empty




bed contact time.   Ozone dosages applied to the settled water (20 minute plug




flow contact time) and dissolved ozone concentrations prior to filtration or




adsorption are shown in Figure 30.    Because of an increase in ozone demand




within the system and the resultant loss of ozone residual, the applied ozone dose




was increased following ozone contactor cleaning early in the 9th month of operation.




     Figure 31 compares the TOG concentrations in both effluents.  Figure 31




illustrates the extension in bed life afforded by the combination of ozone followed




by granular activated carbon. Operating time was extended from about 3-3/4 months




to about 7-1/2 months according to Criterion 2, increase of 0.5 mg/£ of TOC concen-




tration.




     Figure 32 shows a similar trend for the removal of trihalomethane formation




potential, although the MCL requirement was never exceeded in either effluent




during this study.  The results of this study are encouraging enough to prompt




additional in-house and extramural research on biological activated carbon during




FY78, particularly to determine whether or not the use of ozonation prior to




adsorption improves granular activated carbon performance with respect to Criterion




1, the control of low molecular weight halogenated organic compounds.

-------
         -ASS-








CO
UJ
o
<
cc
UJ
>

>- CO
-1 Z
fs
li
«5
Ujg
ttg
:
<=l
X
.0
_i
-iu_
<0







1
1
N
t^
"V.
00
•s^
N
Q
UJ
CO
<
UJ
cc
o
z

UJ
CO
o
Q
UJ
Z
0
N
O
o
UJ
_l
CL
Q.
<











I i
O
(N
I
DC
UJ
h-
_J
>>.

CO
0
5

£E
UJ
h-
<
^
UJ
_J
1-
1-
UJ
CO
o
h-
o
UJ
_l
Q.
Q.
<
UJ
CO
o
Q t
tt 1
UJ
OQ
K
0
CO
Q f
1
1-
<

0
UJ 1
OC T
^M I
j
CO
<
ml
1
1
I
•>>
O)
£
UJ 1.
ZT
01
N
O
' Q <-
_l
0
CO
CO
' Q 
-------
                  -A89-
l/Bui 'NOIlVdlN30NOO OINVOdO "IV1O1

-------
-A90-
1/BrT  'NOI1VU1N30NOO
'Oo93 'AVQ Z

-------
                                      -A91-
Economic Analysis Data




     This text supplements the economic analysis information presented in




the Interim Guide on pages 36 to 54, and is divided into five sections.  The




first section contains basic cost information that can be used to present a




generalized framework for considering the cost impact of using granular




activated carbon adsorption.  In the next two sections, specific operational




configurations and their influence on system costs are considered.  For




example, the tradeoffs that exist between Empty Bed Contact Time (EBCT) and




Reactivation Frequency (RF) are examined in one section, and the effect of




activated carbon cost and reactivation frequency is considered in the other.




In the fourth section the impact of inflation on the choice of systems is




evaluated.  Because reactivation is a significant portion of the cost of a




granular activated carbon treatment system, it is considered separately in




the final section.  Another source of supplementary economic analysis information




is the Interim Treatment Guide for the Control of Chloroform and Other




Trihalomethanes.




     Basic Costs




     The data utilized in this section are the same as used to develop costs




in the Interim Guide.   Tables XI, XII, and XIII, and Figure 13 in the




Interim Guide pages 37, 38, 40, and 42, respectively provide examples of the




use of this baseline information.

-------
                                    -A92-







     Influence of Empty Bed Contact Time




     As discussed in the Interim Guide examining the cost tradeoffs that exist




between bed depth (EBCT) and reactivation frequency (RF) for a given quality of




water is important.   This relationship is examined using an "R" ratio.  The "R"




ratio allows the changing of two variables at one time (EBCT and RF).   For example,




in a sand replacement system, if the removal efficiency of granular activated




carbon is proportional to EBCT and RF, then a system with an EBCT of 10 minutes




and an RP of one month (R = 0.1) might be assumed to be equivalent in cost to one




with an EBCT of 20 minutes and an RF of two months (R = 0.1).  As shown on page 44




in the Interim Guide, however, although the R's are the same, the costs are not.




Table XXI duplicates the data in Table XIV (page 41 of the Interim Guide) with the




exception that values for "R" have been added.




     Another way of examining the tradeoff between cost, RF, and EBCT is shown in




Figures 33, 34, 35,  and 36.  In these figures the relationship that exists between




the cost, EBCT, and RF for 10 and 100 mgd sand replacement systems and for 10 and




100 mgd postfilter adsorbers can be seen.  The problems of non-proportionality




between cost and performance can also be seen in these data.  For example, in




Figure 35 if the required EBCT was 10 minutes with a one month reactivation




frequency, the unit cost would be 6.6 cents/1000 gal.  If the EBCT were increased




to 18 minutes, the period between reactivations would have to be increased to




approximately 2.5 months or greater to achieve a favorable economic tradeoff.




Therefore, to merit an 80 percent increase in empty bed contact time, the period




between reactivations would have to increase by at least 150 percent.  This important




concept is examined more fully both in the Interim Guide (see pages 27, 34, and 35)




and earlier in Appendix A, pages A69 to A76.

-------
-A93-


















C/l

w
H
0
O
O
o
rH
- — .
O-



CU
e
T3 1H
CQ en
>* CJ U
p J l r^
S C 1-1
W O S
u •-'
^
4-J
•H ,—,
.*: cj TJ
rO CO OC
a) a. E
a. co ^
u
C 3 T3
00 0 00


01
a
a
a
en a
:>N >-
en H
o
r-.
O
O

.
a.
CN
in
0
C3
in
_J
rH


O
O
00
in
rH

rH
O
O
a*
r^
CN

O
O
O

00

o
m
o
o
r^
rH*
rH
m
CN
O
o

in
rH
m

o
o
rH
rH
CM
CN
rH
O
O
r>-
r*

m
o
o
o
rH
in
00
0
CM
O
O













0
o
^r
r^
rH
in
CM
0
d
r^
rH
CN

rH
O
O
rH
O

00
o
o
o
CO
^
in
co
O
o
o
rH
^
CM
rH
0

O
O











a
o
rH
d
r^
00

in
r^
0
d
en
<*

O
in
o
o
00
rH
rH
m
CM
0
o
^.
r^
rH
O

O
0
r^
CN

O
rH
O
O
u
CU

rH IH
•H CD
tt4 ^
1 U

en in

0- <
o
in
O
o
o
CO
rH
OO
CO
o
o
e*
^r
r-t
m
CM
o
o
>o
OO
rH
CN

O
o
ON
OO
CN
in
o
o
o
00
,v|
in
O
CN
O
o












CO
O
o
0>
vO
rH
in
CM
O
o
10
ON
rH

rH
O
O
ON
<^
CM
CO
o
o
o
OO
ON


O
O
d
o
CN
CO
O

0
o












-------
-A94-
























,-v
•o
4J
C
o
CJ


CO


£_4
CO

CO

s
o
03
Oi

CJ

a
u
H
^
rH
H
CJ


as

3
1
o
u

o

^
z
o
CO

03

p-l
s
o
CJ

H
CO
o
CJ



M

y


rH

CO
H







































en

4J
c
0
s
c
'H

o
c
41
3
cr
CO



d
o
•H
4J
CO


4J
O
(0
01
at






































































o
e

o







c
S
o

in











•
0
S

o














0
B

o

en







o
B
O

CN














•o
0)
03

^

Q
g







-a:
re
01



d
0
•H
d
a










OS

,_(
CD
cm

o
o
0
rH
O


OS


CO
00

o
o
o
I— f
o-





OS



CTJ
00

o
o
o
rH
-— ^







a!


rH
10
00

o
o
o
rH
o



*
OS


rH
CO
00

o
o
o


o-



01
_B

H ^-~
01
4-1 CO
CJ 4-1
CO 3

C 'r
0 g
CJ -^



4_(

u t
CO O
a. e
CO ^
o

3 ?
1 O o
^ s
Ct, N^









^ 0)
c/i a
>. >
o o o
CN r- 1  •
in r-« ej^





O o en
O O en
-3" CN rH

O O O




O*i rH rH

m ao o
rH







O O O
o m o
en i-H rH
o o d






^O CO O
1-1



O O r^

CN rH O
O O O




cj\ o^ in

\O CT\ CN













O O O
rH CN en









O O C




O O O




S

rH V4
•H cy

i M
U 0
CO  -3
•H 4-1
4J
O 4J
CO U

G
G O
OJ O
0)
3 'O
D .£3
•^
QJ 4-1
•S a
11

cd
-K








-------
        -A95-
                 Hl-Ht- H
                 OOOU Q
                 OQ OQ OQ OQ 00
                 UJ LLlUJUJ LU

                 (/> h-

                             o|
                             < 5
                             Z Q.
                             < UJ
                             03 <
                             "iS>
                             UJ T3
                             5 o>
                             c E
                                              o w
                                              S8
                             0.0)

                             lj(0

                             u-z
                             00


                             o§

                             SE
                             LL UJ
                             LLlQu

                             CO
                             CO

                             UJ
                             CC

                             O
W31SAS JO 1SOD

-------
      -A96-
 I-  I- h-   -
 O  O OOO
 00  OQ 00 CQfiQ
 UJ  LU UJ WQJ
 UJ  UJ UJ
 h-  I- l-
 D  3 ZD
 (D  •* CMOro
 eg  CM CMCM£:
X
I-
                                           CNI
                                           ^   (/)
                                                z
                                                g


                                           s   I
                                                I-
                                                o
                                           00
                                           (O
                                                UJ
                                                cc
UJ
UJ


UJ
CQ

O
g
cc
UJ
Q.
                                           CM
1
O
10
CNJ
1
O
d
CM
1
O
in
1
o
d
^
1
0
in

0
d


      O  H
      m  c/)
      LU  O
      s_x (L

      ILI  T3
z u.

O i-
O (/>


UJ O


> UJ
I- I-
Q. (/)

UJ (/>

U. Z
O O

I- Q

U^

t LJ
UJ CL


CO

UJ
CC
IAI31SAS dO 1SOO

-------
                            -A97-
                                                                     LU
                                         00000
                                         CQCQ CQQQOQ
                                         LULU LULULU

                                         (/)(/>(/> COW
                                         LULU LULULU
                                                          CO
                                                          CM
                                                              (0
                                                              X
                                                              h-

                                                              o
                                                              (0
                                                              z
                                                              g


                                                              <



                                                              o
                                                              <
                                                              LU
                                                              CC

                                                              Z
                                                              LU
                                                              LU
                                            LU
                                            CQ

                                            Q

                                            O

                                            QC
                                            LU
                                            Q.
                                                                  n'S
                                                                  

                             S o
                             CQS

                             > LU

                             Q- CO
                             S >•
                             LU CO

                             LL Z
                             00
                                                                  LL
                                                                  LL
                                                  CC
                                                  LU
q
in
q
o
q
in
q
d
                                                                  LU Q.


