OCR error (C:\Conversion\JobRoot\000003AT\tiff\2000AI7D.tif): Unspecified error

-------

                                   TECHNICAL REPORT DATA
                            (Plccse read tn.fnicnoii* on '/it1 /•< ictse bt'h'nc completing)
?..•
 1. REPORT MO.
   EPA-600/4-82-072
                              2.
                                                           3 RECIPIENT'S ACCtSSIC!»NO.
 4. TITLE AND SUBTITLE
  Analysis  of Chlorinated Organic Compounds  Formed
  during Chlorination of Wastewater Products
                                                               a. REPORT DATE
                                                                December 1982
                                                               6. PERFORMING ORGAN ZATION CODE
 7. AUTHOR^) William H. Glaze, Jimmie L. Burlcson,  James E.
  Henderson IV, Priscilla C. Jones, Warren  Kinstley,  Gary
     Peyton. Richard Rawlev, Farida \. Saleh,  Garmon  Sraiti
 3 PERFORMING ORGANIZATION NAME AND ADDRESS
  Nortn Texas State University
  Denton,  Texas 76203
                                                           8. PLRFORMING ORGANIZATION REPORT NO.
 12. SPONSORING AGENCY NAME AND ADDRESS
  Environmental Research Laboratory—Athens  GA
  Office of Research and Development
  U.S.  Environmental Protection Agency
  Athens,  Georgia 30613
                                                               10. PROGRAM ELEMENT NO.
                                                                  CBNC1A
                                                           II. CONTRACT/GRANT NO.
                                                             Grant R803007 changed  to
                                                             Coop, Agree. R-805822
                                                               13. TYPE OF REPORT AND P
                                                                  Final,  1/74-8/79
                                                                              ERIOD COVERED
                                                               14. SPONSORING AGFNCY CODE
                                                                  EPA/600/01
 15. SUPPLEMENTARY NOTES
 16. ABSTRACT
        Chemical byproducts produced -luring  the chlonnation of municipal wastewater
 were examined in a study that employed  several specially modified analytical methodo-
 logies.   Volatile byproducts were examined by the use of gas chromatography with  se-
 lective  detectors and gas chromatography/mass spectrometry (GC/MS).  Using XAD  resins
 for concentration of trace organics  in  the wastewater samples before and after  chlor-
 ination, a number of chlorinated aromatic  and aliphatic compounds were found after
 chlorination and superchlorination.
        A  rapid and convenient microextraction method was developed that is suitable
 for the  analysis of trihalomethanes  and other volatile halogenated organics at  the
 laicrogram-per-liter level in water.  Also,  a computer program was developed that  may
 be used  in conjunction with a GC/MS  computerized data system for the identification
 of polyhalogenated compounds present as minor components in a complex chemical  mixture
 A  procedure also was developed to determine the concentrations of amino acids in  waste
 waters,  sludges and septage, before  and after chlorination.  Two chlorinated deriva-
 tives of tyrosine were found in a superchlorinated septage sample.  Nonvolatile com-
 pounds in natural waters and municipal  wastewaters, before and after chlorination,
 were studied by high performance liquid chromatography.  Fractions collected before
 chlorination  of the sample showed that trihalomethane formation potential was  spread
 throughout the natural polymer.
                               KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                             b.IDENTIFIERS/OPEN ENDED TERMS
                                                                          COSATI 1-ield/Group
 3 DISTRIBUTION STATEMENT

 RELEASE TO PUBLIC
                                                   9. SECURITY CLASS (TinsReport!
                                                    UNCLASSIFIED
                                                                             21. NO. OF PAGES
                                             20 SECURITY CLASS jlhispage)
                                                UNCLASSIFIED
                                                                        22. PRICE
EPA Form 2220-1 (9-73)

-------
                                                              X

                     s*^^1,^^r^.^*n^'*i!<>^'^^
                           DISCLAIMER
      Although the research described  in  this  report  has  been
funded wholly or in. part by the United States  Environmental
Protection Agency through interagency  agreement  numbers R-803007
and R-805822 to North T'.ixas State University,  it  has  not  been
subjected to the Agency's required peer and policy  review and
therefore does not necessarily reflect the views  of the Agency
and no official endorsement should be  inferred.

-------
                          FOREWORD
      Nearly every phase of environmental protection depends on
a capability to identify and measure specific pollutants in the
environment.  As part of this Laboratory's research on the oc-
currence, movement, transformation, impact and control of environ-
mental contaminants, the Analytical Chemistry Branch analyzes
chemical constituents of water and soil and develops and assesses
new analysis techniques.

      The Federal Water Pollution Control Act of 1970 requires
the disintection of all wastewater effluents.  In most treatment
plants in the United State.^, disinfection is achieved through
chlorination.  Recently, concern has been expressed concerning
the formation of chemical by-products when chlorine is used as
a disinfectant or biocide.  In a five-year study, separation and
identification methods were developed for volatile and nonvola-
tile byproducts of chlorination of natural waters and waste-
waters and a number of chlorinated comoounds were characterized.
                              David W. Duttweiler
                              Director
                              Environmental Research Laboratory
                              Athens, Georgia
                             111

-------
                            ABSTRACT

      Chemical by-products produced during the chlorination cf
municipal wastewater were examined in a study that employed
several  specially modified analytical methodologies.  Volatile
by-products were examined by gas chromatography with selective
detectors and gas chromatography/mass spectrometry  (GC/MS)-
Using XAD resins for concentration of trace organics in the
wastewater samples before and after chlorination, a number of
chlorinated aromatic and aliphatic compounds were found after '
chlorination and superchlorination.

      A  rapid and convenient microextraction method was developed
that is  suitable for the analysis of trihalomethanes and other
volatile halogenated organics at the microgram-per-liter level
in water.  Also, a computer program was developed that may be
used in  conjunction with a GC/MS computerized data system for
the identification of polyhalogenated compounds present as minor
components in a complex chemical mixture.  A procedure also was
developed for determining the concentrations of amino acids in
wastewaters, sludges and septage, before and after chlorination.
Two chlorinated derivatives of tyrosine were found in a super-
chlorinated septage sample.

      Non-volatile compounds -in natural waters and municipal
wastewaters, before and after chlorinaticn, were studied bv high
performance liquid chromatography.  Fractions collected before
chlorination of the sample showed that trihalomethane formation
potential was spread throughout the natural polymer.  After
chlorination, "total" organic halogen of a non-volatile nature
was determined by adsorption of the organics on either XAD resin<-,
or powdered activated carbon (PAC) followed by elution of the
resin and combustion of the eluate or by direct combustion of the
PAC.  In both cases, it was four.d that organic halogen was spread
throughout the natural polymer, although chlorination at -the levels
used (20-30 mg/L)  did not much aEfect the average molecular weight
of the polymer.

      This report was submitted in fulfillment of Grant No. R-803007
and Cooperative Agreement No. R-805822 by North Texas State Uni-
versity under the sponsorship of the U.S.  Environmental Protection
Agency.   The report covers the period 1 January 1974 to 1 August
1979,  and work was completed as of January 1980.
                               IV

-------
               •
                           CONTENTS
Foreword	iii
Abstract	iv
Figures	vii
Tables	xii
Abbreviations and Symbols  ..... 	  xv
Acknowledgment 	 xvi

   1.  Introduction  	   1
   2.  Conclusions 	   3
   3.  Recommendations	   6
   4.  Background  ,	   7
          Chemistry of Chlorine  	   7
          Chemical Reactions of Chlorine 	  11
          Soil and 7\quatic Humic Substances	15
          Separation and Identification of Specific Organic
             Compounds in Wastewaters Before and After
             Chlorination  	  25
          Toxicity of Chlorinated Wastewater Effluents ...  26
          Surrogate Methods Versus Specific Compound
             Identification  	  28
   5.  Computer Assisted GC/MS Analysis of Or']onic Compounds
       in Municipal Wastewater Products Before and After
       Chlorination  	  30
          Introduction	•	30
          Experimental 	  35
          ResultJ and Discussion 	  44
   6.  Chlorination of Amino Acids in Municipal Waste
       Products	60
          Introduction 	  60
          Summary of  Previous Work	Gl
          Experimental Procedures  	  70
          Results and Discussion 	  76
   7.  The Analysis of Non-Volatile Organic Compounds in
       Wacor and Wastewater after Chlorination 	  92
          Introduction 	  92
          Background	92
          Experimental Procedures  	  93
          Results and Discussion 	 100
          Conclusion	121
   8.  Analysis of Volatile Chlorinated Organics in Water
       by Liquid-Liquid Extraction 	 122
          Introduction 	 122

-------
          Experimental 	   125
          Results and Discussion 	   130

References   	   144

Appendices
   A.  Limited Cluster Search Mass Spectroscopy  	   156
   B.  Program:  Limited Cluster Search  (LCS)  	   178
                            VI

-------
BT*
                                      FIGURES

           Number                                                       Paqe
             1    Haloform mechanism  for methyl  carbonyl  compounds
                     after Morris  and Baum  (34)	14

             2    Structural units identified by Morris and  Baum
                      (34) as THM precursors	14

             3    Separation scheme for soil and aquatic  humic
                     substances  (42)	17 ^

             4    Structure of fulvic acid as proposed by
                     Schnitzer  (42)	22

             5    Structure of humic  acid as proposed by  Christman
                     and Ghassemi  (49)	23

             6    Rook-Moye mechanism for the aqueous -chlorination
                     of resorcincl (51)	24  '

             7    Total ion  (top)  and LMS chromatograms at m/e  149
                      (bottom) for  XAD.extracts of  superchlorinated
                     septage,         	32

             8    Total ion  (top)  and LMS chromatograms at m/e  35
                      (bottom) for  XAD extracts of  superchlorinated  .
                     septage-.         	33

             9    Schematic of the Denton, Texas municipal wastewater
                     treatment plant  	   36

            10    Scheme for the extraction of organic compounds .
                     from wastewaters	39

            11    Glass apparatus  containing XAD-2 resin  for
                     extraction of organic compounds from water.  ...   41

            12    Gas chromatograms  (FID,16xlO~l-1;amp/fs)  of  Denton,
                     Texas wastewater extract.   Bottom, before
                     chlorination;  top, after 2,000 mg/L
                     chlorination  for 1-hr, contact period.
                     Analytical conditions described in text .....   46
                                          Vll

-------
.Number      ';                  :                            Page
•             !                   ;
j   13   Gas chromatograms (Coulson electrolytic
•           conductivity detector,  halogen mode, X 4)
.           on Denton,  Texas wastewater extract.  Bottom
           before;  top, after 2000 mg/L chlorination  for
           1-hr,  contact time.   Analytical conditions
1  .........  described in text ....... .  ..... ...   "47"
   14   Reconstructed GC/MS chromatogram of Denton,  Texas
i           wastewater extract after 2000 ppm chlorine
:           treatment for one hour.   Vertical line markers
           show the positions of new chlorinated organics
           not present in control samples.   Peak numbers
i           correspond to compounds  identified in Table 7. .    48

   15   Reconstructed GC/MS chromatogram of superchlo-
•           rinated septage extract  (reference Table  8
;           and text) ....................    51

\   16   Reconstructed GC/MS chromatogram of purgeable
           organic compounds in superchlorinated septage. .    53,

   17   Reconstructed GC/MS chromatogram of purgeable
           reference compounds ...............    54

   18   Total ion chromatogram (top)  and LCS chromatograms
           for two chlorines (middle)  and three chlorines
           (bottom)  for CALSTI data set ..........    56

',   19   Scheme for the separation of amino  acids in
           wastawater products ...............    73

   20   Fragmentation of aromatic amino acids (N(O)-
           heptaf luorobutyryl propyl esters) .  All species
           have +1 formal charge ..............    78

   21   Fragmentation of aliphatic  amino acids (N(O)-hepta-
           f luorobutyryl, n-propyl  esters).  All species
           have +1 for.nal charge  .............    7r

i   22   Reconstructed GC/MS chromatogram of amino acid
           standard.   a-alanine;  b-glycine; c-valine;
           d-threonine; e-serine; f-leucine; g-isoleucine;
           h-proline; i-cysteine; j-methionine; k-aspartic
           acid;  1-uriknown impurity;  m-phenylalanine; n,o-
           glutamic acid, ornithine;  p-lysine; q-tyrosine;
           r-arginine; s-hJ stidine ; t-tryptophane;
           u-phthalate ......... ' .........    80
   23   Mass spectrum of chlorotyrosine ..........    8^

-  24   Mass spectrum of dichlorotyrosine ...... .  . .    85

                               viii

-------
Number                                                     Page

  25   Reconstructed GC/MS chromatogram of amino acid
          extract from superchlorinated septage extract.
          a-alanine; b-glycine; c-valine; d-threonine;
          f-serine; g-leucine; h-isoleucine;
          i-phenylalanine; j-glutamic acid; k-lysine;
          1-tyrosine; m-chlorptyrosine;
          n-dichlorotyrosine  	    88

  26   Calibration curve for Partisil 10/glycophase
          size exclusion colunns.  A Proteins
          (ovalbumin, M = 45,000; jhymotrypsinogen A,
          M = 25,000). D Sodium polystyrene
          sulphonates (M = 16,000; 6,500; 4,000,
          1,600).  O Methanol.	    96

  27   Scheme for the study of the effect of
          chlorination on non-purgeable organics 	    97

  28   Size exclusion chromatograms of Cross Lake
          sample, freeze dried water soluble fraction.
          Top, unchlorinated; bottom, chlorinated
          at 20 mg/L for five days	101

  29   Size exclusion chromatograms of water soluble
          fractions  (unchlorinated).  Right: reinjected
          fractions; left: superposition of
          chroTtiatograms of reinjected fractions
          over original trace	   104

  30   Trihalomethane formation potential  (TTHMFP)and
          total organic halogen  (TOX/GAC) formation
          potentials for Cross Lake water; 20 mg/L
          dose for five days	   105

  31   VJeak anion exchange HPLC chromatogram of Cross
          Lake water, acid soluble fraction of freeze
          dried sample (unchlorinated).  Reference
          compound code: a-phenol, b-3-methylcatechol;
          c-vanillic acid; d-2,4-dihydroxybenzoic acid;
          e--2, 4 ,6-trihydroxybenzoic acid.  Dotted line:
          solvent gradient (100% = pH 6.2; 0% = pH 3.2).  .   112

  32   Size exclusion chromatoqrams of water soluble
          fraction of freeze dried Denton municipal
          wastewater. Top: unchlorinated; bottom,
          chlorinated	   117"

-------
Number
                                                    Paqe
  33
  34
  35
 A-4

 A-5



 A-6


 A-7
Size exclusion chromatoqrams of acid  (TOP) and
   base  (BOTTOM) soluble fractions of freeze dried
   Denton municipal wastewater- ... 	
Sample bottle used for collection of water
   samples for analysis of purgeable volatile
   organic compound3	.-	
Scheme for analysis of volatile organics by
   liquid-liquid extraction 	
       Procedure for removal of water and addition of
          pentane for volatile purgeable organic
          analysis 	

       Modified purge and trap apparatus with liquid
          sample loop injector 	
  38   Electron capture ras chrcmatogram of VCOs from
          pentane LLE extraction (conditions in text).

  39   Coulson electrolytic conductivity gas
          chromatogram of VCOs from modified purge/
 118



 126


 128



 129


 131


 132
40
A-l
A- 2
A- 3
Electron capture gas chromatogram of VCOs by
direct aqueous injection (conditions in text) . .
Flowchart of limited cluster search program ....
Flowchart of GC/MS data 	
Relationship between Z-value and percent deviation
j. ~> j
140
157
161

   of the M+2 peak in a dichlorinated isctopic
   cluster	

Relationship between cumulative fit and Z-value

Mass spectra (no background subtracted)  f
   dichlorobenzene and trichlorobenzene
   (HALSTI data set)	
163

164
TIC and limited cluster search chromatograms for
   mixture of halogenated compounds 	
Mass spectra of dichlorobenzene and bromoform
   (no background subtracted) 	
170
                                                            171

-------
Number       .                                              Pac
 A-8   Weighted fit vs_. precision estimate for three
          values of the variation estimate parameter.
           f» dichlorobenzene;   A*  bromoform .......  173

 A-9   Limited cluster search chromatoarams for HALSTI
             data set with various precision and
             variation estimate parameters 	  174

 A-10  Mass spectrum of hexachlorobutadiene (no
             background subtracted)	176
                              XI

-------
                          '  TABLES


Number                                                      Page
   1   Properties of Components of Humic Substances . . .  .  18

   2   Total Organic Carbon Ranges for Several Water
          Types  (44)	19

   3   Elemental Analyses of Aquatic Humic Fractions  (41)  .  20

   4   Chlorinated Products Identified in Chlorinated
          Water by Jolley (gl>	27

   5   Chlorinated Products Produced by Water Chlorinaticn,
          by Glaze, et. al. ,  (36,60)	27

   6   Names and Sources of Compounds in Halsti Sample. .  .  44

   7   Summary of New Chlorinated Organics Found in
          "Superchlorinated" Municipal Wastewater
          (Ref. Fig. 14)	  49

   8   Compounds Identified in Ventura, California,
          Superchlorinated Septage Supernatant, XAD-
          Diethyl Ether Extract (Ref. Fig. 15)	  52

   9   Total Amino Acids Found in the Sewage Samples from
          Four United States Cities (122)	  66

  10   Amino Acid Content of the Soluble Fraction in
          Untreated Domestic Wastewater  (125) 	  67

  11   Minimum Detectable Limit of N(O)-Heptafluorobutyryl
          Alkyl Esters of Amino Acids3	'.  .  81

  12   Reaction Products Identified from the Reaction of
          Amino Acids with HOC1	'	  83

  13   Recovery Efficiencies of 20 Amino Acids by the
          Combination of the Cation and Ligand Exchange
          Resins	  86

-------
Number                                                      Page

  14   Comparison of GC/MS Analysis of Derivatized Amino
          Acids with Analysis by a Beckmann Amino Acid
          Analyzer using a Wastewater Matrix  	 -89

  15   Amino Acids Present in Municipal Wastes, ug/1. ...  90

  16   Characteristics of Partisil 10/Glycophase HPLC
          Columns	94

  17   General Chemical Characteristics of the Secondary
          Treated Wastewater	98

  18   Characteristics of Cross Lake Water (0.45 y Filtrate).  99

  19   Residual Chlorine at Different Treatment Levels and
          Contact Time	99

  20   Characteristics of the Water Soluble Fractions of
          Cross Lake Water Collected by Size Exclusion
          HPLC (Unchlorinated)	102

  21   Characteristics of the Water Soluble Fractions of
          Cross Lake Water Collected by Size Exclusion
          HPLC (Chlorinated)	107

  22   Characteristics of Acid Soluble Fractions of
          Unchlorinated Cross Lake Water Separated by
          Size Exclusion HPLC	108

  23   Characteristics of Acid Soluble Fractions of
          Chlorinated Cross Lake Water Separated by
          Size Exclusion HPLC	109

  24   THMs Formed by Chlorination of Acid Soluble
          Fractions of Cross Lake Water Separated by
          Size Exclusion HPLC	110

  25   Ka and Retention Volumes for Standard Compounds.  .  . 113

  26   Trihalomethane Formation Potential Data	114

  27   Jon-Purgeable TOX Formation Potential  of a Secondary 115
          Municipal  Wastewater Effluent 	
  28   Molecular Size Distribution for the Freeze Dried
          Unchlorinated Sample:  Denton Secondary
          Wastewater	119

  29   Response Factors for VCO's Using Different
          Analytical Techniques 	 138
                              xna

-------
Number                                                     Page
  30   Precision of Analytical Methods for Analysis of
          VCOs at S/N 20-200  (Per cent Relative Standard
          Deviation)	138

  31   Minimum Detectable Limits for the Analysis of
          VCOs by LLE,  Bellar D/T and DAI Methods
          (S/N =2)	139

  32   Extraction Efficiencies of VCOs by the
          Pentane LLE Method	139

  33   Analysis of VCOs in Denton, Texas Tapwate^	142

  34   Time to Complete Multiple VCO Analyses by LLE
          and Modified Bellar Methods -(HRS)	142

 A-l   LCS Dialogue	159

 A-2   Cumulative Fit Values  for HALSTI Data 3et
          (cf. Fig. A-10) Hexachlorobutadiene	177
                              xiv

-------
                                                             V* **^w5*"V >*-*• ' «•v
                   ABBREVIATIONS AND-SYMBOLS
CALCLI

CECD
DAI
EC
GAC
GC
GC/MS

HA.LSTI

HPLC
LLE
LCS
S
THMs
TOC
TOC1
TOX
VCO
X
XAD
-data set used to evaluate LCS program (California
 waste after superchlorinat.ion)
-Coulson electrolytic conductivity detector
-direct aqueous injection
-electron capture detector for gas chromatography
-granular activated carbon
-gas chromatography
-gas chromatography with mass spectroscopic
 detection
-data set used to evaluate LCS program (synthetic
 mixture of organic compounds)
-high pressure liquid chromatography
-liquid-liquid extraction
-limited cluster search
-standard deviation (n based)
-trihalomethanes; CHC13, CHCl^Br, CHClBr2,  CHBr3
-total organic carbon
-total organic chlorine
-total organic halogen
-volatile chlorinated organic compounds
-experimental value
-trade name for Rohm and Haas macroreticular
 resins for adsorption of organic compounds from
 water
                              xv

-------
                        ACKNOWLEDGEMENT
     The special assistance of Dr. A. W. Garrison to this
project is gratefully acknowledged.  In addition, the assistance
of the staff of the North Texas State University Institute of
Applied Sciences and The University of Texas at Dallas was
important to the completion of this report.  Special thanks go
to Sharon Dumas and Julie Kerestine for typing and graphics
work (NTSU) and to Shirley Price for typing (UTD).

     The sustaining support of the NTSU Faculty Research
Committee is acknowledged with gratitude.
                              xvi

-------
                                       SECTION  3.

 ;                                    INTRODUCTION
 1

 ;                This  report describes  a  series of  research efforts  whose
 !           general  theme was  to  study  the  chenical by-products  formed during
 ;           the  chlorination of municipal wastewater  effluents.   At the
            initiation of the  research  in 1974  little  was  known  of this
            subject; since  then many  research groups worldwide have
            investigated various  aspects of thg question,  or more  broadly,
            the  question of by-products formed  during  the  chlorination of
            all  types  of waters.   It  is now well established that  by-
 1           products are produced whenever  chlorine is used as a disinfec-
            tant or  bio-cide.  Among  these  are  the  trihalomethanes,  now the
            subject  of regulations which  limit  their concentration in
            drinking water.

                 Th"> resear-h  which is  reported here had several objectives,
            some of which evolved over  the  project  period.   They were as
            follows:

              1.  to  develop  separation  and identification methods  for
            the  determination  of  the  types  and  quantities  of volatile by-
            products prcduced  by  the  chlorination of water,  particularly
            municipal  Wcist.2water  after  secondary treatment.   Central to this
 i           effort was the  evaluation of XAD resins for concentration of
            trace organics  in  water,  and the use of gas chromatography
            with selective  detectors, and gas chromatography/mass  spectro-
            metry (GC/MS) for  the elucidation of the structures  of these
            substances.  In the course  of the study a  rapid and  convenient
            microextraction method was  developed which is  suitable for the
            analysis of trihalomethanes and other volatile  chlorinated
            organics at the pg/L  level  in water.  Also, a  computer program
            was  developed which may be  used jn  conjunction  with  a  GC/MS
,           computerized data  system  for the identification of polyhalogen-
I           ated compounds  which  may  be present as  minor components  in a
j           complex chemical mixture.

              2.  to  develop  separation methods based on  high performance
            liquid chromatography (HPLC) for the study of wastewater
i           effluents  and natural waters before and after  chlorination.
|           The  purpose of  these  studies was to extend our  knowledge
|           concerning the  non-volatile compounds in water,  and  in particu-
            lar,  to determine  whether chlorinated non-volatile halogen-
            ated products are  produced  by chlorination of water.

-------
   3.  to investigate the chlorinated by-products produced
when very large doses of chlorine (1000-3000 mg/L) are used
for the treatment of wastewaters/ sludges and septage.  The
use of such high doses has been proposed as a method for dis-
infection and stabilization of septage and sludge and as a
possible alternative wastewater treatment scheme for small
systems.
                              2

-------
                                       SECTION 2

                                      CONCLUSIONS


                 The conclusions derived from this study are as  follows:

                 1)   Chlorination of natural waters and municipal waste-
                     waters causes the formation of many new halogenated
                     organic compounds.

                 2)   These new halogenated organic compounds may be
                     classified as purgeable-volatile;  non-purgeable-
                     vclatile; and non-volatile.  Based on the content
                     of organic-bound halogen, the yields of purgeable-
                     volatile and non-volatile organic  halides are larger
                     than those in the non-purgeable volatile category.

                 3)   Purgeable-volatile by-products usually are  dominated
                     by the chlorine-,  bromine- and iodine-containing
                     trihalornethanes.   The yield of these compounds  is
                     approximately 2-5 mole per cent based on the carbon.
                     content of the original water, providing no ammonia
                     nitrogen is present.   When ammonia nitrogen is  present,
                     yields of trihalomethanes are correspondingly lower.

                 4)   Iodine-containing compounds are found in very small
                     quantities relative to the bromine and chlorine
                     containing compounds.

                 5)   The yield of non-volatile organic  halides  (as measured
                     by GAG adsorption/pyrolysis/microcoulometric procedure)
                     is usually 2-5 tines  the yield of  trihalomethanes.

                 6)   The nature of the non-volatile organic compounds in
                     water and wastewater,  either the halogenated by-
                     products or their precursors,  is not known.   However,
                     size exclusion HPLC studies have shown that the
                     molecular size of the natural organic matrix is changed
                     only slightly upon  chlorination.

                 7)   Fractionation of  the  natural aquatic organic polymer by
                     size exclusion HPLC has shown that trihalomethane
                     formation potential is spread through the molecular
                     size range of the polymer.
L-.

-------
 8)  Fractionation of the polymer after chlorinatior. hs.3
     shown that the non-volatile organic-bound halogen in
     the polymer is also spread throughout the polymer.

 9)  Fractionation of the natural polymer with a weak anion
     exchange resin with pH gradient elution proc^ices
     fractions in three separate pH regions.  Mo<3t>l compound
     studies suggest that one of these fractions I as a pKa
     value similar to that of phenols, and another the pK
     value of phenyl-carboxylic acids.  The nature; of the
     third fraction, which occurs to various extents in
     waters from different sources, is unknown.

10)  Non-purgeable volatile organic halides are increased
     in yield and in number by the use of high concentrations
     of chlorine (2000 - 4000 mg/L).

11)  The structures of the non-purgeable volatile organic
     halides isolated from "superchlorinated" municipal
     wastewater and other domestic waste products, suggests
     that they are derived from oxidative degradation of
     humic- or fulvic-acid-like precursors.

12)  A special computer program has been developed for the
     analysis of GC/MS data.   The program searches for
     isotope clusters in the mass spectra of GC fractions,
     thus assisting in the detection of new halogenated
     compounds.

13)  Application of vhe new GC/MS computer program to a
     superchlorinated septage sample confirmed the presence
     of a larg'2 number of new chlorinated compounds as
     contrasted to the unchlorinated control sample.

14)  A combined ion exchange/ligand exchange procedure has
     been used for the isolation of free amino acids from
     wastewater matrices.  After isolation the amino acids
     are derivatized and quantified by combined gas
     chromatography/mass spectroscopy.

15)  Analysis of a secondary treated municipal wastewater
     sample for amino acids has revealed only low levals of
     free amino acids (<5-10  yg/L).

16)  Analysis of municipal sludge and septage supernatants
     revealed substantial amounts of free amino acids which
     upon chlorination are converted in some cases to
     chlorinaced by-products.

-------
I
t
(                 17)   In one case, tyrosine, the amino acid is found with
|\                     ring-chlorination; a mono- and di-chlorotyrosine were
i'      .               confirmed by GC/MS analysis.
f
\                 18)   Further study or the non-volatile components cf
!.                      natural waters and wastewaters is required to under-
\                      stand the mechanics of the chlorination process.  In
;:                      particular, spectroscopic methods, and new HPLC
'-                      column modes and detectors should be applied to this
                      study.

-------
IIS
• :'   •      ~  •„	           .     SECTION 3

                         i            RECOMMENDATIONS


; ;          :      Further  study of the non-volatile organic constituents in
' |           natural surface and ground water sources,  and in treated
 I          • waters and wastewaters is recommended.  Particular attention
f I           should be given to the development of chromatographic and other
 |           separation processes; to the invention of  new chromatographic
• '           detectors with element specificity or spectroscopic capabilities;
. [          , to studies which elucidate the mechanism of  the formation of
 ;           chlorinated by-products during water chlorination; and to the
; ',           investigation of the fate and effects of non-volatile organic
 |           compounds before and after chlorination in the aquatic
: *           environment and upon consumption by man.
 I I
 ; i

-------
                                       SECTION 4

                                      BACKGROUND


               Chlorine and hypochlorite solutions have been used for the
           disinfection of wastewaters since the middle of the ninete;enth
           century  (1).  In 1384, Brewster, New York  became the first
           American city to use chlorine as a wastewater disinfectant to
           protect against potential contamination of the Nev York City
           drinking water supply.  By 1945, Enslow aad Symans  (2) had
           defined properly effective wastewater treatment as a three-
           stage process including terminal chlorinaticn for disinfection.
           Active interest in terminal wastewater chlorination grew rapidly
           across the United States.  This interest culminated in Amendments
           to the Federal Water Pollution Control Act  (1970)  (3) requiring
           all wastewater effluents to be disinfected.  Although chlorina-
           tJon was not specified in the Act as the only acceptable disin-
           fection technique, it is the one used almost exclusively in the
           United States.

               Terminal chlorination of wastewa^er effluencs is emphasized
           to achieve the following objectives  (4):

               1.  prevention of the spread of disease,

               2.  protection of potable water supplies, bathing beaches,
                   receiving waters used for boating and water contact
                   sports, and

               3.  protection of shellfish and other aquatic life forms.

           The minimum effective chlorine dosage necessary to achieve the
I           above objectives was evaluated by the California State Depart-
           ment of Health, Bureau of Sanitary Engineering and later by
           the Environmental Protection Agency  (5).  The studies were
           based on fecal coliforms contained in residual effluents.  The
           Environmental Protection Agency's temporary commitment of 200
           to 400 bacteria per 100 ml of effluent depending on the nature
           of the receiving waters is the national guideline, although
           some states have set more stringent requirements  (6).

           CHEMISTRY OF CHLORINE

               When molecular chlorine is dissolved in water, it is
           hydrolyzed  (7) according to the equation:

-------
                      C12 + H20 --- S**-HOC1 -f- H+ + Cl~

            This reaction is 99 per cent complete in only seconds, and thus
            the aqueous chemistry of dilute chlorine solutions is in fact
            the chemistry of hypochlorous acid.  Indeed, chlorination
            reactions can be effected with equal efficiency by the use of
            solutions of hypochlorite salts.

                Hypochlorous acid is a weak acid which dissociates accord-
            ing to the equation:

                            HOC1

            The dissociation constant for HOC1 is 2.95 x 10~8at 18°C (8) ;
            thus, at a pH of 7.5 there are approximately equimolar amounts
            of HOC1 and OC£ ~ present in aqueous solution.

                Kinetic studies have shown aqueous chlorination reactions
            to be extremely complicated.  Reviews have been published by
            Jolley (9) , Morris (10) , Carlson and Caple  (11) , and Pierce
            (12).  For many reactions the pH dependence of reaction rates
            suggests that hypochlorous acid HOC1 is the reactive species or
            is involved directly in the generation of the principal
            reactive species.  Hypochlorous acid is reported  (12) to be
            104 times more reactive than the hypochlorite ion CO .  However,
            various authors have attributed the reactive species to be the
            hypochloronium ion H2OC1"1", the free chloronium ion Cl+ (14) , and
            the chlorine radical Cl- (11) .   Carlson and Caple  (11) state that
            chlorine containing organic products will be derived from the
            attack of electropuilic species such as K2OC1+ or Cl+ or by a
            free radical process.   The former process will generate products
            by aromatic substitution or addition reactions, while the less-
            likely radical process may occur with reactants which will give
            a stable radical intermediate.   De la Mare et. al. (15) studied
            aromatic halogen substitution using low concentrations" of HOC1
            and added perchloric acid and silver perchlorate.  They con-
            cluded that the measured rate was determined by the generation
            of Cl+ according to the following sequence:
                               HOCl  -- 55*- Cl  + H0

                               HOC1 + H+ --- 2*«- H2OC1+

                                    +
                               H2OC1  --- S*»» Cl  + H20

                               Cl+ + ArH --- ^»- ArCl + H+
            However,  in the absence of low pH and silver ion, the system
            became much more complicated.  Hypochlorite and chloride ions
fe_ L i

-------
which are present in solution could react  to  form new chlori-
nating species C12 and C12O:
H2OC1+

H2CC1+
OC1
 C1OC1
                                            H20
These products have been shown to be potent chlorinating  species
 (16).  Furthermore, at concentrations above about  0.001 M HOC1,
the kinetic form showed partial dependence on the  square  of  the
HOC1 concentration indicating other reactions produce even more
powerful chlorinating species.  Hine  (14) has discussed evidence
that the hypochloroniuw ion H20C1+ is less reactive than  the
chloronium ion Cl"1", but may still be an important  species in the
chlorination of very reactive substrates such as anisole  and
phenol.  Still, it is safe to say that at this point few
definitive mechanistic studies have been 'reported  and much
uncertainty remains regarding the mechanism of aqueous chlorin-
ation processes.

    At greater concentrations of chlorine, e.g. above one
gram per liter  (lOOOppm) and at low pH, it is possible that
molecular chlorine may represent an important kinetically
active species.  Such solutions have a characteristic yeQlow
color which may be attributed to molecular chlorine (17) .

    Other important reactive species in the chlorination  of
natural water and wastewater are chloroamines (18) ; the products
of reactions o^ ammonia (and its derivatives) with HOC1:
              NH3+HOC1-

              NH2C1+HOC1-

              NHC12+HOC1-
NH2Cl-f-HOH

NHC12+HOH

NCl-j+HOH
    Studies of the reactive nature of chloroamines have been
extensive.  This work was reviewed in detail by Kovacic  (18)
and Jolley (19).   Jolley calculated the concentrations of
active chlorinating species based on equilibrium constants for
the appropriate reactions.  For example, a.system at a pH of
7.5 containing C12 and Cl~ at concentrations of 1 and 10 mg/1 as
Cl equivalent and containing ammonia at 1 mg/1 resulted in
the following concentrations of the chlorinating species:

-------
r
                        Species                  Concentration,  mg/1 (as Cl
                        	                 .	equivalent)	

                        HOC1+OC1"               .          0.0004

                        NH2C1                             0.9729

                        NHC12              '       '        0.0264
                        NC13                              trace

                 ,  However,  it  is important to  note  that the  specific reactivity
               of monochloramine is  approximately 104  less than  that of HOC1
               (20)  thus offsetting  much  of its concentration advantages.
               Also, monorhloramine  is  an effective  aminating reagent (21) ,
               which may effectively compete with chlorination reactions.
               The  following (22) may also be a competitive reaction:

                        NH2C1  + NHC12  +  HOC1	J**- N2O +  4H+ + 4C1~

               This type of  reaction occurs when the free available  chlorine,
               HOC1 and OC1~, approaches  a value of  approximately eight times
               by weight that of the available  nitrogen.   The mechanistic
               chemistry for this reaction is unclear.   One should recognize
               that, while chloramine formation reactions produce new potential
               chlorinating  reagents, break point chlorination results in a
               reduction in  the total amount of chlorinating  species.  Break
               point chlorination  (23)  is defined as the point at v.hich the
               amount  of chlorine added is equal to  the stoichiometric quantity
               required for  complete conversion of t'iiiinonia to nitrogen accord-
               n.g  to  the following  equation:

                         3C12  + 3H2O + 2NH3 	~SSSi»^2 +  6HC1 +  3H2°

               In practice break point  chlorination  refers to the introduction
               of chlorine until free available chlorine is observed by some
               analytical method, indicating the complete conversion of ammonia
               to chloroamines  or other forms.   Thus,  break point chlorination
               results in a  reduction of  the "total  available chlorine"
               whereas chloramine formation simply produces a shift  in the ratio
               of "free available chlorine" versus "combined  available chlorine".
               In systems containing ammonia and chlorine the following defini-
               tions apply to active chlorine species  present:

                         Free  available  chlorine:   hypochl.rous acid
                         in its various  forms  including hypochlorite
                         ion and chlorine,  if  present.

                         Combined available chlorine:   chloramines  in
                         all their  forms which will oxidize  iodide  ion
                         to iodine.
                                             10
 L,

-------
r
I   !
              CHEMICAL REACTIONS OF CHLORINE

                  In discussing the chemical reactions of chlorine, the primary
              reactive species considered here is hypochlorous acid, HOC1.
              The reactions of this species have been reviewed recently by
              Jolley (24)  who classified them into three categories:

                    1.  • Oxidation
                    2.   Addition
                    3.   Substitution

                        a.   Formation of N-C1 compounds
                        b.   Formation of C-C1 compounds
                        c.   Haloform Reaction

              Oxidation Reactions

                  Some of  the most important oxidation reactions occur with
              other inorganic species.  This class includes the following
              reactions:

                        HOC1 + S03~ - ^s» 304  -i- HC1

                        H20 •*- 4HOC1 + S203~ - J>^2SO4  + 4HC1 -<- 2H+


              Sulfite (25)  and thiosulfate  (26)  react instantaneously and
              quantitatively with all chlorinating species and thus are often
              used as "quenching" reagents in chlorination studies.  Nitrite
              (27)  reacts with aqueous chlorine  to form nitrate according to
              the following equation:

                        N02~  + HOC1 - ^^Cl~  + NO3" + H+

              Other species which serve as reducing agents for hypochlorous
              acid include  Fe++,  Mn++, and H202  (27) ,  and organic compounds.
              The chemistry of these reactions is not straightforward and has
              not been  described in detail.  These oxidation reactions
              probably  constitute the largest category in terms of total
              chlorine  consumption.  As Jolley indicated, about 99 per cent
              of the reacted chlorine ends up as the  reduced inorganic
              chloride  (19).

