DEVELOPMENT AND EVALUATION OF ANALYTICAL PROCEDURES …

DEVELOPMENT AND EVALUATION OF ANALYTICAL PROCEDURES FOR BROAD

SPECTRUM ANALYSIS OF SYNTHETIC ORGANIC CHEMICALS IN

SOURCE AND FINISHED DRINKING WATERS

Gregory S. Durell Russell F a Christman Andrea M. Dietrich

Department of Environmental Sciences & Engineering School of Public Health

University of North Carolina Chapel Hill, NC 27514

Project No. 70034

The work on which t h i s p u b l i c a t i o n i s based was supported by funds provided by t h e Water Resources Research I n s t i t u t e of The Un ive r s i t y of North Ca ro l ina .

ACKNOWLEDGEMENT

This work was made possible by a research grant from the Water Resources Research Institute of the University of North Carolina, and its support and cooperation have been greatly appreciated.

Special thanks are due Susan Ossoff and Kathy MacKinnon for their assistance in preparing this manuscript. The willingness of Dr. J. Donald Johnson and Dr. Francis A. DiGiano to review this manuscript was also greatly appreciated.

The staff at the Water Treatment Plant in Eden, NC were very cooperative and deserve thanks for their help with water sampling and providing pertinent information on water quality and water treatment at Eden.

DISCLAIMER

The c o n t e n t s o f t h i s p u b l i c a t i o n d o n o t n e c e s s a r i l y r e f l e c t t h e v iews and p o l i c i e s o f t h e Water R e s o u r c e s Resea rch I n s t i t u t e , n o r d o e s men t ion o f t r a d e names o r commerc ia l p r o d u c t s c o n s t i t u t e t h e i r endorsement o r recommendat ion f o r u s e by t h e I n s t i t u t e o r t h e S t a t e o f Nor th C a r o l i n a .

ABSTRACT

Based on the scientific literature, continuous liquid-liquid extraction (CLLE) and closed loop stripping analysis (CLSA) were chosen for development and evaluation as the most appropriate extraction and concentration procedures for broad spectrum analysis of synthetic organic chemicals (SOCs) in natural waters. These procedures complemented each other better than any other combination of methods available.

A set of 16 SOC standards was selected which would be representative of the wide range of SOCs which are encountered as pollutants in natural waters. Method optimization experiments were carried out by determining the percent recoveries of these representative SOCs under a variety of analytical conditions. Using the optimized analytical conditions, the CLLE and CLSA procedures were applied to natural waters so that water matrix effects could be studied and broad spectrum SOC analysis, identification, and quantification performed.

A final evaluation of the developed methods was carried out by comparing results from four water samples, which were analyzed by the CLLE and CLSA methods, to results obtained at a qualified laboratory using the currently recommended EPA methods. The developed methods were found to be powerful for the analysis of a broad range of SOCs in natural waters. The CLLE/CLSA combination of procedures was applicable to a wider range of SOCs and had greater sensitivity than currently recommended EPA methods.

TABLE OF CONTENTS

Paqe

ACKNOWLEDGEMENTS. . . . . . . . . . . . . . . . . . . . . . i i i A B S T R A C T . . . . . . . . . . . . . . . . . . . . . . . . . . i v

LISTOFFIGURES . . . . . . . . . . . . . . . . . . . . . . v i i i

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . X

SUMMARY . . . . . . . . . . . . . . . . . * . . . . . . x i i

CONCLUSIONS AND RECOMMENDATIONS . . . . . . . . . . . . . . xv

CHAPTER I. INTRODUCTION

1-1. RATIONALE FOR DETERMINING SOCS IN WATER . . . . . . . . . . . . . . . . .

. . . . . . 1-2. ANALYTICAL METHODS AVAILABLE

1-3. RESEARCH OBJECTIVES

CHAPTER 11. LITERATURE REVIEW

. . . . . . . . . . . . . 2-1. INTRODUCTION.

2-2. BRIEF OVERVIEW OF SOC REMOVAL IN WATER . . . . . . . . . . . . . . . TREATMENT 7

. . . . 2-3. HISTORY OF SOC ANALYSIS TECHNIQUES 8

2-4. IDENTIFICATION METHODS IN SOC A N A L Y S I S . . . . . e . . e . . . . . . . 9

2-5. SOC EXTRACTION AND CONCENTRATION . . . . . . . . . . . . . . . . METHODS

2-5-A. MEMBRANE METHODS IN SOC ANALYSIS. . . . . . . . . . . .

2-5-B. RESIN METHODS - XAD AND . . . . . . . . . . . . TENAX-GC.

2-5-C. LIQUID-LIQUID EXTRACTION METHODS . . . . . . . . . . . .

2-5-D. HEADSPACE METHODS - PURGE AND TRAP AND CLOSED LOOP STRIPPING ANALYSIS (CLSA) . . . . . . . .