                                                                  LO
                                                                  CO

                                                                  LU
                                                                  CC

                                                                  3

                                                                  O
                              JO 1SOO

-------
                           -A98-
                  CQ CQ OQOQCQ
                  LU UJ LULU HI

                  W W (/)
                  UJ LU LULU HI


                  55556
                                               (f)
                  5 5555
                                           CM
                                           00
                                            CM
q
in
q

O
o
in
o
d
                                                  g cc
                                                  i- u

                                                  >§

                                                  o
                                    CC DC

                                    II
z  b £
o  m O
5  HI Q.


*  u7?>

s  1E
P  po
<  I- O
>  o ^
I-  < CC



It*
•"  / \ ^^
z  ^ CO
LU  Q O
LU  UJ O

^  ffl5
lii  ^" "J
OQ  h- H
Q  ^ CO
o  S >-
£  UJ CO


s  feg


   «s
   UJ ^
   LL DC
   U. UJ
   UJ Q.

   (O
   CO

   UJ
   CC
   D
   g
   u.
                  (|BB OOOI-/0)

                W31SAS dO ISOO

-------
                                      -A99-
     Influence of Granular Activated Carbon Cost




     Because of the wide variations in the types of activated carbon available and




their prices and performances, a water purveyor must have some understanding of




the relationship between activated carbon cost and performance.  If a lower cost




activated carbon also has lower performance, requiring more frequent reactivation,




then its use may result in higher total system cost as opposed to using a higher




cost carbon with better performance.  Figures 37 through 39 show the effect of




activated carbon cost and RF on the cost of 10 and 100 mgd sand replacement systems




and for a 100 mgd post-filter adsorber system, see also page 48 of the Interim




Guide.




     Influence of Inflation




     In the Interim Guide, Table XV (page 53) contains a summary of the "present




value" of the expenditures for 10 and 100 mgd sand replacement and 10 and 100 mgd




post-filter adsorber systems.  These data have been calculated for two discount




rates (6 percent and 8 percent) and three inflation rates (5 percent, 7 percent,




and 9 percent).  Capital expenditures were assumed to be amortized over 20 years




at 7 percent interest, and reflect a continuous, constant expenditure pattern over




the life of the facility.




     Tables XXII, XXIII and XXIV show the costs over time for 10 and 100 mgd sand




replacement systems and 10 and 100 mgd post-filter adsorption systems at assumed




inflation rates of 5 percent, 7 percent, and 9 percent.  Table XV (page 53) in the




Interim Guide summarizes present value calculations for the given investment




streams for each combination of system configuration and inflation rate.




     Figures 40 and 41 supplement Figures 18 and 19 in the Interim Guide




(pages 50  and 52) by showing the effect of 5 percent and 7 percent inflation




rates for two 10 mgd systems (one of each type.)  In contrast to the LOO mgd




systems, the post-filter adsorbers never become less expensive than the sand

-------
         -A100-
                                       O

                                       Q
                                       O
                                       oc
                                       111
                                       Q.
                                   en

                                   H
                                   Z
                                   O
ot-
o

01-
                                       u->-
                                       UJCO
                                       CO

                                       UJ
                                       oc
                                       D
                                       O

                                       EE
  (|B6

INH1SAS JO 1SOO

-------
                  -A101-
                                                               UJ
                                                               I-
                                                            ,_CL
                                     000
                                     N into
                                     606
                                     <*
-------
         -A102-
                                            DC
                                            UJ
                                            CO
                                            CC
                                          OQ
                                          H<

                                          IE
                                CM
                                    o
                                     Z
                                     g
                                    o
                                    <
                                    m
                                    cc

                                    Z
                                    LLJ
                                    UJ


                                    H
                                    UJ
                                    CQ

                                    Q
                                    g

                                    cc
                                    UJ
                                    0.
     DO
     ZO-

     
     I- O)
     (/) E


     8§
                                          CO Q

                                          CC LL
     UJ
     I-
        UJ
l_l    — mm

q    Oo
1
0
10
M
1
O
0
CM
1
0
in
v—
1
0
6
1
o
in

o
d
                                          LL DC
                                          LL UJ
                                          UJ Q.


                                          0>
                                          CO

                                          UJ

                                          CC


                                          O
(|B6 OOOi-/U>)

      dO 1SOO

-------
-A103-
Table XXII. INFLATIONARY IMPACT FOR INVESTMENTS IN POST FILTER ADSORPTION
AND SAND REPLACEMENT (5%)
Costs in C/1000 gal
,,£ars 10 mgd 100 mgd
After Post-Filter Adsorption Sand Replacement Post-Filter Adsorption Sand Replacement
Construction Capital O&M Total Capital O&M Total Capital O&M Total Capital O&M Total
CO >3- rH
OO rH 
CN r~
0 0
rH rH
ON ON
in ON
CM CM
rH rH
ON CO
r- oo
rH CO
CM rH
CN CO
CM CM
CO O
CM CM
OO ON
rH rH
rH VD
CJ ON
CM CM
r~ CM
O CO
in m
rH rH
sr in
rH rH
rH P^
o m
CO CO
rH r~-
CO CO
rH rH
rH rH
O -3-
3-
•^ -* m
vO ON ^~

-------
                                                                 -A104-
fr"
r-
H
!3
W
s
w
o
•<
rJ
PH
W
Pi

P

5
C/3
53
o
H
H
PH

O
tt)
Q
g  ?"
H

O
   4-1


i*
W
s
H
en
W
>
H
O
K^ '


|


O
t-l
H
X
X
43
<0
OO mgd
        T3
         M
         S

        O
 C. H
 M
 O
 W
T3
             fn
              a
              Q)
              P.
              M
              O
              CO
              -H
              P-i
                 O
                 •H
                 a.
Cap
                  ti
                  O
                 •H
                 4-1
           CO      O

           rfl  Q)  (-1
           C  4J 4-1
          p-i  MH  co
              <  e
                  O
                 CJ
                           oo
                           CO
                           CXI
                           O
                           r-
                           vD
                           0\
                           in
                                CM
                                CTi
                                          O


                                          00
                                                                                                                           CMCM
                                *3"   CO   ^
                                m   o*>   co
                                     in   vr>
                                                                                                                                CM
                           vO   *«^*   *^   vO
                           vo   o\   csi   m
                           oo   oo
                           rH   O^   C^*   rH
                           O   CM   m   o^
                                                                                       rHrH
                                                                                                                                CM
                                     o   
-------
-A105-


H
TD REPLACEMI
Q
ORPTICM
*$ rH
cd
H O
i-J O
it 3
H ~<>
CO
0 d
Pi -H
Z, CO
M 4J
CO O
w
CO
55
M
O
my IMPACT
fNOIIVIi
M
M
CU
rH
,0

icement
P.
cu
&
$
CO
•o
0 g
0 -H
O.
)-i
O
CO
TJ
-Filtei
CO
o
Pi
uacement
o.
cu
w
TJ
00
1(
: Adsorptic
:-Filtei
CO
o
CO
n) cu
0) 4J
rH
rd
H
2
rH
«0
•H
a
rd
u
"ca
4-1
H
rH
cd
4-1
P.
cd
u
rH
cd
o
H
a
i— (
•H
O.
CJ
rH
cd
0
H
1— 1
cd
4-1
Construction Cap:
oo sr
CO en
oo -*
rH VO
m m
o
rH
VO rH
vO O
00 ON
rH VO
O en
in
-*
rH ON
en cn
rH rH
"•3" **3"
CM O
0> O
rH
en
r^ i~i
vo r^
Q\ p*-.
m CM
rx. co
ON
O rH
•vf ON
CO en
p- CO
-3- ON
rH vo
vO vo
ON rH
cn co
ON ON
n cn
P^- CO ON
ON rH r —
O r-. 3"
in OO O
CN CM en
rH ON VO
CM cn P~
sr vo co
CM CM CM
-3-00
rH CO VO
cn 
-------
           -A106-
                                                      I


                                                     00
                                                o
                                                13
                                                oc
                                                t-
                                                (0
                                                z
                                                o
                                                O
                                                OC
                                                UJ
                                                (0
                                                oc
                                                <
                                                UJ
                                                     U_ UJ
                                                     O 5
                                                     h- u
                                                     5)0
                                                     o <
UJ O
I- Z
< <
CC 00

Z 0
oz
Zh-
— a.
 SCC
                                                      u.
                                                      UJ U.
                                                      UJ
                                                      tr
                                                      =)
                                                      o
                                                      \L
(IBB
       JO 1SOO

-------
         -A107-
                                       Z
                                       o
                                       i-
                                       u
                                       D
                                       QC
                                       I-
                                       o
                                       o

                                       QC
                                       LU
                                       V)
                                       CC


                                       LU
                                             05
                                             Q.UJ
                                             O H
                                             i- z
                                             LL LU

                                             OS


                                             fcS
OS

7UJ
CCO)
ZQ
OZ


ii

~Qu

u^O
                                             LLJ u_
                                             LLJ
                                             QC


                                             O

                                             LL
IN31SAS dO 1SOD

-------
                                         -Ales-







replacement systems.  For small systems,  post-filter adsorbers are less desirable




from an economic point of view than for larger systems.   A possible alternative




that minimizes the economic impact of using granular activated carbon adsorption in




small systems would be to consider the use of truck transport and regional




reactivation systems.




          Cost of Reactivation




     The most common (but not the only) method of reactivating carbon is by the




use of multi-hearth furnaces, and a significant portion of the cost of using granular




activated adsorption treatment is associated with on-site reactivation.  Figures 42 and




43 show the annual capital (amortized over 20 years at 7 percent interest) and




annual operating and maintenance cost for reactivation based on reactivation rate




in Ibs/day and can be used to calculate this cost.