              Addition  Reactions

                  Hypochlorous acid may add to olefinic double bonds to
              yield chlorohydrins as shown in the following equation:
                                                 H   H
                                   HOC" 1           '   '
                        R1CH=CHR2       ag»- RI - C -  C - R2
                                                HO   Cl
                                            11

-------
Carlson and Caple  (11) have  shown that aqueous  chlorination  of
oleic acid produces a mixture of 9-chloro-lO-hydroxystesric
acid  (III) and 10-chlcro-9-hydroxystearic acid  (IV).

           •Ill:     CH3-(CH2)7 - CH - CH -  (CH2)7  - COOH
                                 OH   Cl

            IV:     CHo-(CH?)7 - CH - CH -  (CHo)-7  - COOH
                    CH3-(CH2)7 - CH - CH -  (CH2)7

                                 Cl   OH
Other authors (12) have considered this type of reaction too
slow to be important in dilute aqueous solution, but the data
of Carlson and Caple do not confirm their expectations.
However, it should be noted that few examples of addition
products have been observed in "real world" surveys of chlorina-
tion products from wastewaters or municipal drinking waters.

Substitution Reactions

    Reactions of HOC1 with ammonia have been discussed previously.
HOC1 also reacts with organic amines to displace a proton and
form the corresponding N-C1 bond.  The reaction rate depends, in
general, on the nucleophxlicity of the nitrogenous substrate
(20) .   Reactions of HOC1 with amides (13)  usually require more
vigorous conditions than those available under normal wastewater
chlorination.  Most N-C1 bonds are relatively unstable in
aqueous media.   For example, dichloroam-'.ne and trichloroamines
are reported to decompose to nitrogen and hvpochlorous acid
(23):

            2NHC1  -1  H0 - &RB.  N  + HOC1 + 3H+ + 3C1~
                 2 -   2 -  RB.   2


and dichloroamino acids decompose to nitriles and/or aldehydes
depending on the ratio of amino acid to chlorine, the pH,
and other factors (28, 29, 30):
            R - CH-COOH - ^B»» RCN + RCHO

                NC12

    The second, and most important, group of substitution
reactions are characterized by displacement of a proton in a
carbon-hydrogen bond to form a carbon-chlorine bond.  These
reactions require activation of the leaving proton before the
reaction will proceed under typical wastewater chlorination
conditions.  Substrates such as activated aromatic systems or
alpha, alpha1 -diketomethylene groups are required for success-
ful reaction.  Soper (31)  was one of the earliest to study
such a mechanism, describing phenolic substitution by chlorine
in 1926.  Aromatic substituents such as hydroxy, alkoxy, and

                              12

-------
amino which are strongly electron donating activate  the  ring  to
chlorination, while electron withdrawing groups  such as  nitro,
carbonyl, cyano, and positively charged ions retard  the  chlorin-
ation process  (32).

Haloform Reaction

    An especially important group of substitution reactions
known as haloform reactions have been known since 1822  (33).
These are complex reactions which ultimately yield trihalo-
methanes such as chloroform.  Morris  (10), Morris and Baum
(34), and Pierce (12) have reviewed the basic chemistry  of
haloform formation.  Several typeb of substrates are known,
most commonly those which contain the methyl keto group
CH3C = O, and those which can be easily oxidized to  methyl
keto forms; such as secondary alcohols.

    Morris and Baum  (34) have discussed the clasrical reaction
pathway for the haloform reaction shown in the following scheme .
(Figure 1).  This scheme includes several observations known
about the reaction, such as the fact that it is base catalyzed,
and that bromide and iodide may react with hypochlorus acid to
be incorporated into halogenated products.  Morris and Baum give
six functional atom group arrangements  (Figure 2) they feel
would give haloform products, each of which easily forms the
carbanion intermediate.  They speculate that many of these
groupings are in humic substances found in water.  Major
chloroform yields were obtained for twelve model compounds
with the six functional groups, including chlorophyll which
contains the pyrrole group.  Hoehn (35) et al. also  suggested
that the addition of algae or more specifically the  chlor<5phyll-a
accounts for the additional THM's observed during summer months.
It should be noted that rapid electrophilic substitution
produces the initial intermediate I which opens under the
influence of base to form the Carbanion II.  Morris  and Baum
(34) have pointed out that stable carbanion formation is a
prerequisite for haloform production, and that precursors such
as m-dihydroxy aromatic compounds are more reactive  than simple
methyl ketones for this reason.

    The presence of chloroform generated by chlorination of a
wastewater was first reported by Glaze and co-workers (35).
The first extensive treatment of the presence of haloforms
was conducted by Rook (37).  It should be noted that, following
Rook's lead, the term "haloform reaction" is applied to any
series of aqueous chlorination reactions which produces halo-
forms, in this case, CHC13, CI!Cl2Br,  CHClBr2 and CHBr3.   This
not only includes the traditional haloform reaction  characterized
by the chlorination of alpha-keto methyl groups but  also the
polychlorination of aromatic systems, followed by ring rupture
to result in haloform production.  Rook (38)  initially recog-
nized the correlation of the bleaching effect of water

                              13

-------
                   0    I
                   II
                 R-CCH3
                         H20
           >0
            V_u  0
            RC-CH2
             RC=CH2
                         OH©
                              HOX 5?=* H2OX©
                 R-CCH2X
                         HoO
             0  _
            v>»  0         '
            RC-CHX «-> RC=CHX
                         OH©
                 R-CCHX2
                         H20
 HOX ^=^ H2(5x

V° ^
 v." ©
 RC-CX2
 c-e   -
 I
RC=CX2
            RCCH2X
                         RCCHX2
                                     RCCX3
0
                                0
                                            (THM'S)

                                 X=CI,Br,or I

            Figure 1.  Haloform mechanism for methyl carbonyl compounds
                           after Morris and Baum (34; .
           (A)
             CH3-C-
                 ii
              !BWc-ct°
                    0   OR
                          (C)R-C-C?C-R'
                               II    II
                               0   0
           (D)  OH
            H
              (E)
                                          CH
                                          C = 0
                                       c
                                       n
                                         II     II
                             -c
                                                c-
            Figure 2.  Structural units identified by Morris and Baum  (34)
                               as THM precursors.

                                       14
k	-.-

-------
chlorination with the appearance of haloforms.
the coloration of the water was caused by
                             He noted that
              "humic substances which are very stable  to
              biological decay and do not appreciably
              diminish in concentration during impoundment
              (of the water supply).   These substances are
              the products of plant decay and include  macro-
              molecules which are condensation products of
              quinones and polyhydroxybenzenes,  with
              substituent NH2 groups".

His laboratory experiments demonstrated that the chlorination
of purified humic substances dissolved in doubly distilled
water produced haloforms.  Recently,  Glaze et^ al.  (39)  and
Schnoor e_t al. (40)  have shown that fractionated fulvic acids
fron; natural waters  yield trihalomethanes.   In view of the  .
importance of aquatic humic substances in this regard,  the
following sections discuss in more detail the properties of
these substances as  they relate to the formation of haloforms
and other halogenated organic by-products.

SOIL AND AQUATIC HUMIC SUBSTANCES

    Much of the earth's carbon is found in the form of woody
tissue, a major component of which is lignin.   Lignin  is a
mixed polymer which  appears to have only three structural
units, guaiacyl,  syringyl, and p-hydroxyphenylpropane.
        -OCH3
   OH
GUAIACYL
   OH
SYRINGYL
p-HYDROXYPHENYLPROPANE
Apparently, these units are incorporated  into  the  lignin
polymer by carbon-carbon or carbon-oxygen-carbon linkages with
0-4'-ether linkages predominating  (41).
                               15

-------
                               OH

EXAMPLE OF 0--4-ETHER  LINKAGE BETWEEN SYRINGYL AND

p-HYDROXYPHENYLPROPANE  UNITS.

    It has also been proposed that  most  €3  side chains are
methyl ketone, allyl, and secondary alcohol configurations  (42).

    Although lignin is a relative refractory material towards
biodegradation, certain microorganisms,  particularly fungi, are
capable of degrading the lignin polymer.  This process, which is
not fully understood, is a part of  the so-called humification
process which results in the deposition  of  organic substances
called humus in soil and water.  Among the  various hypotheses
regarding the synthesis of humic substances, Martin and Haider (43)
prefer the following.  Lignin molecules  are degraded to smaller
phenolic units which together with  simple phenolic substances
synthesized by microorganisms,  plant and microbial proteins,
carbohydrates, and other substances in the  soil, are combined by
autoxidative and enzymatic polymerization to form humus.  A
portion of this humus is sufficiently water soluble so that it
is eventually leached into ground and surface waters and provides
the bulk of the carbon content of these  waters.

    Historically, humic substances  have  been separated into
four substances by the scheme shown in Figure 3.  Table 1 gives
the properties of the three  soluble fractions as determined by
several different methods.   It is clear  that the terms "humic
acid", "fulvic acid", and "hymatomelanic acid" do not refer to
monodisperse substances.  Rather these are  strictly operational
terms referring to the products obtained by the scheme shown in
Figure 3.   The amount and type of each fraction will depend on
several parameters such as the  amount and type of vegetation
contiguous to the origin of  the humic substance, ambient
factors such as temperature,  soil or water  type, the presence
of soil or aquatic microorganism..?,  etc.  (42).
                             16

-------
r
                                  SOIL OR FREEZE  DRIED
                                      WATER SAMPLE
                                          BASE
                              ir
                      INSOLUBLE FRACTION   SOLUBLE FRACTION
                           (HUMIN)
                                                 ACID
                                                         II
                              SOLUBLE FRACTION   INSOLUBLE FRACTION
                               (FULVIC ACID)            |
                                                      ETHANOL
                   Figure 3.  Separation scheme for soil and aquatic humic substances (42)
                                    SOLUBLE FRACTION   INSOLUBLE FRACTION                  I
                             (HYMATOMELANIC ACID)         (HUMIC ACID)                     j
                                                                                           it

-------
r
                 TABLE  1.   PROPERTIES  OF  COMPONENTS OF HUMIC SUBSTANCES
             Group Name        Solubility  (44)        M.W.  Range of Average


             Fulvic Acid       Sol.  in  NaOH &         200 - 1,000 (44)
                               Mineral  Acid          200 - 300 (45)  a.
                                                     951 (41)  a.
                                                  •   688 (41)  a.
                                                     668 (41)  c.

             Hymatomelanic     Sol.  in  NaOH &
             Acid              Alcohol;  insol.
                               in  Mineral  Acid

             Humic Acid        Sol.  in  NaOH          up to 200,000 (44)
                               insol. in Mineral     700 - 26,000 (45) c.
                               Acid  and Alcohol      1,300 - 13,000 (45)
                                                     30,000 - 80,000  (45) d.
                                                     ^1,000 (41; e.
                                                     1,684 (41) b.
                                                     4,500 - 26,000 (41) c.
                                                     14,000 - 200,000  (41) f,
                                                     ^53,000 (41) g.
                                                     ^36,000 (41) h.
                                                     47,000 - 53,000  (41) a.
              a.   Osmoir.etry
              b.   Freezing Point  Depression
              c.   Diffusion
              d.   Ultracentifugation  &  Light Scattering
              e.   Isothermal  Ditillation
              f.   Gel Filtration
              g.   Sedimentation
              h.   Viscosity
                                           18

-------
r
                  The precise chemical composition of humic and fulvic acids*
              is still largely unknown.  Schnitzer and co-workers  (42) have
              contributed most to present knowledge in this area, but there
              remains much doubt regarding not only the nature of the building
              blocks which make up these natural polymers, but also the
              secondary and tertiary nature of the polymers themselves (46).
              The following is not meant to be a thorough'review of the
              available information on this subject; rather it is a summary
              of knowledge pertinent, to the discussion of water treatment
              practices which are effected by the presence of "humic substances"
              in water.

              Structure of Aquatic Humic Material

                  A general measure of the amount of aquatic humic substances
              is dissolved organic carbon (DOC) or total organic carbon  (TOG).
              Some typical TOG values in (mg/L) are shewn in Table 2  (44),
              for four types of water:  ground, sea, surface, and wastewaters.


              TABLE 2.  TOTAL ORGANIC CARBON RANGES FOR SEVERAL WATER TYPES
                        (44).
                 Water Type                     TOG (mg. per licer)


                Ground                          0-2

                Sea                             0.5-5

                Surface (NORS)         '          3.5 (average)
                                                1-20  (range)
                                                300 (maximum)

                Waste                           10 - 20 (average)
                                                1000 (maximum)
                  Aquatic organic matter is largely of the "fulvic" type,
              i.e.  it is soluble •> ri both acid and base (47), ranging from
              58 - 90 per cent in seventeen samples studied by Christman and
              co-workers.  The elemental composition of aquatic fulvic acid
              (FA)  is shown in Table 3,  taken from Alexander and Christman
              (41).   Schnitzer and Khan  (42)  propose that aquatic FA contains
              more oxygen and less nitrogen.than typical soil humus, but as
              Table 3 shows,  there is considerable variation in FA composition
              depending on the source water.
              *
              Hymatomelanic  acid is often combined with humic acid in this
              discussion.

                                            19

-------
TABLE 3.  ELEMENT7vL ANALYSES OF AQUATIC HUMIC FRACTIONS  (41)

Fraction %
Fulvic
Fulvic
Fulvic
Fnlvic
Fulvic
Fulvic
Fulvic
Fulvic
Fulvic
Acid
Acid
Acia
Acid
Acid
Acid
Acid
Acid
Acid
55.
54.
59.
58.
57.
57.
••
58.
41.
46.
C
61
87
32
42
91
08
39
50
2
•
5.
5.
6.
6,
6.
6.
5.
5.
5.
H
91
56
75
18
11
47
61
72
9
«
2.
2.
1.
1.
3 .
2.
0.
1.
2.
N
13
41
22
26
34
17
57
38
6
% O
a
a
a
a
a
a
a
50.80
45.3
Ref .
27
27
27
27
27
27
27
18
29

a:  By Difference
                            20

-------
   i               Characterization of the chemical content of aquatic fulvic
   ?           acids has been carried out by two general methods; chemical
   }           degradation and spectroscopic studios.  Christman (41) and
   ?           Schnitzer and Khan (42) have reviewed both methods.   Separation
              of aquatic fu'lvics has been carried beyond the scheme shown in
   '           Figure 3, using various chromatographic techniques.   Until
              recently, it was generally agreed that FA and HA are largely
   •           aromatic in carbon type.  This conclusion was based on products
   -           obtained by oxidatl/e and reductive degradations  (42).  Recent
   .           spectroscopic data (48) suggest, however, that the aliphatic
              content of FA may be much higher than expected, due possibly
   "*           to the loss of aliphatic units during the workup of degradation
   '           products.
   f
   ••               Schnitzer and co-workers (42)  have preferred a hydrogen
   -           bonded structure shown in Figure 4 as the principal structural
   ;           type in fulvic acid.   Christman and Ghassemi (49)  prefer the
              structure shown in Figure 5.  The latter includes more non-
              carboxylic units and more aliphatic components of an unspecified
   (           type, and is a covalently bonded macromolecule.  Warshaw and
   :           co-workers (50)  have discussed a structure for FA in terms of a
   :           hierarchy of moieties, in the lowest level of which are the
   •'           simple phenol, quinoid, and other small molecular units.   These
   .           are grouped together by covalent bonds into small polymers with
   ,           molecular weights of a few thousand or less.   Groups of small
   •           polymers can then be linked together into aggregates by inter-
   !           molecular forces such as hydrogen bonds.   The degree of aggre-
   \           gation is a function of water pH,  the oxidation state of the
   ,           molecules, etc.   Wershaw's model may be as precise as one can
   '           be regarding the generalised structure of aquatic humic matter.
              To be more specific,  one must specify the precise origin of the
              FA, its pH,  and other ambient factors.   Whether a more specific
   !           generalized structure of FA can be written as preferred by
f   s           Christman and Ghassemi (49) , must await further research.

'   '           Reaction of Chlorine  with Aquatic Humic Substances

|   ;               Several research  groups continue to investigate  the reaction
f   '           of chlorine with carbonaceous substances  in v/ater.  Rook (51,
;   -           52)  has suggested that the active  sites within FA molecules are
i   -           1,3-dihydroxybenzenes,  and the  mechanism  shown in Figure 6 is
|              suggested to account  for the formation -f THMs and other
              halogenated compounds.
                                           21

-------
HO—C



   HO
    //
0
          OH
                   CH OH 0
                Px
                "
       OH    OH    OH C-OH
                     I!
                     0
        JP            H

   Ov   C  OH      0   0   OH

   ^C  A  OH   HO-C.
 HO'
 OH-C
       01 J
            r  c   HO
    0  C = 0  OH   OH    C = 0  OH    OH
       OH
                      OH
Figure 4.  Structure of fulvic acid as proposed by Schnitzer
                      22

-------
COOH

 (C)v
                   C = 0
                    1
                                           CO.OH

C
                                                      ~f
Figure 5.   Structure  of  humic  acid as proposed by Christman
                     and Ghassemi  (49).
                             23

-------
                                        HOC1
                                       in  H20
                    COOH
                V*


                V
                 '^
                                            COOH
                                                                          Cl
   i''R/ci] o !  \i            R-   C1

   c      bo

           1
      Figure 6.  Rook-Moye mechanism for the aqueous chlorination
                          of resorcinol  (51).
                 Structures I and II yield upon cleavage, dichloromethans  (la),
            dichloroacetic acid (Ib), trichloroacetone  (Ic), chloroform  (Ila),
            trichloroacetic acid (lib) and tetrachloroacetone  (lie) respec-
            tively.  Further chlorination of fragments was proposed to account
            for polychlorinated acetones  (52).

                 Christman and co-workers have isolated 3,5,5-trichloro-
            cyclopent-3-ene-l,2-dione (III) and a number of other chlorinated
            species from the chlorination of resorcinol  (53) and aquatic
            humic material (54),  They point out that  (III) is not consistent
                                            0
                                      Cl
            with Rook-Moye mechanisms,
            no clear evidence that 1,
                               III
                            In fact, it should be noted there  is
                          3-dihydroxybenzene moieties are  the
principal precursors of THMs.  As shown by Morris and Baum
(341, the haloform reaction is possible with any one of a  number
of substrates such as the 6-diketones and pyrolles. '"Morris and
Baum have shown that acetogenins (natural pigments) such as
phloroacetophenone are potent haloform precursors.  More recently,
Arguello et al. (55) list a total of thirty-four substances which
give low to high yields of chloroform upon aqueous chlorination.
                                           .24
Eta^,-.'

-------
SEPARATION AND IDENTIFICATION OF SPECIFIC ORGANIC COMPOUNDS IN
WASTEWATERS BEFORE AND AFTER CHLORINATION.

    The tremendous number of organic species with widely
divergent structures combine to produce an incredibly complex
organic matrix in wastewater effluents.  Feng  (56) indicated
that the study of the chlorination of sewage would perhaps be
impossible because of the complexities of such systems.  Geiger
(57) recently emphasized the point:

         "Such investigations are hindered by two intrinsic
         properties of organic water constituents.  First,
         the organic assemblages in environmental samples
         are of an extra-ordinarily high compositional
         complexity; and second, single components occur in
         trace quantities only".

    Organic content of wastewaters was historically evaluated in
terms of gross parameters such as volatile solids, suspended
solids, dissolved solids, Biological Oxygen Demand (BOD) and
Chemical Oxygen Demand (COD).  Unfortunately, these parameters
do not provide information on the specific chemical compounds
present in the wastewater.  To obtain this type of information,
it is generally necessary first to separate the components of
the sample into major fractions.  Separation methods commonly
used include differentiation by solubility (liquid-liquid
extraction), adsorption  (with activated carbon or "resins),
membrane permeability (reverse osmosis and ultrafiltration),
and volatility (purge techniques).  For volatile compounds
further separation may be achieved by gas chromatography; for
non-volatile compounds the corresponding technique is high
performance liquid chromatography (HPLC).

    Identification of specific organic components of the isolated
fractions may be accomplished by GC or HPLC by corralation of
the retention time of the unknown with that of a known compound.
A higher level of assurance may be obtained by the use of selective
GC or HPLC detectors such as the electron capture GC detector.
For ultimate proof of structure, these techniques should be
combined with spectroscopic data on the compounds in the GC
or HPLC eluates.   Most commonly, this is mass spectroscopic
data, although infrared absorption spectra may also be useful
in many cases.

    Several works have been published which utilize combinations
of gas chromatography and mass spectrometry  (GC/MS) for the
determination of specific chemical constituents of natural
waters and wastewaters.   References(58) and  (59) include many
examples.

    In 1S73 and 1975, Glaze e_t al. (36, 60) used these techniques
for the determination of by-products produced by the chlorination

                               25

-------
               and superchlorination of municipal wastewaters.  About the same
               time, Jolley  (61) developed an analytical procedure using
               high pressure liquid chromatography and 36Cl-labeling to identify
               polar compounclr containing carbon-hound chlorine  (9, 61)
               formed by water chlorinacion.  Tables 4 and 5  list the small
               halogenated molecules identified by these two  groups.

                   Several chlorination studies have been done whereby
               compounds known to occur in municipal wastewaters  (secondary
               effluents) were subjected to chlorination in a distilled water
               matrix under various treatment conditions.  For example,
               Carlson et al.  (32) demonstrated facile chlorine uptake by such
               compounds as phenol, anisole and acentanilide.  In the sa:ne
               report, it was also demonstrated that biphenyl formed various
               chlorinated analogues upon aqueous chlorination.  Increased
               chlorine doses resulted in formation of increasing polychlor-
               inated analogues.  Also found were chlorohydrins formed upon
               chlorination of some olefinic systems such as  oleic acids.

                   Four conferences have been held to discuss the impact of
               water chlorination and the proceedings of the  first three
               conferences are published (62-64).  The contents of these
               publications include several other reports on  the by-products
               formed during water chlorination.  In addition, several papers
               on this subject are included in references (58) and  (59).


               TOXICITY OF CHLORINATED WASTEWATER EFFLUENTS.

                   It is beyond the scope of this report to discuss the
               potential toxicities of municipal wastewater effluents before.
               and after chlorination.  It is clear, however, that chlorination
               of water causes the formation of some new compounds which have
               toxic properties.  The danger of this situation as pointed out
               by Bunch (65)  is that drinking water supplies  originating from
 r:              surface water sources may be composed of some  fraction of
} .              reconditioned sewage.  To compound this problem, some areas of
               the world face the prospect of declining water supplies and will
               be forced to consider direct reuse of treated  wnstewater.

 :                  In addition to the potential effects of chlorinated by-
 ;              products on humans, there is the concern that  chlorination by-
 )              products may be detrimental to aquatic life in streams receiving
 ;              wastewaters (62-64).

                   Severc.1 reports have appeared on the toxicity and mutagen-
               icity of chlorinated compounds and extracts from chlorinated
               waters and wastewaters (62-64, 66).  These studies generally
               take one of two paths.  In one, the wastexvater is tested
               directly,  or after only gross fractionation.    No attempt is made
               to determine the active chemical species,  only to ascertain if
               the test water or fraction is toxic, mutagenic, teratogenic,  etc.


                                              26

-------
               TABLE 4.  CHLORINATED PRODUCTS IDENTIFIED IN
                          CHLORINATED WATER BY JOLLEY  (61)
       5-Chlorouracil
       8-Chlorocaffeine
       8-Chloroxanthine
       5-Chlorosalicylic acid
       2-Chlorophenol
       4-C'hlorobenzoic acid
       3-Calorobenzoic acid
       4-Chlororesorcinol
       4-Chloro-3-methylphenoi
5-Chlorouridine
6-Chloroguanidir.e
2-Chlorobenzoic acid
4-Chloromandelic acid
4-Chlorophenylacetic acid
4-Chlorophenol
3-Chlorophenol
3-Chloro-4-hydroxybenzoic acid
       TABLE 5.  CHLORINATED PRODUCTS PRODUCED BY WATER CHLORINATION',
                             BY GLAZE ET AL., (36,60)
       Chloroform
       Dichlorobutane
       Chlorocyclohexane
       o-dichlorobenzene
       p-dichlorobenzene
       Pentachloroacetone
       Dichloroethylbenzene
       N-methyl-trichloroaniline
       Trichlorophenol
       Di cnloromethoxytoluene
       Trichloroethylbenzene
       Dichloro-bis(ethoxy)benzene
       Trichloro-a-methyl benzyl
          alcohol
       Trichlorocumene
       Dichloroaniline derivative
       Dichloroacatate derivative
       Tetrachloropthalate derivative
Dibromochloromethane
3-chloro-2-raethylbut-l-ene
Chloroalkyl acetate
Tetrachloroacetone
Chloroethylbenzene
Hexachloroacetone
Chlorocumene
nichlorotoluene
Chloro-a-raethyl benzyl alcohol
Trichloromethylstyrene
Dichloro-a-methyl benzyl alcohol
Trichloro-N-methylanisole
Tetrachlorophenol

Tetrachloroethylstyrene
Dichloroaroma-cic derivative
Trichlorophthalate derivative
                                      27
..9*--

-------
     The second approach  seeks  to  identify the  individual  species
in the effluents, then applies  various  toxicological  tests on
these compounds to evaluate  the ha2ard  associated with  their
presence.  Garrison ejt al.  (67) , who  used this  approach, pointed
out:

         "Knowledge of the specific compounds discharged
         .is needed to study  health effects of pollutants,
         to help determine the  sources  of compounds found
         in drinking water surveys and  to establish effluent
         guidelines".

Geiger  (57) also used this approach stating:

         "Since most biochemical reactions show a very
         pronounced structural  dependence, studies on
         chemical ecology necessitate analyses  for single
         constituents".

SURROGATE METHODS VERSUS SPECIFIC  COMPOUND IDENTIFICATION.

     While it is desirable to determine the precise composition
of molecular types in a wastewater effluent, at least two
factors make this goal unachievable.

   (1)  Raw and treated waters  may contain thousands of
        compounds at trace levels;

   (2)  Many of these compounds  cannot  be identified and
        quantified by available  analytical methods such as
        GC/MS.

     For these reasons, water quality is often  measured by
surrogate or group parameters such as total  (or dissolved)
organic carbon TOC (or DOC); total organic nitrogen (TON),
chemical oxygen demand (COD), etc.  In view of  the toxic
nature of many organohalogen compounds  (68), a  TOX parameter
would seem to be of value in assessing the quality of re.w  and
treated waters.

     Three approaches have been proposed for TOX measurement
(69-72), each of which involves a  preconcentration step and
the subsequent determination of organic halogen in the pre-
concentrate.  Since organic halogen may not be  determined  non-
destructively, the sample preconcentrate is usually pyrolyzed
to convert the halogen to halide ion  in aqueous solution.  The
halide may then be titrated by potentiometric or coulometric
methods or determined with a halogen  sensitive  electrode (73).
Pyrolysis of organohalogen compounds  to form hydrogen halides
may be done reductively or oxidatively; in either case the
product is almost entirely hydrogen halrdte (71)  if the pyrolysis
temperature is above 300°C.

                              28

-------
     Preconcentration of the water sample has buen carried out
by adsorption of the chlorinated orqanics on activated carbon
(69, 70, 73) or XAD resins  (71, 7^, 75) or by extraction into
a non-polar organic solvent  (72).  Each method has its
limitations due to inefficient adsorption or interferences
from inorganic halogen or other species.

     Several recent investigations have shown by the use of
these methods that non-volatile compounds which cannot be
determined by GC or GC/MS &ra formed in abundance by the
chlorination process.  Ku'hn, Sontheimer and co-workers (69,
70) have shown that non-volatile activated carbon adsorbable
organohalogen compounds exceed volatile compounds (in halogen
content) by a factor of two to four in typical chlorinated
surface water supplies in Eux'cpe.  Gl.-ze et al.  (75) and
Oliver  (74) used XAD-resin adsorbents to obtain similar results.
The relative yields of THMs and non-volatile TOX has been
shown to be a function of pH and dose of halogen  (70).
Indirect evidence for the occurrence of non-THM products, such
as those proposed by Rook (51) comes from the results of
direct aqueous injection THM analysis discussed in Section 8.
The larger yield of THMs found by this procedure, as compared
to other methods, has been taken as evidence for decomposition
of chlorinated non-volatile THM precursors in the GC injec-
tion port  (76, 77).  Thus, while THMs are the major volatile
products which arise by the chlorination process, und^r certain
conditions they represent a minor yield of orgainic halogen
when non-gas-chromatographable material is taken into account.
                              29

-------
                           ' SECTION 5

    COMPUTER ASSISTED GC/MS ANALYSIS OF ORGANIC COMPOUNDS IN
              MUNICIPAL WASTEWATER PRODUCTS BEFORE
                     AND AFTER CHLORINATION
INTRODUCTION

     This study of municipal wastewater products before and
after application of chlorine was conducted in order to deter-
mine the extent to which chlorination causes the formation of
volatile organohalogen compounds.  Because of the harmful
potential of many chlorinated organic compounds reviewed in
the previous section, any source which discharges these
compounds into the environment poses a possible threat to man
and other forms of life.

     Gas chromatography  (GC) is the most powerful separation
method available for determination of volatile organic compounds.
Combined with element specific detectors, GC is capable of
detecting extremely small quantities of organic substances.
For example, with the electrolytic conductivity detector one
can detect as little as 50 picograms (10~12 g) of a chlorinated
organic compound such as•carbon tatrachloride.  For the separa-
tion and identification of unknown volatile organics in complex
mixtures, the most advanced method available is GC combined
with on-line mass spectrometric detection  (GC/MS).  Both GC/MS
and GC with element-specific detection have been used in this
work.

     The advancement of GC/MS technology was aided immensely
in 1968 with development of the first fully computerized
GC/MS system by Hites and Biemann (78).  Their work showed
examples of how an on-line computer can acquire and process
GC/MS data from complex samples which have more than one-
hundred GC peaks.  Today specialized computer programs have
been developed to simplify the data processing and/or to extract
grossly obscured, relevant information from the data.

     One such computer program has received various names in
the literature:   "Limited Mass Search (LMS)", "Extracted Ion
Current Profile", or "Mass Chromatography".  This is a data
manipulation technique ..applied to GC/MS data subsequent to itc
acquisition and storage.  The technique is used to identify the
                               30

-------
J           locations of specific compounds or classes of compounds within
*           a total ionization chromatogram  (TIC).  The computer program
]           extracts the ion current intensities  from each spectrum in
           t? e TIC at a specific mass which is characteristic of a com-
           pound or class of compounds.  These intensities are then
           plotted as relative ion current intensity versus spectrum
           number.  The LMS can then be compared to the TIC to determine
           which peak(s) is (are) due to a particular compound or class
           of compounds.  An example is shown in Figure 7.  The top
           chromatogram is a TIC of a septage extract.  The bottom chrom-
           atogram is the corresponding LMS at m/e 149.  Almost all
           phthalate esters have a characteristic base peak.at m/e of
           149.  After LMS, the phthalate esters are conveniently marked,
           and each specific ester can be identified by its complete
           fragmentation pattern.  Thus, comparison of the chromatograms
           facilitates the location and identification of the phthalate
           esters.

                This technique has become common during the past several
           years.  Recent literature cites examples of its application to
           determine phthalate esters  (79), polynuclear aromatic
           hydrocarbons (80),  mononuclear aryl hydrocarbons (81), and
           many other types of compounds.  Another important application
           of this technique,  recently reported, was to determine the
           location of chlorinated organics in water extracts (82).  The
           m/e of 35 was used, a mass which is highly specific for
           chlorine.  It is unlikely that any other elemental combination
           with an m/e of 35 would form, but unfortunately not all
           chlorinated compounds produce this fragment (i.e.,  some
           chloroaromatics).  Figure 8 shows an example:  the TIC for a
           "superchlorinated"  septage extract with the corresponding LMS
           at m/e of 35.  The peaks shown in the LMS chromatogram show
           a high correlation with the presence of chlorinated organics
           in the TIC.

                This section describes a new GC/MS/DS computer program
           which manipulates acquired and stored data to produce a
           "Limited Cluster Search" (LCS).  The resulting LCS chromatogram
           indicates the peaks in a TIC which possess a specific number
           of ch3orir.es and/or bromine atoms.   The program is similar to
           Limited Mass Search programs in that it extracts specific
           information from each and every mass spectrum in a TIC.   This
           information can then be plotted on a relative basis versus
           spectrum number for comparison with the TIC.  Peaks appearing
           in the LCS chromatogram have a high probability o€ containing
           a given number of chlorine and/or bromine atoms.

                As Asron (83)  recognized in 1920, chlorine contains a
           mixture of j jotopes with masses,  respectively, of  35". and 37
           and occurring naturally in a ratio of approximately 3 to 1
           (35C1 to 37C1).   This is reflected in the mass spectra of
           chlorinated compounds.   An ion possessing a chlorine  atom will

                                          31

-------
U>
to
                                                                                                                      .-w*o<"f *v; ^"
        CflL. PFX. SMPLE.i XBO. CHLORO.
                TOTflL IONI2HTION CHROfWTOGfWI
        108
i ,. I i I I , . | : | , , .-,-r. . .p., . ,. i , , . ! . ,. | , , . | . , , ; , , , , , | , | , , i | , , , | , , , | , | , | , | i , , , , , ,, , | , , , , , , .

SO    100     153    200    2SO    380    3SO    400    450
        CPL. PFX. SHPUE.i XRO. CHLORQ.
                LIMITED WISS 5ERRCH RT M/E=149
        ICO
                                                                       SSO    580
                                                                                  'sso"
                                                                                         700
IT"
750
                                                                                                    888
                                                                                                          eso
                                                                                                                903
              50    100    150    200   250    300    330    400    4EO   500    SSO    600    650    700   750   800    8SO    900    950
                                                                                                                                        I
                                                   SPECTRUM    NUMBER
          Figure  7.   Total  ion   (top)  and  LMS chromatograms  at  m/e  149   (bottom)   for  XAD
                                  extracts  of  superchlorinated  septage.

-------
         CM-. PFX. SMPLE.i XRO. CHLORO.
                   TOTHL IOWZRTION CSTOIflTOGRfiH
         ;ao
                            JL/W.
                so     ion
                                      ZOO    2EO     300
                                                            3SO  .   403     450
                                                                                   500
                                                                                           sso
                                                                                                  600     650
                                                                                                         •' " I • "
                                                                                                          700
U)
U)
ca. PFX. SHPLE.I xno. CHORD.
           imiTFn MUSS SEBRCH m
100
                      tea      iso     203
                                            250     300     3S3
                                                                   400
                                                                  "I"'
                                                                   450
                                                                                  BOO     SSO     600     660
                                                                                                        T-,,1-

                                                                                                         7CO
                                                               SPECTRUM      NUMBER
                                                                                                                                       SSO
                                                                                                                                                      950

-------
t v-'N
m
lli..           produce two peaks at masses X and X + 2 corresponding to the
11;            respective masses of the ions with the 3 5C1 and 37C1 isotopes.
I i .           The ratio of intensities of. the X to X + 2 peaks will be
f *'•'           approximately 3 to 1.  An ion possessing more than one chlorine
|   '           atom will produce a "cluster" of peaks.  The number of peaks
|;            in the cluster will be n + 1, where n is the number of chlorine
•   .           atoms.  The intensities of each peak in the cluster can be
j,  ',            calculated by expanding the simple binomial expression,
              (a + b)n where a and b are the relative abundances of the
              light and heavy isotopes.  Tables have been published  (84)
              which show the relative ion intensities of the peaks in clusters
              for varying numbers of chlorine atoms.

                   Bromine is also a mixture of two isotopes, with masses
              .of 79 and 81, respectively.  The naturally occurring ratio
              for these isotopes is approximately 1:1.  Tables for the
              clusters of fragments containing multiple bromine atoms are
              calculated in an analogous manner to chlorine clusters a.id
              have been published  (34).  The relative intensities of ion frag-
              ments containing mixtures of bromine and chlorine atoms also
              can be calculated, and the resulting tables have been published
              (84>-

1 \                The "Limited Cluster Search" program searches a pre-
I  '-           determined range of masses for the appropriate isotopic
I  Y           cluster.  An important characteristic of this program is that
I  *           it is not mass specific, that is, the recognition of a
|  V           chlorinated or brominated compound does not require that the
f  T.           isotope cluster occur at specific masses in the mass spectrum.
              This ineans that a compound can possess a wide variety of
              additional functional moieties and still be recognized as
              containing chlorine and/or bromine.