2-6. COMPARISON OF HEADSPACE AND OTHER . . . . . . . . . . . . . . . . METHODS

2-7. SUMIJlARY OF LITERATURE AND CHOICE OF . . . . . . . . . . . . . . . . METHODS

CHAPTER 111. MATERIALS AND METHODS

3-1. SELECTION OF ANALYTICAL METHODS . 27

3-2. STANDARDS SELECTION . . . . . . . . . 27 3-3. PRELIMINARY GC WORK - RELATIVE RESPONSE

FACTOR DETERMINATIONS . . . . . . . . . 3-4. EVALUATION AND OPTIMIZATION OF

ANALYTICAL METHODS - CLLE . . . . . . . 33

3-5. EVALUATION AND OPTIMIZATION OF ANALYTICAL METHODS - CLSA . . 43

3-6. APPLICATION OF METHODS TO SOURCE AND FINISHED DRINKING WATERS . . . . . . . . 46

3-6-A. IN-DEPTH SOC ANALYSIS OF SMITH RIVER - BACKGROUND AND EXPERIMENTAL DESIGN . 50

3-6-Be COMPARISON OF CLLE/CLSA TO EPA'S RECOMMENDED METHODS . 51

a - .

CHAPTER IV. RESULTS AND DISCUSSION

4-1. CLLE OPTIMIZATION EXPERIMENTS . 53

4-1-A. EXTRACTION TIME OPTIMIZATION . . 53

4-1-B. EXTRACTION VOLUME OPTIMIZATION . . . . . . . . . . 59

4-1-C. pH OPTIMIZATION . . . . . . . . 59 4-1-D. p H VARIATIONS AND

COMPLEMENTARY pHs . . . . . . . 64

4.1.E . PRELIMINARY COMPARISON OF . . . . CLLEANDEPAMETHOD625

. . . . . 4.1.F MINIMUM DETECTION LIMITS

. . . . . . 4.2 CLSA OPTIMIZATION EXPERIMENTS

4.2.A . PURGE TEMPERATURE OPTIMIZATION . . . . . . . . . .

. . * . . 4.2.B PURGE TIME OPTIMIZATION

. 4.2.C EXTRACTION SOLVENT CHOICE . . . 4-2-Dm MINIPIUM DETECTION LIMITS . . . .

4.3 . APPLICATION OF CLLE AND CLSA METHODS TO SOURCE AND FINISHED DRINKING WATERS . .

. 4.3.A WATER MATRIX EFFECTS . . . . . .

. 4.3.B PRELIMINARY SOC SURVEY . . . . . SOLVENT ARTIFACTS

4.3.D . IN-DEPTH SOC ANALYSIS OF THE . . . . . . . . . . SMITH RIVER

4-3-Em COMPARISON OF CLLE/CLSA TO EPA'S RECOMMENDED METHODS . . . . 100

. . . . . . . . . . . . . . . . . . . . . . . . . REFERENCES- 105

LIST OF PUBLICATIONS . . . . . . . . . . . . . . . . . . . . 116

. . . . . . . . . . . . . . . . . . . . . . . . . APPENDICES 117

Note: trichloroethylene is also-referred to as trichloroethene; tetrachloroethylene is also referred to as tetrachloroethene .

vii

LIST OF FIGURES

comparison of Typical Separation from Packed and capillary Columns . . . . . . . . . . . . . . . . . . 10

Purge and Trap Apparatus . . . . . . . . . . . . . . 18

CLSA - Purge Mode . . . . . . . 20

CLSA - Solvent Extraction Mode . . . . . . . . . . . 20

GC Run of Primary and Secondary Internal S t a n d a r d s . . . . . . . . . . . . . . . . . . . . . . 35

Relative Response Factors . . . . . . . . . . . . . . 36

~ilution Adjusted Relative Response Factors . . . . . 3 7

~iquid-Liquid Extractor for Liquids Lighter than W a t e r . . . . . . . . . . . . . . . . . . . . . . . . 39

Liquid-Liquid Extractor for Liquids Heavier than W a t e r . . . . . . . . . . . e . . . . . . . . . . . . 40

~uderna-Danish Evaporator . . . . . . . . . . . . . . 41

Micro-Snyder Column Evaporator . . . . . . . . . . . 41

CLSAApparatus . . . . . . . . . . . . . . . . . . . 44

Natural Water Sources Used in Methods Evaluation Work . . . . . . . . . . . . . . . . . . . . . . . . 47

CLLE - Typical GC Run . . . . . . . . . . . . . . . . 54

C L L E - T y p i c a l G C R u n o f B l a n k . . . . . . . . . . . 55

CLLE - Solvent (Methylene Chloride) GC Run . . . . . 5 6

Methylene Chloride Run on GC After Kuderna-Danish Concentration . . . . . . . . . . . . . . . . . . . . 56

CLLE - Percent Recovery vs. Extraction Time CLLE - Percent Recovery vs. Extraction Solvent Volume . . . . . . . . . . . . . . . . . . . . . . . 61

CLLE - Percent Recovery vs. Extraction pH .

