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
-------
------- |