                   Two papers have been published which describe computer
              programs that can identify chlorinated and/or brominated
              isotopic clusters regardless of mass.  Mun, et al.  (85) have
              extended the McLafferty "Probability Based Matching" program
              for this purpose.  The goal is to identify the number of
              chlorines and/or bromines in the compound, and the c^  ->uter
              program in its present state is designed to be applied to
              individual spectra of relatively high quality.  For GC/MS
              data, the appropriate background spectrum should be subtracted
              from the spectrum of interest before it is subjected to computer
              analysis.  And, since there is no quantitative measure of the
              ion clusters, the program could not be easily adopted to produce
              a chromatographic profile showing relative peak intensities.
                   Regnier and Canada  (36) have published the description of
              a computer program which does produce a chromatographic
              profile of the chlorinated and/or brominated compounds in a
              GC/MS data set.  However, their goal was to use the relative
                                             34

-------
heights of corresponding peaks in the different profile as a new
identification technique.  The relationship between these peak
heights tends to be specific  for a particular compound.  It
should be noted that the goal of the LCS computer program
described in this chapter is  to improve the ability of the
analyst to find mass spectra  of chlorinated and/or brominated
compounds produced from the GC/MS analysis of matrices.
Regnier and Canada have never reported the application of their
computer program to such a sample for such a purpose.  In fact,
the examples they use in their report would not adequately
test the ability of their program to process data in a com-
plicated matrix which contains many interfering nonchlorinated
analytes.  The most complex samples which they analyzed were
standard polychlor.inated biphenyl mixtures.  Although these
samples produce a complex chromatogram, the individual peaks
are closely related homologues of each other.  This means that
the fragmentation process for the different analogues will be
very similcir, and the spectra of overlapping GC peaks will
tend to reinforce the isotopic clusters as opposed to inter-
fering with them.

EXPERIMENTAL

     It is well known that GC/MS, even with the most sophisti-
cated data analyses, usually cannot be applied directly to
the analysis of dilute water  solutions.  The technique generally
is not sensitive enough to directly detect compounds of
interest which are present in the parts-per-billion (ug/L)
concentration range or below.  Moreover, many compounds such
as certain phenols, carboxylic acids and amines must be
derivatized to neutral forms before high resolution GC is
possible.  Thus, various preconcentration, derivatization, and
separation  processes are utilized on environmental samples
before GC/MS analysis.   To separate relatively non-polar
compounds from water, one commonly uses liquid-liquid extrac-
tion (87), purge-and-trap (88) , or adsorption techniques
(89).  This section emphasizes the use of adsorption with
synthetic macroreticular polymers as the preconcentration method
for volatile compounds.   Liquid-liquid extraction and purge-
and-trap techniques for highly volatile ("purgeable")  compounds
are compared in a later section.

Chlorination of Deriton,  Texas, Municipal Wastewater

Sample Collection—
     Most of the water samples used in this study were collected
at the De«ton Municipal Wastewater Treatment Plant.   This plant
utilizes a biologically activated sludge treatment process as
shown in Figure 9.   The treatment process includes the following:
                               35


-------
*
I

                PRIMARY
               SETTLING
                  TANKS
                FINAL
              CLARIFIERS
















n

ACTIVATED
SLUDGE
BASINS
                                                  CHLORINATION
                                                      BASIN
                                         f  TO  PECAN  CREEK
                 Figure 9.  Schematic of the Dentor., Texas municipal
                             wastewater treatment plant.
                                        36

-------
                       1.  Primary clarification;
                       2.  Digestion by activated sludge;
                       3.  Secondary clarification;
                       4.  Terminal disinfection by chlorination.

                 During most of the sampling, the plant was operating near
            capacity, at approximately 5.7 million gallons/day.  BOD
            levels of the final pl_nt effluent before chlorination     • -•
            averaged approximately 19 mg/L during most of the sampling
            period, although daily records were not kept.

                 In the earliest studies,  samples were collected before
            and after the chlorination process.  The specific sampling
            point for the chlorinated sampla was at a spillway at the
            effluent end of the chlorination basin.  The unchlorinated
            control sample was collected at the effluent point of one cff-*
            the secondary clarifiers.

                 Later, however, only the  unchlorinated sample was collected
            and transported in a steel container to the laboratory where
            the actual chlorination was conducted.  This alternative pro-
            cedure allowed the chlorinated and control sample background
            matrices to be identical prior to beginning the experimental
            procedure.   The alternative procedure also allowed quality
            control procedures to be performed on all reagents used in
            the experiments.

                 The genera.1  approach used to demonstrate the formation
            of new organochlorine compounds was to analyze concentrated
            extracts of the chlorinated wastewater by gas chromatography
            using a halogen-specific detector, the Coulson electrolytic
            conductivity detector.   This profile chromatogram could then
            be compared to a  chromatogram  of the unchlorinated control
            sample extract.   The additional peaks seen in the chlorinated
            extract chromatograra represented the formation of new halogen-
            containing species.

                 The^e experiments  published in 1973 in the Journal of
            Chromatographic Science (36) clearly demonstrated that new
            chlorinated organic  compounds  were generated using waste-
            water disinfection procedures.   Th-2 experiments were followed
            with another series  of  experiments to identify the new
            chlorinated species.   The  same sampling procedure was used,
            but the chlorine  doses  were in the 1,000 to 4,000 mg/L range.
            These large doses were  used to increase the concentrations of
            the chlorinated species, making "their identification easier.

                 A second reason for conducting these studies at such high
            chlorine doses was to simulate a treatment process that is
            described  in a 1969  patent (90) .   This process uses pressurized
            chlorine at similar  concentrations as a means of oxidizing and
            disinfecting wastewater and sludge by-products.   The purpose

                                         37
¥nUUUHM*A

-------
             of the process is to increase the settling ability of the
             suspended materials, as well as to disinfect the &st.ire
             substrate.   The process operates by pumping the substrate into
             a chamber that is pressurized with chlorine.  A portion of
             the supernataat rS then punped off into a second chamber
             where oxidation by chlorine is completed.  A portion of the
             reactor mixture,  approximately. 75 to 80 per cent, is recycled
             into the first chamber where it is subjected to further
             oxidative chlorination.  The process can be used in place of
             anaerobic digesters for sludge treatment, or it can even be
             used on the entire plant influent as an alternative to
             conventional treatment systems (i.e., activated sludge,
             trickling filter).

             Chlorination and XAD-Extraction of Effluent Samples—
                  When the wastewater samples arrived in the laboratory,
             they were processed according to the scheme shown in Figure
             10.  The water was first filtered using coarse, fluted filter
             paper (Whatman §41, Whatman  Limited),  followed by suction
             filtration using Gelman £61694 glass fiber filters.  Two
             10-liter aliquots were then transferred into 20-liter glass
             vats that were equipped with glass stirring mechanisms.

                  Chlorination of one of the aliquots was effected by
             bubbling chlorine gas (Dixie Chemical Company, 99.5 per cent
             purity)  into the  sample as it was stirred vigorously.  The
             chlorine concentration was monitored by removing aliquots,
             making pj-aner dilutions, and assaying for chlorine by using
             the orthotolidine arsenite method (91).  The chlorine contact
             time was one hour, after which tine excess granular sodium
             sulfite was added to both chlorinated and control water
             portions.   The pH of both was then adjusted to 2-to-3 with
             concentrated 112804.

                  The organic  compounds in the sample were  concentrated using
             Amberlite  XAD-2 resin (20-50 mesh,  Rohm and Haas Company).   The
             resin is a  polystyrene/divinyl benzene  copolymer that is
             manufactured in the form of spherical beads (92).   The porous
             nature of  the beads results in a  fairly high surface  area  of
             300 m2/gm  (89),   The surface of the  resin is extremely hydro-
             phobic,  which accounts  for its affinity for non-polar organics
             in waters.   Upon  contact with o^anic solvents,  the surface  is
             said to "relax",  readily releasing most adsorbed organics.
             This system has a  major advantage over  conventional liquid-
             liquid extraction  methods  in that a  much higher  water-to-
             extractant  ratio  can be  achieved,  resulting in higher concentra-
             tion factors and  therefore higher sensitivity.   The system has
             a  similar advantage over carbon adsorbants.  The higher adsorp-
             tion efficiency of  the  XAD resin  for most non-polar organics means
             that a relatively  smaller  amount  of  adsorbant  can be  used  as
             compared tc the amount  of  carbon  required.   This means that  less
                                           38
L«, «_-... vwitllW	*_.	
                                                                                    \

-------
                                         t  i
                                           SAMPLE COLLECTION
                                          Gravity Filtration
                      L
                                      Division into Two Aliquots
Chlorination
                        Quench (Na.
       VS03>|
Quenc'i
                        XAJ) Extraction
                                                             XAD Extraction
                         Et?0 Elution
                                         Elution
                        Concentration
                                                             Concentration
                                               ANALYSIS

                                  Chroraatographjic:   FID> CECD, GC/MS
                               Total Halopv.i:   Pyrolysis/Microcoulometry
                       Figure  10.   Scheme  for the extraction  of organic
                                    Compounds from wastewaters.

                                                   29
- iUWv.if.


-------
organic solvent can be used to remove the organics from the
adsorbant, resulting in higher wator-to-solvent ratios and thus
higher sensitivity.

     Careful attention was paid to the cleaning and preparation
of the resin prior to its use as an analytical adsorbant.  The
cleaning techniques have been described in detail previously
(111) .  Approximately 25 gra of resin was transferred from its
shipping container to an erlenmeyer flask.  The resin was then
covered with approximately one inch of reagent-grade methanol
(Fisher, AC'S Certified).  The erlenmeyer was then swirled
several times, which dissolved unreacted monomer and suspended
polymeric fines that were generated during manufacturing and
shipment.  The excess methanol was then removed by suction using
a glass tube attached to an aspirator.  Another portion of
methanol war, then added to the erlenmeyer, and the process was
repeated until the methanol remained clear upon vigorous
swirling.

     Following the washing, the resin was transferred to a
soxhlet extractor and extracted sequentially with methanol
(Baker, b.p. - G5°C.), acetone 'Baker, b.p. = 56°C.), and diethyl
ether  (Mallinekrodt,b.p. = 34°C.) for 24 hours with each solvent.
According to Junk, et al.  (89) , the decreasing boiling points
are important in avoiding cracking the resin beads with sudden,
large changes in temperature.

     After the extraction process had been completed, the
resin was washed into a clean erlenmeyer flask using methanol,
and it was t.tored under a portion of methanol in the stoppered
flask.

     The apparatus used to contain the resin during the extrac-
tion process is shown in Figure 11.  After cleaning the apparatus
with chromic acid, a small plug of glass wool was placed in the
bottom of the column, as shown in the Figure.  Approximately
1 gm of the resin, in a methanol slurry, was then transferred
into the column, and the methanol was removed by allowing
approximately 0.5 L of distilled water to flow through the
column.  Next, the column was slowly and carefully backflushed
with Milli-Q water by stoppering the top of the column, placing
the effluent end of the column in a beaker of distilled water
and attaching an aspirator to the takeoff arm on the column.
The backflush flow.cate could be controlled by manipulating the
stopcock on the takeoff arm.  Thio procedure suspended the resin
beads in the water in the column.  The beads could then be
packed in an extremely regular matrix with virtually no
"channeling" effects, by quickly switching the aspirator from
the takeoff arm to the effluent end of the column, removing the
stopper at the top of the column, and momentarily opening the
effluent stopcock.  The column was then ready for use in the
extraction process.

                               40

-------
   24/40
STANDARD
   TAPER
   OUTER
    JOINT
TEFLON   STOPCOCK
                               XAD-2 RESIN  BED
                                GLASS  WOOL  PLUG
                                   TEFLON  STOPCOCK
   Figure 11.  Glass apparatus containing XAD-2 resin for
         extraction of organic compounds from water.

                          41

-------
                   The resin columns were attached to a glass siphon that
              descended from the glass vats.  The water samples were allowed
              to flow through these columns at a rate of approximately 30
              ml/miri.  The excess water was then forced out of the column
              using a pipette bulb.  The resin was eluted with 25 ml of
              diethyl ether (Mallirrkrcut, prepurified) , which was glass-
              distilled in the laboratory prior to use.  The diethyl ether
              extracts were concentrated to 1 ml using a modified Kuderna-
              Danish flask and a Snyder three-ball distillation column, as
              described by Junk
                   The survey chromatographic analyses were then performed on
              the concentrates, followed by further concentration to approxi-
              mately 100-50 yl for gas chromatographic-mass spectrometric
              analysis.

              Survey Gas Chromatographic Analysis —
                   The survey chroma tographic analyses were performed on a
              Varian 1800 gas chromatograpn that is equipped with a flame
              ionization detector (FID) and a Coulson electrolytic conduc-
              tivity detector  (CECD) .   The glass chromatographic column was
              6 feet by 2-mm i.d. , packed with 3 per cent Dexsil 300 GC, coated
              on 100/120 mesh Supelcoport.  The helium carrier gas flow rate
              was 30 ml/min.  The temperature-program conditions were, as
              follows:

                  1.  Isothermal at room temperaturs (approximately 27°C.)
                      for four minutes after injection;
                  2.  Program ballistically from room temperature to 50°C;
                  3.  Program from 50° to 300' at 6°/minute;
                  4.  Isothermal at 300°C. uncil no further peaks eluted.

                   The injector temperature was 225°,  and the detector
              temperature was 300°C.  The Coulson block temperature was 300°C.
              and the furnace temperature was 830°C.   The detector was
              operated in the reductive mode with 80 mL/min. of hydrogen
              added to the gas chromatographic column effluent before pyrolysis.
              The bridge current was 30v; the attenuation was X4 to X32 (as
              indicated in, the Figures).   The FID detector was run at a
              range of 10"*1 amp/mv, and an attenuation of X8 to X32 (as
              indicated in the Figures) .

              GC/MS Analysis —
                   The GC/MS system was a Finnigan Model 3200, controlled by a
              Finnigan Model 6100 computerized data system.   The chromatographic
              conditions were the same as chose described for the survey
              analyses.   Successive  mass spectral scans were acquired from m/e
              = 35 to m/e =  451- at a scan rate of approximately I sec/decade.
                                             42
L       >^..._. ,_
BtMJjr^jf-TJc ,y^frV»
               i^

-------
Analysis of Superchlorinated Septage and Sludge Supernatants

      In the second portion of this study, samples of raw and
superchlorinated septage pumpings were obtained from Ventura,
California.  The septage had been treated at the plant site by
a Purifax reactor for disinfection and stabilization of the
septage.  Samples were shipped to the North Texas State
University  (NTSU) laboratory through arrangements by officials
at USEPA/MERL (Cincinnati).

Sample Treatment—
      Raw and superchlorinated samples of septage pumpings, the
latter treated in Ventura, California, with excess Na2SO3 to
quench were centrifuged with a Sharpies centrifuge to obtain a
clear centrlfugate.  Samples were analyzed by a modified Bellar
purge and trap procedure  (88) coupled to the GC/MS system.
Quantification of the purge and trap data is no!: reported,
however, since the samples were not shipped in headspace-free
containers, nor were precautions taken to preserve purgeable
compounds at the NTSU laboratory.  Control (raw) and super-
chlorinated samples were also extracted by the XAD-2 resin
procedures, as described previously.  Ether eluates were
analyzed by GC/FID, GC/CECD, and GC/K3.  GC/FID data are not
reported and GC/CECD data were used only to indicate halogen-
containing species in the analysis of GC/MS data.  Samples of
XAD-2/Et2O eluates were also combusted in the Dohrmann MT-20
pyrolysis furnace, and the furnace effluent titrated by micro-
coulometry.  Total organic halogen values so obtained were
compared with the sum of GC/MS-identified peaks.

Limited Cluster Search Mass Spectroscopy

Test Samples—
      Two samples were run for demonstration of the new computer
program.  The first sample, HALSTI, was an artificial mixture
of halogenated compounds and nonhalogenated normal alkanes.
The names and sources of the compounds are listed in Table 6.
The sample was prepared by adding approximately 250 microliters
of each compound to 2 milliliters of acetone.   0.5 microliter
aliquots of this mixture were then injected into the GC/MS for
analysis.  This sample was used to study the decision parameters
in the computer program.
                              43
                                                          • ejft'f Au&.'t&A.&ji.^ _,„.,•

-------
                 TABLE  6.   NAMES  AND  SOURCES  OF COMPOUNDS  IN HALSTI SAMPLE
f
                Name                          Source,  Grade

                 •- - ~ ' "--" ~~    Lr-T--._i-.  , —._._ —r   .-..-- - ---^f— ' -       —  -

                Bromoform                    Fisher,  Certified

                n-Decane                      Fisher,  Certified

                1,3-Dichlorobenzene          Mallinckrodt, 98% or

                1,2,4-Trichlorobenzene       J.T.  Baker,- Practical

                1,3-Dibromobenzer.e           Fisher,  Practical
                                         * •*•*
                Hexachlorobutadiene          MCB,  Practical

                n-Tetradecane                Fisher,  Reagent
                The  second sample,  CALCLI,  was the XAD-2 extract of super-
            chlorinated  California  septage  as described earlier in this
            section.   The  etner eluant was  concentrated to approximately
            100 microliters,  and a  2.5 microliter aliquot of that concentrate
            was injected into the GC/MS system.   Only a portion of the
            TIC (scans 290 through  510;  was processed by the computer program.
            The first spectrum of these data was selected so that-enoUgh
            data would be  processed to accurately reflect the ability of the
            program  to extract important information.  The entire RGC was
            not processed  bacause of the computer CPU time limitations.  The
            evaluation of  eaoh spectrum required about 0.75 minutes of
            computer time.   Thus, for the limited data selected, the complete
            computer run took almost 2-1/2  hours.

                The  GC/MS  data were transferred to Finnigan magnetic tape
            cassettes for  long-term storage and transportation to the
            computer system described below for the actual computer
            processing.
l
)           LCS Computer Program—
I               Details  of the Limited Cluster Search computer'program are
'$           given in Appendices A and B.


§           RESULTS  AND  DISCUSSION

I            Denton,  Texas  Municipal Wastewater Extracts
                Figure  12  shows typical FID survey chromatograms of the
            chlorinated and control  wastewater extracts (chlorine dose,

                                          44

-------
            2,000 mg/L).  The chrpmatograms  are  typical of what one would
            expect from such a complicated matrix  as wastewatp.r.  Well over
s            100 peaks are identifiable, many of  which  are mixtures of two or
[^          more components.  Figure  13 shows typical  Coulson  (CECD) survey
            chromatograms.  More than thirty halogenated species detected
            in the chlorinated portion are not present in the  control, or they
            are present at distinctly lower  concentrations.  Some of these
            chlorinated compounds are the same as  those generated at chlorine
            concentrations at 10 mg/L (36) .   The GC/MS total ion chromatogram
             (Figure 14) is similar  to the FIC chromatogran.  Inspection of
   •-        the mass spectra reveals  some thirty or more chlorinated com-
            pounds not present in the control or present in distinctly lower
\            concentrations.  Table  1  lists the compounds that  are identified
            as species generated by chlorination.

;             _  Most of the compounds identified thus  far are  aromatic
I           derivatives.  The compounds are  by no  means derivatives of
I           "activated" aromatics in  every case, however.  The chloro
>,           derivatives of benzene, toluene,  ethyl benzene and benzyl
v           alcohol do not necessarily suggest a mechanism of direct
I           chlorination, as we shall note later.  Moreover, of particular
iJ           interest is the formation of several nonaromatic derivatives,
:i           such as chlorocyclohex^ne, a chloroalkyl acetate,  and, perhaps
;           most significant, three chlorinated  acetone derivatives.  The
'•!           latter may be precursors  of chloroform, which was  shown in a
,',           previous work (60) to be  formed  in wastewater chlorinatier.s, and
J           which has been shown by other workers  to result from the chlor-
','           ination of organics in  drinking  waters (51).
I
'-I      •         Finally, the concentrations  of the compounds listed in
!J           Table 1 are in the microgram-per-liter range; hov;ever, it should
;\           be noted that the sum cf  the concentrations shown  in the right"
,!'           hand column of Tabla 7  (786 yg/L)  does not represent the total
            concentration of chlorinated organics.  Rather, this sum is a
            lower limit in view of  the inefficiency of the various steps
            in the analytical procedure, and,  in particular, because the
'-.           procedure does- not detect non-volatile species.

]           Ventura, California 3eptage Extracts                   *
I                                                           *'
I               Figure 15 is the reconstructed total ion chromatogram for
            the extract from a sample of superchlorinated Ventura,
            California septage.   Organics in  the sample were extracted by
            XAD-2 resin previously and eluted with diethyl ether, as
            described previously.  Peaks in the  chromatogrsms that show
            evidence of halogen clusters are  shaded.   Peaks that give a
            response on the Coulson electrolytic conduct!/ity detector are
            indicated by a "C";  and those that have an ion at m/e 35 are
            indicated by the number 35 (cf.  Figure 8).   Figure 15 also
            designates peaks found in the control  (unchlorinated)  sample
                                           45

-------
$•'
                                                           I   i
                                                        TIME
              Figure  12.  Gas  chromatograms  (FID,  16xlO~   amp/fs)  of  Dentcn,  Texas wastewater extract
                          Bottom, before chlorination,;  top,  after  2,000  mg/L  chlorination for
                          1-hour contact period. Analytical  conditions described in text.          !  '
                                                                                                              •'*
                                                                                                              '•ft

-------
f*~""
                    J
                              CONTROL
                                                         TIME

               Figure 13.  Gas chromatograms  (Coulson electrolytic  conductivity detector, halogen mode,
                    X 4) on Denton, Texas wastewater extract.   Bottom before;  top, after 2000 mg/L
                    chlorination  for 1-hour contact time.  Analytical conditions described in text.

-------
                                ^«.^«V--'rZ'^«T--"""'?" V-'-. -.' ^~«*r\v''"555^-%""'^^^^^
                                 •-'//,' .1   '•  *,:'  ~   •  iv    , •' .v-j-v ,  . .. '. ,• j   .  : ' .„.- '/•',-•  -.  ' -,~ - >-a,,~. -r..~j..Vi '^."," > .«-"•>--^.'^
00
                                 -.20 130 110 ISO IGO 110 180 190 200 2 If Z3! 230 2« 235
   Figure 14.  Reconstructed  GC/MS chromatogram of Denton, Texas wastewater extract after
         2000 ppm  chlorine  treatment  for one hour. Vertical line  markers  show the  positions
         of new chlorinated organics  not present in control samples. Peak numbers  corre-  >
   ;_  '    spond to  compounds identified  in Table 7.

-------
TABLE 7.  SUMMARY OF NEW CHLORINATED ORGANICS FOUND
    IN "SUPERCHLORINATED" MUNICIPAL WASTEWATER

Compound „ , „ a
Number Compound Name
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

Chloroform
Dibroraochlorome thane
Dichlorobutane
3-chloro-2-ineLhylbut-l-ene
Chlorocyclohexane (118)
Chloroalkyl acetate
o-Dichlorobenzene
p-Dichlorobenzene
Chloroethylbenzene
Tetrachloroacetone
Pentachloroacetone
Hexachloroacetone
Trichlorobenzene
Dichloroethylbenzene
Chlorocumene (154)
N-methyl-trichloroaniline (209)
Dichlorotoluene
Trichlorophenol
Chloro-a-methyl benzyl alcohol
Dichloromethoxy toluene
Trichloromethylstyrcne (220)
Trichloroethylbenzene (208)
Dichloro-a-methyl benzyl alcohol
Dichloro-bis (ethoxy ) benzene (220)
Dichloro-a-methyl benzyl alcohol
49
Identifi- Concenr
cation tration
status yg/l
f,g
f,9
d,g
f
d,g
d
f
f
e
e
f
f
f
f
d,g
d,g
e,g
e
e,g
e,g
d,g
d,g
(190) d
d,g
(190) d

-
-
27
285
20
-
10
10
21
11
30
30
-
20
-
10
-
-
-
32
10
12
10
30
-


-------
                     (TABLE 7.—Continued)

, Identifi-
Compound a cation
Number Status
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Trichloro-IJ-methylanisole •
Trichloro-i-methyl benr.yl alcohol
Trichloro-a-methyl benzyl alcohol
Tetrachlorophenol
Trichloro-a-methyl benzyl alcohol
Trichiorocumene (222)
Tetrachloroethylstyrene (268)
Tr ichlcrodimethoxybenzene (240)
Tetrachloromethoxy toluene (258)
Dichloroaniline derivative (205)
Dichloroaromatic derivative (249)
Dichloroacetate derivative (203)
Trichlorophthalate derivative (296)
Tetrachlorophthalate derivative (340)
e,g
e
e
f
e
d
d
d
d
c
c
c,g
C
c
Concen-,
tration
pg/i
-
25
25
30
50
-
-
-
4-
13
15
20
-
—

aCompounds may be listed more than once if GC retention times
 indicate distinct positional isomers.
 Quantitative values should only be considered as estimates
 since response factors and recovery data are not available
 for our extraction system.

°Mass spectral information too incomplete to propose a structure;
 probable molecular weight indicated in parentheses.
 Fragmentation pattern tentatively suggests proposed compound;
 probable molecular weight indicated in parentheses.
eProbable identification based on mass spectral interpretation.
 Completed identification based on MS interpretation and
 confirmed by comparison with a reference spectrum.
Identified in runs other than 11-12-74  (Figure 14).
                              50

-------
50
ISO  200  250  300  350 • 400  450   500  550  6CO  650  700  750
                              SPECTRUM   NUMBER
                                                                    803  850  9
                                                                         95C  1033
Figure 15.   Reconstructed GC/MS chromatogram of superchlorinated septage
                   extract  (Reference Table  8 and ttixt) .

-------
|#:           with an "R".   Compounds identified by a priori interpretation
\l-            of their mass spectra,  or comparison wTth standard reference
!|            tables, are numbered in Figure 15 and listed in Table 8.
             Numerous spectra indicate chlorine isotopic clusters but cannot
             be identified.   Also shown in Figure 15 are notations of the
;             retention times of n-hydrocarbons C-JQ  C-j^, C^g,  C2Q, ^22*
!-.'           and C25» in order to give one a sense'of the temperature program
i'            used." The program was  essentially the-same as the one described
I             previously in this section.
! '•
|                  Purge and trap (P/T) analysis of the septage supernatant
'             was also conducted, using a modified P/T apparatus in conjunc-
: •            tion with the GC/MS system.   Figure 16 shows the chromatogram
If           of a superchlorinated sample using this method.  The principal
ji;           features of the purge and trap modification are the use of a
| -        .    long, narrow tube to contain the water sample  (ca. 20 ml)
ij            during purging, and the swfestitution of Chromosorb 102 as the
 ;         .   GC column material.  Figure 17 shows a six-conpound standard
 I   '        used to evaluate the procedure.

j »'                                                        —
                  TABLE 8.  COMPOUNDS IDENTIFIED IN VENTURA, CALIFORNIA,
                     SUPERCHLORINATED SEPTAGE SUPERNATANT, XAD-DIETHYL
                              ETHER  EXTRACT (REF. FIGURE 15)
Code
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Compound
Chloroform
Dichlorobc-rzene'
Tetrachloroacetone
Dichlorobenaene
Chlorobutandione
Pen tachloroace tone
Dichlorotoluene
Trichloro toluene
'Spectrum
Number
40
310
317
324
337
348
405
484
pg/L
(a)
67
317
179
-
95
24
37
                Quantification not attempted due to losses in workup.

                Tentative identification.
                                            52
f ;
i I

-------
ui
00
               ICQ   150   200   250   300   350   400   450   500   550    600
                Figure  16.  Reconstructed GC/MS chromatogram of purgeable
                     organic compounds in superchlorinated septage.

-------
                                                                                              \
                   100
                         50    100   150   200   250   300   350   400   450   500   550   600

                    Figrre 17.  Reconstructed GC/MS chromatogram of purgeable reference compounds.
  •I
/i

-------
 LCS Analysis  of  GC/MS  Data—
      The  computerized  analysis   of  the  data  shown  in  Figure  15
 was hindered  by  the  poor  quality of the raw  spectra.   This was
 partly  due  to the  complexity  of  the sample matrix  and partly
 due to  a  basic shorccorning  of che F-inniqan 6000  data  system.
 The system  acquires  data  for  each mass  using a single,  fixed
 interpretation to  allow reasonable  precision when  measuring  a
 pe.v.k  of low absolute intensity.   The  fixed time  period limits
 thu dynamic range  of the  mass spectral  peak  intensities which
 the spectrometer can record before  reaching  a saturation point.
 This  in turn  limits  the range of analyte  concentrations which
 will  produce  accurate  representative  spectra.  The concentra-
 tions of  analytes  in the  CALCLI  sample  cover a range  of at
 least four  orders  of magnitude.   This clearly exceeds the
 dynamic range of the data system.

      One  way  to  get  usable  spectra  for  all the analytes in
 such  a  sample is to  make  several GC/MS  runs  of the same sample
 at different  dilutions.   Mass spectra of  analytes  of  high
 concentrations can be  selected from the GC/MS run  of  the
 diluted sample;  mass spectra  of  the trace components  can be
 produced  from the GC/MS run of the  sample at higher concentra-
 tion.   Another alternative  is to concentrate the sample until
 spectra for the  smallest  GC peaks are interpretable,  and then
 select  spectra from  the sides of GC peaks whose  apical  spectra
 contain saturated mass spectral  peaks.  Thus, the  analyst
 can adjust  for the saturation problem of  the larger GC  peaks
 by the  proper selection of  the representative mass spectrum.
 In situations where  the quantity of sample is limited,  the
 latter  procedure is  preferable,  and this was the procedure
 which was used to produce the CALCLI data set.

      Unfortunately,  the gross distortion of  some mass  spectra
 due to  saturated mass  spectral peaks can cause problems  for  a
 computerized  data analysis  system which is forced  to  Analyze
 all spectra as though  each  were  produced within  the dynamic
 range of  the  instrument..  Nevertheless, the  data xvere analyzed
 using the LCS  computer program to search  for ion fragments
 containing  two,  three, and  four  chlorines.   For  all runs  the
 precision estimate parameter  was  0.1, and the variation
 estimate parameter was -50.

      Figure 18 shows LCS chromatograms  for the portion  of
 Figure  15 between spectrum  numbers  290  and 510 (referred  to
 hereafter as  the CALSTI data  set).  It  is seen that the  total
 ion chromatogram is  simplified by the limited cluster search
 procedure.  In the LCS for  two chlorine atoms (middle trace)
 one sees a  resolution of the  overlapping peaks in  the 300-330
 scan number region,  the simplification  of the region around
 350, and the  confirmation of  halogen content  in the region
between 450 and  490.

-------
                                       I'l " " ' I 'I • " I "'' • I
                                      350     400     450
                                              CRL. PFX. SMPLE.: XRD, CHLORD.
                                            LTD. CLSTR. SRCH.:  2 CL'S
                                                         (81
                                       ii I'lVi i | i | i I i i i l» I i I "V''l i I ' I—I ij '
                               300     350     400     450     503
                              CflLCLl           CRL. PFX. SUPLE.:  Xfffl. CHLORD.
                                            LTD- CLSFR. SRCH-: 3 a'S
                                      350     400     450    500
            Figure 18.   Total  ion chromatogram (top)  and LCS  chromatograms
                          for two  chlorines (middle) and  three  chlorines
                          (botto.'n)  for CALS^i data set.

                                                56
L.

-------
     Note that  (3),  (6) and  (8),  tetrachloroacetone,  penta-
chloroacetone and  trichlorotoluene,  are  intense  in  the
trichloro- and dichloro-  cluster  searches, while dichloroberizene
 (2)  isomer fades in  the trichloro-search.  Two other  dichloro-
compounds  (4) and  (7) appear  also in the trichloro-search but
an analysis of the TIC suggests   that  overlapping,  polyhalo-
genated peaks may  be present..  In the  high scan  number  region,
peaks  (a),  (b),  (c) and  (d)„  not  identified  as yet, probably
contain at least three halogen atoms each.


     Further analysis of  the  CALt'TI  data set has shown  that data
 interpretation assisted by LOS ii.:rt=ases the number of  halogen-
containing spectra identified for e\ery  halogen  combination tried.
While  the LCS program does not give  a  dramatic simplication of the
TIC  chromatogram,  it would seem to be  helpful in identifying
peaks with halogen content for further study.  A more detailed
analysis of the characteristics of the-program suggest  that
further refinements may improve its  usefulness for  the  analysis
of complex mixtures for chlorinated  organius,

Microcoulometric Analysis of  XAD-Resin/Ether Extracts—
     The ether XAD-2 extracts used in  the GC/MS  analysis  were
also analyzed by a pyrjlysis/microcoulometric procedure (71)
which converts organic halogen to halide ion which  is titrated.
The detected value for the raw septage supernatant  was  85
pg/L as chloride.  After  superchlcrination the supernatant
showed a value of  870 yg/L.   Given the fact  that the  XAD
procedure does not quantitatively adsorb or  descrb  all  of the
organics in the sample, the observed, value indicates  a  high
level of organic halogen  in the chlorinated  supernatant.   A
similar procedure was applied to  a sample of combined sludges
from an east coast waste  treatment plant.  The supernatant of
the sludges showed an organohalogen  level by the XAD/ether
method of 39±4 ug/L  (two  determinations).  After two  and  four
hours of contact time with a high level  of chlorine  (>1000
pg/L), the sludge supernatant showed values  of 414+1  and
507±4 ug/L, as chloride..

     These data illustrate that total  organically bound halogen
is often much higher than the sum of volatile organohalides
which may be determined by GC (69, 70).  Moreover, the  high
levels suggest that caution should be  used in the disposal of
"superchlorinated" sludges and other waste products,  particularly
when leaching into ground water may  be possible.

Interpretation of Volatile Chlorinated Organics  Formed  from
Superchlorinatioii of Wastewater Products—
     It is now well established that chlorination of wastewaters
and other municipal waste products produces  new  chlorinated
organic compounds.   Tables 7 and  8 list  some of  these compounds;
others are listed by Glaze et al.   (36,  60, 75)  and Jolley'

                              57



-------
 (9, 61, 62), but many others remain unidentified  at this  time.
The precursors of these new substances and their  effects  on  the
environment  (including man) are unknown at this time.  The former
may prove more difficult to determine, primarily  because  of
our ignorance of the rcolecular types present in municipal
wastes.  Undoubtedly, these wastes are extremely  complex  and
one may never know wit'i surety the source of .the  new chlorinated
compounds that are lis:ed in-Table 2 and io the other references
mentioned.

     It is striking to note, however, that many of these  newly
formed halocarbons are similar to those found from laboratory
chlorination of humic and fulvic acid  (51, 52, 53, 54).
However, in the case of municipal wastewater and  septage, the
subjects studied in this project, one would not expect humic
substances to be major components.  That the by-products  observed
in this work are structurally similar and in many cases identical
to the compounds foind by Rook and by Christman et al. from
the chlorination of humic material and surface wa~ter~~organics
may not indicate identical precursors.  Rather, they may  siir.ply
reflect the  fact that similar organic structural units are
found throughout the biological world.

     Three types of compounds listed in Table 7 deserve special
mention:  the polychlorinated acetones, 3-chloro-2-methybut-l-
ene, and several chlorinated alkyl benzenes.  The latter  include
chlorinated benzene, toluene,  ethylbenzene, and cumene isomers.
Direct chlorination of the corresponding aromatic compounds  may
be possible, but is unlikely since some type of activating
substituent is usually required for facile chlorination in
aqueous systems (10).  Mere likely, these neutral chloro-
aromatics result from the decarboxylation of the corresponding
aromatic acids that either are originally present in the waste
product or are formed during the oxidation/chlorination process.
It is not clear whether chlorination occurs before or after
oxidative degradation; however, Larson and Rockwell (93)  have
shown that chlorination of p-hydroxybenzoic acid  (I)  gave
4-chlorophenol:
                      HOC!
                                                 Cl
                              58



-------
The absence of phenol, 2-chlorophenol, or 2,6-dichlorophenol^
suggested to the authors a two-step process with chlorination
as the initial step.  Vanillic acid  (II) yielded 4~chloro-2-
methoxylpheno  (III), apparently by a similar process (93).
Recently, Sievers and co-workers  (94) have observed the forma-
tion of toluene and several other aromatic hydrocarbons from
the chlorination of municipal wastewater, arid saturated aliphatic
hydrocarbons from the ozonation of the same water.  The latter
have been shown to occur as the result of cleavage ancl decarboxy-
lation of oleic acid  (95).
                        OCH
OCH.
     The major product observed in this work, 3-chloro-2-
methylbut-1-ene,  may be related to the chloroisopentanol
observed by Rook (51).   Both are isoprenoids and most probably
result from the degradation of some aliphatic component of a
humic-like polymer.  Likewise, the polychlorinated acetone
deri natives may arise from a cleavage of aliphatic side chains
on the polymer before or after partial
                              59

-------
                                       SECTION 6

                       CHLORINATION OF AMINO ACIDS IN MUNICIPAL
                                   WASTE PRODUCTS
           INTRODUCTION

                The purpose of the experimental investigation described in
           this section is to determine to what extent chlorinated
           compounds are formed during the aqueous chlorination of amino
           acids in municipal waste treatment planx.3.  Chlorination of
           amino acids through the use of various chlorination agents has
           been studied by several previous workers ( 96-110 ), but no docu-
           mentation currently exists of these reactions in the treatment
           of waste products which may contain amino acids or polypeptides.
           The relevance of this research may be inferred from recent
           toxicological data on one halogenated amino acid, namely,
           3, 5-dibromotyrosine (I).
|           Fortunate  111  has shown that this compound inhibits the initial
I           synthesis of thyroid hormones from inorganic iodide with
-,           resulting effects on the hormone concentration in both the
!           thyroid and vascular spaces.  One may postulate detrimental
I           effects from other halogenated amino acids which may be formed
!           in water treatment and discharged into the environment for
•           possible consumption by man or other species.