     Using the data from Table X (page 34) in the Interim Guide, an example will




be constructed to demonstrate a convenient method for estimating a unit cost based




on the activated carbon use rate for reactivation.  Assume a 100 mgd treatment plant




with an activated carbon use rate as given in line 3, column 4 in Table X (page 34)




(530 mg/£).  The use rate converted to pounds per day as follows:




     Use Rate = 530 mg/£ x 8.34 Ibs/gal =




         4,420 Ibs/mil gal




     Multiplying by 100 mgd yields a reactivation rate of =




        442,000 Ibs/day




     Entering Figures 42 and 43 at 442,000 Ib/day yields an approximate annual




capital cost of 430,000 $/yr and annual operating cost of 6,500,000 $/yr.  The




total annual cost, therefore, is 6,930,000 $/yr for this rather high adsorbent




use rate.  The total yearly cost is divided by the annual flow (100 mgd x 365 days/yr =




36,500 mgy) to yield:

-------
                   -A109-
                                                           O
                                                           m
                                                           cc
                                                           <
                                                           o

                                                           o
                                                           UJ
                                                           I-
                                                           o
                                                           <

                                                           cc
cc
UJ
Q.
O
Z
o
Q.
LU
h-
CC
Ul
(/>
^
CC
0
z
0
p
<
_
V-
o
UJ
cc
z
0
CO
cc
o
Q
UJ
t—
^-
o
^
cc
o
X
cc
UJ
I
H~

3
5
cc
o
LL.
1-
co
0
o
UJ
0
z
z
UJ
H
2

^
5
Q (0
55
1»
H CO
< _
                                                             O
                                                           3 O
                                                           Z <
                                                           z ui
                                                           < cc


                                                           oi


                                                           UJ
                                                           cc


                                                           O
aad savmoa 'isoo 30NVN3iNivw QNV Nouvuado IVHNNV

-------
    -A110-
uv3A
savmoa 'isoo iviidvo IVHNNV

-------
                                     -Alll-


     Unit Cost = $6-93 x 10 /vr  = $l90/mg or 19 cents/1000 gal.
                 36,500 mgy

The above unit cost is for buying and operating the reactivation furnace only.

Summary

     This material is intended to supplement the material on pages 36 to 54 of

the "Interim Treatment Guide for Controlling Organic Contaminants in Drinking

Water Using Granular Activated Carbon."  It presents data that can be used to

provide additional understanding of the cost of using granular activated carbon

adsorption treatment for controlling organic contaminants in drinking water.

-------
                                        -A112





                                   ACKNOWLEDGMENTS




     Because of the size and complexity of this report,  its importance,  and




the short time that was available for its completion,  the authors judge  that this




is one of the more difficult assignments given to them.   It could not have been




completed without help.  The authors wish to acknowledge the contribution of




many members of the Physical and Chemical Contaminant  Removal Branch (listed




at the end of the Interim Treatment Guide) who helped  gather the data contained




in Appendix A, as well as Dale W. Dietrich and his staff of artists,




Nancy J. Quilhot, Stephen E. Wilson and Newell J. Maston who prepared all of




the Figures in the Interim Treatment Guide and Appendix A.  They also wish to




acknowledge the efforts of the Principal Investigators of the extramural projects




from which additional data were obtained.  Finally, this difficult project




could not have been successfully completed without the dedicated and skillful




efforts of Ms. Maura M. Lilly who typed all of the drafts and the final




manuscript so willingly and rapidly.

-------
                                    -A113-

                                REFERENCES

1.   Woodward, R.L. , Dostal, K.A. and Robeck, G.G., "Granular-Activated
     Gabon Beds for Odor Removal," JAWWA, 56_, No. 3, 287-295 (May 1964).

2.   Symons, J.M, "Summary of Granular Activated Carbon Practice Data,"
     Water Supply Research Division, U.S. Environmental Protection Agency,
     Cincinnati, Ohio, February 17, 1976, mimeo, 10 pp.

3.   Dostal, K.A., Pierson, R.C., Hager, D.G. and Robeck, G.G., "Carbon
     Bed Design Criteria Study at Nitro, West Virginia," JAWWA, 57, No. 5,
     663-674 (May 1965).

4.   Love, O.T., Jr., Robeck, G.G., Symons, J.M. and Buelow, R.W.,
     "Experience with Activated Carbon in the U.S.A.,"  In:  Papers
     and Proceedings of a Water Research Association Conference at the
     University of Reading, United Kingdom, Paper 13, 279-312 (February 1974)
     398 pp.

5.   Love, O.T., Jr. and Symons, J.M., "Evaluating the Need for a Sand
     Layer in a Granular Activated Carbon Filter," Water Supply Research
     Division, U.S. Environmental Protection Agency, Cincinnati, Ohio,
     April 22, 1977, Mimeo, 10 pp.

6.   Ford, D.B., "The Use of Granular Activated Carbon Filtration for
     Taste and Odour Control," In:  Papers and Proceedings of a Water
     Research Association Conference at the University of Reading,
     United Kingdom, Paper 12, 263-278 (Feb. 1974) 398 pp.

7.   Stevens, A.A. and Symons, J.M., "Measurement of Organics in Drinking
     Water," In:  Proceedings of the AWWA Water Quality Technology Conference,
     pp. XXIII-1 to XXIII-25 (Dec. 1973), American Water Works Association,
     Denver, CO (1974).

8.   Buelow, R.W., Carswell, J.K., and Symons, J.M., "An Improved Method
     for Determining Organics by Activated Carbon Adsorption and Solvent
     Extraction,"  JAWWA, 65_, No. 1, 57-72 and No. 3, 195-199 (Jan., March, 1973)

9.   Heymann, E., "Practical Experience in the Use of F.locculation Filtration
     Connected in Series to Granular Activated Carbon Filters," In:
     Translation of Reports on Special Problems of Water Technology -
     Volume 9 - Adsorption, Conference held in Karlsruhe, Federal
     Republic of Germany, 1975, pp. 112-127, EPA-600/9-76-030,  U.S.
     Environmental Protection Agency, Cincinnati, Ohio, (Dec. 1976), 454 pp.

10.   Middleton, P.M. and Rosen, A.A., "Organic Contaminants Affecting the
     Quality of Water," Public Health Reports, 7J., 1125-1133 (Nov. 1956).

11.   Rosen, A.A., Middleton, P.M. and Taylor, N.W., "Identification of
     Anionic Synthetic Detergents in Foams and Surface Waters," JAWWA,
     48, 1321-1330 (October 1956).

-------
                                        -A114-

12.   Rosen,  A.  A.,  Skeel,  R.T.  and Ettinger,  M.B.,  "Relationship of
     River Water Odor to Specific Organic Contaminants," JWPCF,35,
     No.  6,  777-782 (January 1963).

13.   Rosen,  A.A., Mashni,  C.I., and Safferman, R.S.,  "Recent Developments
     in the  Chemistry of Odour  in Water:  The Cause of Earthy/Musty
     Odour," Water Treatment and Examination, 19,  No.  1, 106-119 (1970).

14.   Anon.,  "Industrial Pollution of the Lower Mississippi River in
     Louisiana," Surveillance and Analysis Division,  U.S. Environmental
     Protection Agency, Region  VI, Dallas, Texas (April 1972),  Mimeo,  146 pp.

15.   Kloepfer,  R.D. and Fairless, B.J., "Characterization of Organic
     Components in a Municipal  Water Supply," Environmental Science and
     Technology, jS, 1036-1037 (Nov.  1972).

16.   Robeck, G.G.,  Dostal, K.A., Cohen, J. and Kreissl, J.F., "Effectiveness
     of Water Treatment Processes in Pesticide Removal." JAWWA, 57, No. 2,
     181-199 (Feb.  1965).

17.   Stieglitz, L.  , Roth,  W., Kuhn, W.  and Leger,  W.,  "The Behavior
     of Organohalides in the Treatment  of Drinking Water," Vom Wasser, 47,
     347-377 (1976).

18.   Grob, K. and Grob, C., "Organic Substances in Potable Water and in
     Its Precursor.  Part II.  Applications in the Area of Zurich,"
     Journal of Chromatography, 90, 303-313 (1974), Grob, K., Grob, K. and
     Grob, C.,  "Part III.   The  Closed-Loop Stripping Procedure Compared
     with Rapid Liquid Extraction," Journal of Chromatography,  106,
     299-315 (1975).

19.   Kolle,  W., Sontheimer, H.  and Stieglitz, L.,  "Quantification Test
     of Activated Carbon for Waterworks Based on Its Adsorption Properties
     for Organic Chlorine," Vom Wasser, 44_, 203-217 (1975).

20.   Suffet, I.H., Radziul, J.V., Cairo, P.R., and Coyle, J.T., "Evaluation
     of the Capability of Granular Activated Carbon and Resin to Remove
     Trace Organics from Treated Drinking Water," In:   Proceedings of
     Second Conference on Water Chlorination - Environmental Impact and
     Health Effects, Gatlinburg, Tennessee, October 31 - November 4, 1977,
     In Press.

21.   McCarty, P.L., Reinhard, M. and Argo, P.G., "Organic Removal by
     Advanced Wastewater Treatment," Presented at 97th Annual AWWA
     Conference, Anaheim, California, May 8-13, 1977.

22.   Anon.,  "National Interim Primary Drinking Water Regulations," Office
     of Water Supply, U.S. Environmental Protection Agency, EPA-570/9-76-003,
     Washington, D.C., 23 pp + 2 Appendices.

23.   Sorg, T.J., Logsdon, G.S.  and Love, O.T., Jr., "Manual of Treatment
     Techniques for Meeting the  Interim Primary Drinking Water Regulations,"
     Water Supply  Research Division, U.S. Environmental Protection Agency,
     EPA-600/8-77-005, Cincinnati, Ohio (May 1977), 73 pp.

-------
                                       -A115-

24.   Cohen,  J.M.,  Kamphake,  L.J.,  Lemke,  A.E.,  Henderson C.  and Woodward,
     R.L., "Effects of Fish  Poison on Water Supplies,  Part 1.   Removal of  Toxic
     Materials,"  J. Amer.  Water Works Assoc., 52,  1551-1566  (1960).

25.   Goodrich, P.R. and Monke,  E.J.,  "Insecticide  Adsorption on
     Activated Carbon," Trans.  Amer.  Soc. Agr.  Engr.  Gen. Ed.,  13,
     No.  1,  56-60 (Jan.-Feb. 1970).

26.   Hyndshaw, A.Y., "Activated Carbon to Remove Organic Contaminants
     from Water," J. Am. Water Works  Assoc., 64, No.  5,  309-311 (1972).

27.   Moergeli, B., "Removal  of Pesticides from Drinking Water," Sulzer
     Tech. Rev.,  54, No. 2,  91-96 (1972).  Also see Gas, Wasser, Abwasser,
     _52,  No. 5, 142-146 (1972).

28.   Shevchenko,  M.A., Marchenko,  P.V., Taran,  P.N.,  and Kravets, E.V.,
     "Removal of  Triazine-Series Herbicides from Natural Waters,"
     Vodopodgotovka Ochistka Prom. Stokov (USSR),  No.  9, 103-112, (1972).

29.   Spahn,  H., Brauch, V.,  Schluender, E., and Sontheimer,  H., "Active
     Carbon Filters for Water Purification, I.   Adsorption on a Single
     Particle," Verfahrenstechnik (Mainz), j?, No.  8,  224-231 (1974).

30.   Parkash, S.,  "Adsorption of Cationic Pesticides  (Diquat and
     Paraquat) from Aqueous  Solution by Activated Carbion,"  Carbon,  12,
     No.  5,  483-491 (1974).

31.   Shevehenko,  M.A., Marchenko,  P.V., Taran,  P.N.,  and Kravets, E.V.,
     "Removal of  Some Pesticides from Drinking Water by Adsorption Methods,"
     Vodosnabzh.  Sanit. Tekh. (USSR), 10, 31-32 (1974).

32.   Dawson, V.K., Marking,  L.L. and Bills, T.D.,  "Removal of Toxic
     Chemicals from Water  with Activated Carbon," Trans. Am. Fish Soc.,
     105, No. 1,  119-123 (1976).                                      ~~

33.   Thielemann,  H., "Identification and Removal of Odor and Taste-Imparting
     Substances in Filtered  Groundwater," Mikrochim Acta (E. Ger.)  No. 1,
     79-80 (1972).

34.   Willey, B.F., "Surveillance and Control of Phenolic Tastes and Odors
     in Water to  Prevent Their Effects on Taste and Flavor of Foods,"
     Amer. Chem.  Soc. Symp.  Series 26: Phenolic, Sulfur, Nitrogen Compd.
     Food Flavors Symp., 71-84 (1975).

35.   Engels, C. "The Removal of Extra Light Fuel Oil from Polluted
     Water," Vom Wasser, 45, 45-51 (1975).

36.   El-Dib, M.A., Ramadan,  F.M. and Ismail, M., "Adsorption of Sevin
     and Baygon on Granular  Activated Carbon," Water Res., 9,  No. 9,
     795-798 (1975).

37.   Allen,  D.H.,  "Evalution of Activated Carbon for Fuel Oil Adsorption
     from a Potable Water  Supply," Natl.  Tech.  Info.  Ser. (NTIS), AD Rep.,
     AD-A037478 (1976).

-------
                                     -A116-

38.   Koganovskii,  A.M.  and Leuchenko, T.M., "Calculation of  Isotherms  for
     Adsorption of Organic Nonelectrolytes  from Aqueous Solutions,"
     Colloid J.(USSR).  2i» N°-  3>  305-307 (May-June 1969).

39.   Bernardin, F.E.,  Jr., "Selecting and Specifying Activated Carbon
     Adsorption Systems," Chem.  Engr. (NY), 83, No. 22, 77-82 (Dec. 18,  1976).

40.   Zogorski, J.S.,  Faust, S.D.  and Haas,  J.H., Jr., "The Kinetcs of
     Adsorption of Phenols by Granular Activated Carbon," J.  Coloid.
     and Interface Science, 55,  No.  2, 329-341 (1976).

41.   Kondo, M., Kishi,  T., Shimamoto, T., Nohihara, T., and  Watabe, K.,
     "Studies on the Adsorption Effect of Activated Charcoal on Polychlorinated
     Biphenyls in Water," Memoirs  of the  Faculty of Pharmaceutical Sciences,
     (Japan), 18,  135-138 (1974).

42.   Snoeyink, V.L.,  Lai, H.T.,  Johnson J.H.  and Young, J.F.  "Activated
     Carbon:  Dechlorination and the Adsorption of Organic Compounds."
     Amer.  Chem. Soc.  Symp. on the Chemistry of Water Supply, Treatment
     and Distribution,  Dallas,  Texas (April 1973).

43.   Whitehouse, J.D.,  "A Study of the Removal of Pesticides from Water,"
     Univ.  of Kentucky Water Resources Institute, Research Report No.  8,
     Lexington, Ky. (1967).  (Also available from NTIS  No. PB-242-971).

44.   Zogorski, J.S. and Faust,  S.D., "Removal of Phenols from Polluted
     Water," Natl. Tech.  Info.  Ser., No.  PB-237-859 (June 1974).

45.   Gauntlett, R.B.  and Packham,  R.F., "The Removal of Organic Compounds
     in the Production of Potable Water," Chemistry and Industry, No.  17,
     812-817 (1973).

46.   Rambow, C.A., "Effect of Flow Rate on Carbon Filter Performance,"
     J. Amer. Water Works Assoc.,  55, No. 8,  1037-1043  (Aug.  1963).

47.   Aly, O.M. and Faust, S.D.,  "Removal  of 2,4-D Derivatives from Natural
     Waters," J. Am.  Water Works Assoc.,  57,  221-230 (1965).

48.   Morita, M., Nakamura, H. and Mimura, S., "Phthalic Acid Esters in
     Water," Water Res.,  8, No.  10,  781-788 (1974).

49.   Schmidt, K.,  "Effectiveness of Standard Drinking Water  Preparation
     for Eliminating Pesticdes and Other  Pollutants," Gas. Wasserfach  Wasser-
     Abwasser,, 115,  No.  2, 72-76 (1974).

50.   Vrochinskiy,  K.K., "Removal of Pesticides from Water by Various
     Water Preparation Methods," Gig. Sanit.  (USSR), W, No.  1, 76-78
     (1973).

51.   Calderbank, A.,  "Environmental Considerations in the Development  of
     Diquat and Paraquat on Aquatic Herbicides," Outlook Agr., ]_, No.  2,
     51-54 (1972).

-------
                                      -A117-

52.  Greve, P.A. and Wit., S.L., "Endosulfan in the Rhine River," J. Water
     Poll. Contr. Federation, 43, No. 12, 2338-2348 (1971).

53.  Stoltenberg, D.H., "How to Reclaim a Poisoned Lake," Pub. Works,
     103. No. 3, 59-61 (1972).

54.  Hansen, R.E., "Organics Removal of Mount Clements, Michigan," Presented
     to Michigan Section AWWA, Ann Arbor, Michigan (Jan. 1977).

55.  Giusti, D.M., Conway, R.A., and Lawson, C.T., "Activated Carbon
     Adsorption of Petrochemicals, JWPCF, 45, No. 5, 947-965 (May 1974).

56.  Juhola, A.J., "Regeneration of Activated Carbon, In:  Papers and
     Proceedings of a Water Research Association Conference at the
     University of Reading, United Kingdom, Paper 9, 177-202, (Feb. 1974),
     398 pp.

57.  Smith, S.B., "Techniques of Activated Carbon Regeneration," In:
     Proceedings of the Physical-Chemical Treatment, Activated Carbon
     Adsorption in Pollution Control Technology Transfer Seminar,
     Sponsored by Environment Canada and Pollution Control Association of
     Canada, 1975 Report M 1000.01, Chemical Division, Westvaco Corporation,
     Covington, Virginia, (1975), Mimeo, 50 pp.

58.  DeJohn, P.B. and Hutchins, R.A., "Treatment of Dye Wastes with Granular
     Activated Carbon," Textile Chemist and Colorist, 8^ No. 4, 34-38,
     (April 1976).

59.  Juntgen, H., "Phenomena of Activated-Carbon Regeneration," In:
     Translation of Reports on Special Problems of Water Technology -
     Volume 9 - Adsorption," Conference held in Karsruhe, Federal
     Republic of Germany, 1975, p. 269-283, EPA-600/9-76-030, U.S.
     Environmental Protection Agency,  Cincinnati, Ohio (Dec. 1976), 454 pp.

60.  Weissenhorn, F.J., "Testing of Activated Carbon Filters in Waterworks,"
     In:  Translation of Reports on Special Problems of Water Technology -
     Volume 9 - Adsorption," Conference held in Karlsruhe, Federal Republic
     of Germany, 1975, pp. 238-250, EPA 600/9-76-030, U.S. Environmental
     Protection Agency, Cincinnati, Ohio (Dec.  1976), 454 pp.

61.  Herzing, D.R., Snoeyink, V.L, and Wood, N.F. , "Activated Carbon
     Adsorption of the Odorous Compounds 2-Methylisoborneol and Geosmin,
     JAWWA, 69_, No. 4, 223-228 (April 1977).

62.  Snoeyink, V.L.,  McCreary, J. and Murin, C.J., "Activated Carbon
     Adsorption of Trace Organic Compounds,"  Final Report on U.S.  EPA
     Research Grant No.  R803473,  EPA-600/2-77-223, Cincinnati,  Ohio,
     March, 1978.

63.  Stevens, A. A. and Symons, J.M.,  "Measurement of Trihalomethane and
     Precusor Concentration Changes," JAWWA, 69,  No.  10, 546-554 (Oct. 1977).

64.  Love, O.T., Jr.,  Carswell, J.K.,  Miltner,  R.J.,  and Symons,  J.M.,
     "Treatment for the Prevention or Removal of  Trihalomethanes in
     Drinking Water,  JAWWA,  In Press.

65.  Symons,  J.M.,  "Interim Treatment Guide for the Control of Chloroform
     and Other Trihalomethanes,"  Water Supply Research Division,  U.S.
     Environmental  Protection Agency,  Cincinnati,  Ohio (June 1976),  mimeo.
     48 pp.  plus 4  Appendices.

-------
                                        -A118-

66.   Symons,  J.M.,  et al.,  "National Organics Reconnaissance Survey  for
     Halogenated Organics in Drinking Water," JAWWA,  67,  No.  11,  Part  1,
     634-647  (Nov.  1975), Update,  67_, No.  12, 708-709 (Dec.  1975).

67.   Sylvia,  A.E.,  Bancroft, D.A.  and Miller, J.D.,  "Detection and
     Measurement of Microorganics  in Drinking Water," In: Proceedings
     of the AWWA Water Quality Technology Conference, pp. XXVII-1 to
     XXVII-8  (Dec.  1974), American Water Works  Association,  Denver,  CO (1975).

68.   Stanier, R.W., Adelberg, E.A. and Ingraham,  J.,  The Microbial World,  4th
     Edition, Prentice-Hall, Chapter 29, "Microbial  Pathogenicity,"  p. 791,  (1976)

69.   Jorgensen, J.H., Lee,  J.C. and Pahren,  H.R., "Rapid Detection  of  Bacterial
     Endotoxins in Drinking Water  and Renovated Wastewater," Applied and
     Environmental Microbiology, 32, No. 3,  (Sept.  1976).