•                This work has examined analytical methods for the identi-
.;           fication of. amino acids in aqueous solution, and has applied
{           these methods  to  the determination of amino acids and their
I           chlorinaticn products in chlorinated municipal waste products.
|           Particular e.ttention has been given to a study of

I
                                          60

-------
"superchlorinated" sewage sludges and other products, it being
assumed these extreme conditions would favor the formation of
new halogenated products.  Superchlorination as used in this
context means the use of high concentrations of chlorine
(0.2-0.4%) for the oxidative stabilization of sludges and other
sewage products.

SUMMARY OF PREVIOUS WORK

     Langheld (96) allowed a-amino acids to react with sodium
hypochlorite solutions and found that an aldehyde, carbon
monoxide, and ammonia were produced.  Dakin ',97) showed if one
equivalent of the N-chloro compound, chloramine-T  [N-chloro-o-
toluene-sulfonamido) sodium] was allowed to react with a-amino
acids, the corresponding aldehydes were produced; and if two
equivalents of the reagent were used (99) the corresponding
nitrile lesulted.  Wright (100) showed that sodium hypochlorite
when reacted with glycine produced complete chlorination of the
amino group, forming dichloroamino acetic acid (II).
                        Cl
Cl
                                    OH
                              I!
Norman (.101) studied the action of sodium hypochlorite on
glycine and expressed doubt as to the formation of chlorinated
intermediates in the oxidation of glycine by hypochorite;
however,  Wright (102) provided further evidence for the formation
of chloroamino derivatives as intermediate products from the
oxidation of amino acids by hypochlorites except in extreme
alkaline  solutions.  He also found that aldehydes w^re formed
from several amino acids when reacted with sodium hypochorite
and that  nitriles formed as the predominant product at lower
pH values:
              H O
              i  ii
            R-C-C-OH
              i
             NH2
 3 NaOCl-
RCH + RCsN
(1)
Ingols (103-4)  stated that when hypochlorous acid reacts with
alanine,  pyruvic acid is formed by oxidative deamination:
                              61

-------
1
                          H O

                      CH3-C-6-OH  +  5 HOC1	 ***"—


                           2                  00
                                              IIII
                                        2  CH3CCOH  +  N2 + 3  H2O +  5  HC1
             Aleksiev (105)  showed that 3,5-dichlorotyrosine (III)  is  produced
             when L-leucyl-L-tyrosine or the hormone oxytocin reacts  in
             85% formic acid with chlorine  saturated carbon tetrachloride  at
             -20°C.  However, no spectroscopic confirmatory data was  given.
             Kantouch and co-workers (106)  allowed sodium hypochlorite to
             react with a few a-amino acids and found that most of  the amino
             acids reacted quickly causing  oxidation and/or formation  of
I            chloro derivatives.  At pH 2,  substitution occurred mainly by
             forming the N-chloro and N-dichloro derivatives.  By acylation
             of the amino group, the oxidation rate was reduced.



                                  0

                          CHo-CH-C-OH         HO-COVCHgCH-C-OH

                                                               NH2


r
                            in                              iv
             It was stated that 3-chloro (IV)  and 3,5-dichlorotyrosine (III)
             were produced from tyrosine at pH 2 and could be detected by two
             dimensional paper chromatography, but no spectroscopic evidence
             was cited as proof of structure.   Pereira and co-workers  (107)
             studied the reaction of hypochlorous acid with several a-amino
             acids and found that L-phenylalanine,  reacted with hypochlorous
             acid to produce phenylacetaldehyde and phenylacetonitrile.
             Glutamic acid reacted with hypochlorous acid to produce 0-alde-
             hydopropionic acid.  Cysteine reacted with either one or  two
             equivalents of hypochlorous acid  at room temperature to produce
             cysteic acid or cystine.   L-tyrosine produced 3-chloro (V)
             and 3,5-dichloroaldehyde (VI)  along with the 3-chloro (VII)
             and 3,5-dichloronitrile (VIII).
                                           62



                                                                                ,*fc*»«k**.W.*>a#

-------
                                    0
                                CH2-CH
                                           CH2-CH
                                        V

                                        VI!
                                CH2C=N
                 It  should be noted that no evidence for the formation of
                 3-chloro and 3,5-dichlorotyrosine was reported in this paper.
                 Chlorination of dipeptides gave the corresponding N,N-dichloro-
                 peptides, with no cleavage of the amide bond reported.

                     Halogenation of compounds similar to the aromatic amino
                 acids have been studied as models of the type of reactions to
                 be  expected with the aromatic amino acids.  Smith (112) reacted
                 aqueous sodium hypochlorite with benzoic acid and obtained a
                 mixture of three isomeric monochlorobenzoic acids with some
                 dichloro acid:
„.- :
                 r    \
                 o
 9   NaOCI
•COH
                     Hopkins and Chisholm  (113) stated that substances in which
                the orientation of substituent groups is most favorable give
                monochloro derivatives in almost theoretical yield when they
                react with aqueous sodium hypochlorite.  These substances
                include vanillin, anisic acid, and piperonylic acid.  Morton (114)
                stated that chlorination by mechanisms similar to aromatic substi-
                tution seems to occur with pyroles and indoles.  Lawson and co-
                workers  (115) found that indole-3-propionic acid reacted with
                                              63

-------
three moles  of N-bromosuccinimide  (NBS) in aqueous media to
produce spirolactonc dioxindole-3-propionic acid with  a bromine
in the 5 position:
                           0
                    CH2CH2COH
                                T 
-------
The addition of water to the reaction mixture apparently is
responsible for benzene ring bromination of indole.  Green and
Witkop  {117; proposed the folloving mechanism for the formation
of oxindole with N-bromosuccinimide (NRS):
                                                   -HBr
                                                     pH 6
Habsrfield and Paul  (118) reported evidence that N-chloro-N-
methylaniline is the intermediate in the chlorination of N-methyl-
aniline by calciun hypochlorite in carbon tetrachloride solvent.
The infrared spectrum of the reaction mixture shows a loss of
the N-H stretching band.  Analysis of a sample which had been
treated with potassium iodide and sodi urn thiosulfate revealed
the principal product to be unreacted N-uiethylaniline.  With
omission of the potassium iodide, the principal products were
found to be o-chloroaniline, p-chloroaniline, and some
dichloroaniline.

Amino Acids in Treated and Untreated Sewage

     Sastry and co-workers  (119) examined sewage effluents from
chemical clarification, mechanical and biological filtration,
septic tanks, activated sludge process and from natural purifi-
cation of flowing sswage obtained from the Indian Institute of
Science sewage works at Bangalore, India and from the city of
Bangalore.  Using a circular paper chromatographic technique for
the analysis of the amino acids, their work showed that raw
sewage contained practically all the essential amino acids.  Raw
srywags obtained from the sewaga works at the Institute was found
to have 0.36 to 1.G1 ir.illigriims of free amino acids and 65.1 to
91.3 milligrams of total amino acids per gram of solid.  Trypto-
phan was not present in the free form and acid hydrolysis de-
stroyed it in the solid material.  The effluents that were
btained by chemical clarification using alum and mechanical fil-
tration were found to contain considsrable amounts of amino acids.

                               65

-------
     Almost all the amino acids from Bangalore sewage were
completely removed as it ran over r. short distance of 1.29 miles
along a natural channel, p?-es^:nab]y due to the presence of
bacteria in the channel.  Vha effluents from the activated sludge
process were also found to he almost free frcn amino acids.

     Painter and Viney (120) indicated that free amino acids were
about 16% of the total amino acids in -whole sewage.  Work by
Subrahmanyam and co-workers (121) showed that the activated
sludge method of sewage purification removed nearly all of the
amino acids.  Raw sewage was reported to contain 86 milligrams
of amino acids per gram of dried solids and after four hours of
aeration only a trace was found.  Six hours of aeration gave no
detectable amino acids.  Cystine, lysine, histidine, and
arginine were said to be present only in trace amounts in the
raw sewage, and no proline was found.  Kahn and Wayman  (122)
obtained raw sewage and sewage effluents i rom Denver, Colorado
wastewater treatment plant which had only primary treatment
facilities.  Other samples were obtained from three other cities
each of which had some forn of the activated sludge process.
The total amino acids found by two-dimensional chromatography
are presented in Table 9.  Phenylalanine, tyrosine, and tryp-
tophan were not found in any of the samples.  Hunter and
Heukelekian (123) showed that wastewater from Highland Park, New
Jersey contained ci particulate fraction which contained 19%
amino acids.  Hans;on and Lee (124) indicated that nearly all the
amino acids in domestic wastewater are present in the combined
state, in disagreement with Kahn and Wayman (122).  Hunter  (125)
compiled the known information about the occurrence of amino
acids in untreated domestic wastewater; his compilation is
presented in Table 10.


    TABLE 9.  TOTAL AWINO ACIDS FOUND IN THE SEWAGE SAMPLES
              FROM FOUR UNITED STATES CITIES ',..22)

Total Free Amino Acids in Raw Denver Sewage:            115 mg/1

Total Combined Amino Acids in Raw Denver Sewage:        165 mg/1

Total Free Amino Acids in Denver Primary Effluent:       30 mg/1

Total Combined Amino Acids in Denver Primary Effluent:   35 mg/1

Total Combined Amino Acids in Chicago Activated Sludqe   , ,,
                                                         10
        Effluent:
Total Combined Amino Acids in Trenton Activated Sludge
        Effluent:                                         5 mg/1
Total Combined Amino Acids in Hamilton Township,
     New Jersey, Activated Sludge Effluent:                5 mg/1
                              66
                                                                        . _,- .4

-------
   TABLE 10.  AMINO ACID CONTENT OF THE  SOLUBLE FRACTION  IN
              UNTREATED DOMESTIC WASTEWATER  (125)
     Amino Acid
Concentration (mg/1)
Cystine

Lysine and Histidine

Histidine

Lysine

Arginine

Serine, glycine, and aspartic acid

Three-nine and glutamic acid

Alanine

Proline

Tyrosine

Methionine and valine

Phenylalanine

Leucine

Tryptophan
   0 - Trace

   Trace

   Present

   Absent

   Trace

   0.02 - 0.13

   0.01 - 0.18

   0.02 - 0.09

   0

   0.06 - 0.09

   0.05 - 0.24

   0.02 - 0.33

   0.06 - 0.28

   Present
Gas Chroraatography of Amino Acids

     The identification and quantification of amino acids by gas
chromatography has been the subject of numerous papers over the
pest several years.  A search of the literature indicates that
the N(O)-trifluoroacetyl-alkyl esters, N(O)-heptafluoro-alkyl
asters, and the trimethylsilyl (TAS) amino acid derivatives have
received the most attention.

     Stalling and co-workers (126) synthesized bis (trims thy Is ily])
trifluoroacetamide (BSTFA) for the silylation of amino acids,
since previous silylation reagents interfered with the separation
of glycine and alanine.  Their work with BSTFA showed that it

                               67


-------
I
b
           had increased volatility and appeared with the solvent front,and
           had greater solubility in some solvents  than former  silylation
           reagents such as bis (trir.lethylsilyl)  acetamide.   The fluorine in
           BSTFA resulted in less silica  deposits and thus  decreased
           detector noise.   No interference occurred with glycine and
           alanine; however, no reproducible chromatographic peaks could be
           obtained for arginine.  Moreover, several of the araino acids gave
           nore than one trimethylsilyl derivative.   The following reaction
           is a typical derivatization of arcino  acids using BSTFA:
            NH-C-COOH   *
                                            /SI(CH3>3        CH3CN
               2                      3

                  R
                                      0

                    (CH3)3Si-NH-CH-C-Q-Si (CH3)3

                                  R
                Gehrke and co-workers (127) made  a study  of  the  BSTFA
           derivatives of the twenty protein  amino acids and  emphasized  the
           ohromatographic separation of the  derivatives, as  well  as  the
           precision and accuracy of the method.   Attention also was  given
           to silylation as a function of time,  reaction temperature  and
           stability of the derivatives.  As  indicated above, several of the
           amino acids, including glycine,  histidine, arginine, lysine,  and
           tryptcphan gave more than one triraethylsilyl   derivative.   The II-
           (O)-trimethyl silyl (TMS)  amino  acid  esters were found  to  be  com-
           pletely stable for a period of five to seven  days  when  stored at
           room temperature in a tightly capped  vial.  Glycine  was the only
           exception and deteriorated in three hours.  Gehrke and  Leiraar
           (12S) reported the silylation and resolution of the twenty  protein
           amino acids on a single 10% w/w GV-11 column  using BSTFA with
           acetonitrile as the solvent.

                One of the earliest works with N-substituted  ester deriva-
           tives of amino acids for gas chromatography was  by Young (123)>
           who reported the N-acetyl-n-butyl  esters of glycine, alanine,
           valine, leucine, isoleucine,  and proline.  Lamkin  and Gehrke  (130)
           studied the N (O) -trif luoroacetyl-n-butyl esters  and  the N(O)-TFA-
           methyl esters of the amino acids.  They found that serious losses
           were involved in concentrating the N (O) -TFA-methyl esters.
           Single chromatographic peaks  were  obtained for all the  common
           protein amino acids as the N (0) -TFA-n-butyl esters except  for

                                          68


-------
tryptophan and arginine.  Tryptophan gave  two peaks and could not
be converted into a. single peak with longer  acylaticn.  The
esterification and acylation reactions  were  given as fellows:
       HO                           HO
        i   H      BuOH                •   ,i
     R-C-CO"	>-  R-C-CO-Bu    *   H00
                 Anhydrous             •   A               *•
                   HC1
   HO             0          CH  C]           HO          0

 R-C-CO-Bu + (F3CC)20  	*- R-C-CO-Bu + CF3COH
   1   +                       Room          j
   ^^3                     Temperature         ^
                                           H-N-CCF3
     Coulter and Hann (131) prepared  the N-acetyl-n-propyl esters
of the amino acids.   They stated  the hydrochloride of histidine
had to be neutralized using lithium  or sodium carbonate before it
could be acylated.   Arginine had  to  be converted to ornithine
while histidine was  either converted to aspartic acid or the
hydrochloride neutralized with  lithium carbonate.  Their proce-
dure required two GC columns for  separation.

     Zumwalt and co-workers (132)  reported work using the N(O)-
TFA-n-butyl esters of the amino acids for the quantitative
analysis of the amino acids in  complex biological substances.
Cation or anion exchange cleanup  procedures were used before
derivatization.  Their work also  reported an optimum molar ration
of 50:1 (trifluoroaceric anhydride to amino acid) maximizes
reproducibility of acylation, stability of the derivatives and
maintenance of small sample volume.   Two columns were again
required for separation of the  derivatized amino acids.  Gehrke
and co-workers (133)  reported a  dual  column system in which
hishidine, arginine, tryptophan,  and cystine would eJute quanti-
tatively and resolve as single  peaks from the other N(0)-TFA-n-
butyl esters. They observed strong substrate-derivative inter-
action with the diacyl histidine  derivative when using OV-120
columns.
                               69


-------
                Moss and Lambert (13$ reported the GLC separation of twenty
           protein amino acids on a single column using the N (0)-heptaf luoro-
           butyryt-n-propyl ester derivatives.  Their work showed that the
           mono-heptaflaorobutyryl-n-propyl derivative of histidine could be
           converted to a diacyl derivative if acet.ic anhydride was injected
           simultc.neously with the sample into the gas chromatograph.  This
           resulted in a sharp single peak for histidine rather than a broad
           peak for the monacal derivative.

                Hardy and Kerrin (135; prepared the trireethylsilyl-n-butyl
           esters of the amino acids and reported that these derivatives
           could be separated in less thin 35 minutes on a lightly loaded
           textare glass GC column.  They also fourd that these derivatives
           gave the best separation when acetonitrile was used as tne
           solvent for trimethylsilylatim.  Close examination of their
           chromatogram showed numerous small peaks which correspond in
           retention time to the fully trimethylsilylated amino acid
           derivatives, suggesting th=it esterification of the amino acids
           was not complete.

                Zanetta and Vincendon (136) reported the preparation of
           N(O)-heptafluorobutyryl-isoamyl esters of amino acids and
           separation on a single column.  Jonsson and co-workers (137)
           studied the N(O)-HFB-n-propyl derivatives with emphasis on
           variation of the esterification, acylation, and chromatographic
           procedures.  They determined the relative molar response, lin-
           earity of response and the stability of the derivatized samples.
           Their findings showed that vr.he two step esterif ication and
           acylation procedure of Moss and Lambert (134) gave the best
           results.  Hiscidine and lysine were shown to require two ester-
           ification periods since they have very low solubility in high
           concentrations of HC1.  The amino acid derivatives were all
           found to be stable for at least three days.

 '<               The use of the mass spectrometer as a highly specific and
: |          versatile detector for the gas chromatcgraph has been documented
'I          by a large number of workers.  Among others, Oro and co-workers
 '          (138) have utilized GC/MS for the analysis of derivatized amino
. f          acids.  The principal advantage of this method is that it allows
J          for specific confirmation of an analysis based on computerized,
|j          mass specific techniques such as multiple ion detection  (139).
\\          Thus, GC/MS is particularly suited for analyses in complex
;jj          matrices such as biological and environmental samples.
 $
 |          EXPERIMENTAL PROCEDURES
' £
 I          Materials and Instrumentation
 ?
 I          Purified Water
 i                                          .
 I               The water used in this work ivas prepared from the tap by
|          passage through a Continental deionizer unit (Model 200), a

                                          70
                                                                                  ISSI^.  ,/

-------
j|2         Calgon  filtrasorb  400  activated  carbon  column and then through a
; 1'         Whatman qualitative  filter.

 \-         Solvents

;.';"              Ethyl  acetate was purchased from Pierce  Chemical  Company,
 t.         Rockford, Illinois.  Acetonitrile -  (Baker Analyzed reagent)  was
           purchased from  3.  T. Baker,  Philadelphia.

. >..         Amino Acids
 /{
 •               Amino  acid standards  were obtained from Jack Graff Associates,
 f?         Santa Clara, California.
•I
 V         Other Reagents
 f.
 :|              Heptafluorobucyric anhydride was purchased from Pierce
 'I         Chemical Company,  Rockford,  Illinois.   Cupric chloride was
 .'I         obtained from Matheson, Coleman  and  Bell,  Norwood,  Ohio.   Dry HC1
 ,|         gas was purchased  xrrom Union Carbide.   Anhydrous sodium sulfite
 s          (Analytical Reagent  grade, Mallinckrodt) was  heated for two hours
 ;|         at 130°C in *n  oven.   Hydrochloric acid,ammonium hydroxide,
 ,5         chloroform  and  sodium  metal  were obtained  from various suppliers
 jj         and taken from  available laboratory  stocks.   The jDrthotolidine-
 •j,         arsenite  (OTA)  reagents for  the  residual chlorine determinations
x. |         were prepared according to Standard  Methods  (140).  The chlorine
 |         gas used to prepare  the hypochlorous acid  was purchased from
 \         Dixie Chemical  Company, Houston,  Texas,  and was claimed by the
 |         manufacturer to contain 100% active  chlorjne.   The  n-propanol,
 |         J-soamyl alcohol, methanol, and n-butanol were purchased from
 I         Fisher  Company, Falrlawn,  New Jersey and were Certified Grade.
 I         All alcohols were  re-distilled in all glass apparatus  after
 I         refluxing two hours  over magnesium turnings,  th^n stored under
           dry conditions  at  low  temperature.

           Resins

                The cation exchange resin,  Dowex 50W-X8  (sodium ion form,
           100-200 mesh) and  Chelex 100 resin (200-400 mesh) were obtained
           from Bio-Rad Laboratories, Richmond, California.

           Gas Chromatography/Mass Spectrometry

                The Finnigan  Model 3200 gas  chromatograph/mass spectrometer
           system  with a Model  6100 data system was used to separate,
           detect, and identify the emino acid  derivatives.  The  system was
           also used to separate,  detect, and identify the reaction products
           of the  amino acids with HOC1.  All test ivixtures and extracts
           were chromatographed using a 5 foot  by  2 mm I.D. glass column
           packed  with 10% SP-2100 on 100/120 Supelcoport (Supelco,  Inc.,
           Bellafontf-, Pennsylvania) .   The  data system was used to quantify
           the-SC  peaks and to separate  any  unresolved peaks using the

                                          71


-------
limited mass search feature,
Spectrophotometry

     A Coleman Model 295 spectrophotometer was used to measure
the color produced in the OTA method for determining residual
chlorine.                -            .         '

Amino Acid Analyzer

     A Beckrnann Model 120-C ami no acid analyzer, wnich is
located in Robert VI. Gracy's laboratory in the Chemistry Depart-
ment of North Texas State University, v;as used in this work.
It was operated under standard conditions.

Glassware

     All glassware was well cleaned using chromic acid cleaning
solution, then rinsed with tap water, deionized water, and
finally with redistilled acetone.  The glassware was then placed
in an oven at 130-150 C for 8 to 10 hours.

Methods of Analysis

     The method used to isolate, concentrate, and analyze free
amino acids in waste waters is shown in Figur^igand is described
below.


Isolation and Concentrationof Amino Acids in Aqueous Solution

     A procedure very similar to that used by Gardner and Lee(141)
was used in this work.  The procedure consists of isolation and
purification of amino acids by a combined ion exchange/ligand
exchange method (Figure 19).

     A glass column (36 cm x 1.5 cm I.D.) vitli standard taper
ground glass joint at the top was slurry packed with 30 cm of
Dowex 5QW-X8 (hydrogen ion form, 200-400 mesh) (141-143).
Glass wool plugs of one centimeter length were used at both ends
of the Dowex resin.

     The ligand exchange column is a modification of that used by
Siegol and Degens  (142). A glass column  (36 cm x 1.5 cm I.D.)
with standard taper ground glass joint at the top and glass wool
plug  was slurry packed with three cm of Chelex-100-NH3 resin.
Twenty seven cm of Chelex-100-Cu-NH3 resin was slurried on top of
the Chelex-100-NH  resin and capped with a 1 cm glass wool plug.

     Three  liter  separator'/ funnels with  standard taper ground
glass joints at the efflusnt end were used to hold  the samples.

                                72


-------
                        SAMPLE
               CENTRIFUGATION/FILTRATION
                   pH ADJUSTED  (2.2)
              DOWEX CATION EXCHANGE  RESIN
               DOWEX EI.UTION  (3  N  NH.OH)
               AMMONIA REMOVAL  (ROTOVAC)
                     CHELEX  RESIN
              CHELEX ELUTION  (2  N  NH4OH)
              ROTARY EVAPORATION  TO 5  ML
              EVAPORATION TO  DRYNESS
              ESTERIFICATION  (n-PrOH/HCl)
                   ACYLATION  (HFBA)
                    GC/MS ANALYSIS
Figure 19.  Scheme for the  separation  of amino acids in
                 wastewater products.

                          73
     a^iflrAfe^^

-------
if-
             Nitrogen pressure applied at the top of the separatory funnel
             was used to control the flew rate at 10 ml per minute.  Waste-
             water samples were treated with excess sodium sulfite to quench
             any chlorine residual, and then thg samples were centrifuged,
             filtered, adjusted to pH 2.2, and passed through the Dowex
             column.  The column was eluted with 100 ml of 3N NI^OH and
             the ammonia removed in a Rotovac evaporator.  The concentrate
             was passed through the Chelex resin column.  The amino acids
             were eluted with 100 ml of 2N amir, on i a solution, and ammonia
             was removed fron the eluate by rotary evaporation at 60°C and
             then transferred to a 5 ml reaction vial where the drying
             continued at 60°C under a gentle stream of nitrogen.

             Derivatization Procedure

                  The amino acid derivatives used in this study are the N(0)-
             heptaf luor obuty ry 1- i soamy 1 and the N (0) -heptaf luorobutyryl-n-
             propyl esters  (134, 136, 137).

                  The N (O) -heptaf luorobutyryl-isoamyl esters were prepared
             using the method of Zanetta and Vincendon  (136) , with some
             major variations.  The esterif ication reagent was triiso-
             amoxymethane which was prepared in the laboratory according
             to the procedure of Gilman (144).  One ml of the esterif ication
             reagent was added for each 10 uroles of dried amino acids in
             a reaction vial.  Fifty ul of concentrated HC1 was added and
             the via] capped and heated in a sand bath at 110°C for one
             hour.  The esterif ication reagent was then evaporated under a
             stream of nitrogen at GO°C.  One half milliliter of acetonitrile
             was then added along with 100 microliters of HFBA for each
             10 umoles of amino acid.  The vial was capped and heated in a
             sand bath for 10 minutes at 150°C.  The derivatized amino
             acids were dried under a gentle stream of nitrogen at room
             temperature, and dissolved in an appropriate volume of ethyl
             acetate for GC/MS analysis.

                  The N (0) -heptaf luorobutyryl-n-propyl esters of the amino
             acids were prepared in a reaction vial by adding 3 ml of 8M HC1
             in n-propanol.  The reaction vial was capped and heated in a
             sand bath for 10 minutes at 100°C.  The esterif ication reagent
             was evaporated under a gentle stream of nitrogen.  The propyla-
             tion procedure was repeated and the n-propyl esters were dried
             under a gentle stream of nitrogen at 60°C.  After drying, 1 ml
             of acetonitrile was added with 1 ml of HFBA.  The vial was
             capped and heated at 150°C for ten minutes in a sand bath and
             then cooled to room temperature.   The derivatized amino acids
             were then dried under a gentle stream of nitrogen.  After dry-
             ing, the remainder of the procedure was the same as for the iso-
             amyl derivatives.
             Gas Chromatography/Mass Spectroscopy Analysis

                  The GC/MS conditions utilized for amino analyses were as
             follows:
                                            74


-------
,f
                     Mass spectrometer sensitivity

                     Electron energy .

                     Injector temperature

                     Column temparature

                          Initial

                          Final

                     Temperature programming rate

                     Carrier gas

                     Gas flow
10~  ams/volt
70 eV

240°C
 50°C

280°C

  4 /min
Helium

 20 ml/min
                     Identification of the amino acids was confirmed using known
                fragmentation iiiechanisms  (L38) .  Quantification was accomplished
                by comparing the samples against a standard mixture of 20 amino
                aoids of known weight.  Any unresolved peaks were quantified by
                limited mass search to determine peak areas.  The limited mass
                search procedure allows one to obtain a chromatogram repre-
                senting the response for a particular mass fragment.

                Reaction of Some Amino Acids with Hypochlorous Acid
                                     "**
                     Serine, threonine, alanine, valine, tyrosine, and tryptophan
                were chosen to test the reaction of amino acids and HOC1.  The
                HOC1 solutions were prepared by bubbling chlorine gas into
                organic free water and checking the HOC1 concentration by the
                OTA method ( 140) ,  One ml of a 20 ymole/ml solution of each of
                the above amino acids except tryptophan was placed in separate
                5 ml reaction vials and adjusted to pH 1 to 2 with 6N HC1.  One
                ml of 2000 mg/1 (28 umoles Cl2/ml) of aqueous chlorine war added
                to each vial.  The vials were capped, shaken, and allowed to
                stand for 30 minutes at room temperature.  Controls were also
                run on each of the amino acids.  The control consisted of 1 ml
                of the amino acid solution and 1 ml of the water used to make the
                HOC1 solution.  After 30 minutes, each vial was extracted by
                adding one ml of ether, capping and shaking vigorously by hand
                for one minute.  The ether extract was analyzed by GC/MS for
                identification of any ether soluble reaction products with the
                only change in GC/MS conditions being the initial temperature of
                the column  (30 C) .  Three microliters of the aqvieous layer were
                also run by GC/MS using direct aqueous injection and the condi-
                tions previously listed.  The remainder of the aqueous solution
                was dried under a stream of nitrogen and the volatile n-pronyl
                derivatives of the amino acids were prepared and analyzed by
                GC/MS .
                                               75

-------
      A solution which contained 76.8 mg (376 yinoles)  of trypto-
 phan was  reacted at ambient pH with 3.3 ml of a HOC1  solution
 which contained 13.3 mg (188 umoles)  of chlorine for  30 minutes
 at room temperature.  The  reaction mixture was then extracted
 with ether,  using the same procedure as above.   The ether extract
 which contained a red colored product was  analyzed by GC/MS
 under the same  conditions  as previous ether extracts.   The
 ether material  did not yield any gas chromatograph peaks.   The
 amino acid derivatization  procedure was carried out on the red
 product in ether and the derivative mixture analyzed  by GC/MS.
 The aqueous  portion of the reaction mixture was dried under a
 stream of nitrogen and derivatized using the procedure described
 above.  In both cases N(0)-heptafluorobutyryl n-propyl esters
 were prepared  (see below).

 Analysis  of  Sewage Sample

      Sewage  product samples from four cities were  analyzed for
 amino acids.  The cation and ligand exchange procedure was used
 to concentrate  the amino acids as described in  Figure  19.
 Derivatives  were prepared  using  the HFB7\ acylation and esterifi-
 cation  procedures as previously  described.   City A sample
 consisted of 100 ml of raw sewage.   The sample  from City B was
 25 liters of anaerobic digestor  supernate  which had been "super-
 chlorinated" at the City plant.   This sample was quenched  on
 site with sodium sulfite to remove  residual chlorine.   The
 sample  from  City C consisted of  combined primary and  secondary
 sludges which were also "superchlorinated1   at the  City's plant.
 Two liters of this sample  were quenched with sodium sulfite after
 two hours.   Another two liters of the sample from  City C was
 quenched  with sodium sulfite after  four days.   Samples from
 City D were  one  liter of "superchlorinated"  septage which  was
 chlorinated  at  the City's  plant  and ono liter of raw  septage.
 Both samples were  quenched  on  site  with sodium  sulfite.  Any
 particulate  matter was  removed from the samples  by filtering
 through 50 an dia.  Whatman  prefojded  filter paper  or centrifuga-
 tion and  filtering.

 RESULTS AND  DISCUSSION

Analysis  of Amino  Acid  Standards

     Analysis of the  amino  acids  was  first  attempted using  the
procedure described  above in which  the  isoamyl esters  were
prepared  using triisoamoxymethane,  followed  by acylaticn with
HFBA.  It was found  that histidine  and  tryptophan  did  not
derivatize by this method,  and that arginine  gave  very  low
yield's.   It was decided then to use the method of  Moss  and
Lambert (134);  that  is, esterification  with  n-propyl alcoho]
which was 8M in HC1,  followed by  acylation with  HFSA.
                              76

-------
     Since cysteine is partially oxidized  to cystine  (134,
136) by this procedure,  the quantification of  cysteine  or
cystine must be taken with reservation.  Moreover,  the  analysis
of histidine showed reproducibility problems  (133)  since the
compound was partially destroyed during derivatization.

     Analysis of the derivatized amino acids can be done by
GC alone; however, interfering peaks  (141)  or  poor  resolution
severely limit this method.  Since the GC/MS system can be
used to separate unresolved or interfering peaks by the limited
mass search procedure and identification of the amino acids
can be made using known  fragmentation mechanisms  (138), it
was the system selected  for this work.  Fragmentation patterns
for aromatic and aliphatic araino acids are shown in Figures
20 and 21 (133).  A total ion chromatogram of  the 20 amino
acids derivatized by this method is shov.-r  in Figure 22.

Minimum Detectable Limits of the N(O)-Heptafluorobutyryl
Alkyl Esters by GC/MS

     Since the minimum detectable limit  (MDL)  of the GC/MS
system for the derivatized amino acids was not known, a study
was performed to determine thia limit.  These  studies were
performed using the HFB-n-propyl and HFB-isoamyi esters of
the amino acids.  A sensitivity setting of 10~'  amps/volt
on the GC/MS system was  used since the 1C"8 amps/volt setting
produced too much noise.  Constant volumes of  solutions of the
derivatized amino acids  at different concentrations were
run on the GC/MS system  and a signal to noise  ratio of  2 was
used as the criterion for minimum d
-------
R f
 ''
(M-300)


       -CO



0CH=CHC=0

(M - 272)


       -OH
                   0
                   it
               (M-255)
                                   J2TCH = CHNH

                                    (M -284)
                                       I-AH
                                    (M - 87)
                                        -C02E
                                             0
JST-CH2-CH-C
         I
        N
       /  \
     H     A
       (M)
                                            \
                                             0-E
                  |-NH2A
                                        it
                                        0
                                    C7H7
                                    (m/e  91)

                                (ring expansion)
               (M-213)
          r'ig ire 20.
                                  A  =  C3F?C

                                  E  =  C3H?


Fragmentation of aromatic amir.o acids (N(0)-
hoptafluorobutyryl propyl esters).  A]1 species
have +1 formal charge.

                78

-------
                      ANH=CHCO
                      (M/e 253)
                  -H20
           ,OH
ANH=CH-C''
(M/e-271) ^
   t
     -(E-H)
      OH
(M/e 313)
           -ANH2>
        CH=R
       (M-300)
                         -R
                                       -CO
A-N-CHCO
  H R
    (M.-59)

       -OE
                           X
                               (R-H)
 I  i
H  R
  (M)
                               OE
                        J-
                         C02E
                       H   R
                     (M-87)
                       i-(R-H)
                    ANH==CH2
                    (M/e 226)
                                         ANH=CHCO+H
                                          (M/e 254)
                                     -H20
                        H R
                       (M-41)
                                                 OH
                                        A =  C3F?C=0
                                        E '  C3H7
Figure 21.  Fragmentation of aliphatic amino acids
                         79

-------
      N  PROPYL RR STD
      130
03
O
                                                         m
                          100
                     150
200
                                               ' ' ' ' ' ' ' I ' [ ' I I I I I I I ' I I I I M I I I I I i I I I I I M
250
300
350
   Figure  22.
Reconstructed GC/MS  chromatogram of amino acid standard:  a-alanine-
b-glycine; .c-valine;  d-threonine; e-serine; f-leucine; g-isoleucine'
m'nh^Ti  17c^steine'  J-methionine; k-aspartic acid; 1-uHkFown impurity;
Tlrai^    ^  ":°TglUtamiC aCid' ornithine; p-lysine; q-tyrosine;   ^
r arginine;  s-histidine; t-tryptophane; u-phthalate. •

-------
TABLE 11.  MINIMUM DETECTABLE LIMIT OF N (0)-HEPTAFLUOROBUTYRYL
                ALKYL ESTERS OF AMINO ACIDSa
Amino Acid
Alanine
Glycine
Valine
Threonine
Serine
Leucine
Isoleucine
Proline
Cysteine
Tryptophan
Hydroxyproline
Mechionine
Histidine0
Phenylalanine
Ornithine
Lysine
Tyrosine
Aspartic Acid
Arginine
Glutamic Acid
MINIMUM DETECTABLE LIMIT (nq)
n-Propyl Ester
Run 1
3.6
2.4
4.1
1.2
1.1
4.0
3.9
4.0
3.2
17.0
4.1
6.9
151
1.4
3 7
4.4
4.5
4.4
7.0
5.8
Run 2
4.4
2.8
3.2
1.6
1.3
3.8
3.6
4.4
3.3
19.5
3.7
6.6
167
1.6
3.9
4.0
5.4
3.8
7.3
6.3
Run 3
3.9
2.6
2.8
1.5
1.2
3.7
3.5
4.6
3.9
20.0
3.5
6.9
182
1.5
3.8
3.9
5.2
3.7
8.7
6.5
Avg.
4.0+0.3
2.6+0.2
3.4+0.5
1.4+0.2
1.2+0.1
3.8+0.1
3.7+0.2
4.3+0.3
3.5+0.3
18.8+1.2
3.8+0.3
6.9+0.2
.167^10
1.5+0.1
3.8+0.1
4 . 1+0 . 2
5.0+0.4
4.0+0.3
7 . 7+0 . 7
6.2+0.5
Isoamyl Ester
10.8
13.5
10.8
2.6
4.0
3.1
5.6
2.1
18.9
2.8
6.3
—
8.9
7.9
8.5
7.7
6.5
9.5
    GC/MS conditions as indicated in text
    Derivatization as isoamyl ester does not recover tryptophan
   GDerivatization as isoamyl ester does not recover histidine
   ^Arginine not in standard for isoamyl ester
                             81

-------
     The reaction products of HOC1 and amino acids in general
were a confirmation of previous work  (96-107).  Major exceptions
are the chlorotyrosine and the dichlorotyrosine which had not
been confirmed before in reactions of HOC1 and tyrosine.  The
oxindole derivatives of tryptophan were also very unusual since
these compounds have not been reported by other workers and
the fact that further carbon-chlorine bonding was shown.

     The mass spectra of N(O)-HFB-n-propyl esters of mono and
dichlorotyrosine are shown in Figures 23 and 24 respectively.
Figures 20 and 21 rationalize the major peaks in terms of a
plausible fragmentation pattern.  The spectra in Figures 23
and 24 actually were obtained on compounds found in the sample
from C after four days of "superchlorinatibn" but are identical
to those of products found in the laboratory chlorination of
tyrosine.  Peaks at m/e of 436, 394, 377, 349, and 337 for
chlorotyrosine and m/e of 470, 428, 411, 383, and 372 for
dichlorotyrosine indicate that chlorination is on the aromatic
ring, presumably at the 3 and 5 positions, but the precise
position cannot be ascertained from mass spectrometry alone.

Cation and Ligand Exchange Recovery Studies

     To determine the recovery efficiency of 20 amino acids
from the cation and ligand exchange procedure described earlier
in this section, 50 yliters of a standard amino acid solution
at the 10 ymoles/ml concentration were spiked into two liters
of water and passed through the exchange procedure.  The final
effluent was dried and derivatized as the N (O)-heptafluorobu-
tyryl n-propyl esters as before.  Fifty pinoles of the same
mixture were derivatized along with the recovery samples as a
control.  The recovery efficiencies from the runs listed in
Table 13 vary from 101.9% for glycine to 59.2% for hiotidine.