70.   McCreary, J.J. and Snoeyink,  V.L., "Granular Activated  Carbon in
     Water Treatment," J. Am. Water Works Assoc., 69, No. 8, 437-444,
     (1977).

71.   Guirguis, W.A., Melnyk, P.B.  and Harris, J.P.,  "The Negative
     Impact of Industrial Waste on Physical-Chemical Treatment,"
     Presented at 31st Purdue Industrial Waste  Conf., Lafayette, Indiana,
     (May 1976).

72.   Directo, L.S., Chen, C.-L., and Kugelman,  I.J.,  "Pilot  Plant Study
     of Physical-Chemical Treatment," J. Water  Poll.  Control Fed.,  49,
     No. 10,  2081-2098 (1977).

73.   Monscvitz, J.T. and Ainsworth, L.D., "Detection and Control of
     Hydrogen Polysulfide in Water," Public Works, 101. 113  (1970).

74.   Eberhardt, M. , "Experience with the Use of Biologically Effective
     Activated Carbon," in Translation of Reports on Special Problems
     of Water Technology, Vol. 9—Adsorption.,  H. Sontheimer, Ed.
     EPA Report EPA 600/9-76-030,  (Dec. 1976).

75.   Sontheimer, H., "Biological Treatment of Surface Waters in Activated
     Carbon Filters," presented at Seminar on Current Status of Wastewater
     Treatment and Disinfection with Ozone,  Cincinnati, Ohio, September 15,
     1977, Intl. Ozone Inst., Cleveland, Ohio.

76.    Kuhn, W., Sontheimer, H. and Kurz, R., "Use of Ozone and Chlorine
     in Water Works in the Federal Republic  of  Germany," Presented at
     Workshop on Ozone Chlorine Dioxide Oxidation Products of Organic
     Materials, Cincinnati, Ohio, Nov.  1976.  Intl.  Ozone Inst.,
     Cleveland, Ohio.

77.   Guirguis, W.A., Jain, J.S., Hanna, Y.A., and Srivastava, P.K.,
     "Ozone Application  for Disinfection in  the Westerly Advanced
     Wastewater Treatment Facility," in Forum on Ozone Disinfection,
     E.G.  Fochtman, R.G. Rice  and M. E. Browning, editors.  Intl. Ozone
     Inst.,  Cleveland, Ohio, pp.  363-381 (1976).

-------
                                     -A119-
78.   Guirguis,  W.A.,  Cooper,  T.,  Harris,  J. and Ungar,  A.,  "Improved
     Performance of  Activated Carbon by Preozonation,"  Presented at
     49th Annual Conf.  Water  Poll.  Control Fed., Minneapolis,  Minn.,
     October 1976.

79.   Scheidtmann, W., "Investigations of the Optimization of Pretreatment
     When Using Ozone," in Translation of Reports on Special Problems of
     Water Technology,  Vol. 9 -  Adsorption, op. cit.,  pp. 98-111.

80.   Schalekamp, M.,  "Use of  Activated Carbon in the Treatment of Lake
     Water," in Translation of Reports of Special Problems of Water
     Technology, Vol. 9 - Adsorption, op. cit.,, pp.  128-159.

81.   Van Lier,  W.C.,  Graveland,  A., Rook, J.J.  and Schultink,  L.J.,
     "Experiences with Pilot  Plant Activated Carbon Filters in Dutch
     Waterworks," in Translation of Reports on Special Problems of Water
     Technology, Vol. 9 - Adsorption,  op. cit., pp.  160-181.

82.   Sontheimer, H.,  "Considerations on the Optimization of Activated
     Carbon Use in Waterworks,"  in Translation of Reports on Special
     Problems of Water Technology,  Vol. 9 - Adsorption, op. cit.,
     pp. 208-214.

83.   Van der Kooij,  D., "Some Investigations into the Presence and Behaviour
     of Bacteria in Activated Carbon Filters," in Translation of Reports
     on Special Problems of Water Technology, Vol. 9 - Adsorption, op. cit.,
     pp. 348-354.

84.   Cornelia, C., and Versanne,  D., "Le Role de 1'Ozone dans la Nitrification
     Bacterienne de 1'Azote Ammoniacal — Cas de 1'Usine de la Chapelle
     Banlieue Sud de Rouene (Seine Maritime), France," Presented at 3rd
     Intl. Symp. on  Ozone Technol., Paris, France, May 1977.  Intl.
     Ozone Int., Cleveland, Ohio.

85.   Sontheimer, H.,  Heilker, E., Jekel, M., Nolte, H.  and Vollmer, F.-H.,
     "The 'Mulheim Process' - Experience with a New Process Scheme for
     Treating Polluted Surface Waters," Submitted for publication in
     J. Am. Water Works Assoc.

86.   Wachs, A., Narkis, N., and Schneider, M.,  "Organic Matter Removal
     from Effluents  by Lime Treatment, Ozonation and Biologically Extended
     Activated Carbon Treatment," Presented at 3rd Intl. Symp. on Ozone
     Technology, Paris, France,  May 1977, Intl. Ozone Inst., Cleveland, Ohio.

87.   Hopf, W.,  "Versuche mit  Aktivkohlen zur Aufbereitung des Dusseldorfer
     Trinkwassers",  Wass er/Abwasser, 101, No. 14, 330-336 (1960).

88.   Sontheimer, H.,  Univ. Karlsruhe, Germany.   Private Communication (1977).

89.   Kuhn, W. and Fuchs, F.,  "Untersuchungen zur Bedeutung der
     Organischen Chlorverbindungen und Ihrer Adsorbierbarkeit," Vom Wasser,
     45, 217-232 (1975).

90.   Kuhn, W. and Sontheimer, H., "Einige Untersuchungen zur Bestimmung
     von Organischen Chlorverbindungen auf Aktivkohle," Vom Wasser, 4_1,
     65-79 (1973).

-------
                                     -A120-
91.  Kuhn, W.  and Sontheimer,H.,  "Einfluss Chemischer Umsetzungen auf die
     Lage der Adsorptionsgleichgewichte an Aktivkohlen," Vom Wasser,
     40,  115-123, (1973).

92.  Kuhn, W., "Untersuchungen aur Bestimmung von Organischen
     Chlorverbindungen auf Aktivkohle," Dissertation, Fak.  f. Chemie-Ing.
     Wesen, Univ. Karlsruhe, Fed.  Rep.  Germany (1974).

93.  Wolfel, P.  and Sontheimer, H., 1974,  "Ein Neues Verfahren zur
     Bestimmung von Organisch Gebundenem Kohlenstoff in Wasser Durch
     Photochemische Oxidation," Vom Wasser, 43, 315-325 (1974).

94.  Poggenburg, W.,  Wasserwerk Dusseldorf.  Private Communication (1977).

95.  Steiner IV, J. and Singley,  J.E.,  "Methoxychlor Removal from Potable
     Water," Engineering Report,  Department of Environmental Engineering
     Sciences, University of Florida, Gainesville,  Florida  (1977).

-------
          APPENDIX B










    ANALYTIC METHODOLOGY




             FOR




MONITORING PILOT COLUMN TESTS
          Written by




       Alan A. Stevens









         Reviewed by




      Gordon G. Robeck

-------
                                 APPENDIX B

                             TABLE OF CONTENTS


Introduction                                                          Bl

Low Molecular Weight Halogenated Organic Compounds  Excluding
  Trihalomethanes - Performance Criterion 1                            Bl

Organic Carbon - Performance Criteria 2 and  3                         B5

     Analytic Method                                                  B5

     Sampling                                                         B6

Terminal Summation Trihalomethanes - Maximum Contaminant Level
  Requirement                                                         B7

     General                                                          B7

     Effect of Time                                                   B8

     Maintenance of Chlorine Residual                                 B9

     Effect of Temperature                                           BIO

     Effect of pH                                                    BIO

     Loss of Volatile Species                                        BIO

     Effect of Bromide or Iodide Contamination                       BIO

     Effect of Precursor Contamination                               Bll

     Procedure for Terminal Trihalomethane Determination             Bll

     Summary                                                         B13

Acknowledgments                                                      B13

References                                                           B14

-------
               ANALYTIC METHODOLOGY FOR MONITORING PILOT COLUMN TESTS

                                     INTRODUCTION

     The methods needed to monitor the granular activated carbon treatment pilot

studies can be divided into three specific categories:

     1.  Low molecular weight halogenated organic compounds, excluding

          trihalomethanes

     2.  Organic Carbon

     3.  Terminal "Summation Trihalomethane"

     Some inherent flexibility exists in each of these categories with regard to

specific measurement methods, instruments selected, and so forth. This discussion

will attempt to address some of the options available and relate them as closely

as possible to present, proposed, or future U.S. EPA Standard Methods that are

ultimately needed to meet compliance monitoring requirements.

LOW MOLECULAR WEIGHT HALOGENATED ORGANIC COMPOUNDS EXCLUDING TRIHALOMETHANES -

                             PERFORMANCE CRITERION 1

     This group of organic compounds has an operational definition. The intent is

to include such low boiling (low molecular weight) halogenated organics as carbon

tetrachloride, tri- and tetrachloroethylene, vinyl chloride, and others often

found in contaminanted drinking waters.   This group is selected because of:  1)

relative ease of analyses, 2) they are common contaminants, 3) evidence suggests

that they break through granular activated carbon adsorption systems earlier than

higher molecular weight toxic materials and therefore as a monitor, place more

stringent operating requirements on the process giving a higher level of protection

to the consumer.
^"Summation Trihalomethanes" is the arithmetic sum of the concentration of the
individual species of trihalomethane found in a given sample.  This parameter is
called "Total Trihalomethane" in the proposed Regulation.

-------
                                        -B2-
     EPA researchers have determined that over 30 purgeable halogenated organic

compounds can be detected and most resolved by use of the two gas chromatographic

columns under the conditions listed for use in the approved purge and trap trihalomethane

method  and when the "multi-purpose" trap containing silica gel is used.  Table I

lists these compounds in order of increasing retention time when a 0.2 percent

Carbowax 1500 on Carbopak C column is used.

                                    TABLE I

         COMPOUNDS DETERMINABLE ON 0.2% CARBOWAX 1500 ON CARBOPAK C

                    In Order of Increasing Retention Time

     chloromethane
     bromomethane
     vinyl chloride
     chloroethane
     methylene chloride
     1,1-dichloroethylene
     bromochloromethane
     1,1-dichloroethane
     trans-1,2-dichloroethylene
     chloroform
     1,2-dichloroethane
     1,1,1- trichloroethane
     carbon tetrachloride
     bromodichloromethane
     1,2- dichloropropane —  __
     2,3-dichloro-l-propene	
     trans-1,3-dichloropropene
     trichloroethylene
     1,1,2- trichloroethane
     dibromochloromethane —
 unresolved
   :unresolved
     cis-1,3-dichloro-l-propene•
     1,2-dibromoethane
     2-bromo-l—chloropropane
     bromoform—	__
     1,1,1,2-tetrachloroethane**'
     tetrachloroethylene
     1,1,2,2-tetrachloroethane
     1,4-dichlorobutane
     chlorobenzene
unresolved

-------
                                          -B3-
     A second column (N-octane on Porasil-c) changes the retention order (see

Table II),  gives different separations, and allows detection of some organic

compounds with higher boiling points.

                                  TABLE II

               COMPOUNDS DETERMINABLE ON N-OCTANE ON PORASIL-C

                   In Order of Increasing Retention Time
                       j-unresolved
vinyl chloride ..
chloromethane —
bromomethane
1,1-dichloroethylene
chloroethane
trans-1,2 dichloroethylene
methylene chloride
carbon tetrachloride
chloroform— .	
     cis-1,2-dichloroethylene
     1,1-dichloroethane
     bromochloromethane
     1,1,1-trichloroethane
     trichloroethylene
     bromodichloromethane
     dibromomethane
     tetrachloroethylene
     1,2-dichloroethane
     dibromochloromethane
     trans-l,3-dichloro-l-properie
     1,2-dichloropropane
     cis-1,3-dichloro-l-propene
     1,1,2-trichloroethane
     2-bromo-l-chloropropane
     chlorobenzene
     1,2-dibromoethane
     bromoform
     1-chloro-l-hexene
     chlorohexane
     1,1,2,2-tetrachloroethane
     pentachloroethane
     £-chlorotoluene
     m-dichlorobenzene
     hexachloroethane
     £-dichlorobenzene
                                 ..unresolved
                            •unresolved
                              •unresolved
                         unresolved
                                  :unresolved
                                •unresolved
                               .unresolved
                                 unresolved
     1,4-dichlorobutane.
     j>-dichlorobenzene
     hexachlorobutadiene
     1,2 ,4-trichlorobenzene
                        :unresolved

-------
                                        -B4-


     Therefore, the operational definition of those low molecular weight halogenated


compounds excluding trihalomethanes related to performance Criterion 1 is;  those


compounds that can be detected by use of the U.S. EPA approved trihalomethane


method - purge and trap version.  Note:  Determining these compounds as part of the


granular activated carbon adsorption system process control program should not be a


large analytic burden when the water utility required to use granular activated


carbon treatment has selected the proper approach for the required measurement of


trihalomethane concentrations.


     The purge and trap method is selected for this analysis mainly because gas


chromatographic conditions have not been well defined for resolution of as many

                                                 2
compounds by the liquid-liquid extraction method,  and because of the difficulty of


solvent interference in the liquid-liquid extraction method with the detection of


many compounds eluting sooner than chloroform.  This is not to say, however, that


these conditions could not be developed.  Additionally, care must be taken in selecting


a particular purge-and-trap apparatus for introduction of the sample into the gas


chromatograph, as a well designed apparatus is needed to eventually accomplish resolutio


of the same lower boiling compounds (vinyl chloride, bromomethane, and so forth).


Details of this are to begin in a supplemental writeup extending the coverage of the


EPA approved trihalomethane method to this larger group of compounds.


     In spite of the resolving power of the two columns used for the purge and trap


trihalomethane method,   all possible purgeable organohalides and isomers cannot be


identified unambiguously.  Identifications by dual column GC-specific halogen detection


are still presumptive.  In order to avoid escalating analytical costs (e.g. - mass


spectrometric  (MS) detection),  these presumptive identifications are considered


sufficiently accurate and granular activated carbon adsorption treatment should be


adjusted to reduce any of these contaminants to  less than 0.5 yg/& based on a


comparison with an analytical standard  of the compound of presumed identity, "Performan


Criterion 1."  Except in the  case of some isomer differentiations, MS confirmation


of compound identity might be useful enough  to warrant the effort and could ultimately


lead  to a better designed treatment strategy.

-------
                                         -B5-





                ORGANIC CARBON - PERFORMANCE CRITERIA 2 and 3




Analytic Method




     A method for the determination of organic carbon must include a step to




remove or adjust for the presence of inorganic carbonate.  Instruments on the




market today, using manufacturers recommended procedures, do this adequately and




efficiently.  Many procedures, however, effect the removal of carbonate by acidification




with mineral acid followed by exhaustive purging of carbon dioxide to waste.  In




this process, a purgeable fraction of organic carbon (POC) is lost.  Ideally, this




purgeable fraction should be included in any total organic carbon (TOC) measurement.




Research in WSRD laboratories, however, has shown that, in general, the POC concentration




is only a small fraction of the TOC concentration and little organic carbon (< 10




yg/&) would be lost by measuring only the non-purgeable organic fraction (NPOC).




Only occasionally have POC concentrations been found in the range of 0.1 - 0.3




mg/£. When the insignificance of POC can be demonstrated, only NPOC need be measured.




Note:  For these cases NPOC determinations can be used to satisfy the TOC requirements




of "Performance Criteria 2 and 3."  This can simplify the sampling requirements and




analysis and expand the options for methods and instruments that can be used.




     Because of the requirements of the Performance Criteria 2 (measure a change of




0.5 mg/£) and the commonly found low concentrations of organic carbon in a granular




activated carbon column effluent (< 0.1 mg/£) the method used to measure organic




carbon must be more precise and accurate when compared to methods and instruments




used in the waste water treatment field, where concentrations of organic carbon are




often in the hundreds of milligrams per liter.  Required precision and accuracy for




measurement of organic carbon in granular activated carbon column effluents therefore




must be in the range of 0.1 mg/£ + .05 mg/& or better.  Although no EPA standard




procedures now exist for this purpose, instruments are available that can be operated

-------
                                           -Be-
according to manufacturers instructions to achieve these goals.   At least one of




these is capable of measuring both POC and NPOC independently.




     The instrument (and method)  must also be capable of measuring TOG concentrations




as high as those expected to occur in the granular activated carbon column influent




(2-10 mg/£).   This is well within the range of instruments capable of the required




low level analysis.




Sampling




     Because of the variability with time of concentrations of  organic carbon in




actual treatment situations, determining what constitute a representative "baseline"




(TOG ) sample for all systems is  difficult. Additionally, activated carbon fines




emanating from new granular activated carbon adsorbers have been observed to influence




organic carbon determinations. For these reasons,  the first organic carbon measurements




are recommended to be performed after one week of  adsorber operation and be performed




at least in triplicate twice during a 2-day period (six samples each on adsorber




influent and effluent).   The average of effluent values not influenced by activated




carbon fines (activated carbon are not visible, or samples do not have a NPOC value




considerably higher than others of the set) is then the "baseline" (TOG ) effluent




concentration against which future effluent samples are compared, Performance




Criterion 2,  and determines the initial organic carbon removal  percentage, Performance




Criterion 3,  when compared with a "baseline" influent concentration.




     Sample containers should be  the same as those used for sampling for trihalomethane




analysis when the POC fraction is to be measured.   Headspace requirements are not




as important when the POC fraction has previously  been shown to be insignificant.

-------
                                            -B7-


        TERMINAL SUMMATION* TRIHALOMETHANES - MAXIMUM CONTAMINANT LEVEL REQUIREMENT


General


     A detailed method for determining Terminal Trihalomethane (Term THM) concentrations

                                         3
has been described by Stevens and Symons.   The principle of Term THM measurement


"Terminal Summation Trihalomethane" is the special case where the individual species


Term THM concentrations in the units yg/£ are arithmetically summed.


     Because their formation is not instantaneous, trihalomethane (THM) concentrations


increase in the water as it flows through a water treatment plant (unless removed


during treatment) to reach some value higher than that which would be observed if


an analysis for THM species was performed immediately after sampling at the first


point of chlorination.  Further, the consumer is likely to receive water with THM


concentrations higher than those leaving the plant because the reaction proceeds in


the distribution system.  Additionally, not only are the concentrations of THM time


dependent, but the rate of the reaction is dependent on pH, precursor concentration,


nature of precursors, temperature and to some degree free chlorine concentration


early in the chlorination process. Finally, the ratio of chloroform to other trihalomethanes


is highly dependent on the bromide content of the source water.


     The Term THM concentration is defined as the concentration of THM that occurs


at the termination of the measurement of this parameter.  To measure Term THM


concentration, chlorine-precursor reaction conditions are selected according to the


treatment practiced at the particular plant being evaluated.  In general, to


determine Term THM a sample of water is chlorinated under these plant conditions,


and chloroform and other THM species concentrations are determined after some


preselected holding period or periods.
*As noted previously "Total" is used in the proposed Regulation, rather than
"Summation" to mean the arithmetic sum of individual trihalomethane concentrations.

-------
                                       -38-






     Th e Term THM concentration is an important parameter for evaluating drinking




water quality because it is an estimate of the concentration of THM reaching the




consumer at various points (residence times) in the distribution system.  Therefore,




determining the change with time in the Term THM concentration in the effluent of a




granular activated carbon adsorption system is a good estimate of whether or not




the effluent would meet the summation trihalomethane maximum contaminant level




required by the Interim Primary Drinking Water Regulations.




     The selected conditions for the Term THM measurement must be the same as




those experienced at the water treatment plant or distribution system and must be




reproducible from sample to sample.  Critical conditions to consider are time of




reaction (time elapsed before halting the halogenation reaction with a reducing




agent) maintenance of a chlorine residual, temperature, pH, prevention of loss of




the volatile products during the time of contamination and avoidance of contamination




of reagents.




Effect of Time




     Although a single measurement of THM concentrations after a storage




period of several days in a bottle under appropriate conditions can give a




useful determination of the Term THM concentration for that specified time, much




more information can be gained from the reaction-rate curves obtained by plotting




THM concentrations vs. time.  The rate curves obtained by periodic measurement of




THM concentrations in properly stored finished water can be used to estimate the




future THM concentrations at any given time after water leaves the treatment plant,




as required by the Interim Primary Drinking Water Regulations.  This is particularly




important when the goal is to estimate ultimate consumer exposure to THM at different




points along the distribution system.  The THM concentration-vs-time curve

-------
                                        -B9-

is especially useful where the utility has a large variation in the time that water

is in various parts of the distribution system.