Comparison of GC/MS Method for Amino Acid Analysis with a
Beckmann Model 120-C Amino Acid Analyzer

     To compare the two methods, using a wastewater matrix,
two liters of the final effluent from a city sewage plant wr>re
obtained.  The final effluent was quenched with excess sodium
sulfite to remove residual chlorine and then filtered.  The
two liter sample was spiked with a standard mixture of 20
amino acids at the 5 nmoles/ml level.  Two one ml samples of
this solution were run on the .T3eckmann Model 120-C amino acid
analyzer and the remainder of the sample was split and analyzed
by the GC/MS procedure.  The cation and ligend exchange for
clean up and concentration of the amino acids, and the n-
propyl esterification and HFB acylation procedures described
earlier in this section were used.

     In general, both methods gave good results, as. can be
seen in Table 14.  The amino acid analyzer gave a low value

                               82

-------
                          J;jf^1;f , *%rf^!Pf^£J$fffi;-,!*' ^^77-on^^^^tSs^Jf^'t-*^^
  TABLE 12      REACTION   PRODUCTS   IDENTIFIED   FROM


  THE   REACTION  OF   AMINO  ACIDS   WITH   HOC!
AMINO
ACID
SERINE
ALANINE



VALINE


THREONINE
LEUCINE
PRODUCTS FN
ETHER EXTRACT
NONE
CH3CHN
0
it
CH3CH
(CH3)2CHC = N
0
II
NONE
(CH,.)_ CHCH.C5N
PRODUCTS IN
AQUEOUS SOLUTION
SERINE
ALANINL



VALINE


THREONINE
i Fiiriwp
|
__i








 PHENYLALANINE
           9

 (CH3)2CH-CH2CH


„.. _ = ll             PHENYLALANINE
                                               1
 TYROSINE
                    Cl
                   TYROSINE


                   A2       2
                HO/O)-CH2-CH - COH
TRYPTOPHAN
                   r~\    °
                 HO
-------
ps


00


30
00
'80
O'J

L
30
rn

316



50

3100



550
^




,1 I





i

g^MMfi&!





100
1, L . ,L
350



600
^^




	 ll,

'S ^










A
150

4^0

IE

1 "
650





I,,







•w-^w™-**



U
1. n |
200 250
IE
,1 , - -
'450' ' ' 500 '




' 	 :;„ ., 1
i


-


)





Figure 23.  Mass spectrum of chlorotyrosine.
                      34

-------


tt
1C
1
1
3
30
30
00
280

36
50
3bo'

t lL.iL.UJL
100
350
f.,1 . 41 ( 1^1
150
400


10
r,|l j, ^JL. — L. 	 L 	 Lu.

200 250
1C
I 1 - „ 1,
J
3
450 ErOO
too
                 10
530 550
600
 Figure  24.   Mass spectrum of dlchlorotyrosine.




                        85

-------

TABLE 13. RECOVERY
; COMBINATION OF

^^W^^^PP,^
. . . _ 	 	 j, . 	 , . . ^j.- -,.... •S.-..T' * -&L i''* 'S
EFFICIENCIES OF 20 AMT.NO ACIDS BY THE
THE CATION AND LIGAND EXCHANGE RESINS

Amino Acid
Alanine
Glycine
Valine
"" Threonine
Serine
Leucine
Isoleucine
Proline
Cysteine
Hydroxyproline
Methionine
Aspartic Acid
Phenylalanine
Glutamic Acid
Ornithine
Lysine
Tyrosine
Arginine
Histidine
Tryptophan
Average
Per cent recovery
88.9
101
95
93.9
93.4
94.7
91.2
91.7
91.8
95.1
94.6
99.4
93.5
82.8
90.6
81.3
89.3
80/6
69.2
82.4
± 0.2
± 2
+ 1
+ 1
+ 0.4
+ 0.3
± 2
+ 1
+ 2
+ 2
+ 2
± 6
+ 2
± 3
± -1
+ 2 ;
+ i •
+ 2
+ 2
+ 3

Based on duplicate runs
                         86

-------
                                 gW^W»fBiKff^^W^!*!«lWl>ftP
for cysteine, while  the  GC/MS  system obtained  a  value  close
to theoretical; also,  the  ainino  acid analyser  would  not  elute
hydroxyproline.  The GC/MS value for ?iistidine was relatively
large and  is  likely  related to the  problems  with histidine
mentioned  earlier  (131,133).  Overall, the  comparison  was  very
good with  both methods having  about a 90% efficiency.  The
average difference between the two  methods was 12.8% with a
maximum and minimum  of 48.8% and 1.40% respectively.   In
general, precision was superior  by  the GC/MS method.

Analysis of Municipal  Waste Samples for Amino  Acids

     Four  types of municipal waste  samples from  four cities
were analyzed for free amino acids.   The  catio'n  and  ligand
exchange procedure was used for  cleanup and  concentration as
described  earlier in this  section.   Derivatives  were prepared
using the  HFBA acylation and the n-propyl esterification
procedure.  The total  ion  chromatogram of the  derivatized amino
acid extract of the  superchlorinated peptage is  shown  in
Figure 25.

     Table 15 shows  the  concentrations of amino  acids  found
in these samples.  It  is interesting to note the levels  found
in raw sewage of City  A  differ in a few cases  from those
reported in the literature.  Presumably,  this  is due to  the
nature of  the plant  influent.  As expected,  amino acid levels
are higher in septage; however,  it  can be seen that  "super-
chlorination" is an  effective  method for  destroying  the  amino
acids.  Since chlorine doses used in this method are typically
in the 0.2-0.4% range, under these  conditions  one would  expect
extensive  oxidation  of organic materials  to  take place.  In
addition,  one would  expect to  observe considerable formation of
chlorinated organic  products,  which has been confirmed by
Glaze and  Henderson  (75).   Parallel work  on  the  samples  listed
in Table 15 by other members of  the  North Texas  State  University
research team has shown  that numerous chloro-organics  are
present in superchlorinated samples  from  Cities  B, C,  and D (Cf.
section 5  of this report).

     Table 15 shows  that these samples also  contain  c-nloro-
tyrosine,  and in one case  dichlorotyrosine.  This observation
represents the first coni'^ rrr.ation of a chlorinated airino acid
in a wastewater product.   Tne  occurrence  of  chlorinat?d  amino
acids in chlorinated wastes  was  predicted quite  earlier  by
persons such as Robert Dean  at the  United Scates Environmental
Protection Agency, Cincinnati, Ohio,  in a personal communication
to William H. Glaze.

     The significance  of these chlorinated products  in waste-
water plant products cannot  be stated at  this  time.  The fate
of the compounds in  the  receiving environment  undoubtedly
will depend on the choice  of disposal  method,  as well  as the

                               87

-------
00
00
    Figure 25.   Reconstructed GC/MS chromatogr am cf amino acid extract from superchlorinated
    septage extract,   a-alanine;  b-glycine;  c-valine; d-threonine; f-serine; g-leucine;
    h-isoleucine;  i-phenylalanine;  j-glutamic acid; k-lysine; 1-tyrosine; m-chlorotyrosine;
    n-dichlorotyrosine.                                                                   ,  .
                                                                                                  I
                                                                                                     •;

-------
03
VD.
      TABLE 14.  COMPARISON OF GC/MS ANALYSIS  OF  DERIVATIZED  AMINO ACIDS WITH ANALYSIS
                BY  A BECKMANN AMINO ACID ANALYZER USING A WASTEWATER MATRIX.
Amino
AciJ

Alanine
Glycine
Valine
Threonine
Serine
Leucine
Isolencine
Proline
Cysteine
Hydroxyprolinea
Methionine
Aspartic Acid
Phenylalanine
Glutamic Acid
Ornithineb
Lysine
Tyrosine
Arginine
Ilistid: ne
Tryptophan
Spiked
Concentration
(Vim/1)

5.08
5.00
5.00
5.18
. 5.00 '
5.00
5.19
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.11
5.00
Amino Acid
Analyzer
Concentration Found
(ym/1)
1
5.29
5.02
4.20
5.04
4.43
4.43
5.12
5.18
2.70
-
4.11
5.29 .
4.52
5.55

5.19
4.41
4.94
4.47
5.06
2
5.29
4.58
5.01
4.70
4.51
4.38
4.81
5.11
2.13
-
3.67
5.16
4.25
5.38

4.8C ,
3.89
4.72
4.32
5.43
Avg..
5.29+0.01
4.80+0.22
4.61+0.41
4.87+0.17
4.47+0.04
4.41+0.02
4.97+0.15
5.15+0.03
2.42+0.28

3.89^-0.22
5.23-J-0.06
4.39+0.13
5.47+0.08

5.04+0.15
4.15+0.26
4.83+0.11
4.40+0.07
5.25+0.18
GC/MS
Concentration Found
(ym/1)
1
4.98
4.51
4.59
4.73
4.74
4.57
5.07
4.72
5.24
4.64
4.57
4 45
4.64
4.95
4.52
5.02
4.76
4.80
6.84
4.67
2
4.83
4.33
4.89
4.66
4.58
4.54
5.02
4.70
5.38
4.56
4.51
4.38
4.54
5.2R
4.58
4.91
4.65
4.68
7.65
4.55
Avg.
4.91+0.08
4.42+0.09
4.74+0.15
4.70+0.03
4.66+0.08
4.56+0.01
5.04+0.03
4.71+0.01
5.31+0.07
4.60+0.04
4.54+0.03 '
4.41+0.04
4. 59 + 0. OS
5.12+0.16
4. 55+0. 03
4.96+0.06
4.^1+0.05
4.74+0.06
7.24+0.41
4.61+0.06
          Kydroxyproline  not eluted  from  amino  acid analyzer column.

          Ornithina  not in  standard  used  on  amino  acid  analyzer.

-------
         TABLE 15.  AMINO ACIDS PRESENT IN MUNICIPAL WASTES, yg/1







Alanine
Glycine
Valine
Threonine
Serine
Leucine
Isoleucine
P^oline
Cysteine
Hydroxyprol.ine
Methionine
Aspartic Acid
Phenylalanine
Glutamic Acid
Lysine
Arginine
Histidine
Tyros ine
Chlorotyrosine
Dichloro-
tyrosine




Typical3
Domestic
Sewage
20 - 90
20 - 130
50 - 240
10 - 180
20 - 130
60 - 280

0


50 - 240
20 - 130
20 - 330
10 - 180
tr
tr
tr
60 - 90


CITY A




Raw
Sewage
250
44
200
120
37
380
130
42
<7
52
19
82
38
52
110
<15
130
150,
<10b
<10b
CITY B

Super-
Chlorinated
Anaerobic
Digestor
Supernatant
4.C
24.0
0.9
3.0
36.9
tr
8.0
3.0
<0.07
<0,08
<0.14
<0.08
5.0
3.0
5.0
<0.15
<3.3
3.0,
3.0b
<0.10b
CITY C
COMBINED SLUDGE

2-hour 4-day
Super Super
Chlori- Chlori-
nation nation
3.1
8.9
11.1
0.9
2.4
7.9
7.1
<0.11
<0.09
<0.10
<0.17
<0.10
10.5
15.1
3.5
<0.19
<4.2
1.4
1.0b
<0.12b
3.2
5.1
3.2
tr
1.2
tr
5.8
<0.04
<0.03
<0.04
<0.07
<0.04
3.6
4.5
3.6
<0.07
<1.6
1.1,
1.3b
0.5b
CITY D
SEPTAGE



Super-
Ra'A' chlorinated
1,220
1,410
780
450
400
960
390
100
60
100
110
130
300
150
130
750
370
450^
<5b
<5b
8.3
5.6
4.0
0.9
1.2
2.7
1.3
<0.07
<0.05

-------
stability of the particular compound.  'In  the case  of  "super-
.c-ilorinated" sludge or septage products, there may  be  some
concern for the leaching of compounds  into receiving r.treans
of ground water if sludge beds are not adequately sealed.   In
any case, these data and the results of earlier works  point
out the need for caution in bhe disposal of  "superchlorinated"
waste products to avoid possible contamination of the  environ-               |
ment.
                              91

-------
                                     ^
                                                                    "H" '-*1* &» •' --"-*,
                           SECTION  7

     THE ANALYSIS OF NON-VOLATILE ORGANIC COMPOUNDS  IN WATER
                AND WASTEWATER AFTER CHLORINAT1ON
INTRODUCTION

     This section focuses on the non-volatile organic constit-
uents of natural waters and municipal wastewaters before and
after chlorination.  Of particular  interest  is the development
of high performance liquid chromatography  (HPLC) approaches to
the study of these systems.  New information is presented on
the irolecular size dispersion of non-volatile organic substances
in water and their role in the formation of  trihalomethanes and
other chlorinated by-products.
BACKGROUND

     Section 4 summarizes current knowledge on the structure
and composition of naturally occurring organic substances in
water.  As noted there, soluble aquatic organic matter is largely
of the "fulvic acid" type, that is, it is soluble in D'oth
mineral acid and base  (Figure 3, Section 4).  There is still
controversey over the chemical structure of fulvic acid;
indeed, ic is almost certain that a single "structure" cannot
be written for the material.  The structures proposed in
Figures 4 and 5 in Section 4 represent composite structures
based principally on chemical degradation products, and the
persons responsible for the structures are among the first to
point out that the actual composition of a natural organic
matrix is much more complicated.

     The approach taken in this study was to concentrate on
the development of new HPLC-based methods to separate the
natural matrix into its components.  This approach was based
mostly on the intuitive judgement that more could be learned
about natural organics and their by-products if the systems
were first simplified by the use of modern separation methods.
High performance liquid chromatography offers several modes
for the study of aqueous organic compounds (145).  While the
study reported here is by no means complete, significant new
data is reported on the fractionation of aquatic organic
matrices, and on the effects of disinfection agents on these
fractions.

                              92

-------
                                       -^^
EXPERIMENTAL PROCEDURES
Purified Water

     The water used to prepare reagents and for TOX and LLE
blanks came from one of two systems.  The first system, which
produced water referred to as "D" water, was described in the
work of Glaze and others  (71) .  This process began with
ozonated,. high-purity-bottled water supplied by Ahlfingers
Water Corporation.  This water was further purified by passage
through a 1.5 cm i.d. glass column that was filled with 50 cm
Filtrasorb F-400 activated carbon  (12/40 mesh) , followed by
25 cm of XAD-2 (20/50 mesh) , and finally a Millipore glass-
filter disc.  This purified water was analyzed for total organic
carbon (TOC) and was found to contain less than 200 ppb.  The
total organic halogen (TOX) value for this water by the GAC
method (146) was found to be 15 to 20 ppb.  No detectable
THMs were found.

     The second type of water was supplied via a commercial
water-treatment apparatus from Millipore.  This system pro-
duced so-called "Q" water; it consists of a tap-water feed to
a series of cartridges consisting of Regard prefilter, reverse
osmosis,  granular activated carbon, and two icn-exchange
cartridges.  Typically,  this water gave the following measured
values:  IOC — 300-500 ppb? TOX — 5-10 ppb; and no detectable THMs.

High Performance Liquid Chromatography

     Two instruments were utilized.  The first is a Waters
ALC-201 with 6000A pump and built-in refractive index detector.
The second is a Micromeritics 7000B.  Three detectors were used
with both instruments:  a Tracer 970-A scanning UV detector;
a Schoeffel FS970 f luorometer; and a modified Coulson electro-
lytic conductivity detector originally designed as a GC detector.
Off-line detection of halogen content of HPLC effluents was
accomplished by manual collection of fractions,  followed by
pyrolysis/microcoulometry with a Dohrmann MCTS-20 system.

     HPLC exclusion  (molecular size) separations were carried
out with two columns containing Partisil 10 (Whatman, 11 n
particle size, 60 A pore size) deactivated with bonded glyceryl-
propylsilane according to the procedure of Regnier and Noel
(147).   The characteristics of the two columns are shown in
Table' W3.  Columns were packed by the upward slurry technique
of Bristow et al. (148)  and coated with Carbowax 20M in most
cases.   Carbon~TTiltered "Q" or "D" water was used as carrier
solvent,  but 2% isopropanol/water also was used in earlier work.
The working range of the column was determined with a combin-
ation of proteins and sodium polystyrene sulfonates of known
                               93

-------
                                                        i?«» -f^fpfvr^v^^^^'^Se'^f



                                                        »-*>'. JB-*il«'k~*,i i^S',"*"!}-**1*!*
             TABLE 16.  CHARACTERISTICS OF PARTISIL
                   10/GLYCOPHASE HPLC COLUMNS


Length fcm)
ID (nun)
Number of Theoretical Plates N
HETP (mm)
Linear Velocity (cm/sec)
Void Volume V (mL)
Permeation Volume V (mL)
Interstitial Volume V. (mL)
Analytical
25
4.6
2500
0.1
0.185
2.25
4.9
2.65
Preparative
25
9.4
3500
0.07
0.097
3.1
7.4
4.3
Packing  Material:  Partisil 10  (Whatman 11  y  particle size,
                    60  A pore size)  deactivated  with Glycercl-
                    propyl silane  and treated  with Carbcwax 20M.
                                 94

-------
molecular weight.  As shown in Figure  26, the exclusion limit                f
of the column is 45,000  (k'=0), but the practical range is                   1
probably from 30,000 - 1,500.                                                ]
                                                                             j
                                                                             •4
     Weak anion exchange chromatography was carried out using a              '
98 cm X 1.5 mm I.D. column containing  AL Pcllionex WAX, a weakly
basic anion exchanger  (Whatman).  Gradient elution was used
beginning with an acetic acid solution (pH3) and ending with a
triethylamine-acetic acid buffer  (pH6.5).  Both linear and
concave N^ gradients were utilized.

     Reverse phase HPLC used a preparative column (25 c.n X 9.4 ran.
ID) and an analytical column  (25 cm X  3 mm ID),  The former was
prepacked with Whatman Magnum-9 Partisil-10 ODS-2 and the latter
was slurry packed in the laboratory with Partisil-10 ODS-2.  A
linear gradient was used for elution beginning with 5% methanol
in acetic acid (pH 3) and ending with  50? methanol in water.

Water Samples

     Secondary treated wastewater from the city of Denton, Texas
was collected before chlorination and  after final clarification.
General characteristics of the samples are given in Table 17.  It
is noted that from 67 to 89 percent of the total organic carbon
is passed through a 0.45 yfilter (dissolved organic carbon, DOC).

     Lake water was taken from Cross Lake, Louisiana.  Some
characteristics of this water are shown in Table 18, although it
should be noted that variations are expected on a seasonal basis.

     Samples were collected in glass bottles which had been
rigorously cleaned with detergent, chromic acid cleaning solution,
carbon^iltered wa>ter, and then dried  at 150°C.  No attempt was
made to collect head space free samples.

Analytical Scheme

     Tht: procedure used to isolate and characterize nor-volatile
organics is shown in Figure 27. The procedure involves freeze-
drying of the purged, filtered samples.  Redissolutdon into
water, base and acid soluble fractions was attempted initially,
but later more emphasis was placed on  water and acid soluble
fractions.  Following redissolr.tion the fractions were examined
by various HPLC modes, including characterization of fractions
collected by preparative HPLC.  Analytical parameters used to
characterize the fractions included total organic carbon  (TOC)
using the Dohrmann DC-54 analyzer, and total organ: ? halogen
(TOX)  using the Dohrmann GAC method (Uo) or the XAD-2 method  (71).
It should be noted that recent work has shown that the latter
gives values of TOX substantial! ly lower than the former method,
when a chlorinated surface water source is used (1^9). The reason
for this discrepancy presumably is ;.he inability of XAD-2 or

                               95

-------
                               K
Figure 26.  Calibration curve for Partisil 10/gl^cophase size
exclusion columns. A Proteins (ovalbumin, M=45,000; chymotryp-
sinogin A,M=25,000).a Sodium polystyrene sulphonates (M=16,000;
6,500; 4,OOC, 1,600).   Q Methanol.

                              96

-------
                           Sample
                                      N? Purge 1,  2,  4, 5
Fraction I1
Water Sol .
pH 7


1

Fraction II
Acid Sol.
HN03, pH 2

Fraction. I
Water Sol .
pH 7


1

Fraction II1
Acid Sol.
HN03, pH 2

       ize Sep. HPL
        Size  Sep. HPLC
 ollect  High Mole Size Fracr:
reeze Dry to Original  Volunj
       (1, 2, 4, 5)
  ollect High Mole Size Fract
  reeze Dry to Original Volum
         (1, 2, 4, 5)
   Chlorinate Each Fractioi
         (1, 4, 5)
Analytical  Parameters

1.  TOC
2.  TOX
3.  Free  and Combined Chlorine
4.  THMFP 3 days
5.  TOXFP 3 days
  Figure  27.   Scheme  for the  study of  the effect of
        chlorination on  non-purgeable organics.
                             97

-------
          TABLE 17.  GENERAL CHEMICAL CHARACTERISTICS
             OF THE SECONDARY TREATED WASTEWATER

Sample Collection Date
pH at 25°C
Specific Cond. (pmho's
at 25°C )
Total TOC fng/L)
Soluble TOC, before purging
(mg/L )
Soluble TOC, after purging
(mg/L )
Percent Soluble. TOC
(before purging)
Percent o4: Soluble
TOC Hot Purgeable
12/7/1978
7.6
877
X~ S
10.8 0.3
7.2 0.2

6.8 0.1

66.7
94.4

1/9/1979
8.2
910
X S
12.7 0.4
11.3 0.1

11.1 0.1

89
98.2

1/17/1979
7.8
868
X S
11.2 0.4

9.8 0.2

-
87.8


X = mean of three injections each 16 ml                                    5
S = standard deviation                                                     1
                              98

-------
         TABLE 18™  CHARACTERISTICS OF CROSS LAKE WATER
                        (0.45 u Filtrate)
Color  (units)                                 110

pH  @ 25°C                                       7.5

Ammonia - N (mg/L)                               0.2

Organic - N (mg/L)                               0.6

Total   - P (mg/L)                               0.1

COD         {mg/L)                              35

Soluble TOG (mg/L)                              11.0 + 0.4
       TABLE 19.  RESIDUAL CHLORINE AT  DIFFERENT TREATMENT
                    LEVELS AND  CONTACT TIME
Chlorination Level
20 mg/L,
30 mg/L,
30 mg/L,
Ih contact time
Ih contact time
24 hr contact time
Residual
Free
1.6
6.8
4.4
Chlorine*
Combined
0.7
2.0
1.8
mg/L
Total
2.0
8.8
6.2
Sample:  Filtered, purged secondary treated wastewater
         collected on 12/7/1978

*Analysis by amperometric titration
                               99

-------
XAD-8 resins to trap oxidized natural1 humics,  although  they
perform satisfactorily on  less polar, low molecular weight
organics.

     Trihalomethanc formation potential  (TKMFP,  150)  a-.d  TOXFP
were determined by THM or  TOX analysis -after chlorination of
samples at a dose of 20 mg/L at pH 6.5 and  25°C.  THM analysis
was made by the liquid-liquid extraction  (LLE) procedure
described in Section 8.  Residual chlorine  was measured by an
amperometric procedure  (151, Table 19).

RESULTS AND DISCUSSION

HPLC Investigation of a Municipal Drinking  Water  Source

     Fractionation and analysis of the 0.45 M  filtrate  of Cross
Lake, Louisiana water before and after chlorination followed
the schemes shown in Figure 27.  Chlorination  of  the  unbuffered
lake water at 26°C with a  dose of 20 mg/L causes  the  formation
of trihalomethanes (THMs)  the yield of which reaches  a  plateau
after three days.  At a buffered pH of 6.5  the relative yields
of THMs after this period  are 81% CHC13, 16% CHCl2Br  and  3%
CHClBr2  (by weight).  Expressed as halogen, the yield of  total
trihalomathanes does not represent the majority of bound  halogen
in chlorinated natural waters.  This observation, which has
been extensively documented in this laboratory, is consistent
with the works of Sontheimer, KUhn, and co-workers  (152)  and
of Oliver (74).

     As expected, the carbvn.matrix of unchlorinated  Cross Lake
water is largely a non-volatile fraction, presumably  consisting
of a mixture of fulvic acid, and other components.  After micro-
filtration and freeze drying, the residue obtained is mostly
soluble in purified water  (50% of original  TOC) and in  dilute
nitric acid (24% of remaining TOC).  A darkly colored solid
remains which is partially base soluble and which probably
represents humic c^cids and clay particles which presumably
are colloidal in f-.^ze in the original sample.  The water  and
acid soluble portions are of interest in this work, but it
should be noted that the role of fine particulates may  be
crucial in the transport of :nicropollutants and in water  treat-
ment processes.

     After freeze-drying, the water soluble fraction of the
Cross Lake sample was analyzed by size exclusion HPLC.  Figure
28 shows chromatograms of the water soluble fraction before and
after 20 mg/L chlorination for five days (refractive  index
detector).   Apparent average molecular weight has shifted
slightly downward upon chlorination as much be expected.  Table
20 shows some of the characteristics of five fractions of this
water soluble portion collected by preparative HPLC before
chlorination.   Exclusive chromatograms of the reinjected

                              100

-------
                                   A
           r~
           6
                                                    UJ
                                                    CO
                                                    z
                                                    o
                                                    Q_
                                                    CO
                                                    LJ
                                                    o;
                                                    LJ
                                                    c
                                                    cr
                                                    o
                                                    o
                                                    LU
                                                    cr
  4   ~~3      2      I      0

RETENTION  VOLUME (mL)
Figure 28.   Size exclusion chromatograms of Cross Lake sariple,

     freeze dried water soluble  fraction.  Top, unchlorinated;

     bottom, chlorinated at 20 mg/L  for five days.
                              101

-------
TABLE 20. CHARACTERISTICS OF THE WATER SOLUBLE FRACTIONS OF
CROSS LAKE WATER COLLECTED BY SIZE EXCLUSION
HPLC (UNCHLORINATED)
S (
t \

THMFP f, '•?•• I
Fraction
: No.
1
2
: 3

4

5
Average

Sura

Mole Wt.
Range
Mole
Wt. at
Pk Maximum
31.6 x IO3 -
22.3 x IO3 -
19.1 x IO3 -

15.9 x IO3 -

6.3 x IO3 -
10.5



15. y x IO3
7.9 x IO3
7.1 x IO3

5.1 x IO3

0.2 x IO3
x IO3



22.4
14.2
10.3

7.9

3.9




x IO3
x IO3
x IO3

x IO3

x IO3




'DOC
X
0.63
0.53
1.28

1.22

0.91


4.97

(mg/L)
S
0.04
0.00
0.04

0.20

0.20




THMFP* (pg/L)
X
31
73
95

43

62


304

S
7
2
10

5

2


(252
as Cl)
X
.049
.078
.074

.035

.068
.061



iX3C ; ;_
s s i
.011 !"-. |
.002 |
f •> A
.008 \
i
.004 ;'
r,
.002 i
.005 1
I
;

-------
fractions are shown in Figure 29.  The data indicate that
THMFP is evenly distributed throughput the molecular weight
range of the polymer, with the possible exception of fractions
1 and 4.

     Table 21 shows characteristics of water soluble fractions
collected by size exclusion HPLC after chlorination of the
original sample.  Again, it is noted that average molecular
weight has decreased as measured by che refractive index
detector (compare with Table 20}.  Non-volatile organic halogen
values in these samples were measured by the XAD method  (71).
The results show that the total yield of halogen in the non-
volatile fraction is 260 yg/L (as Cl) as compared to a yield
of trihaJomethanes of 252 yg/L  (as Cl).  An exhaustive search
for other halogen-containing volatile compounds by GC/MS
revealed only small amounts (<10 yg/L) of trichloroacetic acid
and other compounds (see Section 5).  Thus, non-volatile
compounds represent the majority of chlorinated products in
the lake water.

     Later experiments using the GAC/TOX procedure showed that
the preponderance of .non-volatile organohalides over THMs was
greater than shown by the XAD procedure.  Figure 30 shows the
formation of trihalomethanes and TOX as measured by the LLE
and GAC procedures.  Final yielus of THMs were 280 yg/L,
whereas TOX/GAC and TOX/XAD values were 1400 and 500 yg/L
(all as Cl).

     Tables 22 and 23 show characteristics of size exclusion
fractions of different samples of Cross Lake water.  In this
case only acid soluble fractions were taken, i.e. the freeze-
dried residue was extx'acted directly with dilute nitric acid.
Table 22 shows that molecular size distribution  (this time
determined with the UV detector) is similar to that shown in
Table 20.  Number average Mn and weight average M^ molecular
weights of the acid soluble fraction are 3.9 x 10™ and 8.2 x
103 respectively as determined by analysis of the size exclusion
chromatograms.  Table 22 also shows that THMFP and TOXFP are
spread throughout the polymer, but on a per gram of carbon basis
more formation potential is present in the lower molecular
weight fractions.  Taking, the TOX/GAC as more accurate than
TOX/XAD values, the average ratio of TOX/DOC is 0.232 mgCl/mgC
(using the values for TOXFP and DOC obtained in the unfraction-
ated material, which more closely represent-conditions used in
water treatment).  On a molar basjs this ratio is 0.080 Cl/C.
If one assumes that, the average molecular weight of the polymer
is 3900  (Mn), and that the polymer contains 50% carbon on a
weight basis, the observed Cl/C ratio predicts the average
polymer will contain thirteen atoms of chlorine.
                              103

-------
                   37m!
  876543210

  RETENTION  VOLUME  (mL)
                                   765432   I    0


                               RETENTION   VOLUME  (mL)

Fjgure 29.  Size exclusion cfirumatograms of water soluble fractions
(xricl-ilorinated).  Itighh:  re^Jijected fractions; left:  superposition of
chromatograitis of reinjected fractions over original trace.

    ;                           104

-------
o
U1
                                                            0 TOX/GAC

                                                            A TTHMFP
                                                                                   (O
                                                                                   N
                                                                               - 3CO
                                                                               H 200
                                                                               -  100
                                    48         72

                                        TIME,    HOURS
120
   Figure  30.   Trihalomethane formation potential  (TTHMFP)  and total organic halogen
    (TOX/GAC)  formation potentials for Cross  Lake water;  20  mg/L dose for five days.

-------
                    ,.^^,-,,,-~~
       (3900)   (0.50)         n
            __^         x    u


      moles of C in        ratio of        moles of Cl  in
      polymer per mole     moles of        polymer per  mole
      of polymer           C1:C            of  polymer


Or in other words, the typical polymer  has approximately 162
atoms of carbon and thirteen atoms of halv)gen.  It should be
noted thab this ratio is approximately  twice the value  deter-
mined earlier from XAD/TOX measurements (39).

     Table 24 reports the values of  individual THMs formed by
the chlorination of the five fractions  described in Table 22.
An examination of the relative molar yields of the three THMs  in
the five fractions suggests a somewhat  lower yield of brominated
THMs in the first two (high molecular weight)  fractions.
Schnoor and co-workers (40) suggested that brominated organics
were formed from lower molecular weight precursors.  However,
their size fractions were collected  by  Sephadex gel chromato-
graphy which has bean shown to be subject to severe adsorption
effects when used on polymeric electrolytes  (153).  It  should
be noted that only proteins were used to calibrate their column.

     Schnoor et aJL. (40)  also report that the  overall yield of
trihalomethanes from their source (Iowa River  water) was 2.3-
7.2 ug/L of TTHM per mg/L TOC.  Tables  20 and  22 show that
values obtained in this work range from 35-227 pg/L of  TTHM
per mg/L TOC.  The lower yields observed in the work of Schnoor
et a1. are presumably due to the shorter reaction times (10
hr. vs. 24 hr), Icwer chlorine dose  (6  mg/L vs. 20 mg/L), and
perhaps differences in sample  type.  It should also be noted
that the pH of the Iowa samples was  adjusted to 10.8 before
chlorination, whereas the Cross Lake samples were run at 6.5.

V7eak Anicn Exchange Chromatography—

     MacCarthy et al.  (154) have reported recently the  use of
XAD resins in an HPLC mode for the separation  of natural humic
materials into two fractions.  No structural evidence was
presented but it was suggested from  the values of pH at which
elution occurred, that the two peaks corresponded to carboxylic
acids and phenolic polymers.

     Figure-31 is a chromatogram of  Cross Lake water using a
weak anion excTTaag^ (WAX)  resin with gradient pH elution and
UV detection.  Weak aiid-strong solvent  refer to boric acid -
borate buffer at pH 3.2 and 6.2 respectively.  Also shown in
Figure 31 are elution times of five  low molecular weight model
compounds listed in Table 25.  lonization constants Ka and
elution volumes of the standards suggest that  the first two

                              106

-------
            TABLE 2L   CHARACTERISTICS OF THE WATER SOLUBLE FRACTIONS OF
                   CROSS  LAKE WATER COLLECTED BY SIZE EXCLUSION
                                HPLC. (CHLORINATED*)

Fraction Mole Wt Range p^LximSm ^ ^^ "™* ^^
No. Pk Maxlmum X S X S
1 31.6 x 103 - 14.2 x 103 19.1 x 103 1.04 0.48 65 11
2 19.1 x 103 - 6.9 x 103 12.6 x 103 0.98 0.16 47 8
3 15.9 x 103 - 6.3 x 103 7.1 x 10* 1.00 0.36 86 13
4 12.6 x I®3 - 2.2 x 103 3.9 x 103 0.64 0.08 20 9
5 7.1 x 103 - 0.15 x 103 2.5 x 103 0.52 0.04 42 8
Average 8.2 x 10
Sum 4.18 260


) NVTOX/DOCt
X S
.063
.048
.086
.031
.081
.062

.010 •'
.008
.013
.014
.015
.01?


 *Origxnal sample chlorinated at 20 mg/L for 5 days before freeze-drying and
  redissolving.
**NVTOX = non-volatile TOX determined by the XAD procedure (71)  after purging
  sample to remove THMs and other purgeable organohalides.

t NVTOX/DOX in mg Cl/mg C.

-------
           TABLE 22.  CHARACTERISTICS  OF ACID SOLUBLE  FRACTIONS OF  UNCHLORINATED
                      CROSS LAKE WATER SEPARATED BY SIZE EXCLUSION  KPLC
          Molecular Weight  Molecular Wt.     ^?     THMFP
Fraction       Range        at Peak max.   J"9^      Mg/L
                                         X    S    X   S
                                                                          --TOXFP--
                                                            TI1MFP  XAD   yg/L  GAC )ig/L     TOX/GAC

                                                             000    X     S     X    3        "a*
         31.6x10-15.9x10    22.4 x 10   0.66 0.01  36   1

         22.3xl03-  7.9xl03   14.2 x 103  1.74 0.03  118 1.1

         19.1xl03-  7.1xl03   10.3 x 103  1.88 0.04  218 0.9
         15.9xl03-  S.lxlO3

          6.3xl03-  0.2xl03
   1

   2

   3

   4

   5


Average

  Sum
Value
Before   31.6x10 -  0.2xlOJ
Fraction-
 ation
7.9.x 10   1.14  0.02  201 0.4
3.9 x 10   0.70   C
159 1.4
                                        6.12 0.02  732*  2
0.055   58   14     142   13

0.068  347   15     856  110

0.116  354   13   1064  182

0.176  281   16     612  104

0.227  187   12     340  113


0.128   .  	

 	  1227*  31   3014* 262
9.2 x 10J  6.1   0.1   239* 1.6  0.039   425* 21    1419*
0.215

0.492

0.566

0.537

0.485


0.459
                                                                                             0.232
*TOXFP and THMFP of fractions were carried out a higher [HOC1]/(C)  ratio for  fractions than for the
 combined sample before fractionation; TOX values expressed as yg/L of chlorine.

-------
Molecular
Fraction Weight Range


1 31.6xl03-14.2xl03
2 19.1xl03-6.9xl03
3 15.9x10 -6. 3xl03
4 12.6xlOJ-2.2xl03
5 7.1xlOJ-0.15xl03
Average 	

Sum 	

Value
Before 31. 6xlO?-0. 16xl03
'Fraction-
ation

*TOX values expressed as ua/r
Molecular
.
max.
19.1xl03
12.6xl03
7.1xl03
3.9xl03
2.5xl03



"" "™ ™

8 . 4xl03



DOC '.

XO
o
0.82 0.1
0.96 0.1
1.74 0.1
0.9
1.0 0.2



5.42 0.3

5.7 0.1





A o
17 4
81 3
96 10
72 6
26 5

™» « «_

292 14

346 10



TO:

x
68
157
625
410
124

—

1384

771



k" vJ/VG
J/L*
S
13
22
136
33
15

— —

144

47



TOX-+
XAD
~DOC~
21
84
55
80
2C

53

— — —

61



TOX-GAC+ , ;
DOf ;

0.083
0.164
0.359
0.456 '
0.124

0.237



0.135

i

+TOX/DOC in mg Cl/mg C.
                                                                                                 i   i
                                                                                                 *   •
                                                                                                 I   i
                                                                                                 *?   j
                                                                                                 u

-------
        TABLE 24.  THMs FORMED BY CHLORINATION OF ACID SOLUBLE FRACTIONS OF
                  CROSS LAKE WATER SEPARATED BY SIZE EXCLUSION  HPLC*
**
Fraction
; l
2
3
4
5
0.
0.
1.
1.
1.
CHCI3
X S
32 0.006
97 0.009
43 0.006
39 0.001
16 0.006
THM (ymole/L)
CHCl0Br
&
X S
0.02
0.17
0.53
0.43
0.30
0 ,
0 .
0.
0.
0.
001
001
OQ1
001
OOo
TTHMFP
CHClBr (ymole/L)
X S X S
N.D. - 0.34 0.007
0.02 0 1.16 0.01
0.09 0.002 2.05 0.01
0.05 0.001 1.87 < 0.01
0.02 0.001 1.48 0.01
	
 *Mo broraoform detected
**Those listed in Table 22

-------
peaks in Figure 32 correspond to phenols and carboxylic acid
polymers as suggested by MacCarthy et a^.  (154).  The third peak,
not reported by MacCarthy et al. is of unknown molecular type.