     When analyzing a granular activated carbon adsorption system effluent, simply

measuring a single point on the THM growth curve such as one representing maximum

residence time in the distribution system would place a requirement on adsorber

performance more stringent than required under the Interim Primary Drinking Water

Regulations.  To be consistent with that part of the Regulation, generation and use

of a complete THM growth rate curve is recommended to properly evaluate adsorber

performance.  Specifically, the effluent water from the granular activated carbon

adsorption unit should be chlorinated in a manner consistent with treatment plant

practice where free chlorination is used (see below).  Aliquots of that water are

then stored for various times from "0" (corresponding to clear well) to "T" (corresponding

to maximum time in distribution system) . The intermediate time samples _(T_, _T, and
                                                                       ~2  4"
so forth) determining the shape of the THM growth curve and, therefore, expected THM

concentrations at intermediate distribution system residence times.

     A minimum of five points on the curve (0, T, and 3 in between) are selected

consistent with the sampling plan for THM MCL compliance monitoring. The average

Term Summation THM concentration obtained from those five values (of Term Summation

THM at respective times) is then used to judge adsorber performance.

Maintenance of Chlorine Residual

     In conventional U.S. water treatment practice, maintenance of a free

chlorine residual throughout the distribution system often is recommended

or required.  The continued reaction of precursor with chlorine to yield

trihalomethanes depends on the maintenance of a free chlorine residual.

Thus for evaluation of systems, where free chlorination is practiced, a

chlorine-residual measurement always must be performed at the time of

THM analysis to ensure that a free residual is present.

-------
                                          -BlO-





Effect of Temperature




      Upon chlorination of a natural water, approximately twice as much chloroform




can be formed in a given period of time at 25°C as is formed at 3°C.  This




range of temperature is not uncommon, summer to winter, in U.S. surface




waters.  A need for close temperature control during the determination of Term




THM concentration, therefore, is indicated.  Because temperature largely is




controlled seasonally, this temperature effect must be taken into account, if




extrapolations to summer operating conditions from winter pilot studies and




vice versa are made.




Effect of pH




     The trihalomethane formation rate has been shown to increase with an




increase in pH.  Because pH is a factor determining rate of THM formation and




therefore Term THM values, the pH should be controlled near that found in the




distribution system.




Loss of Volatile Species.




      To prevent misleading losses of trihalomethanes produced during the




reaction period, the reactions must be carried out in sealed, head-space-free




containers.  Container materials should be all glass or glass with PTFE-lined




caps.   Standard glass-stoppered reagent bottles filled to overflowing so as to




wet the stopper surface or the PTFE-septum-sealed serum vials, used for sampling




for Inst. THM determinations have been found suitable.




Effect of Bromide or Iodide Contamination




     As mentioned earlier, bromide or iodide present in the water can cause formation




of THM species other than chloroform, as a result of first reacting with chlorine.




In the case of bromide, the relative amounts of THM species formed has been shown




to be highly dependent on the bromide content of the water and the chlorine dose,




presumably because these determine the ratio of bromine to chlorine available for




completing reactions.   This change of product ratio, because bromine-containing

-------
                                          -Bll-







species are heavier than chloroform, could cause a dramatic change in calculated




Terminal Summation Trihalomethane values.




     Preliminary work indicates that equal amounts of bromine and chlorine substitution




as trihalomethanes would be expected if the original bromide concentration is as




little as 2 percent of the chlorine dose.  Clearly, any bromide or iodide contamination




of reagents used will cause a different ratio of THM species to be formed than




would occur normally on chlorination of that water under plant conditions and must




be avoided.




Effect of Precursor Contamination




     In the WSRD laboratory, distilled, deionized, activated carbon-filtered water




has been used for "blank" water for reagent preparation.  At pH 7, the contribution




of precursor in reagents has been small.  At higher pH, however, blank values tend




to be higher.  Care should be taken to minimize volumes of reagents used in Term




THM measurements in order to avoid this contribution to the THM concentrations




obtained.




Procedure for Terminal Trihalomethane Determination




     A test for Term THM concentrations can be standardized in approach, but the




conditions for sample treatment and storage will vary widely from system to system




depending upon distribution-system residence time, total clorine demand of the




sample, ambient temperature of the system, and pH of the finished water in the




particular system under investigation.   These variables must be chosen to match




those in the system.

-------
                                             -B12-
     In work at the WSRD laboratory,  a large (1-3 liter)  sample of water is collected,



and the pH adjusted to that selected  with an appropriate  inorganic (e.g. phosphate



or borate) buffer.   The final buffer  strength is about 0.01 M,  although the strength



is not critical as  long as the desired pH is maintained.   The sample then is chlorinated,



if needed, by the addition of a previously standardized chlorine or hypochlorite



solution.  Sufficient chlorine is added at this time to maintain a free residual



for the duration of the test period.



     Several sample bottles are filled and capped head-space-free, two bottles for



each point to be determined on the THM growth rate curve.  For example, ten bottles



are needed for the five points on the curve to be determined. One of the zero-time



sample bottles contains sodium thiosulfate to immediately reduce the chlorine so



that the "0" time THM concentration is measured. The other zero-time sample has no



reducing agent and is used for measurement of the chlorine residual.  This entire



sequence from sample collection to the capping of the bottles should be done as



quickly as possible to avoid loss of  THM during the manipulations.



     The samples, except zero-times,  are stored at the selected temperature.



After the preselected times one sample bottle is opened and an aliquot is transferred



by pouring into a smaller bottle containing sodium thiosulfate to prevent further



reaction of precursor with chlorine.   This smaller bottle is quickly sealed head-



space-free to await THM analysis.  This measurement determines a THM concentration



for the respective time on the rate curve.  A second bottle is opened at the same



preselected time and the chlorine residual is measured.



     The actual measurement of THM concentrations that are arithmetically summed to



produce the "summation" or "total" value can be performed by use of either of the



two U.S. EPA approved procedures for that determination  (purge-and-trap or liquid-


                   1 2
liquid extraction).  '   The calculated sum of the THM concentrations by the EPA



approved procedure is the Terminal Summation Trihalomethane concentration for the



respective time period.

-------
                                         -Bis-
     Brief ly, for the purge and trap  analysis, the sealed sample is brought to


25°C prior to opening in order to obtain reproducible purging efficiencies. A 5-m£


aliquot then is removed and transferred to a glass purging apparatus, wherein the


trihalomethanes are stripped from the aqueous phase by passage of a flow of helium


upward through the sample.  The trihalomethanes stripped in this manner are collected


on a sorbant, porous polymer material contained in a stainless steel trap that is


placed in series with the purging device. The trihalomethanes are desorbed thermally


from the trapping material onto a gas chromatographic column.  Finally, temperature-


programmed gas chromatography is carried out, and the concentrations of trihalomethanes


are measured by use of a halogen-specific detector.  The liquid-liquid extraction

      2
method  involves the extraction of a small volume of water with an even smaller


volume of organic solvent, followed by gas chromatographic analysis of the extract


using an electron-capture detector.


Summary


     Weekly analysis of the effluent from a granular activated carbon adsorption


system by this procedure will provide a good estimate of whether or not the delivered


drinking water will be in compliance with the trihalomethane maximum contaminant


limits as specified in the Interim Primary Drinking Water Regulations.


                               ACKNOWLEDGMENTS


     The author and compiler wish to express their sincere thanks to Ms. Maura M.


Lilly, who typed this Appendix so quickly and accurately, such that the entire


Interim Guide could be finished on time.

-------
                                        -B14-


                                   REFERENCES
1.   "The Analysis of Trihalomethanes in Finished Waters by the Purge and
     Trap Method," U.S.  Environmental Protection Agency, Environmental
     Monitoring and Support Laboratory,  Cincinnati,  Ohio 45268, Sept. 9,  1977.

2.   "The Analysis of Trihalomethanes in Drinking Water by Liquid/Liquid
     Extraction," U.S.  Environmental Protection Agency, Environmental Monitoring
     and Support Laboratory, Cincinnati, Ohio 45268.

3.   Stevens, A.A. and Symons,  J.M., "Measurement of Trihalomethane and
     Precursor Concentration Changes," JAWWA, 69, No.  10, 546-554
     (Oct. 1977).

-------
        APPENDIX C
DESIGN OF PILOT GRANULAR




ACTIVATED CARBON COLUMNS
       Written by




  0. Thomas Love, Jr.




          and




  Kenneth L. Kropp
       Reviewed by




    Gordon G. Robeck

-------
                                 APPENDIX C






                              TABLE OF CONTENTS









Introduction                                             Cl




Selection of Materials                                   Cl




Adsorption Columns                                       C2




Establishing Test Conditions                             C2




Operation                                                C7




Parts List                                               C9




Acknowledgments                                         Cll

-------
             DESIGN OF PILOT GRANULAR ACTIVATED CARBON COLUMNS







Introduction




     The "Interim Treatment Guide for Controlling Organic Contaminants in




Drinking Water Using Granular Activated Carbon" describes two adsorption




schemes.  The first, is to retrofit an existing water treatment plant by




replacing the filtration media with granular activated carbon, and the




second, called post-filter adsorption, requires separate contactors following




te filtration step.  The ability of either adsorption scheme to reduce




the concentration of trace organics should be investigated on-site using




a small manageable system before expending large sums of money on a process




design that may give marginal performance.  Trial experimentation (i.e.




pilot studies) permits selecting the appropriate concept through improved




engineering judgement.  This Appendix describes the design and operation




of the EPA Water Supply Research Division's experimental adsorption systems




used in organic removal studies.  This material is presented for guidance




as an example of one approach to gathering adsorption performance data.




Selection of Materials




     One precaution taken by EPA in trace organic studies is to construct




pilot scale equipment with stainless steel, Teflon, and glass whenever




possible to minimize contamination from structural materials during




experimentaiton.  Whether or not materials such as rubber, acrylic resin,




polyvinylchloride (PVC), polyethylene tubing, or similar products would




compromise the experimental results is not known.  To avoid this possibility,




however, these materials should not be used if possible.