HPLC Investigation of Municipal Wastewater •

     Secondary effluent from the' Denton, Texas municipal waste-
water treatment plant was collected before final chlorination.
Characteristics of the water are shewn in Table 17.  Aliquots
of the filtered, purged sample of 12/7/78 were chlorinated at
different treatment levels c.nd contact times.  Free and combined
chlorine were determined by air.perometric titration according to
procedures described in Standard Methods (151).  Results are
presented in Table 19,  Based on these results, 30 mg/L chlor- .
ination level was selected for this type of water to ensure free
available chlorine in the sample.

     Two liters of the same filtered, purged wastewater were
chlorinated at 30 mg/L level.  At time intervals corresponding
to 0, 1, 24, 48, 72, 96 and 120 hours, duplicate samples were
collected, quenched with Na2SC>3, and analyzed for THMs using
the modified liquid-liquid extraction procedure (155).  At the
same time intervals, samples were collected and quenched and
analyzed for TOX using the XAD procedure (71).  After 120 hours
the sample was purged with N2 for one hour and samples were
taken for THMs and TOX determinations.  The free and residual
chlorine in the sample was monitored for 48 hours.  Also initial
and final samples were analyzed for TOC.

     Data on THMs are presented in Table 26 and data on TOX
are presented in Table 27.

     Examination of Table 26 shows no detectable concentration
of CHC13 or TTHM in the unchlorinated purged sample.  Upon
chlorination and after one hour contact time 29.5 pg/L CHCL-,
and 48.6 yg/L TTHM were detected.  This indicates that a fraction
of the nonpurgeable organics react rapidly with chlorine to
produce CHClj and TTHM.  Concentrations of CHCl-> and TTHM
showed a gradual increase and reached a plateau after 72 hours.
The changes in CHCl^ end TTHM concentrations after 72 hours do
not seem to be statistically different.

     Examination of Table 27 also indicates several important
features.  While the limitation of the TOX method are well
recognized, the data indicate the presence of a measurable
concentration (23.9 pg/L)  of halogenated non-purgeable organic
compounds in the purged sample before chlorination.  Total THM
concentration in the same sample was below the detection limits
of 0.5 ug/L.  Presumably these non-purgeable compounds are
present in the wastewater influent.
                               Ill

-------
    "It STRONG
    SOLVENT
    100
     80 -
     60 -
     40 H
                                        e   d
                                                       c b a
     20 -
          SOLVENT BLANK
          32  3O 28 26  24 22 2O 18  16  rt  12  IO   8   6  4  20
                  RETENTION TIME —	:—r-:	«	
                                      ( mm )
Figure 31.  Weak  anion exchange HPLC chromatogran of Cross
Lake water, acid  soluble fraction of freeze dried sample
(unchlorinated).  Reference compound code: a-phenol;  b-3-methyl-
catechol; c-vanillic  acid; d-2,4-dihydroxybenzoic acid;
e-2,4,6-trihydroxybenzoic acid.   Dotted line:  solvent gradient
(100%=pH 6.2;  0%=pH 3.2).
                               112

-------
   TABLE 25.   K  AND RETENTION VOLUMES FOR STANDARD COMPOUNDS
     Compound
   K
Retention
  Volume*
Retention
   Time**
a.   phenol
1.1 X 10
                                 -10
    3.0 mL
  1.5 min.
b.   3-methylcat->chol   ^1.0 X 10
                                 -10
                   4.0 mL
                2.0 min.
c.   vanillic acid
8.3 X 10
                                 *- C
    6.0 mL
  3.0 min.
d.   2,4-dihydroxy-     1.05 X 10
       benzoic acid
                                 -3
                  21.4 mL     10.7 min.
e.   2,4,6-trlhydroxy-   2.1 X 10
       benzoic acid
                                 -2
                  26.3 mL     13.2 min.
 *N  concava gradient pH 3.2 - 6.2 boric acid - borate buffer
  in 25 minutes.
**See figure 31.
                               113

-------
                    TABLE 26.  TRIHALOMETHANE FORMATION POTENTIAL DATA


Sample Designation

Unchlorinated
Unpurged
Unchlorinated

CHC13
X S
4.1 0.31

<0.5
ug/L
CHCl^Br
X S
1.1 0.06

<0 . 5 -

CHBr2Cl CHBr3
X S X S
0.85 0.11 <0.5

< 0 . 5 - .< 0 . 5

TTHM
X S
6 0.48

<0. 5
      Purged
   Chlorinated 1 h      29.5  2.S6  10.8  1.25   5.3   '0.57   <0.5    -     48.6  0.35
   Chlorinated 24 h    . 60.4  4.50  27    0.51  10.7   0.21   <0.5    -     98  •  4.80
M  Chlorinated 48 h     67    4.90  29.2  2.60  11.3   1.10    0.76  0.18  108.3  8.82
*•  '
   Chlorinated 72 h     77.9  3.78  29.9  1.46  11.3   0.65  .  0.89  0.06  119.9  5.96
   Chlorinated 96 h     74.3  1.45  29.1  0.50  11.3   0.22    1.0   0.09  115.6  2.26
   Chlorinated 120 h    72.1  3.08  28.9  1.19  11.5   0.43    1.0   0.01  113.5  4.70
   X  Mean of duplicate analysis
   S  Standard deviation
   TTHM  Total trihalomethane
                                                                                                  I

-------
       *!"^
     Upon chlorination the TOX values also increased after one
hour contact time indicating the instantaneous formation of
halogenated organic compounds from the non-purgeable precursors
present in the sample.  The TOX formation potential follows
essentially the same pattern as the TTI'M.  Tables 26 and 2"? show
that throughout the chlorination process, non-purgeable organic
halogen which is XAD-adsorbable is equal or greater than the
halogen in the lorm of THMs.  As noted in the first, part of this
section, XAD adsorption does not measure total organic halogen
(149); thus, the TOX value of the non-purgeable fraction may
be assumed to outweigh the purgeable fraction (THMs) by at
least a factor of 2 to 3.
         TABLE 27.  NON-PURGEABLE TOX FORMATION POTENTIAL OF A
              SECONDARY MUNICIPAL WASTEWATER EFFLUENT
                               (ng/L)
Sample Designation                          X
Procedure Blank
Unchlorinated Unpurged
Unchlorinated Purged
Chlorinated, Ih
Chlorinated, 24h
Chlorinated, 48h
Chlorinated, 72h
Chlorinated, 96h
Chlorinated, 120h
8.8
28.5
23.9
52.7
187
211.8
233.7
22i
205.3
1.2
3.8
5.2
2.2
5.3
13
11.2
12.3
13.4
 X   Mean of two injections 20 pi each

 S   Range x 0.89
                              115

-------
Examination of Freeze Dried Concentrates—
     Four hundred mL of filtered, purged wastewater sample was
freeze dried under vacuum.  A fluffy residue was left upon the
evaporation of the water.  The freeze dried  (FD) residue was
separated :'nto three portions and dissolved in organic-purified,
deionized water  (WATER SOLUBLE FRACTION), 0.1 N HN03  (ACID
SOLUBLE) and 0.1 N NaOH  (BASE SOLUBLE).  The measured pH of the
ACID and BASE fractions were 2 and 11 respectively.  The dilute
acid was found to dissolve almost all of the residue.

     The supernatent solutions from each of the solubility tests
were filtered through 0.45 u pore size filter and the base
fraction was neutralized to pH 7 with few drops of C N HNOo*.
One mL of the ^O soluble fraction was qualitatively chlorinated
with about 50 mg/L Cl2 at. contact time of one hour.  All
filtered fractions of the FD solubility tests were subjected to
size exclusion HPLC under the conditions described in the
experimental part of this section.  Figure 32 shows the size
excljsion chromatograms of the unchlorinated and chlorinated
water soluble fractions.  The chromatograrc of the chlorinated
sample shows a shift of peak £1 towards the lower molecular
v/eight end and a formation of peak #11 beyond the total permea-
tion volume.  Since the sample was not purged, it is likely
that peak #11 corresponds to the volatile halogenated compounds
which partition with the glycophase coating on the silica beads.
Earlier experiments showed that peak #11 could be removed by
purging the sample before HPLC.

     The HPLC effluent corresponding to peak #1 in both the
chlorinated and unchlorinated fractions (Figure 32) was
collected and subjected to TOX analysis by direct injection
of 20 pL into the Dohrmann micrccoulometer.  Presumably these
fractions contain no inorganic halide ions since these are
aluted along with the small molecules at the total permeation
volume.  The TOX valiias for the '1PLC effluent of the unchlorin-
ated and chlorinated samples were 23.3 pg/L and 147 pg/L,
respectively.  These results are based on the original volume
of the sample before freeze drying.

     Figure 33 shows the size exclusion chromatograms of the
acid and base solxible fractions.  The size exclusion chromato-
grams of the II2°' acid an<3 base soluble fractions were analyzed
for molecular size distribution based on a standard curve of
sodium polystyrene sulfonate and proteins in molecular weight
ranging from 1.6 X 103 to 45 x 103 as described before.   The
results are presented in Table 28.  The water soluble fraction
showed an apparent molecular weight, ranging from 0.63 X 103 to
15.85 X 103 with an average of 2.99 X 103.   The molecular weight
*
 pH adjustment to 7 is necessary in order to perform size
 exclusion HPLC on the glycophase/silica column.

                              316

-------
        '-  ,

      Volume (ml)
       3.3
2.3
 I
2.0
 I
1.3
I
0.5
Figure 32.  Size exclusion chromatograms  of  water  soluble
     fraction of freeze dried Denton municipal  wastewater.
     Top: unchlorinated; bottom: chlorinated.

                              117

-------
                  1
        3.0
I
2.0
 2.5     2.0    1.5

—Volume I ml )
0.5
n
 0
Figure 33.  Size exclusion chromatograms of acid (TOP)  and
        base (BOTTOM)  soluble fractions  of  freeze  dried
             Denton municipal wastewater.

      •      "                  118

-------

TABLE 28.  MOLECULAR SIZE DISTRIBUTION FOR THE
  FREEZE DRIED UNCHLORINATED SAMPLE:  DENTON
              SECONDARY WASTEWATER
Water Soluble
Fraction
Mean
Mole Wt
15350
8910
5010
3160
1990
1000
630 .
Mean
%
3.2
6.96
10.66
15.16
19.67
43.0
1.2
2990
Acid SoJuble
Fraction (pH2)
Mean
' Mole Wt
25120
15850
8913
5012
3162
2239
1995
4
%
1.43
5.71
13.57
17.14 .
26.42
33.57
2.14
?60
Base Soluble
Fraction (pHll)
Ma an
Mole Wt
25120
15850
6913
7079
3981
3162


%
1.2
3.1
13.4
30.48
39.6
12.19

6100
.

-------
of 1000 represents the highest  percentage  of  the water
soluble fraction.  The acid  soluble  fraction  showed  an  apparent"
molecular weight ranging  from 2.00 x 103 to 25.12 X  103 with  an
average of 4.96 X 103.  Apparent molecular weights of 2.24 X  103
and 3.16 X 10  represent  a substantial ,fractior, ca. 60 percent,
of the acid soluble  fraction.   The basa soluble fraction  showed
an apparent molecular weight ranging from  3.16 X 10  to 25.1  X
10  with an average  of 6.10  X 103.   Apparent  molecular  weights
of 3.98 X 103 and 7.08 X  10J represent 70  percent of the  base
soluble fraction.

     Several points  need  to  be  considered  regarding  tne inter-
pretation of the freeze drying  experiment.  One relates to the
results of the solubility tests.  The formation of an only slightly
water soluble freeze dry  residue is  not surprising.  Organic
compounds in the sample could be in  the colloidal form  as well
as in true solution.  At  the low temperature  of the  freeze drying
process and in the presence  of  a. complex matrix of inorganic
salts, agglomeration of colloidal particles is likely to  occur.
Chelation between organic ligands and metal ions to  form water
insoluble molecules  is another  possibility.

     Considering these factors, distilled  water is expected to
solubilize only a small fraction of  the freeze dry residue.
Dilute HNO-j is expected to interact  and solubilize the  inorganic
components of the residue whij;h consist essentially  of  the Coi~,
KCO3~, Cl~, S0|~ and PO43 ~'j  of  the Na+, Ka+,  Ca2+ and Mg2+.
Dilute HNO.J will also dissolve  basic organic  compounds  such as
nitrogenous compounds, substitute the metal ion in the  organic
complexes with H , and affect partial hydrolysis of the  organic
esters.  Dilute NaOH is expected to  precipitate the  hydroxides
of the divalent and trivaleut cations in the  residue.   Also,  it
will dissolve organic acids  and affect partial hydrolysis of
the esters.

     Molecular size distribution data presented in Table 28
seems to confirm the previous discussion to some extent.  The
water soluble fraction consisted essentially  of smaller mole-
cules.  The acid and base soluble fractions contained higher
percentages of the larger molecules.

     There are some limitations to the molecular weight
distribution data.   For one,  the molecular weights are based on
tne retention "olume of compounds of known structure and exact
molecular weights.   The compounds under investigation are yet
to be identified and their accurate molecular weights are yet
to be determined.
                               120

-------
CONCLUSION

     The  study of  the  non-volatile  components of  a municipal
drinking  water source  and a municipal wastewater  reported  in
this section  is by no  means a  completed  story.  The results are
only fragmentary and should be viewed mainly as a portent  of
studies to  come in the future.   Undoubtedly more  elegant
separations and spectroscopic  methods will be developed to
assist in the elucidation of these  materials whose complex,
polymeric  structures make them  so intractable.

     Nonetheless,  the  studies  reported here show  that a
substantial amount of  carbon-bound  halogen occurs in the non-
volatile  matrix which  remains  largely unchanged in the drinking
water or  wastewater processes  which are  in common use.  The need
for further study  of the environmental fate and effects of
these compounds is implied, as well ac further investigations
cf their  molecular structures  before and  after disintection
processes.
                               121

-------
                           SECTION 3

          ANALYSIS OF VOLATILE CHLORINATED ORGANICS IN
                  WATER BY LIQUID-LIQUID EXTRACTION
INTRODUCTION

     One of the most important classes of chlorinated organic  •
compounds generated during water chlorination appears to be
the volatile chlorinated organics  (VCOs).  The four most
prominent compounds in this class are the trihalomethanes
(THMs):  chloroform, bromodichlorometha.ie, chlorodibromo-
methane and bromoform.  They are important because recent data
indicate they comprise a significant portion of the total
chlorinated organic material generated by water chlorination
(156), and because the THMs are all known or suspected to have
toxic and/or carcinogenic  (157) potential.

     The relationship between water chlorination and the
formation of these VCOs has only recently been recognized.
Glaze et al. (36) were the first to demonstrate that VCOs
could be formed by the action of chlorine in water.  They
reported chloroform as a product of the  chlorination of
municipal v,astewaters.  However, the analytical techniques
which they used were not designed to accurately quantitate
such highly volatile pollutants.  J. J.  Rook (37) adapted a
"closed system" headspace analytical procedure originally used
in the flavor industry, which was a more appropriate technique
for volatile analytes.  He later >\sed this technique to clearly
demonstrate the formation of all of the THMs during drinking
water chlorination (158).

     Concern about the presence of VCO£>  increased as more studies
confirmed their occurrence in water supplies and drinking
water in other geographical areas.  In 1974, a new analytical
technique was developed by Bellar, Lichtenbergr and Kroner
(88), which offers several advantages over the headspace
"technique.  In the Bellar method, 5 ml sample aliquots are
purged v:ith inert qas which entrains the VCOs and transports
them to a trap containing an adsorbent resin, usually Tenax
GC.  Following adsorption, the gas flow  is reversed, the trap
heated, and VCOs desorbed orto the head of a GC column for
subsequent analysis.   The headspace procedures used by Rook
requires larger samples for analysis necessitating bulky

                               122

-------
sampling equipment  and  laboratory glassware.  Also,  the  Bellar
method requires only  15 minutes  for  the  concentration portion
of the analytical procedure  versus 12  hours  for  the  analytes to
equilibrate between the liquid and gas phases in Rook's  head-
space technique.

     As a result of these advantages,  the  Bellar method  was
adopted by the United States Environmental Protection Agency,
and subsequently by VCO analysts in  general.  This method was
used by the USEPA for the National Organics  Reconnaissance
Survey  (159) and the  National Organics Monitoring Survey
 (156) which showed  the presence  of various combinations  and
quantities of VCOs  in all of the eighty  U.S. cities  sampled
during the survey.

     The impact of  these studies on  the  water treatment  field
is already significant and will  probably become  more important
in the near future.   The role of chlorine  in the water and
wastewater treatment  field is undergr Lng extensive reevalua-
tion because of the potential harmful  side effects of chlor-
inated organics such  as the THMs which are produced  in the
process.  Mandatory VCO monitoring programs and  maximum
allowable limits for  THMs in drinking  waters have recently
been promulgated by the USEPA  (160).

     These developments imply that great numbers of VCO
analyses will soon  have to be run by water treatment facilities
as well as regulatory and surveillance authorities.  Much
manpower and equipment will be required  if the Bellar method
is used for routine screening.  The  alternative  is to develop
new and even more streamlined analytical techniques.  Nicholson
(76)  recognized this  and underscored the importance of an
analytical method which can "... handle  large  numbers  of
samples in a relatively short period of  time" as well as be
"highly specific for  halogen-containing  compounds".

     In accordance with these requirements, two  new analytical
techniques have been  developed.  The first, developed by
Nicholson (76), is  a  direct aqueous injection technique
(DAI).   This technique takes advantage of  the high sensitivity
of the electron capture gas chromatographic detector to  circum-
vent any concentration step.  The second technique is the
liquid-liquid extraction (LLE) method  (161).   It  utilizes a
closed system extraction step followed by  electron capture
gas chromatography.    This chapter describes in detail the LLE
method and compares it to the other two  contemporary methods,
the Bellar purging ir.echod and the DAI procedure.

     Important parametric critiques of the heedspace, the
Bellar,  and the LLE analytical methods have been published
which have led to modification in equipment and procedures.
Kaiser and Oliver (162)  have reported  a miniaturized headspace

                              123

-------
 system which only  requires  sixty railliliters  of  sample.   Head-
 space analyses  can be performed in  approximately forty-five
 minutes using this system.   The equilibration step  is  carried
 out at a reduced pressure which allow;;  equilibration to  occur
 in only thirty  minutes.  The small  sample  size requires  the
 use of the more sensitive electron  capture detector, but this
 juakes Kaiser and Oliver's headspace method more  specific for
 halocarbons than Rook's  survey procedure.   Kaiser and  Oliver's
 paper also reports the results of their study of the important
 parameters which control the headspace  procedure.   Although
 theso data were collected using their new  miniaturized head-
 space system, most of them  are universally applicable  to
 headspace procedures.

     Kus, et al.  (163) have published a definitive  analysis of
 the important parameters which control  the performance of the
 purging methods.   Although  their study  included  the evaluation
 of several purging systems  of various configurations,  they
 concluded that  the Bellar system was the most effective.
 Therefore, the  majority of  the data reported  was collected
 using the Bellar apparatus.   These  data will  be  the pi'imary
 ones cited as representative of the purging method  in  comparing
 it to the liquid-liquid extraction  procedure  described in this
 section.

     Liquid-liquid extraction techniques have been  and still
 are the classical  means of  concentration of trace organic
 compounds in water.  The most important area  in  which  these
 techniques are  currently being used is  in  the analysis of
 pesticides, herbicides, and fungicides  (Io4).  These procedure
 typically call  for a series  of organic  extractions  followed by
 a several-fold  evaporative  concentration of the  extracting
 solvent before  analysis.  Such extensive handling and  trans-
 ferring of tht>  water sample  and extracting solvents might be
 acceptable for  classical applications,  but new interest  in
 much more volatjle analytes  requires considerable modification
 of the classical procedures  to avoid losses due  to  volatiliza-
 tion.  Grob and Grob (165)  were among the  first  investigators
 to recognize this   In their  liquid-liquid extraction  procedure,
 a high ratio of wat^r to organic solvent is used which
 eliminates the  need for concentration,  and the organic layer
 is sampled directly out of  the extraction  vessel  for chromato-
 graphic analysis which reduces the  losses  due  to  sample
 handling and transferring.

     Since the  LLE method developed in  this work was first
 reported (161),  two other liquid-liquid  extraction procedures
 have been published by Richard and Junk  (166)   and Mieure
 (167), respectively.   Although neither procedure utilizes  a
 closed extraction  system, both procedures  place  heavy
 emphasis on the  importance  of the careful  handling of  both the
water sample and the organic  solvent in  order  to avoid losses

                              124

-------
 due to volatilization./
                       ;
 EXPERIMENTAL          j

      Samples for the LLE, the Bellar and the DAI procedures
 were collected and handled in identical manners.• It has been
 recognized (150) that samples must be collected so as to avoid
 contact with air bubbles within the sampling container which
 could cause analyte losses due to water/headspace partitioning.
 Therefore, samples were  collected in 125 ml serum bottles
 filled to overflowing and then sealed with teflon-lined
 silicone septa crimped in place by an aluminum outer sleeve
 (Figure 34).   Before sampling, the bottles were cleaned with
 chromic acid, water, acetone, and then dried in an oven at
 165°C for several hours.  After sampling was completed, the
-bottles were transported to the laboratory for analysis.  If
 more than a few hours had to elapse between sample collection
 and analysis, the sample bottles were chilled with ice and then
 warmed to room temperature before analysis.

      Two reagents were added to the samples at the time of
 collection.  Sodium sulfite  (Baker analytical grade)  was
 added as a chlorine reducing aqent in varying amounts depending
 on the anticipated chlorine residuals   This reduction of
 residual free chlorine to inorganic chloride prevented
 chlorination reactions from occurring subsequent to sampling
 (150).  A buffer was also added to avoid possible extraction
 anomalies related to pH  effects (1.2 ml of a buffer prepared
 from a 2:3 mixture of 1.0 M NaH2PO4 and 1.0 M Na2HPC>4 pH 6.5).
 Although no pH effects were observed during preliminary tests
 on spiked samples, it was felt that not all possible matrix
 variations could be anticipated;  thus,  the decision was made
 to continue to use the buffer despite the lack of any data
 indicating its usefulness.

 LLB Procedure

      The LLE  method basically involves  v. specialized closed
 system organic solvent extraction procedure followed by
 chromatographic separation and analysis.   A schematic represent-
 ation is shown in Figure 35.

 Pentane Extraction Procedure—
      Normal pentane (Fisher,  pesticide  grade)  is used as the
 extracting solvent.   Chromatographic analysis  of the  pentane
 prior to use  in the LLE  procedure "usually shows this  grade
 of solvent to be of adequate quality as received.   If purifica-
 tion is necessary,  it is effected by fractional distillation
 from sodium metal or by  passing through chromatography
 reagent grade alumina.   The internal standard,  1,2-dibromo-
 ethane (Aldrich,  reagent grade),  is distilled  and added to
 the solvent at a concentration of approximately 20 pg/1.

                               125

-------
      TO?*?St«St»W8B»»MHW-..* • - «»«*«*„
                                                                                   •n
                                        ALUMINUM CRIMPED SEAL
                                         TEFLON-LINED RUBBER
                                            SEPTUM
                       SAMPLE

                   (CHLORINE QUENCH

                  AND BUFFER ADDED)
                                               120 ml GLASS
                                              SEPTUM BOTTLE
Fjgure  34.  Sample bottle  used for  collection  of water  samples
      for analysis  of purgeable volatile organic compounds.
                                 126

-------
     Three ml of this solvent/internal standard mixture is
added in the manner shown in Figure 36 to the 125 ml water
sample using two 10 cc syringes.  One syringe contains the
solvf fit mixture while the other is empty.  As the solvent
mixture is injected into the inverted sample bottle, it rises
to the top of the bottle and an equivalent  amount of water
is displaced into the empty syringe.   The sample bottle is
then strapped to the surface of a platform gyratory shaker
(Junior Orbit. Shaker, Labline Instruments, Inc.) and shaken
at a speed of 400 rpm/for twenty-five minutes.  After shaking,
the samples are ready for immediate analysis.

Chromatographic Analysis—
     A two to five microliter pentane aliquot is removed
through the silicone septum with a Hamilton 801 ten micro-
liter syringe.  This aliquot is then injected into a Tracor
560 gas chromatograph equipped with a 63jji linearized electron
capture detector.  The glass Chromatographic column is 183
cm by two mm I.D. and is packed with ten per cent squalane on
100/120 mesh Supelcoport (Supelco Inc.).  The carrier gas is
a 95/5 per cent argon/methane mixture.  The column flow rate is
20 ml/min with 60 ml/min of makeup gas  (the same argon/
methane mixture) added to the GC column effluent to improve
detector performance.  The respective oven temperatures are:
injector 100°C, column 66°C, detector 300°C.  An electronic
digital integrator  (Supergrator, Columbia Scientific
Instruments) is used for quantitation. - Chromatograms are
recorded on a Perkin Elmer 56 strip chart recorder.

Direct Aqueous Injection

     The same chromatograph, integrator, and recorder are used
for this procedure as were used for the LLE procedure described
above^  The glass GC column used is 122 cm by two mm I.D.
and is packed with Chromosorb 102 60/80 mesh, a polystyrene/
divinylbenzene copolymer adsorbent.  The injector, column
and detector temperatures are 175°C, 135°C, and 300°C,
respectively.  The procedure followed is to remove a three
to five microliter water aliquot directly from the VOA sample
bottle and inject it into the chromatograph.

Bellar Purge and Trap Method

     A schematic diagram of the procedure used in this labora-
tory is shown in Figure 37.  The procedure is essentially the
same as that outlined by Bellar, et al. (88) with three
important modifications (refer to Figure 37):

   1.  A liquid Chromatographic sample loop injector (Altec,
Inst. # 5U8031) is used to accurately introduce reproducible
aliquots of water samples into the purging device.


                              127

-------
                        SAMPLE
                        (120ml)
                        QUENCH
                        (Na2S03)
                        BUFFER
                        (pH 6.5}
                  REMOVE 5ml  WATER
                 ADD  5ml n-PENTANE
                    EQUILIBRATION
                       (Shaking)
          GAS  CHROMATOGRAPHIC  ANALYSIS
                    (GC/Ni -63 EC)
Figure 35.  Scheme for analysis of volatile organics by
               liquid-liquid extraction.
                         128

-------
          Pentane
        Sample
                                                        Pentane
                                                        Being
                                                        Added
                               Water
                               Being
                               Pushed
                                Out
Figure  36.   Procedure for removal of water and  addition of
        pentane for volatile  purgeable organic  analysis.

                                129

-------
   2.  The analytical column serves as the adsorbing trap.
Thus the VCOs are purged directly onto the Chromatographic
column eliminating a significant source of erratic results.

   3.  The Chromosorb 102 analytical column which is used
allowed the separation of several less significant analytes
such as carbon  tetrachloride and 1,2-dichloroethane which
could not be separated using Bellar's original analytical
column.

Purging Procedure—
     The 5.5 ml sample injector loop is filled to overflowing
with the water sample using a 20 cc syringe.  The valve is
switched to the inject position and the sample forced to flow
into the purging device by the pressure of the helium carrier
gas.  As the pressure continues to build, the helium purges
the volatile analytes from the water and sweeps thera into the
analytical column which is cooled to room temperature.  A
carrier gas bypass valve allows the purging equipment to be
circumvented during Chromatographic analysis thus obviating
possible problems due to water-saturated carrier gas.

Chromatographic Analysis—
     The gas chromatograph is a Hewlett Packard 3700 equipped
with a Coulson Electrolytic Conductivity Detector (CECD-
Tracor Inst., Inc., Austin, Texas).  The glass Chromato-
graphic column is 122 cm by 2 mm I.D. and contains Chromosorb
102  (60/80 mesh) .  The helium carrier gas flow is 20 rnl/min.
The injector temperature is 200°C, and transfer lines and
venting valve leading to the CECD are at 275°C.  After sample
introduction, the GC column oven is programmed ballistically
from ambient temperature to 60°C, then from 60° to 220° at
8°/min.  The CECD is operated in the reductive mode with 60
ml/min of hydrogen added to the GC effluent prior to the
influent end of the pyrolysis furnace which is at 850°C.
The recorder and integrator are the same as those used in the
LLE and DAI procedures.

RESULTS AND DISCUSSION

     A comparison has been made in the laboratory between the
LLE method and the Bellar method for VCO analysis.  Although
the Direct Aqueous Injection technique was also evaluated
relative tc the LLE method, the comparison was less extensive
than with the Bellar method for reasons described below.

     Typical chromatograms are shown in Figures 38-40 for the
three analytical methods.  The concentration of chloroform in
all samples was about 40 ppb.  Other analyte concentrations
were adjusted using the appropriate response factors (shown
in Table 2.9}  to produce peak heights approximately equal to
that of chloroform.  It is important to note that the response

                              130            '

-------
                  SAMPLE  OF
                 WATER  II. A
                 20ml SYRINGE
                                \He
                                out to
                                PURGING
                               APPARATUS
         ALTEC  VALVE IN
        EMPTYING  POSITION
              PURGING
             APPARATUS
                                CARRIER  GAS
                                BYPASS VALVE
                                GAS
                           CHROMATOGRAPH
                                                      COULSON
                                                         EC
                                                      DETECTOR
                                                      RECORDER
                                                     INTEGRATOR
Figure 37.   Modified  purge and  trap apparatus with liquid
                      sample loop injector.

                                  131
WWJLJ.tlU i!

-------
  H-

  C

  CD

  U)
  CO

(0
3
ft M
CD !-•
3 CD
CD O
  rt

F O
n 3

CD n
X CD
rfO
M rt-
CD C
O H
ft (B
3 P
   CO

o" n
O 3"
3 i-J
a. o
p. rr
ri" C-
H- rt
O O
3 iQ
W i-(
   CD
H- 3
3
   O
rr M»
(B
X <
n- n
—• o
•  M
            00-
        m
           ro-
           O-
   O
                                                                                             • Internal  Standard
                                                                                                                                       ;S

                                                                                                                                       fe
                                                                                                                                      i
                                                                                                                                      «
                                                                                         CHBrj
                                                                                     	CHBr2CI
                                                                                     CH8rCI2
                                                                                          CHCI3(40yug/L)
                                                                                                                                      t*

-------
                                                                              " ••*** *B|
                                    o
                                    _ro
                                    o
                                    X
                                    o
             CM
             O
             k-
             CD
             X
             o
                                                o
                                                I?1
                                                ffl
                                                o
                                                       IO
                                                       l-
                                                       00

                                                       o
r
0
	1	
    12
 TIME (min.)
16
20
24
  Figure 39.  Coulson  electrolytic conductivity gas chromatogram
              cf. VCOs  from modified purge/tiap procedure
              (conditions in text).

                                133

-------
                    CM
                    O

                    CD
                    X
                    O

                    N
                    I
                    O
                           I
                          16
        12
TIME (run )
I
8
I
4
n
 o
Figure 40.  Electron capture gas  chromatogram of VCOs by
            direct aqueous injection  (conditions in text).
                               134

-------
 r*;«^*^
factors reflect concentrations which  produce  similar  peak
heights as opposed  to equal  areas.  This  approach  was used  for
two reasons:   firstly, the classical  (though  less  reliable)
quantitation of trace analytes by  gas chromatography  has
usually been performed by manually measuring  peak  heights to
calculate concentrations.'" This  is the procedure used hit——•
Bellar  (88 ) in his  first report,  and is  the  procedure still
being widely used today.  Secondly, signal  to noise ratios
 (S/N)^are related to peak heights  rather  than areas.   Since
the minimum detectable limit  is  a  function  of the  S/N,  it is
therefore also a function of  peak  height.

     The digital electronic  integrator used in our laborator-
ies measures both peak heights and areas  automatically.
Therefore, either variable can be  selected  as the  basis of
quantitation.  Th'e estimates  of  quantitative  precision are
shown ir. Table 30.  For precision  studies,  the analyte concen-
trations were  all adjusted to an approximate  20 to 200 S/N
ratio.  Therefore, these precision values probably approach
the optimum which can be produced  for each  respective method.
Of particular  interest is the fact that the internal  standard
quantitation method used for  the LLE  procedure produced values
roughly comparable to the values for  t-he  Bellar method which
used the usually less precise external quantitation method.
This is probably due to the use  of  the liquid chrcmatographic
sample loop injector in the Bellar  method which allows good
reproducibility of sample size for  successive analyses.  The
precision for  all techniques  is  generally at  or below the
5 per cent level which is acceptable  precision for a  trace
analytical technique of this  type.             ---~

     The minimum detectable limits  (MDL) were determined by
analyzing increasingly dilute standard mixtures until a
signal to noise ratio of approximately 2 was  reached.  The
resulting values are shown in Table 31.  The  data  indicate
the LLE technique to be .generally more  sensitive than the
Bellar or DAI methods.  However, it should  be  noted that the
MCL values for the Bellar method were  determined using the
Coulson detector.   Recently,   this detector  has been replaced
with the Hall detector which  is more  sensitive by  at  least
one order of magnitude.   Hence the Bellar method is at least
as sensitive as the LLE method and in  the case of  compounds
such as 1,2-dichloroethane the Bellar method  is much  more
sensitive.   The MCL values of the DAI method  generally are
greater by an order of magnitude than  either  of the other
two methods.

LLE Parameters

Solvent Selection.—
     Pentane  is used as  the organic extraction solvent for
several reasons.   The Fisher  pesticide grade commercial

                              135

-------
pentane is usually pure enough to use as received.  It is
volatile enough to separate easily from the chloroform peak,
the first peak of analytical interest.  It has a  low electron
capture response producing little or no solvent front.  It
is highly insoluble in water; and it has a very favorable
distribution coefficient versus water for the analytes of
interest  (Table 32).

     As noted earlier, Grob and Grob  (165) first  developed a
survey liquid-liquid extraction process using pentane as a
solvent.  Their system uses 200 yl of pentane to  extract 900
ml of water.  Whdle this procedure seems to work  reasonably
well in their hands, it certainly requires a highly skilled
technician with much experience in the technique  to achieve
reproducible results.  The procedure also takes more person-
hours as it requires manual shaking of the extraction vessel
for maximum efficiency.  Grob and Grob indicate a loss of
extraction efficiency for more volatile components which might
be attributed to volatilization of such analytes  due to the
headspace present in their extraction apparatus.

     Richard and Junk  (166) also selected pentane as a solvent
for essentially the same reasons cited above.  For part of
their work, a flame ionization detector was used which
required more polar chromotographic columns for greater
separation cf the solvent front from the chloroform peak.
This requirement was also observed in this work when using
a GC/MS system for analysis.

     Mieure (167) used methyl cyclohexane as a solvent.
Although this solvent has a boiling point of 101°C, h»j
indicates that it separates adequately from the chloroform
peak.  We observed interferences with such solvents as
hexane (b.p. 69°C)  for our chromatographic system, and there-
fore, did not try higher boiling solvents.

Extraction Apparatus—
     The extraction apparatus used in the LLE procedure is
the only closed extraction system having no headspace in con-
tact with the pentane/water matrix.   The three other extrac-
tion systems cited anove do have a headspace in their respec-
tive extraction steps-,-  While this may not be n rigid require-
ment for VCO analysis, one advantage of the closed LLE
procedure is that emulsion problems seem to be eliminated due
to the relatively docile nature of the shaking process.  Even
the dirtiest wastewater samples can still be analyzed by this
technique.  This is not always the case with the more conven-
tional extraction procedures where intractable emulsion
formation may be a major problem.   Extraction systems similar
to Grob's showed a distinct tendency to form emulsions when
tried in this laboratory.


                              136

-------
                                                     '( .C3*B*-J«s~.'*"£*5. ~
Salt Effects—
     Mieure  (167) added sodium chloride to his samples to
increase the ionic strength of the water layer which increased
the extraction efficiency of his system.  The sample shown was
only for a sample/extractant ratio of 5:1.  With this ratio,
an increase in extraction efficiency from 93 to 98 per cent
was observed.  Our experiments showed no trend of increasing
extraction efficiency with increasing ionic strength.  We
therefore, did not add salt to samples in our system for the
purpose of increasing extraction efficiency.

Shaking Time—
     Standard samples for the evaluation of shaking time on
extraction efficiency were prepared by spiking organic free
water and extracting for varying lengths oZ time.  Times of
2-, 5-, 10-, 20-, and 30 minutes were examined.  The resulting
analyses indicated that equilibrium had bten achieved within
two minutes.  Integrals for all the peaks in the sample were
within the limits of precision of the technique relative to
peak integrals of samples shaken for longer time periods.
A shaking time of twenty-five minutes was arbitrarily selected
because heavily polluted samples might take longer to equili-
brate and because such a long shaking time can be used and
still not be the limiting factor for the total analysis time.

Extractant/Water Ratios—
     It is clear that the extractant/water ratio affects the
extraction efficiency (E).  The relationship between E and
the extractant/water ratio is given by the following equation:

                E =         100D
                       D + VW/VQ

Here E is the per cent of the VCO extracted, Vw is the volume
of water used, and Vo is the volume of organic solvent used.
D is the distribution coefficient which for dilute solutions
of solutes  (VCOs)  should be a constant.  Looking at published
data for chloroform, Junk produces distribution coefficients
from 40 to 53 (166).  Data from this laboratory result in
a D of 66; Mieure's data run from 49 to 114 (167).  Although
Grob and Grob did not analyze chloroform per se, their distri-
bution coefficients generally run 1000 or greater.  In
describing their extraction system, Grob and Grob allude to
the difficulty in achieving these remarkably high distribution
coefficients  (165), indicating that careful techniques had to
be rigorously applied.

     With the exception of Grob's data, LLE extraction effi-
ciencies observed in this laboratory and by other workers show
chat increasing water/solvent ratios lead to decreasing


                             137

-------
   TABLE  29. RESPONSE FACTORS FOR VCO'S  USING DIFFERENT
                  ANALYTICAL TECHNIQUES

Compound
CHC13
C2H4C12
cci4
CHBrCl2
CHBr2Cl
CHBr3
LLE
100
0.56
2000
500
333
67
Modifiedb
Be liar P/T
100
67
67
50
40
29
Direct3
Aqueous Injection
100
1
1000
200
100
22

Electron capture detection
Coulson electrolytic conductivity detection
 TABLE 30.  PRECISION OF ANALYTICAL '.lETHODS FOR ANALYSIS
     OF VCOs AT S/N 20-200 (PER CENT RELATIVE STANDARD
DEVIATION)

Compound
CHC13
C2H4C12
cci4
CHBrCl2
CHBr2Cl
CHBR,
LLE
1.9
3.8
2.1
8.1
5.4
4.7
Modified „ . . , _ .
Bellar P/T Direct Aqueous Injection
1.8 1.5
2.2 	
5.2 	
2.4 3.8
3.4 3.1
6.6 	
                          138

-------
    TABLE 31.  MINIMUM DETECTABLE LIMITS FOR THE ANALYSIS  OF
      VCOs BY LLE, BELLAR D/T AND DAI METHODS  (S/N  =  2)


Compound
CHC13
1,2-C2H4C12
cci4
CHCl2Br
CHBr2Cl
CHBr2

Detection Limi
LLE Modified
0.2 0.
36 0.
0.0.1 C.
0.04 0.
0.06 . 0.
0.3 .0 .
t (ug/L in H20)
Bellar
2
3
3
4
5
1
Direct Aq. Inj .
1
90
0.1
0.5
1
4.5

           TABLE 32.  EXTRACTION EFFICIENCIES OF VCOs
                  BY THE PENTANE LLE METHOD  "
Compound
Extraction Efficiency (%)
CHC13
1,2-C2H4C12
ccx4
CHBrCl.,
CHBr2Cl
          62
          41
          87
          69
          72
          66
                             139

-------
             a? •ra»i>>wtr i-
extraction efficiencies thus placing a practical  limit on  the
concentration factor which can be achieved.

Matrix Effects—
     Table 33 shows a comparison of the analytical techniques
when applied to a  "real world" sample  (Denton, Texas tapwater).
Clearly, in this situation the precision of the LLE method is
better than that for the Bellar procedure  (no precision values
for DAI are available).

     Most important is the discrepancy in  the quantitative
results between the DAI method and the other two  methods for
chloroform and bromodichloromethane.  The  DAI values are seen
to be much higher  than corresponding values for the other
techniques.  Nicholson  (76) observed a similar trend with his
DAI procedure.  He contends that the higher values are due to
haloform formation in the injector port of the gas chromato-
graph.  Apparently, the neat catalyzes the chlorination and/or
decomposition of chlorinated haloform precursors  which increase
the apparent haloform concentration.  This is the primary
objection to the method as an instantaneous "CO monitoring
technique.  These  anomalous effects are not observed for the
LLE procedure because the chloroform precursors apparently
are not observed with the Bellar procedure because the
precursors apparently are not purged from  the water sample.
Thus, these latter two methods raore accurately reflect the
instantaneous concentrations of the chlorocarbons.

Procedure—
     It is difficult to compare quantitatively the procedural
advantages which one technique has over another.  However,
analysis time is one parameter which can be accurately
estimated.  Table  34 shows a comparison of time of analysis
for various numbers of samples.  Clearly,  the LLE method has a
distinct time advantage over the Bellar procedure.  The
importance of this in a high sample volume survey program has
been indicated above.

     The LLE method has two other advantages over the Bellar
procedure.  Firstly, considerable special  equipment is required
for the Bellar process.   The Bellar technique requires a
special purge/trap apparatus which can either be  purchased
commercially for aboat $3,000, or can be built in-house,
perhaps requiring several months of development time.   At
least some modifications are required in the injection port of
a commercial chromatograph, and conventional chromatography
via syringe injection is difficult while the Bellar apparatus '
is in place.   Thus, the system is considerably less flexible
than might be desired.   Finally,  an electrolytic  conductivity
detector is almost universally used (although other specialized
detectors might potentially be used).   For most laboratories,


                              140

-------
 this  means  an  additional  $3,000 investment with the accompany-
 ing  installation problems.

      The  extraction/concentration procedure of the  LLE method
 is carried  out in the  sample  bottle,  eliminating the need
 for  cleaning extraction equipment.  Once  extracted, the sample
 is immediately ready for  chromatocraphic  analysis.   The LLS
 procedure uses a conventional electron capture gas  chromato-
 graph in  the configuration  supplied by the manufacturer.   Thus,
 any  laboratory equipped for pesticide analysis already has
 the  necessary  analytical  instrurientation.

      The  other major procedural advantage of the LLE procedure
 is the technical expertise  required by the analysts.   The
 simplicity  of  the LLE  procedure has resulted in competent
 analyses  being performed  by the least trained technicians in
 our  laboratory.   Tha Bellar method, on the other hand,  has
 required  highly skilled analysts in our laboratory  who have
 had  enough  experience  with  this specific  technique  to under-
 stand the idio'syncrasies  of the system.   Generally, the Bellar
 method requires more sample handling,  more hardware manipulation
 and  leaves  more room for  "cockpit"  errors.

      There  are  two principal  disadvantages  of the LLE method
 as compared to  the purge  and  trap procedure.   One is  the  lower
 sensitivity of  LLE method when  GC/MS  confirmation is  attempted.
 The LLE method  is sensitive primarily  because of the  use  of
 the electron capture detector which detects  picogram  quantities
 of organohalides.  If  confirmation  of  the  compounds is  attempted
 by GC/MS, mucb  higher  concentrations  are  required since only a
 portion of  the  LLE solvent  extract  may be  injected  into the
 GC/MS  system.   On the  other hand, virtually  all of  the  VCOs
 in a  sample go  into  the GC/MS system when  the purge and trap
 procedure is used, thus, the  detection limits are comparable
 to the values given  in Table  31.

     The LLE method as described  above also  cannot  analyze
 complex mixtures of VCOs,  particularly very  volatile  compounds
 such as vinyl chloride.  This limitation may  be minimized to
 some extent by the use of capillary columns  and cryogenic
 temperature programming, but  the  purge and trap procedure is
more convenient.

Recent Developments in the Analysis of VCO Compounds  in Water

     The LLE procedure described  above was developed  in 1975
and first reported at the  "First  Chemical Congress  of the North
American Continent" in Mexico City in December, 1975  (161).
Most of the material contined in  this section describes the
results of early work on the development of the method.
Subsequently,  the method was used in a study of approximately
twenty-five East Texas  area  water supplies  (168) and has been

                             141

-------

      TABLE 33.  ANALYSIS  OF VCQs  IN  DENTON,  TEXAS TAPWATER
Method
pg/L Halogen, as Cl*
CHC13 CHBrCl2 CHBr2Cl CHBr.
LLE

Modified
  Bellar

Direct Ag.
  In j .
n.d.  not detected
                  27.3 + 0.3  25.-7  +  0.3   19.4  + 0.4   4.24 + O.C8

                  29.3 + 2.2  32.5  +  1.8   25.4  + 1.0      n.d.
                      71.5
47.1
29.5
n.d.
        TABLE 34. • TIME"TO COMPLETE MULTIPLE VCO  ANALYSES
                   BY LLE AND MODIFIED BELLAR METHODS
                             -  (F.3S)

Number of Samples in Set*
1
2
4
8
16
LLE
1.6
2.5
4
6
10
Bellar
3
4.5
6
13
25

fc
Duplicate assays for each sample and QC samples included,
                              142


-------
used by numerous other groups for VCO analysis.  In addition
the USEPA has adopted a modified version of the Mieure method
(167) as an acceptable THM method  (160).  Alternative tech-
niques for the analysis of THMs and case histories of THM
formation in water treatment plants have been compared
recently by Brass*  (169).

     Recently, a convenient LLE method has evolved which
utilizes a 25 ml sample bottle from which five ml of water
is removed and replaced by one ml of pentane  (155).  The 20 ml
of water and one ml of pentane may be handshaken for one
minute and analyzed directly by EC/GC.  Recent optimization
studies have shown that the extraction efficiencies in this
system are higher than in the original method, and are
relatively unaffected by changes in pH, salt content (up to
moderate values), and the presence of up to 1% methanol.
More significantly, recent studies have shown that n-pentane
allows one to utilize the LLE method combined with glass or
silica capillary GC columns for the analysis of a much broader
range of VCOs.  Lower detection limits of "purgeable organics"
are higher on some cases than with the Bellar method (88) but
quite acceptable for many applications.  It is clear that
capillary GC/LLE methods with either FID or EC detection are
promising VCO analytical methods of the future.
                              143

-------

I                                      REFERENCES
?
I
c
\
{            1.   Bunker,  George,  C.,  The Journal of the American Medical
I                   Association,  92,  1-6 (1929).
|
f            2.   Enslow,  L.H.,  and Symans,  G.E., Sewage Works Journal, 17,
*                   984 (1945).
;
i            3.   United States  Government,  Federal Water Pollution Control
,                   Act,  Public law 92-500, United States Government;
!                   Printing Office,  Washington, D.C.  (1970).

|            4.   White, C.G. , Abstract of_ Technical Papers, Conference on
i                   the Environmental Impact of  Water Chlorination, Oak
!                   Ridge National Laboratory, Oak Ridge, Tenn., October,
i                   1975.

            5.   White, C.G., Journal American Water Works Associacion, 67,
                   410-3 (1975).

            6.   White, C.G., Journal Water Pollution Control Federatidh, 46,
;                   89-101 (1974).
i
1            7.   Connick,  R.E., et al. ,  Journal  of the American  Chemical
                   Society,  £1,  1285-9  "(I960) .

            8.   Handbook of Chemistry and  Physics, 54th Edition, Chemical
                   Rubber Company Press,  Cleveland, Ohio, p. D-130
                   (1973-4).

            9.   Jolley,  R.L.,  Oak Ridge National Laboratory Publications
                   ORNL-TM-4920,  Oak Ridge,  Tenn.  (1973).

           10.   Morris,  J.C.,  United States  Environmental Protection Agency
                   Report No.  EPA-600/1-75-002, Washington, D.C. (1975).

           11.   Carlson,  R.M.  and R.  Caple,  United States Environmental
                   Protection  Agency Report  No. EPA-600/3-77-066,
                   Washington, D.C.  (1977) .

           12.   Pierce,  R.C.,  National  Research Council of Canada,  Publica-
                   tion  No.  16450, Ottawa,  Canada (1978).
                                         144

-------
13.  Morris, J.C., Abstracts  2£ Technical  Papers,  Conference on
        the Environmental  Impact of Water  Chlorination,  Oak
        Ridge National Laboratory, Oak  Ridge,  Tenn.,  October,
        1975.

14.  Hine, F.J., Physical  Organic Chemistry, McGraw-Hill Book
        Co., Inc., New York,  361-3  (1962).

15.  de la Mare, P.B.D., Ketlcy,  A.P.,  and Vernon,  G.A.,
        Journal of the Chemical Society (London),  133,  1290
        (1954).

16.  Shilov, E.A., Kupinskaya,  G.V.,  Yasnikow, A.A.,  Dokiady
        Akademite Nauk, S.S.S. R., 81, 435  (1951).

17.  Fair, G.M., Morris, J.C.,  Chang, S.C., Weil,  I., and
        Burden, R.P. , Journal p_f the  American  Water Works
        Association, _40_, "1051  (1948).

13.  Kovacic, P., Lowery,  M., and Field, K.W. , Chemical  Reviews,
        7J), 639-65  (1970).

19.  Jolley, R.L., Ph.D. Thesis,  University of Tennessee, NTIS
        Publication ORNL-TM-4290, 369 pp.  (1973).

20.  Morris, J.C. , Principles and Applications of Water
        Chemistry, edited  by S.D. Fanst and J.V. Hunter,  John
        Wiley and Sons, New York  (1967).

21.  Fischer, J., Jander,  J, , Zeitschrift  Fur Anorganische
        Allgemeine Chemie. 313,  37-47 (1961).

22.  Colton, E. ,  Journal o_f Chemical  Education,  32, 485-7 (1955).

23.  Wei,  H.I.,  Morris, J.C., Abstracts, American Chemical
        Society,  Division  of Water, Air and Waste Chemistry,
        General Paper No.  13, 100-1 (1973).

24.  Jolley, R.L.,  Abstract of  Technical Papers, Conference  on
        the Environmental-Impact  of Water  Chlorination, Oak
        Ridge National Laboratory, Oak  Ridge, Tenn.,  October,
        1975.

25.  Halperin,  J.,  Taube,  H.,  Journal American Chemical Society,
        ~i±, 375-9 (1952).        .

26.  Latimer, W.M.,  and Hildebrand, J.,  Inorganic Chemistry,
        The MacMillan Co.,  New York, 563 pp. (1940).

27.  Allen, L.A., Journal  of the  Institute of Water Engineering,
        £, 502-32 (3950).


                              145

-------
                              •*',*"-.'?*• ,-V" '„ «*v?* 3rV.
28.


29

30


31.


32.



33.


34.



35.
37.


38.


39.



40.



41.



42.
Burleson, J.L. , Peyton, G.R.,  and  Glaze, W.H.,  Bulletin
  of_ Environmental Toxicology,  19,  724  (1978) .

Wright, N.C. ,  Biochemical J. ,  30,  1661  (1936).

Burleson, J.L., M.S. Thesis, North Texas State  University,
  Dentcn, Texas, 1<<77, p. 1-8.

Soper, F.G., Journal o_f the  Chemical  Society, London ,  127,
  1582-90  (1926).

Carlson, R.M., Carlson, R.E.,  Kopperman, H.L. ,  and  Caple,
  R. , Environmental Science  and Technology,  9,  674
  (19757:

Fuson, R.C. , and Bull, B.A. , Chemical Review, 15, 275-
  309  (1934).

Morris, J.C.,  Baum, B. , Water  Chlorination Environmental
  Impact and Health Effects, Vol.  2,  Ann Arbor  Science
  Publishers,  Inc., Ann Arbor,  Mich., 29-48  (1978).

Hoehn, ".C. , Goode, R.P., Randael,  C.W., and Shaffer,
  P . T . B . , Water Chlorination Environmental Impact and
  Health Effects , 2_, Jolley, R.L.,  Gorcheve,  H. , and
  Hamilton, D.H., eds. , Ann Arbor  Science, Ann  Arbor,
  Michigan, 519-36, 1977.

Glaze, W.H., Henderson, J.E.,  IV,  Bell, Johnny,  and
  Wheeler, Van A. , Journal of  Chromatographic Science,
  1JL, 580-4  (1972).

Rook, J.J. , Water Treatment and Examination,  2J3  (pt. 2),
  234-43  (1974).

Rook, J.J., Abstracts, American Water Workb  Association
  Conference, Minneapolis, Minn.,  June, 1975.

Glaze, W.H., Peyton, G.R. , Saleh,  F.Y., and  Huang,  F.Y.,
  International Journal of Environmental Analytical
  Chemistry.

Schnoor, J.L. Nitzschke, J.L.,  Lucas, R.D. ,  and  Veenstra,
  J.N. , Environmental Science  and  Technology, 13, 1134,
  (1979).

Alexander, H. J. , and Christman,  R.F., Unpublished liter-
  ature review performed in conjunction with EPA Research
  Grant No. R804430,
Schnitzer, M. , and Khan, S.U., liumi c Substances  in the
  Environment, Marcel Dekker, New York^  1972.

                        146

-------
43.  Martin, M.P. , and Haider R. ,  "Microbial  Degradation  and
       Stabilization of C -^-Labeled  Lignins-,  Phenols,  and
       Phenolic Polymers .\n Relation to  Soil  Humus;  Formation",
       Manuscript communicated to  W.  H;  Glaze,  (September,
44.  Kavanaugh, M.C.,  "Modified Coagulation  for  Improved  Removal
       of Trihalouethane Precursors" , presented  at  Nevada
       Section, AAWA Seminar on "Organics  in Domestic Water
       Supplies", Palo Alto, California,  (April  12,  1978).

45.  Gjessing, E.T, , Physical and Chemical Characteristics of
       Aquatic Humus , Ann Arbor Science, Ann Arbor,  Michigan,
       1976.

46.  Schnitzer, M. , Soil Organic Matter . Schnitzer  M. ,  and Knan,
       S.U., eds. , El-sevier Scientific Publishing Co.,  1-58,
       New York,  1978.

47.  Black, A. P. , and Christman, R.F. , Journal of the American
       Water Works Association, 55,  897  ("1963) .

48.  Sposito, G. , Schaumberg, G.D. ,  Perkins,  T.G.,  and
       Holtzclaw, K M. , Environmental Science and Technology,
       1_2, 931-34,  (1978).

49.  Christman, R.F.,, and Ghassemi, K. , Journal  of  the  American
       Water Works Association, 58,  723-41,  (1966).

50.  Wershaw, R.I.., Pinckney, D.J., and Booker,  S.E., Journal
       Research U.S. Geological' Surrey, 5_, 565-69  (1977) .

51.  Rook, J. J. , Environmental Science and Technology,  11,
       478-82 (1977).

52.  Rook, J. J. , Water Chlorination;  Environmental  Impact and
       Health Effects, R. L. Jolley, Brungs, W.A. , Gumming,
       R.B., and Jacobs, V.A. , editors, Vol.  3, Ann  Arbor Science
       Publishers, Inc., Ann Arbor, Michigan, 85-98  (1980).

53.  Christman, R.F. ,  Johnson', J.D. , Hass, J.R. , Pfaender,
       F.K., Liao, W.T. , Norwood,  D.L., and Alexander,  H. J. ,
       Water Chlorination:  Envi~onmental Impact and Health
       Effects, Vol. 2, R. L. Jolley, H. Gorcher and D.H.
       Hamilcon, Jr.,  editors, Ann Arbor Science Publishers,
       Ann Arbor, Michigan,  15-28  (1C78) .

54.  Christman, R.F. ,  Johnson, D. J. , Pfaender, F.K., Norwood,
       D.L , Webb, M.R., Hass, J.R. and Bobenreith, M. J. ,
       ibid. , 75-84 (1980).
                              147

-------
55.   Arguello, M.D., Criswell, C.D.,  Fritz, J.S., Kissinger,
       L.D., Lee, K.W.,  Richard, J.J., and Svec, J.H.,  Journal
       of the American Water Works Association, 71, 504 (1979).

56.   Feng, T.H., Journal Water Pollution Control Federation,
       35., 475-85 (1966).

57.   Geiger, W. , Reinhatd, M. , Schaffner, C., Zurcher, F.,
       Identification and Analysis of Organic Pollutants in
       Water, edited by L.H. Keith, Ann Arbor Science Publishers,
       Inc., Ann Arbor,  Michigan, 433-52  (1976).

52.   Keith, L.H., editor, Identification and Analysis of Organic
       .Pollutants in Water, Ann Arbor Publishing Company,  Ann
       Arbor, Michigan  (1976).

59.   Keith, L.H-, editor, Advances in the Identification and
       Analysis 01 Organic Pollutants in Water, Vol. 1 and 2,
       Ann Arbor Publishing Company,  Ann Arbor, Michigan (1981).

60.   Glaze, W.H., and Henderson, J.E., Journal of Water Pollution
       Control Federation, 47, 2511-15  (1975).

61.   Jolley, R.L., Ph.D. Dissertation, University of Tennessee,
       Knoxville, Tennessee  (1973).

62.   Jolley, R.L., editor, Water Chlorination;  Environmental
       Impact and Health Effects, Vol. 1, Ann Arbor Science
       Publishers, Ann Arbor, Michigan  (1976).

63.   Jolley, R.L., Gorcher, H., Hamilton, Jr., D.H., editors,
       Water Chlovination;  Environmental Impact and Health
       Effects, Volume 2, Ann Arbor Science Publishers, Inc.,
       Ann Arbor, Michigan  (1978).

64.   Jolley, R.L., Brungs, W.A., Gumming, R.B., and Jacobs,
       V.A., editors, Water Chlorination;  Environmental Impact
       and Health Lffects, Volume j,  Ann Arbor Science Pub-
       lishers, Inc., Ann Arbor, Michigan  (1980).

65.   Bunch, R.L., Barth, E.F., and Ettinger, M.B., Journal of
       the Water Pollution Control Federation, 33, 122-6
       (1961).

66.   International Symposium on Health Effects of Drinking
       Water Disinfectants and Disinfection By-Products,
       Proceedings to be published in Fundamental and Applied
       Toxicology.

67.   Garrison, A.W., Pope, J.D., Allen, F.R., Identification of
       Organic Pollutants in Water, edited by L.H. Keith,  Ann
       Arbor Science Publishers, Ann Arbor,'Michigan, 517-56
       (1976).
                              148

-------
68.  "Drinking Water and Health", National Academy  of  Sciences,
       National Research Council, Washington,  D.C.,  Vol.  2,  3,
       1980. .

69.  Kuhn, W., and Sontheimer, H., Vom  Wasser,  43,  327 (1974).

70.  F.uhn, W. , Fuchs, F. , and Sontheimer, H. ,  2.F.  Wasserund
       Abvasser-Forschung,  10, 192-8  (1977).

71.  Glaze, W.H., Peyton, G.R., and .Rawley,  R. ,  Environmer-jal
       Science and Technology, 11, 685-90  (1977) .

72.  Wegmim, R.C.C., and Greve, P.A., The Science of the  Total
       Environment, 7, 235-45  (1977).

73.  Dressnan, R.J., McFarren, E.F., and Symons, J.M.,  "An
       Evaluation of the Determination  of Total  Organic Chlorine
       (TOCl)_in Water by Adsorption onto Ground Activated
       Carbon, Pyrohydrolysis and Chloride-Icn Measurement",
       presented at Water Quality Technology Conference,  Kansas
       City. Missouri  (Dec. 1977).

74.  Oliver, B.C., Canadian Research, 1JL, 21-22  (1978).

75.  Glaze, K.H. , Henderson, J.E., IV and Smith, G., in Water
       Chlorination;  Environmental Imoact and Health  Effects,
       Vol. 1. R.L. Jolley, editor, pp. 139-157  (1973).

76.  IJicholson, A.A., Meresz, O., and Lemyk, B., Anal.  Chem.,
       4_9_^ 814 (1977).

77.  Pfaender, F.K., Jones, R.B., Stevei.s, A.A. , Moore, L. and
       Haas, J.R. , Environ. Sci. Tech_. , 12,  438  (1978).

78.  Kites, R.A., Biemann, K., Analytical Chemistry, 40,  1217
       (1968).

79.  Budde. W.L., Eichelberger, J.W., in Identification and
       Analysis of Organic Pollutants in Water,  adited  by
       L.H. Keich, Ann Arbor Science Publishers, Inc.,  Ann
       Arbor, Michigan, p. 155-176 (1976).
80.  Finnigar., R.E. , Knight, J,G., in Identification
       Analysis of Organic Pollutants in Water, edited by
       L. H. Keith, Ann Arbor Science Publishers, Inc., Ann
       Arbor, Michigan, 185-204  (1976).

81.  Coleraan, K.E., Lingg, R.D., Melton, R.G., and Kloptler,
       F.C., in Identification and Analysis of Organic Pollutants
       in Water, edited by L. H. Keith, Ann Arbor Science
       Publishers, Inc., Ann Arbor, Michigan, 305-28  (1976).


                              149

-------
82.  Geiger, W., Reinhard, M., Schaffner, C., Furcher,  F.,  in
       Identification and Analysis, of Organic Pollutants  In
       Water, edited by L. H. Keith, Ann Arbor  Scienca
       Publishers, Inc., Ann Arbor, Michigan, 433-52  (1976).

83.  Aston, F.K. , Philosophical Magazine', 39, 611-25  (1920;.

84.  McLafferty, F.W. , Interpretation of Mass Spectra,
       W. A. Benjamin, Inc., Reading, Mass.  (1973).

85.  Mun, I.K., Venkatarahavan, R., McLafferty, F.W. ,
       Analytical Chemistry, 4_9_, 1723-6  (1977).

86.  Canada, D.C. , Regnier, F.E., Journal of_ Chromatographic
       Science, 14, 149-54  (1976).

87.  Grob, K. , Grob, G., in Identification and  Analysis of
       Organic Pollutants in Water, edited by L. H. Keith,
       Ann Arbor Science Publishers, Inc., Ann  Arbor, Michigan,
       P. 75-86 (1976).

88.  Bellar, T.A., Lichtenberg, J.J., Kroner, R.C., Jour.
       AWWA, 66, 12 (1974).

89.  Junk, G.A., Richard, J.J., Grieser, M.D.,  Witiak,  D.,
       Witiak, J.L., Arguello, M.D., Vick, R.,  Svec, H.J.,
       Fritz, J.S., Calder, G.V., J. Chromatog., 99, 745
       (1974).

90.  Purifax, Inc., French Patent 1,516,.054.  Chemical  Abstracts,
       7_0, 99472k  (1969).

91.  American Public Health Association, "Standard Methods  for
       the Examination of Water and Wa3tewa*-.er, 13th Edition,
       New York, N.Y.  (1971).

92.  Webb, R.G., "Isolating Organic Water Pollutants:   XAD
       Resins, Urethane Foams, Solvent Extraction, EPA-660/
       4-75-003 (1975).

93.  Larson, R.A.  and Rockwell, A.L.,  Environmental Science and
       Technology, 13_,  325  (1979).

94.  Sievers, R,E., Barkley, R.M., Ficeman, G.A., Haack, L.P.,
       Shapiro, R.H.,  Walton, H.F. in Water Chlorination;
       Environmental Impact and Health Effects, Vol. 2, Ann
       Arbor Science Publishers, Ann Arbor, Michigan, 615-
       624 (1978).

95.  Sievers, R.E., Barkley, R.M., Eicenan, G.A., Shapiro, R.H.,
       Walton,  H.F., Kolonko, K.J., Field, L.R., J. Chromato-
       graphy,  142, 745-754 (1977) .

                               150

-------
 96.  Langheld, K., Chemische Berichtc,  4_2,  2360-74  (1909).

 97.  Dakin, H.D., Cohen, J.B., Daufresne, M. ,  and Kenyon, J.,
        Proceeding of the Royal Society  o*: London. 89  !B),
        232-51  (1916).

 98.  Milroy, T.H., Biochemical Journal,  iCl,  453-64  (1916).

 -99.  Dakin, H.D., Biochemical Journal,  £, 79-95  (1917).

100.  Wright, N.C., Biochemical Journal,  2Q_,  524-32  (1926).

101.  Norman, M.F. , Biochemical Journal,  3_0,  484-96  (1936).

102.  Wright, N.C., Biochemical Journal,  30,  1661-67  (1936).

103.  Ingols, K.S., Wyckoff, H.A., Kethley,  T.W., Hodgden, H.W.,
        Fincher, E.L., Hildebrand, J.C.,  and Mandel, J.E.,
        Industrial and Engineering Chemistry,  45(2) , 996-1000
        (1953).

104.  Ir.gols, R.S., and Jacogs, G.M.,  Sewage and  Industrial
        Wastes, 2£, 258-62  (1957).

105.  Aleksiev, Boris V., Stoev,  S., Doklady Bolqarskoi
        Akademic Nauk, 22(11), 1265-8  (1969).   Chemical
        Abstracts, 72, 90866a  (1970).

106.  Kantouch, A., Abdel-Fattah, S.H.,  Chemicke  Avesti,  25(3),
        2-2-30  (1971).

107.  Pereira, W.E., Hoyano, Y.,  Summons, R.E., Bacon, V.A.,
        and Duffield, A.M., Biochimica et Biophysica Acta, 313,
        181-89  (1973).

108.  Kaminski, J.J., Bodor, N.,  and Higuchi,  T., Journal of
        Pharmaceutical Sciences,  C5 (4) ,  553-57  (1976).

109.  Taras, M.J., Journal of American Watei  Works Association,
        42, 462-74  (.1950).

110.  Taras, M.J., Journal of American Water Works Association,
        j45_, 47-61  (1953).

111.  Scares Fortunato, J., Omnia Medica  et  Therapeutica, 3
        121-33  (1976).  Chemical Abstracts,  8_6, 84019Z  (1977).

112.  Smith, J.C., Journal of the Chemical Society of London,
        213-18  (1934).

113.  Hopkins, C.Y., and Chisholm, M.J.,  Canadian Journal of
        Research, 2_4, 208-10 (1946).

                               151

-------
           114.   Morton,  A.A. ,  The Chemistry p_f Heterocyclic Compounds,
                   New York,  McGraw and Hill, 549"~(1946) .
1
\           115.   Lawson,  W.B.,  Patchornick,  A., and Witkop, B., Journal
I                   of  the American Chemical  Society, 82, 5918-32  (1960).
i
\           116.   Hinman,  R.L.,  and Bauman, C.P.,  Journal of Organic
I                   Chemistry,  2_9,  1206-15 (1964).

:           117.   Green,  N.M. ,  and  Witkop, B., Transactions of_ the New
\                   York Academy of Science,  87, 55C2 (1965).

)           118.   Haberfield,  P., and Paul, D..  Journal of the American
\          .         Chemical  Society,  87,  5502  (1965).
I                   - '  -••-•--

           119.   Sastry,  A.C.,  Subrahamanyam, P.V.R;, and Pillai, S.C.,
i                   Sewage and Industrial Hastes,  30(2),  1241-46  (1958).
i                   	 	            	

;           120.   Painter, H.A. , and Viney, M. ,  Journal o_f Biochemical and
                   Microbiological Techno]r,gy and Engireering, 1(2) ,
                   143-62 (1959).

           121.   Subrahamanyam, P.V.R., Sastry, A.C., Rao, P., and Pillai,
                   S.C.,  Journal of the Water Pollution Control Federation,
                   37^,  1142-63  (1965) .

           122.   Kahn,  L. , and  Wayman,  C., Journal of_ the Water Pollution
                   Control Federation,  30(11),  1368-71  (1964).

           123.   Hunter,  J.V. ,  and Heukelekian, H. , Journal o_f the Water
                   Pollution  Control Federation,  37, 1142-63  (1965).

           124.   Hanson,  A.M.,  and Lee, G.F., Journal of the Water
                   Pollution  Control Federation,  43 (11), 2271-79  (1971).

           1?5.   Hunter,  J.V. ,  Origin of Orc;anics from Artificial Contam-
                   ination, Faust, C.D.,  and Hunter, J.V., Ed. Organic
                   Compounds  in Aquatic Environments_, New York, Marcel
                   Dekker, Inc., pps.  51-94  (1965).

           126.   Stalling, D.L., Gehrke,  C.W.,  and Zumwalt, R.W.,
                   Biochemical  and Biophysical  Research Communicatl'on,
                   31(4), 616-22~Tl968).

           127.   Gehrke,  C.W.,  Nakamoto,  H., and Zumwalt,  R.W, , Journal
                   o_f  Chromatography,  45, 24-51 (1969J .

           128.   Gehrke,  C.W.,  and Leimer, K.,  Journal of Chromatography,
                   5_7,  219-38  (1971).                   ~

           129.   Youngs,  C.G.,  Analytical Chemistry, 31, 1019 (1959).


                                          152
       w

-------
                                                                #.«.•< «s£*wn#<
130.  Larakin, W.M., and Gehrke, C.W., Analytical  Chemistry,
        3_7, 385-89  (1965).

131.  Coulter, J.R., and Hann, C.S., Journal  of Chromatography,
        3_6, 42-9  (1968).

132.  Zumwalt, R.W., Roach, D., and  Gehrke, S.W.,  Journal of
        Chromatography, 53, 171-94  (1970).

133.  Gehrke, C.W. , Kuo, K. , and P'urnwalt, R.W. , Journal o_f
        Chromatography, 57, 209-17  (1971).

134.  Moss, C.W., and Lambert, M.A., Journal  of Chromatography,
        6J3, 134-6  (1971).

135.  Hardy, J.P., and Kerrin, S.L., Analytical Chemistry,
        44(8), 1497-9 (1972).

136.  Zanetta, J.P., and Vincendon,  G., Journal of Chromato-
        graphy, 76_, -91-9  (1973).

137.  Jonsson, Jorgen, Eyem, J. , and Sjoguist, J., Analytical
        Biochemistry, 51, 204-19  (1973).

138.  Gelpi, W.A. , Koenig, F.W. , and <~>ro, J-•  Journal  of
        Chromatographic Science, 7_,  604-13  il969) .      -  ••

139.  McLafferty, F.W., Interpretation of Mass Spectra,
        W. A. Benjamin, Inc.,  Reading, Massachusetts,  201
        (1973).

140.  "Standard Methods for the Examination of Water and
        V7astewater" , 13th ed. , American Public Health  Assoc-
        iation, New York, N.Y., 17-24 (1971).

141.  Gardner, W.S., and Le'e,  G.F.,  Environmental  Science and
        Technology, 7_, 719-24  (1973).

142.  Siegol, A., and Degens,  E.T.,  Sciences,  151, 1098-1101
        (1966).

143.  Harris, C.K., Tigane, E., and  Hanes, C.S., Canadian
        Journal of Biochemistry, 39, 439-450  (1961).

144.  Gilman, Henry, E., Organic Synthesis Collective,  Vol.  I,
        John Wiley and Sens, New York, 258  (1941).

145.  Snyder, L.R., Kirkland,  J.J.,  Introduction to Modern
        Liquid Chromatography, John  Wiley and  Sons, Inc.,
        Faw York, N.Y. (1979).
                              153

-------
146.  Dohrmann Division, Envirotech  Corporation, Technical
        Bulletin, "TOX-1 Adsorption  Module  for Total Organic
        Halogen in Water", Santa Clara,  Calif.  (1978).

147.  Regnier, F.E., Noel, R., J. Chromatographic Sci. ,  14,
        31C-320  (1976).

148.  Bristow, P.A., 3riJtain, P.N.,  Riley,  C.M., Williamson,
        B.F., J. Chromatography, JL3JI,  57-64  (197V).

149.  Kinstley, W., "A Comparison of XAD-Resins  and Granular
        Activated Carbon Methods for Measurement of Organo-
        Halogen Compounds with Water", unpublished thesis,
        North Texas State University,  Denton, Texas  (1980).

150.  Stevens, A.A., Symons, J.M., J.  Amer.  Water Works  Assoc.,
        6_9, 546-554 (1977;.

151.  American Public Health Assoc.,  Standard Methods  for the
        Examination of Water and Wastewater, 14th Edition,
        New York, N.Y.', pp. 278-282,  407-416 (1975).

152.  Sander, R., Kuhn, W., Sontheimer,  H.,  Z. f. Wasser-und
        Abwasser-Forschung, 10, 155-160  (1977),

153.  Schnitzer, M., Khan, S.U., Humic Substances in the
        Environment, Marcel Dekker,  New  York, N.Y.~PP^106-
        107 (1972) .

154.  McCarthy, P., Peterson, M.J.,  Malcolm, R.L., Thurman,
        E.M., Anal. Chem. , 5_1, 2041-2043  (1979).

155.  Glaze, W.H., Rawley, R., Burleson, J.L., Mapel,  D.,
        Scott, D.R. in "Advances in  the  Identification and
        Analysis of Org-anic Pollutants in "Water", L.H. Keith,
        Editor, Ann Arbor Science Publishers, Inc., Ann  Arbor,
        Michigan, 267 (1981).

156.  Federal Register, 4_3, 5746 (1978).

157.  Bowman, F.J., Borzelleca, J.F., Munson, A.E., Toxicology
        and Applied Pharmacology, 44,  213  (1978).

158.  Rook, J.J., Water Treatment and Examination, 23, 234
        (1974) .                                    —

159.  Symons, J.M., Journal of the American  Water Works
        Association,  6_7, "634 (1975).

160.  Federal Register, 44, 68624 (1979).
                             154

-------
                          -,-
161.  Henderson, J.E., IV, Peyton, G.R., Glaze, W.K., in
        "Identification and Analysis of Organic Pollutants in
        Water", L.H. Keith, editor, Ann Arbor Science Publish-
        ers, Inc. , Ann Arbor, Michigan, 105  (1976) .

162.  Kaiser, K.L.E. ,  Oliver, B.C., Analytical Chemistry, 48 ,
        2207 (1976).

163.  Kus, P.P.K., Chian, E.S.K., DeWalle, F.B. , and Kim, J.H.,
        Analytical Chemistry, 49^, 1023  (1977).

164.  Gould, R.F. , editor, "Pesticides Identification at the
        Residual Level", Advances in Chemistry Series, No.
        104, American Chemical Society, Washington, D.C.,
        (1971).

165.  Grob, K.K. , Jr. , and Grob, G. , Journal of Chromatographic
        Science, 106,  299  (1975).

166.  Richard,  J.J., and Junk, G.A. , Journal of the American
        Water Works Association, 69, 62 (1977).

167.  Mieure , J . P . , Journal of the American Water Works
        Association, 69, bO  (1977) .

168.  Glaze, W.H., and Rawley, R. , Journal of American Water
        Works Association, 71, 509-515  (1979).

169.  Brass, H. , Journal of the American Water Works Association,
        74. 107-112 (1982)."
                              155

-------
                           'APPENDIX  A '                            .