-------
                                         -C2-




Adsorption Columns


     Because of a limitation of available water,  the WSRD pilot scale


adsorbers consist of 3.8 cm (1.5 inches)  diameter glass columns 153 cm


(60 inches) in length.  A schematic of the experimental system with


details on fabrication are shown in Figures 1-4.   The columns are arranged


so that both modes, sand replacement  and post-filtration adsorption,


can be studied simutaneously if desired.   Granular activated carbon is


placed in column 1 (see Figure 1) to a depth allowable in the existing


filter boxes at the water treatment plant.  Settled water should be


applied to this unit at an approach velocity similar to that in the


existing plant.  Approximately 15 cm (6 inches) of graded gravel should


be placed in the bottom of the columns as an aid in distributing the


backwash water.  Some type of surface scrubbing (air or water scour)


should be incorporated in the system because with granular activated


carbon, like sand, most of the filtration occurs in the upper few centimeters


of the bed and vigorous scouring insures adequate cleansing. The surface scrubber


shown in Figure 1 (see also Figures 2 and 4) is intended to slide on a Teflon


ferrule so that it can be located close to the granular activated carbon surface


for effective agitation.  If headloss monitoring is desired, a U-tube


manometer or a sensitive pressure gage can be included in the design.  If, however,


headloss is not monitored, a backwashing schedule such as that used within the existing


plant can be employed.


Establishing Test Conditions


     Like disinfection, the process of adsorption is very dependent upon


contact time (see Figure 24, Appendix A, page A72).  The term "empty bed


contact time" (EBCT) is commonly used to characterize this adsorption


variable and is calculated by dividing the volume of media "V" by the


hydraulic loading "Q" (i.e. — = EBCT).  Assuming a filter box contains from


76 to 122 cm (30 to 48 inches) of media and the hydraulic loadings range

                                   2
from 4 to 8 m/hr (2 to 4 gal/min/ft )  then the EBCT in a sand replacement

-------
                                             -C3-
INFLUENT
         (PUMPl







         DETAIL  B
                    •
                    a.
                    UI
                    O


                                       SURFACE SCOUR


                                       DETAIL  B.
                              BACKWASH EFFLUENT
                                   EFFLUENT
(V)  COL. 1
                                   TO WASTE
                                                                                       DETAIL A
r
T
D


1
I
J
(
t
\)
•

!


(;
i)
•
T
1
%
1
y
A
(j
z
0
K
u
o
Ui
u
X
^
z
0
* (
j- ^VGI
                                                                                                 INFLUEN1
                                                                      •-GRADED GRAVEL
                                                  COL. 4
                                                               COL. 3
COL. 2  ®
                                                                               BACKWASH INFLUENT
                         FIGURE 1  PILOT GRANULAR  ACTIVATED CARBON  COLUMNS

-------
                       -G4-
ITEM 1
ITEM 2
ITEM 4
ITEM 5
ITEM 6
ITEM 7
                                Slip  joint  for
                                surface  scrubber
                                              -COLUMN
                                               INFLUENT
                                                 ITEM  3
ITEM 8
ITEM 6
ITEM 10
— ITEM 9
              FIGURE 2  DETAIL  A

-------
                     -C5-
ITEM 3
                     I EDO
ITEM 5
ITEM 6
ITEM 7
ITEM 8
ITEM 6
ITEM 10
— ITEM 9
              FIGURE 3  DETAIL B

-------
         -C6-
                 1/8"  STAINLESS STEEL TUBING
                           "316"
                     SLOTS OR HOLES
                Q^	END OF TUBING CRIMPED
                    CLOSED
FIGURE 4  SURFACE SCOUR

-------
                                        -C7-






mode could typically vary between 1.5 to 15 minutes, with an average of 7 to 8




minutes.




     The additional columns shown on Figure 1 can be operated concurrently in a




post-filter adsorption mode to examine the effects of longer empty bed contact




times.  For example, columns 2, 3, and 4 (Figure 1) can be charged with granular




activated carbon, exposed to filtered rather than settled water, and sampled in




series to monitor the breakthrough (wave front) of specific organics ("Performance




Criterion 1") and the total organic carbon ("Performance Criterion 2 and 3"). On




the other hand an investigator may desire to use column 2 as a sand filter receiving




settled water and columns 3 and 4 as adsorbers.  Another column in series would




further increase the EBCT. Additional columns and the appropriate plumbing modifications




would allow a utility to investigate in parallel, rather than in sequence, performance




of different brands of commercially available granular activated carbon, should




that be desirable.  Finally, the choice as to whether or not to apply disinfected




water to the adsorber will depend on how the final treatment scheme will be arranged.




The engineer should have some experimental "breadth" to select the most satisfactory




adsorption design.




Operation




     The granular activated carbon must be wetted before it is put into service.




This is accomplished by backwashing the material similar to the initial steps in




using any granular media.  Gently tapping or bumping the columns might be necessary




to insure that the media is wetted.   Sufficient freeboard should exist to permit 50




percent bed expansion during backwash.   The frequency of backwashing during the




experimental study will vary, depending upon the same factors influencing full-




scale filtration (e.g., headloss, turbidity,  carryover floe from settling, etc.)




Details on monitoring the system are given in Appendix B, however,  a routine operation




schedule would include the following:

-------
                                        -C8-
          Parameter                   Frequency



     Flow Adjustment, Q               Daily

     Temperature, pH, Turbidity     Investigator's discretion —
                                    Based on variability of ap'plied water

     Low Molecular weight
      halogenated organic compounds Weekly

     Organic Carbon                 Weekly

     Terminal Summation THM         Weekly

     At most field installations no restriction in available water volume would

exist as it does in the WSRD pilot plant.  Therefore larger, at least 10 cm

(4 inch)  diameter columns are recommended for use in the treatability studies

proposed in the Interim Treatment Guide (see pages 23-26.)

     The EPA Water Supply Research Division is currently designing

an experimental adsorption system using 10 cm (4 inch)  diameter glass columns.  The

system will be installed in at least one existing water treatment plant, and the

experimental results along with the problems encountered with pumps, maintenance,

and operation will be reported on when available.  Details may be obtained by

writing to the Director, Water Supply Research Division, Municipal Environmental

Research Laboratory, 26 West St. Clair Street, Cincinnati, Ohio 45268.

-------
ITEM NUMBER

     1
     2

     3
     4

     5
     9


    10
     V
     R
   PUMP
           C-9

        PARTS LIST
        DESCRIPTION*                     QUANTITY

STAINLESS STEEL MALE CONNECTOR           3 EACH
1/8" TUBE x 1/4" PIPE THREAD
BORED THRU WITH TEFLON FERRULES
(EXAMPLE - SWAGELOK FITTING
# SS-200-1-4-BT WITH TEFLON
FERRULES)

1/4" STAINLESS STEEL PIPE  TEE             3 EACH

STAINLESS STEEL MALE BRANCH TEE           8 EACH
1/4" TUBE x 1/4" PIPE THREAD
(EXAMPLE - SWAGELOK FITTING
# SS-400-3-4TTM)

1/4" STAINLESS STEEL CLOSE NIPPLE        3 EACH

1/2" x 1/4" STAINLESS STEEL REDUCING     8 EACH
BUSHING

1-1/2" CORNING CONICAL FLANGE            16 EACH
STYLE 1, ALUMINUM, # 72-9061

1/2" TEFLON SHEET SHAPED LIKE  THE        8 EACH
FLANGE IN ITEM 6, BOLT HOLES
CLEARENCE DRILLED FOR 5/16 X  18
BOLTS, CENTER DRILLED AND  TAPPED
FOR 1/2" PIPE THREAD

1-1/2" GASKET, STYLE 1-2,  SOLID           8 EACH
TFE, TYPE T, CORNING NUMBER 72-9255

1-1/2" CORNING CONICAL MOLDED             8 EACH
INSERT  (HARD), # 72-9057

1-1/2" X 72" CORNING PYREX               4 EACH
CONICAL PROCESS GLASS PIPE
# 72-7501

ALUMINUM 5/16 X 18 x 2" BOLTS  W/NUTS    24 EACH

STAINLESS STEEL FORGED BODY SHUT-OFF    17 EACH
VALVE WITH VEE TYPE STEM AND  1/4"
TUBE FITTINGS  (EXAMPLE - WHITEY
VALVE # SS-1VS4)

STAINLESS STEEL FORGED BODY REGULATING   2 EACH
VALVE WITH 1/4" TUBE FITTINGS
(EXAMPLE - WHITEY VALVE #  SS-1RS4)

STAINLESS STEEL AND TEFLON GEAR  PUMP     2 EACH
WITH MAGNETIC COUPLING AND INTERNAL
BY-PASS  (EXAMPLE - MICROPUMP  # 12-50-316)
* MENTION OF COMMERCIAL  PRODUCTS DOES NOT CONSTITUTE
  ENDORSEMENT BY USEPA.

-------
                           C-10
                    PARTS LIST  (CONT.)

ITEM NUMBER              DESCRIPTION*                    QUANTITY

                STAINLESS STEEL MALE CONNECTOR           4 EACH
                1/4" TUBE x 1/8" PIPE THREAD FOR
                USE WITH MICROPUMPS  (EXAMPLE -
                SWAGELOK FITTING # SS-400-1-2)

                STAINLESS STEEL 1/4" UNION  TEE           8 EACH
                (EXAMPLE - SWAGELOK FITTING
                # SS-400-3)

                STAINLESS STEEL TUBE REDUCER             3 EACH
                1/8" x 1/4" WITH 1/8" TEFLON
                FERRULES (EXAMPLE - SWAGELOK
                FITTING # SS-200-R-4 WITH
                TEFLON FERRULES) TO BE  USED
                WITH THE SHUT-OFF VALVES ON THE
                SURFACE SCOUR LINE.

                1/8" STAINLESS STEEL TUBING "316"        20 FEET
                FOR SURFACE SCOUR. SHAPED AS
                IN FIGURE 4 AND PLACED  THRU
                ITEM 1 (DETAIL A) AND CONNECTED
                TO THE SURFACE SCOUR VALVE
                USING THE STAINLESS STEEL TUBE
                REDUCER.

                1/4" STAINLESS STEEL TUBING "316"        60 FEET
                FOR ALL CONNECTIONS OTHER THAN
                SURFACE SCOUR.

                1/2" TEFLON TAPE USED FOR SEALING        2 SPOOLS
                ALL PIPE THREAD CONNECTIONS
* MENTION OF COMMERCIAL  PRODUCTS DOES NOT CONSTITUTE
  ENDORSEMENT BY USEPA

-------
                                             -Cll-







Acknowledgments




      The authors wish to acknowledge the  assistance  of Ms. Maura M. Lilly




who  typed Appendix C.
                                                       »US GOVERNMENT PRINTING OFFICE 1978—757-140/6652

-------
-'4

-------