            LIMITED CLUS1ER SEARCH  MASS  SPECTROSCOPY

 COMPUTER PROGRAM

      The LCS  computer program w- s  originally written in BASIC
 computer language .on the  North Texas State University Hewlett-
 Packard 200 Timeshare System.   This  system could not access the
 Finnigan GC/MS  data directly,  but  program debugging could be .
 affected by manually inputting masses'-'and intensities for                 |
 single spectra.                 .                                           I
                                                                           I
      Once the program was debugged,  it  was transferred to a  .             «
 second, more  powerful Finnigan 6100  system located at th'j
 University of Texas Health Science Center, Houston, Texas.  This
!system possessed the Finnigan BASIC  interpreter which required
 a total of 16K  of  CPU core.  The system could execute the BASIC
 program and directly access the previously acquired GC/MS data.

 COMPUTER HARDWARE

:      The program is initiated oy a pushbutton interrupt on the
 front panel of  the computer system which activates the BASIC
 interpreter.  Then the program instruction statements are ready
 to be entered manually via teletype  or  to be read in from
 magnetic tape storage.  The GC/MS data  must be stored on disk as
 spectra are sequentially retrieved from there.  GC/MS data which
 has been previously stored on magnetic  tape must be read onto
 the disk prior  to  program execution.  Tabular LCS data is output
 after each spectrum is processed.  Once all the spectra have              ;
 been processed, a "command can be entered which returns the                ,
 system to the assembler-controlled mode deactivating the BASIC            \
 interpreter.  This process automatically transfers any computed           5
 chromatogram or mass spectrum onto the  cathode ray tube  (CRT)             \
 display.  In  this  case, the Limited  Cluster Search chromatogram           7
 is transferred  to  the "GC OR UPPER"  portion of the 'CRT.  Using            f
 the assembler,  the LCS data can be manipulated and plotted as             I
 though it were  data which had been acquired in a conventional             -%
 manner directly from the mass spectrometer.                               :<

 DESCRIPTION OF  THE PROGRAM                                         .        i
      Appendix B is a listing of the  LCS program.  Figure A.I  -    -       ;
 shows a flow chart which highlights  the important programming
 features of the Limited Cluster Search Program.  These features            ;
 are discussed below. .  -   -     , -	  -   -- -  		  - -     '  •   \

     .'"•'.-                 156           '                     -'.••'•'•(

-------
        C    START    J



    INPUT # OF CL'S & BR'S
                                 RETURN TO ASSEMBLER,
                                    DISPLAY GRAPHIC
                                        ON CRT,
                                          END
                                                YES
      CALCULATE  ISOTOPIC
           CLUSTER
              I
                              NO
                                              _l_
     DISCARD SMALL PEAKS
 z
              I
OUTPUT CLUSTER MATRIX
   INPUT SEARCH PARAMETERS
    CALCULATE Y INTERCEPT
                                  ALL SPECTRA CHECKED?
                                   OUTPUT TABULAR
                                    SPECTRUM DATA
                                 SCALE SPECTRUM DATA
                              NO
   RETRIEVE MASS SPECTRUM
                                                YES
                                    ALL CLUSTERS CHECKED?
                                          i

                                 OPTIONAL OUTPUT OF
                                      MASS DATA
                              NO
                                               YES
                                    ALL RATIOS EVALUATED?
                                                    l>
V
CALCULATE INTENSITY
FACTOR
J 	 	 '


EXTRACT "CLUSTER" FROM
SPECTRUM

1
y "^
EXTRACT RATIO FROM CLUSTE
"R

>

»^

k
ADD SCALED FIT TO TOTAL
>
i
SCALE RATIO FI"" |
t
COMPARE DAI
CALCULATE
k
'A RATIO TO
D RATIO
Figure A-l.  Flowchart of limited cluster search program.
                            157

-------
Calculation of Isotope Cluster

     Table A-i shows the typical computer/programmer dialogy and
data output of the program.  The program parameters which are
underlined are input by the programmer in response to computer
querries, and the tabular data are output below.  These data
include the mass data output, which is optional, and the spectral
data output.

     The program begins by requesting the number of chlorine and
bromine atoms for which the data will be searchsd.  A total of
20 atoms in any combination may be input.  Then the program
calculates the relative peak heights for the cluster and outputs
them as percentages with the base peak of the cluster being
assigned 100 per cent.

     The program uses two constants, 3.08664 for chlorii.e and
1.02041 for bromine, which represent the natural abundances for
3^               3'       79               8-
     relative to  'C1 and   Br relative to  - Br, respectively.
These constants were derived empirically in order to produce
resulting isotope peak percentages which are in agreement with
the U.S. Environmental Protection Agency mass spectral tables
(193) to two decimal places.  Either constant can be changed by
altering the corresponding equality instruction statement in
the program.

     Isotope peaks which have an intensity less than seven per
cent relative to the base peak of the cluster were discarded.
This is done in an effort to avoid possible mismatches based on
potentially not finding peaks of relatively low intensity.

Search Parameters

     The following are input consecutive] y in response to com-
puter queries:  first and last masses; first and last spectra;
baseline noise, percent; precision estimate; variation estimate.
The first and last masses define the spread of masses to be
examined in each spectrum.  The same masses must be examined for
all spectra in a given run. but not all masses which were
acquired in a GC/MS run :nust be examined.  The first mass
selected was usually the mass of tne smallest probable fragment
for a given cluster.  Thus, for a two-chlorine search, the
smallest probable fragment was (CCln) at an m/e of 82.

     The baseline noise parameter eliminated some computer
calculations of trivial data.  This was done by not comparing a
data cluster to the calculated cluster if the intensity of the
base peak for that data cluster was below the baseline noise
per cent parameter relative to the base peak of the partial
spectrum as defined by the first and lass mass search parameters.
A baseline noise percentage of seven per cent was arbitrarily
selected for all the data shown below.

                              158

-------
                   TABLE  A-l.   LCS DIALOGU'3
INPUT CL's THEN BR's
?2
CL = 2  BR = 0                  '
ISOTOPIC CLUSTER LISTING
A (1) = 100
A (2) = 64.7954
A (3) = 10.4961

INPUT LOWEST AND HIGHEST MASSES THEN-
FIRST AND LAST SPECTRA TO BE  SEARCHED
735^        -  , - -    -   •
?160
?93~
BASELINE NOISE FILTER PER CENT
77.
INPUT PRECISION ESTIMATE
?2_
INPUT VARIATION ESTIMATE
?-25
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
MASS
=
=
=
=
=
=
=
—
=
=
==
=
=
=
=
=
=
-
=
=
=
35
39
41
46
48
53
55
69
70
71
72
73
107
109
111
142
143
144
145
146
147
SPECTRUM=93
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
CUMULATIVE
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
FIT
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
.-=
^
—
=
0
0.
0.
1.
2.
3.
3.
4.
4.
6.
7.
9.
12
14
16
21
22
27
27
37
38
38

429949
859898
77292
68595
08933
49272
04368
89866
64601
95617
70351
.1219
.5402
.9585
.9585
.3255
.3255
.6925
.6925
.4265
.4265 WEIGHTED FIT=17746
                               159
                                                                           U

-------
                 The  precision  and  variation estimation  parameters  relate  to
            the  decision  process  of determining whether  a  data  cluster is
            that of a chlorine- or  bromine-containing fragment.   That process
            is described  below.

k           DECISION  PROCESS

|.                A chart  describing the  flow of the GC/MS  data  as it is
|            processed is  shown  in Figure A--2. Once the search  parameters have
t            been input, the program retrieves the first  spectrum to be
|            examined  from the data  file.  Then, the first  data  cluster to  be
?           examined  is extracted from that spectrum beginning  at the  first
!,           mass search parameter.   The  total number of  paaks in the data
|            cluster is equal  to the total number of isotope peaks in the
i           calculated cluster.
(
I                The  comparison of  the calculated cluster  with  the  data
I           cluster is a  three  step process.  First, the difference between
|           the  two clusters  is calculated.  This calculation uses  an
I  .         arbitrary formula to  derive  the Z-value which  is  a  measure of
I           the  ratio differences.   The  second step involves  scaling the
'{           Z-value to produce  the  Cumulative Fit (CF) .  The  scaling process
\.           accomplishes  two  objectives.  It introduces  two independent
>           parameters into the calculations -,;hich allow the  arbitrarily
f           derived Z-value to  be empirically optimized.  And,  it allows the
f           establishment of  a  directly  proportional relationship between
J           the  data  and  the  numerical description of that data, the CF.
 I           Tne  directly  proportional relationship causes  the baseline of
I           the  LCS chromatogram  to occur at a Y-value of  approximately
',     •      zero and  the  peaks  to extend in an upward direction frcm the
I           baseline  of the chromatogram.  If the inversely proportional
^           relationship  which  exists between the Z-value  and the data
 I           ratios were used, the baseline would' be located at  an arbitrary
 f           positive  value on the Y-axis and the peaks would  extend in a
 I           downward  negative direction.  Thus, the scaling procedure  is
 [           essentially an inversion of  the Y-axis.

 I                Once the CF  has  been calculated, other necessary factors
 |           are  applied to produce  the Weighted Fit (WF).   This is the actual
 5           value that is plotted against the spectra numbers in an LCS
 |           chromatogram.
 {
 I           Z-value
 !
 |                The  Z-value  is the absolute numerical difference between
 |           the  ratio of  the  peak heights of two adjacent  peaks in the data
 I           cluster versus two  corresponding peaks in the  calculated
 I           cluster.It  is formulated
 j


 |                          z  '  cx/cx + 2 - VDx + 2


 I                                         160

-------

TOTAL DATA SET INCLUDING ALL SPECTRA TO BE ANALYZED

1.9975
 t
         1.9056
        z  o?
          10?
                    EXTRACTED  MASS SPECTRUM
                         TO  BE ANALYZED
                     EXTRACTED DATA CLUSTER
                         TO BE ANALYZED
RATIO OF TWO PEAKS
 IN DATA CLUSTER
RATIO OF TWO PEAKS
 CALCULATED CLUSTER
                            =0    j—
                          CF^IO
                       CALCULATE CF
                  WF
        Figure A-2.  Flowchart of GC/MS data.


                     161
                                                        -i  i
                                                         4

-------
^Sm^fcWtBWWtf^^                                                                         5
       where C is the ion intensity of a peak in the calculated cluster
       at mass X or X + 2, respectively, and D is the ion intensity of
       a peak in the data cluster at mass X or X + 2 respectively.  As
       the relative data cluster peak heights approach the relative              I
       peak heights of the calculated cluster, the Z-value approaches            f
       zero indicating a high probability that the data cluster contains          j
       the same number of chlorines and/or bromines as the calculated            i
       cluster.  Note that this formula compares the peak heights of             j
       only two adjacent peaks at a time.  Therefore, n-1 Z-values will          '
       be generated for each total cluster examination where n is the            [
       number of peaks in the cluster.  If the calculated cluster were           ;
       derived from two chlorine atoms, for example, the cluster would           $
       contain three isotope peaks, and two Z-valuec would be produced.          {
       This is important in analyzing raw data as shown below.                   4
                                                                                 »
            An exanple of the relationship between the 3-value and a             ?
       data peak ratio is shown in Figure A-3. This example refers to the
       first Z-value generated for a dichlorinated isotopic cluster.             j
       In such a calculated cluster, the relative peak heights of the            i
       X and X + 2 peaks are 100 per cent and 64.7954 per cent, respec-
       tively.  The X-axis in Fig.A-3  shows the percent deviation of            ;
       the X + 2 peak in the data cluster from that perfect relative             I
       peak height as defined in the calculated cluster (64.7964 per
       cent).  This can be formulated as                                         :

                 Z = 100.0/64.7954 - 100/(fi4.7954 + % deviation).

       Thus, if the mass spectrometer recorded the DX   _ peak in the
       data cluster with a relative intensity of 62.79545, the per cent          -t
       deviation would be -2.0, and the corresponding Z--value would bt           •
       0.049.

       Cumulative Fit

            The scaling procedure which converts the Z-value to the CF,
       is itself a three step process as indicated in Figure A-2
       Initially, the CF is calculated according the formula

                             CF = VZ + C

       where " is the variation estimate parameter,  and C is a depen-
       dent variable which is a function of both the precision estimate
       and the variation estimate parameters.   As indicated in Fig.A-4,
       the CF has a range of 0.0 to 10.0.   The 10.0  CF value indicates
       a "perfect fit" between the data cluster and  the calculated
       cluster.   Note that this value corresponds to some Z-value which          :
       is always greater than zero and is defined by the precision
       estimate as described below.   The 0.0  CF value indicates the fit          =
       between the data cluster ratio and the  calculated cluster  ratio           |
       are so poor that there is no probability that the data cluster            ?
       could represent an ion fragment which  contains the proper                  '
       number of chlorines and/or bromines.   Fits can be worse than              ]

                                     ] 62           ,                             ]

                                                                                 1
                                                                                 )

-------
       0.20
H
ON
u>
       0.15
N 0.10
       0.05
             -9
                          J	I
                                                    I   I   I
                  -6
-303

PERCENT  DEVIATION
                                                            J	I
   Figure A-3.
          Relationship between Z-value and percent deviation of the M+2 peak
                   in a dichlorinated isotopic cluster.

-------
            PERFECT  FIT
                              CF =  VZ  + C
                    0.3     0.4
                    Z  VALUE

Figure A-4.  Relationship between cumulative fit and Z-value,
                                                                      'f
                                                                      V1
                                                                     1

-------

this and generate negative CF' s.  However, since the probability
of a fit can be no worse than zero, the lowest rational CF value
is 0.0, and negative values are adjusted to that value.

The Precision Estimate Parameter

     The precision estimate parameter reflects the spread of
values  ( precision) which the mass spectrometer produces when the
same species is analyzed repeatedly.  This spread usually pro-
duces a Gaussian-shaped curve when the intensity values are
plotted against the frequency of occurrence of each value (107).

     Within this spread no reliable distinction can be made
between a match and a mismacch.  Therefore, the precision
estimate defines the maximum Z-value below which improvements in
fits cannct be distinguished.  GF values which are greater than
the 10.0 perfect fit are automatically reset to the perfect fit
value.

The Variation Estimate Parameter

     The variation estimate parameter (V) adjusts for the other
primary contribution to the variation of the relative data peak
heights from the theoretical values.  This deviation has several
sources.  They all result in the unequal spurious contribution
of ion intensity to peaks in the data cluster.  These unequal
contributions cause distortions in the Z-values, and consequently
in the CF's.  Instrument background aad chromatographic column
bleed are primary sources of this effect.  Contributions from
nonhalogen atoms can also cause distortions.  Atoms such as
oxygen, silicon, sulfur, and even hydrogen can contribute to the
distortion of X + 2 peak heights due to significant contributions
of the X + 1 and X + 2 isotopes of these atoms.  This effect is
more pronounced in fragments which occur at higher m/e values.
Therefore, a provision was made in the program which allows for
a partial fit that indicates some, though less than certain
probability that the data cluster contains the appropriate
number of chlorines and/or bromines.  The partial fit region of
Fig.  A-4 shows the line which relates to CF partial fits to the
corresponding  Z-values.  Note that the variation estimate
parameter controls that relationship since it is the slope of
that line.  As the variation estimate parameter becomes a larger
negative number, the decision process approaches a "YES/NO"
system.  In such a syscem, the decision depends only on the value
of the precision estimate; below this va^ue a perfect Tit is
indicated, and above this value a perfect miss, no probability
of a fit, is indicated.  Such a system has a high risk of
misinterpreting data which happens to fall close to the decision
boundary.  Examples of such data will be seen later.

     Both the precision estimate and the variation estimate para-
meters were empirically optimized using the HALSTI data set.

                              165

-------
The optimum combination wa« then used to process the CALCLT data
set.(see Section 5  for results).

OTHER SCALING FACTORS

     Three other scaling factors are required in this program
including the halogta number factor, the cluster intensity factor,
and the spectral intensity factor.

The Halogen Number Factor

     For each complete cluster analysis, n-1 CF's are generated
where n is the number of peaks in the calculated cluster.  Since
the total CF for a perfect match between the calculated cluster
and ths data cluster has been defined as 10.0, the CF for each
rc.tio within a given cluster must be divided by  (n-1), the halo-
gen number factor, in order that each CF-contributes the
appropriate fractional contribution to the total CF for the
cluster.

     If a calculated cluster contains three peaks, as it would
in an LCS for two chlorines, two CF's would be generated for
each cluster analysis.  Therefore, each CF must contribute 5.0,
10.0/(3-l), to the total CF for the cluster.  This effect is
important in comparing two different LCS's of the same data set.
Using this factor, a perfect fit of an LCS for two chlorines
would generate the same CF as would a perfect fit for three
chlorines.  This is indicated in Fig. A-5  which shows raw
spectra (no background subtracted) from the HAL3TI data set.
The dichlorobenzene -spectrum was searched for ion fragments
containing two chlorines.  A perfect fit for one dichlorinated
ion cluster was detected as indicated by the CF .of 10.0.  The
trichlorobenzene spectrum was searched for ion fragments
containing three chlorines.  The CF of 9.6 indicates a partial
fit which is almost equal to one trichlorinated ion fragment.
The 0.4 unit error in the CF is due tc the recorded ion intensity
of the X + 2 peak in the data cluster which beginb at m/e 180
(the molecular ion cluster).  The relative value for the peak
was 92.8% versus 97.2% for the corresponding peak in the
calculated cluster.  This -4.4% deviation in the relative peak
height causes the first two Z-values for the cluster to be out-
side the precision limits of a perfect fit.  Therefore, the CF's
for those Z-values will be determined in the partial fit sector
of the decision process, and the values of those CF's will be
less than the perfect fit of 3.3333, 10.0 (4-1).  Note that if
the decision process had been a "YES/NO" system, the total CF
for the cluster would have been only 3.3333 because the first
two CF's would have been zero.
                              166

-------
     HflLOGENflTED STRNDRRD MIXTURE
     »   124     CUMULRT1VE FIT= 9.5428
                                                15
                                  150
     HflLOGENRTED STflNDfiRD MIXTURE
     «    92     CUMULRTIVE FIT= 10.0000

     130
            200
          50
100
                                           20
150
Figure A-5.  Mass spectra (no background subtracted)  oi
dichlorobenzene and trichlorobenjrene (HALSTI data set) .
                          167

-------
 , 	m,,-.,_»__™™ ,_,m,^.m.v=*^^,w::e^'^                    i"\^^?*l*t'-^l?ti*^RW«,>
-------
This  is  important  to  the  computer  internally in  plotting the LCS
data.  The product of the CF  times the  spectral  intensity factor
is  called the  "Weighted Fit"  (WF).  The WF  is the  actual value            I
which is plotted along the ordinate axis in the  LCS  chromatogram.          j"

      Fig.A-6  shows the relationship between the Total  lonization          j
Chromatogram,  the  Cumulative  Fit and the Weighted  Fit which is            i
the Limited  Cluster Search chromatogram. The data set  is being           \
searched for dichlorinated compounds.   The  TIC indicates all of
the compounds  which eluted from the GC-column including chlori-           '
nation and nonchlorinated compounds alike.   The  LCS  chromatogram          ,
indicates only those  peaks which contain dichlorinated  ion frag-          |
ments.   Note the square shape of the CF chromatogram "peak" of            ;
the dichlorobenzene peak  (spectrum numbers  89 through 99).   This
is  also  observed to some  extent for the trichlorobenzene peak             :
'(spectrum numbers  122 through 127) and  the  hexachlorobutadiene            ;
peak  (spectrum numbers 147 through 157). •                                  [

      The "noise" in the CF chromatogram between  spectra 100 and           '
120 is caused  by traces of dichlorobenzene  which are adsorbed to
the ion  source housing of the mass spectrometer.  As it desorbs,
a dichlorobenzene  spectrum of low  absolute  intensity is produced.
However, since no  other chromatographic peak elutes  at  that time,
and since the  instrument  background and chromatographic column            .
bleed are both low, the weak  dichlorobenzene spectrum still               j
accounts for the most significant  ions  in the scans.  Therefore,          ;
the CF had a high  value although "noisy" because the ion inten-           '
sities were  close  to  the  detection limit of the  instrument.
When  this factor is multiplied by  the spectral intensity factor,          ;
the resulting  WF is quite low as observed in spectrum numbers 100
through  120.   This same phenomenon is observed at  spectra numbers
130 through  138.   Note that when a nonchlorinated  analyte,  n-
tetradecane, begins to elute  from  the chromatographic column at
spectrum number 137,  the  residual  adsorbed  trichlorobenzene
becomes  insignificant and the CF drops  to 0.   Thus,  although the          :
spectral intensity factor is  high, the  LCS  response  is  zero               j
because  the  CF is  zero.                                                    r

HALSTI:  EMPIRICAL OPTIMIZATION OF THE  SEARCH PARAMETERS

      A detailed preliminary study  of the relationship between the
precision estimate parameter  and the variation estimate para-
meter was made in  an  effort to select starting values for the
subsequent empirical  optimization  study. Spectra  from  the
apices of two  GC peaks in the HALSTI data set were used for the
preliminary  study.  The peaks were bromoform, spectrum  number 46,
and dichlorobenzene,  spectrum number 92. Spectra  of the raw
data  (no background subtracted) are shown in Figure A-7.  Both
spectra  were subjected to a series of LCS searches for  dichlori-
nated ion fragments.   The bromoform spectru-i shows a dibrominated
ion fragment,  (M-Br)  , of high relative intensity  whose data
cluster  begins at  m/e 171.  The X  + 2 and X + 4  peaks of this

                               169              '                             l
   ufiifeb;

-------
            10OTOTAL IONIZATION CHROMATOGRAM
            5O-
           2
           in
               80
               SPECTRUM NUMBER
               12O

       CUMULATIVE  FIT
8O          ~"12O  12O
SPECTRUM NUMBER  •
160
                                                160
            1OOi
                  LIMITED CLUSTER SEARCH
           o
           2
           m
           2
           cn
               8O
               SPECTRUM  NUMBER
                120
160
Figure A-6. TIC and limited cluster  search  chromatograms for
                mixture of halogenated  compounds.
                              170

-------
fl
n
\ :
HHLQG&tflTED STRNDflRD MIXTURE
*    46     BRQMQFORM

 103
                 J.. .
                               100
                             150
                                                      ,11,
200
250
              HflLOGENflTEO STflNDflRD MIXTURE
              «    92     DICriLOROBENZENE

                 100
U It.. 1 L ^
I
2C
J
58 100 150
                 Figure A-7.  Mass spectra of  dichlorobenzene and
                         bromoform  (no background subtracted).

                   '                     '171

-------

cluster have a similar relative height relationship as that of
the X and X + 2 peaks in the calculated cluster.  Because of
this, bromoform can show a partial fit for the dichlorinated LCS
at ra/e 173.  The value of the CF wi]1 depend on the valuss of
the precision estimate and variation estimate parameters.  Such
a mismatch, as indicated by a bromoform peak in the LCS chromato-
gram, should be considered an interference.  The dichlorobenzene
mass spectrum shows the expected intense dichlorinated data
cluster beginning at m/e 146, the M4" cluster.

     Figure A-8 shows the data from this preliminary study.  The
study comprised three series of comparative LCS's.  Each graph
shows the Weighted Fit versus the Precision Estimate..   A sing la
variation estimate parameter is used in each graph.  In all three
cases, for low values of th'3 precision estimate parameter whei>3
the fit requirements are strenuous, the bromoform spectrum shov.-s
a WF of zero.  This indicates zero probability of the spectrum
containing a dichlorinated data cluster.  The dichlorobenzene
spectrum shows a constant positive WF indicating a perfect match.
As the precision input parameter is increased, the fit require-
ments are less strenuous, and the computer program begins to
incorrectly indicate some probability that the bromoform spectrum
contains a dichlorinated data cluster.  As the value of the
precision estimate parameter is increased further, the value of
the WF for the dichlorobenzene spectrum also begins to increase
due to the interfering mismatch of the data cluster beginning at
m/e 111, an isotopic cluster produced by a monoch.lorinated ion
fragment.

     Comparison of the three graphs shows the effect of the
variation estimate parameter on the decision process.  With a
variation estimate parameter value of -25, the precision estimate
parameter must be set at a low value  (less than 0.01) in order
to avoid possible mismatches for the bromoform spectrum.  How-
ever, for higher precision estimate parameter values at this
variation estimate parameter value, the effect of the mismatch
is diminished because the system is generally less sensitive to
mismatches as indicated by the relatively gradual slope of the
line showing the increasing bromoform interference.  With a
variation estimate parameter of -75, the system is extremely
sensitive to mismatches, as indicated by the steep slope '.f the
increasing bromoform interference, but the latitude of the
precision parameter before a mismatch is detected is much
greater.  The graph showing the data with a variation estimate
parameter value of -50 appears to be a good compromise between
sensitivity and precision estimate latitude.  This value v/as
therefore selected as the initial variation estimate parameter.
It was conpared to the value of -25 in an effort to avoid the
extreme sensitivity observed with the high variation estimate
parameter value.
                              172

-------
    d
    i-t
    (D
_ < s:
• oj ro
  f H'
H-- ro y
o en rt
y   ro
(—» Q Q,
O Mi
t-l   Ml
O rt H-
CT 5" rt
ro ro

3 J^
N < I W
ro &'•
3 h
ro H-T3

  rt a
  H- O
  o H-
    o
  ro
  en
crn-
h H- ro
O 3 en
3 t) rt
O r^ >-•
HI ro 3
O   Oi
H T? t
3 & ro
  3 O
  ro ••<
  rt-
  ro rr
    ro
    ro
              o
             O
8
o
o
O
O
O
                         O
                         O
                         O
INTEGRATOR COUNTS
       ro     4^     O)
       o     o     o
^     o     o     o
o	 o	o     o
n p
Q ro
CO
o
20

CO


|s

m
  O
  01
                                       p
                                       ro
                                       p
                                       Ol
                  O
                  01
                                                          a
                                                          -*•
                                                          o"
                                     I
                                     Ol
                                     o
ro
o
o
o
o
o
o
0)
o
o
o
                                     p
                                     ro


                                     p
                                     OJ


                                     o
                                     p
                                     Ol


-------
-o
.b.
                        HH.OKNRTCD STflNDORO mxtlKE
                      TOIHL ION nONITOR CHROTBIOGRF..
                   50   tlW   150   201!
             108
                        HnLOCtwiiEB srracwD nixiure
                       L.C.S.i 2 CL'3. P=0.2. V=-2S
              HfiLSTl


              100
    _,,™,,.,
HTP.I OeiM»W
                                                                  H«L:TI        MRLOCtHnrtD STWOWD nixiuit
                                                                           L.C.S.I 2 a's. frfs.2. v=-so
                                                                       TT
                                 •v	i	
                                 ISO    200
                                                                  108
                                                                             H«occNmED STHNCWRO rixnmt
                                                                            L.C.3.1 2 Ct'S. P=0.1. v=-6fl
                                                                       IT
                                 ifco	zoo
     Figure  A-9.    Limited  cluster search  chromatograms  for HALSTI  data  yet with  various
                             precision  and variation estimate parameters.
;    \
                                                                                     I   |
                                                                                     *i   \
                                                                                     I

-------

     The initial precision estimate parameters were  selected
based on a second preliminary  study which was an  investigation
of the precision of the mass spectrometer.  A series of  ten 100
nanogram injections of dichlorobenzene were made.  The per cent
deviation of the relative height of the  (M + 2)   ion at  m/e 148
was compared against the 64.7954% relative height'of the X +  ?
peak in the calculated cluster.  The data values  ranged  from
62.11% to 65.69% corresponding to deviations of -2.69% and 0.89%,
respectively.  The Z-values for these deviations  were 0.067 and
0.021  (note that both positive and negative per cent deviations
produce positive Z-values.)  Since this  sample was analyzed under
ideal conditions, the minimum  Z-value selected for use as the
initial precision estimate parameter was 0.1, approximately twice
the average of the Z-value deviations observed in the preliminary
study.  A value of 0.2, twice  the minimum value,  was used for
comparison.

     Figure A-9 shows the TIC and the LCS's for the entire HALSTI
data set.  The four possible combinations of the  two precision
estimate parameters and the two variation estimate parameters
were used.  Chromatograms 12-B and 12-C  show the  LCS's using a
variation estimate parameter of -25.  No interference is observed
in either chromatogram from the normal alkanes in the data set.
Some interference is observed  from the dibrominated  compounds
demonstrating the lack of adequate discriminating ability
associated with the variation estimate parameter  value.  However,
the decreased peak heights of  the interfering analytes relative
to the height of the peaks which contain dichlorinated data
clusters indicate the decreased sensitivity of the system to
the interferences.

     Chromatograms 12-D and 12-E show the LCS's using a  variation
estimate parameter of -50.  In these Chromatograms,  no inter-
ference is observed from the normal alkanes, and  the interference
from the dibrominated analytes is also eliminated.   Although the
two Chromatograms appear to be identical, subtle  differences do
exist as indicated by the differences in the amplitude values.
These values are the WF's of the apical spectra of the tallest
analyte peak in the chromatogram, in this case hexachlorobut?-
diene.  Figure A-9.0 shows the raw spectrum, and TableA-2 shows the
CF data for that spectrum. 'The CF's for m/e's 82, 94, 106, and
118 represent the proper recognition of clusters  formed  by
dichlorinated ion fragments.   The CF for m/e 143  is  actually a
mismatch of the X + 2,  X + 4,  and X + 6 peaks in  a trichlorinated
data cluster which begins at m/e 141.  The CF for m/e 155 is a
mismatch of the X + 4,  X + 6,  and X + 8 peaks in  the  data
cluster generated from a pentachlorinated ion fragment beginning
at m/e 153.   The CF's at m/e 225 and 227 are mismatches  of
various combinations of peaks from a data cluster of  pentachlo-
rinated ion fragment beginning at m/e 223.   Note  that the total
CF's for the clusters which are produced from dichlorinated ion
                              175

-------
-J  '
01
           HflLOGENflTED STflNDRRD MIXTURE

           *   151     HEXflCHLOROBUTflDIENE
100

kM,
ill I L i

L4^
.1
i

L-^J

c
ill — 1L
               SO
IOC
150
200
                                                                 250
       Figure A-10.  Mass spectrum of hexachlorobutadiene (no background subtracted).

-------
fragments are approximately the same for the two different
precision estimates, 6.13 versus 6.04.  However, the higher CF
for the 0.2 precision estimate parameter indicates that it will
exhibit a higher sensitivity to mismatches than will the 0.1
value.  Therefore, values of 0.1 for the precision estimate and
-50 for the variation estimate were selected as optimum values
for the decision process.  These values were then used for the
analysis of the CALCLI data set (Section 5).           	
    TABLE A-2.  CUMULATIVE FIT VALUES FOR HALSTT DATA SET
              (cf.  FIGURE A-10) HEXACHLOROBUTADIENE
                          Precision Estimate Value

  m/e                       0.2              0.1
82
94
106
118
143
155
225
227
0.38
1.21
1.73
6.13
6.79
8.23
13.23
16.35
0.28
1.11
1.64
6.04
0
6.80.
11.80
14.60
                              177

-------
                           APPENDIX  E

             PROGRA!!:  LIMITED CLUSTER SEARCH (LCS)
  1  DIM M(1000),0(1000)
  5  FILES Z7
 10  DIM A(21),B(21)
 20  DIM TS(50)
 30  PRINT "INPUT CL'S THEN BR'S"
 40  INPUT N,L
 45  LET T=-l
 50  MAT A=ZER
 60  MAT B=ZER
 70  MAT G=ZER
 80  MAT M-ZER
 90  REM **********  CALCULATE  CHLORINE CLUSTER ***************
100  IF N=0 THEN 220
110  LET R=3.08664
120  LET A(1)=R
120  LET A(2)=l
140  IF N=l THEN 220
150  FOR 'J--2 TO N
160  FOR 1=1 TO W
170  LET A(W-I+2)=A(W-I+2)*R+A(VJ-I+1)
180  NEXT I
190  LET A(1)=A(1)*R
200  NEXT W
210  RE1 **********  CALCULATE  BROMINE CLUSTER
220  IF L=O THEN 400
230  LET X=L
240  LET R=l.02041
250  IF A(1)>O THEN 300
260  LET A(1)=R
270  L2T A(2)=l
280  LET X=L-1
290  3F L=l THEN 400
300  FOR W=l TO X
310  FOR 1=1 TO 19
320  LET A(21-I)=A(20-I)
330  NEXT I
340  LET A(1)=O
350  FOR 1=1 TO 19
360  LET A(I)=A(I+1)*R+A(I)
370  NEXT I
380  NEXT W

                               178

-------
        :iri'^3 <.;'- : <.
 390  REM ********** BASH PEAK:   ISOTOPIC  CLUSTFR ***********
 400  LET W=O
 410  FOR 1=1 TO 20
 415  LET W1=A(I)
;420  IF WKW THEN 440
'430  LET W=W1
 440  NEXT I
 450  REM ********** CLUSTER ADJUSTMENT-CHUNK  SMALL  PEAKS  ***
 460  LET Q=0
|470  FOR P=l TO 20
1480  IF A(P)>.07*W THEN 560
,490  IF A(P)=0 THEN 590
;500  FOR R=P TO 20
.510  LET A(R)=A(R+1)
-. 520  NEXT R
 530  LET Q=Q+1
'540  IF P>W THEN 560
.550  LET P=P-1                  i
 560  NEXT P
 580  REM **********  PRINT ISOTOPIC CLUSTER   ***************
i590  PRINT
!600  PRINT
}670  PRINT "CL-=";N;"BR=";L
:620  PRINT "ISOTOPIC CLUSTER LISTING"
:630  IF A(W)=O THEN 6PO
 640  FOR 1=1 TO 21
 650  IF A(I)=O THEN 670         :
;660  PRINT "A(";I;")=";A(I)/A(W)*100
i670  NEXT I
 680  PRINT
 690  PRINT
 700  LET N=N+L-Q                '
;710  REM **********  INPUT SEARCH PARAMETERS  ***************
;720  PRINT "INPUT LOWEST AND HIGHEST MASSES THEN-"
 730  PRINT "FIRST AND LAST SPECTRA TO BE  SEARCHED"
 750  INPUT L,H,C,F
'760  REM **********  INPUT DATA BASELINE  NOISE FILTERING  ***
;770  PRTNT "3ASELINE NOISE FILTERING, PERCENT"
.780  INPUT D
.'910  LET D=D/100
 920  REM ********** INPUT PRECIS. AND VARIATION  ESTIMATES  **
 ^30  PRJNT "INPUT PRECISION ESTIMATE
 940  INPUT XI
 950  PRINT "INPUT VARIATION ESTIMATE"
 960  INPUT i41
 970  REM **********  CALCULATE Y INTERCEPT  ****************
;980  LET Bl=10-(M1*X1)
 990  REM **********  SPECTRUM SELECTION LOOP  **************
1000  FOR I=C TO F
1010  DREAD #1,I,M
102C  REM ********** CALCULATE SPECTRUM INTENSITY FACTOR ****
1030  REM *************** AND BAS3 PEAK, SPECTRUM ***********
                               179

-------
' ^? A^^'VA**"* J-t?to- ^£t-l>'ii>&>>
 • -'- - 'JfelWivA'JSj "Vj> /< i*£v£. T
                                                           •fi^ty ^ci-.**-; *JEyTBgjgpg«» ypBj^jti
                                                                             •Of'
        ••  1040
          1050
          1060
          1070
         ,1080
          1090
        .  1100
        ' 1120
         '1130
         • 1140
         ' 1150
         ; 1160
         : 1170
        ; 1180
          1190
         ' 1200
          1210
        > 1220
        : 1230
          1240
        : 1250
        ; 1260
        , 1270
        •• 1280
        ! 1290
        ; 1300
        i 1310
        ; 1320
        ' 1330
          1340
        :  1350
        :  1360
          1370
        .  1380
        '•  1390
        ;  1400
        '  1410
        •  1420
        ;  1430
          1440
          1450
        :  1460
          1470
        '  1480
          1490
          1500
          1510

        >  1520
          1530
          1540
LET R=O            '        '
LET E=M(L)                 ;
FOR J=L TO H
LET E2=M(J)
IF E2>0 THEN  1100
LET M(J)=.01
LET E=E2
LET R=R+E2      '	       	-  -  •-	  	
NEXT J
LET R=R*.G1
LET G(I)=0
REM **********  CLUSTER SELECTION LOOP  *****************
FOR J=L TO H-2**I
REM **********  COPY  DATA CLUSTER INTO  MATRIX B ********
REM **********  CALCULATE BASE PEAK FOR CLUSTER ********
LET B3=0
FOR P=l TO N+l
LET B(P)=M(J+(P-1)*2)
LET B2=B(P)
IF B3>B2 THEN 1260         ......	   ......
LET B3=B2
NEXT P
REM ******•>***  ASSIGN CLUSTER HEIGHT SCALING FACTOR ***
LET X=B3/E
REM **********  TEST:  CLUSTER EXCEEDS NOISE CUTOFF *****
IF B3X1 THEN  1410
LET G(I)=G(I)+10*X/N
GO TO 1430
    y*********  ZERO  <
                                                          	4
                       FIT < TEN ***********************
LET G(I)=G(I)+(M1*Z+B1)*X/N
NEXT P
REM *******•-**  PRINT-DATA ANALYSIS  ********************
IF G(I)<=T4-.2 THEN 1470
PRINT "MASS=';J;"CUMULATIVE FIT =";G(I)
LET T=G(I)
NEXT J
LET T=G(I)
LET G(I)=G(I)*R
PRINT "SPECTRUM #=";I;"CUMULATIVE FIT=";T;"WEIGHTED FIT",

LET T=-l
NEXT I
END                       i
                                        180

-------

-------