v i i i I

CLLE - Percent Recovery vs. pH Variations and Complementary pHs . . . . . . . . . . . . . . . . . CLLE - Percent Recovery of SOCs in a Preliminary comparison of CLLE to EPA Method 625 . . . . . . . . CLSA - Typical GC Run . . . . . . . . . . . . . . . . CLSA - Typical GC Run of Blank . . . . . . . . . a . CLSA Solvent (Carbon ~isulfide) Run on GC . . . a . . SOC Standards Solvent (Acetone) Run on GC . . . . . . SOC Applicability of CLLE/CLSA Analytical Methods . . CLSA - Percent Recovery vs. Purge Temperature . . . . CLSA - Percent Recovery vs. Purge Time . . . . . . . CLSA - Percent Recovery vs. SOC Concentration . . . . GC Run of Fieldcrest WTP Finished Water - CLLE . a . GC Run of Fieldcrest WTP Finished Water - CLSA a a

LIST OF TABLES

Paqe

synthetic Organic Chenical Maximum Contaminant Levels . . . . . * . . . * . . . . * * * * . * * * * 2

Volatile Organic Chemical Recommended Maximum Contaminant Levels . . . . . .

. * . * * * . * * * . 4

Applicability of SOC Concentration/Analysis Methods . . . .

- * . * . m e * * * 2 4

Sources of Information for Selecting Standard SOCS . . . . . . . . . . . . . . . . . . . . . . . 28

List of SOCs Chosen as Standard Compounds and Relevant Chemical Data . . .

* e * e e * a * * * e * 29

Matchup of Primary and Secondary Internal Standards . . . . * . * * * * * * * . . . * * . . * * 3 0

Relative Response Factors of Primary:Secondary Internal Standards . . . . . . . . . . * * * * . * * 3 4

CLLE - SOC Percent Recovery vs. Extraction Time CLLE - SOC Percent Recovery vs. Extraction Solvent Volume . . . . . . . . . . . . 60

CLLE - SOC Percent Recovery vs. Extraction pH . CLLE - SOC Percent Recovery vs. pH Variations and Complementary pHs . . . . . .

~ ~ * ~ * * m ~ ~ ~ * 65 a .

CLLE - Preliminary Comparison of Extraction ~fficiency of CLLE to EPA Method 625 . . . . . . . . 67

CLSA - SOC Percent Recovery vs. Purge Temperature . . 75

CLSA - SOC Percent Recovery vs. Purge Time . . . . . 77

CLSA - SOC Percent Recoveries, Comparing Methylene Chloride to Carbon Disulfide as the Trap Elution Solvent . . . . . . . . . . . .

* * m * * . * e e * e 80

CLSA - SOC Percent Recovery vs. SOC Concentration . CLLE - Water Matrix Effects. SOC Percent Recovery for Natural Waters . . . . . .

* * * * . * * * . * . 84

CLLE - Water Matr ix E f f e c t s . SOC Percent Recovery . . . . . . . . . . . . f o r D i f f e r e n t Natura l Waters

CLSA - Water Matrix E f f e c t s . SOC Percent Recovery . . . . . . . . . . . . . . . . . f o r Natura l Waters

CLSA - Water Matr ix E f f e c t s . SOC Percent Recovery . . . . . . . . . . . . . . . . . f o r Natura l Waters

Pre l iminary SOC Analys is of t h e Deep and Yadkin . . . . . . . . . . . . . . . . . . . . . . . Rivers

SOC Analys is of Smith ~ i v e r Water by CLLE - . . . . . . . . . . . F i e l d c r e s t Water P l a n t I n f l u e n t

SOC ~ n a l y s i s of Smith ~ i v e r Water by CLSA - F i e l d c r e s t Water P l a n t I n f l u e n t . . . . . . . . . . . 92

SOC Analys is of Smith River Water by CLLE - Four . . . . . . . . . . . . . . . . . . S a m p l i n g P o i n t s . 95

SOC Analysis of Smith River Water by CLSA - Four S a m p l i n g P o i n t s . . . . . . . . . . . . . . . . . . . 96

SOC ~ n a l y s i s of Smith River Water - A Comparison o f C L S A t o E P A M e t h o d 6 2 4 . . . . . . . . . . . . . . 102

SOC Analysis of Smith River Water - A Comparison of CLLE t o EPA Method 625 . . . . . . . . . . . . . . 103

SUMMARY

Analyses of synthetic organic chemicals (SOCs) in source and drinking waters at the part per billion (ppb) and part per trillion (ppt) levels demands the use of sensitive and accurate analytical techniques. To cope with monitoring those chemicals with established maximum contaminant levels (MCLs) and to survey water samples for other SOCs that may be present, broad spectrum methods of analysis are employed. These methods should have the ability to detect the target SOCs at the required level, as well as to be able to reasonably detect a variety of other non- targeted synthetic organic chemicals that may be present.

This research evaluated the suitability of selected analytical techniques for broad spectrum analysis of synthetic organic chemicals in source and finished drinking waters. The ob j ectives were f ive-f old:

To review the scientific literature and select appropriate analytical procedures(s).

To select representative SOCs to develop, evaluate, and optimize the chosen analytical procedure(s).

To develop, evaluate, and optimize the selected analytical procedure(s), using %leant1 laboratory solutions and ttnaturaltl water matrices.

To apply the procedure(s) to broad spectrum analysis of SOCs in source and finished drinking waters.

To compare the selected procedure(s) to EPA's broad spectrum methods in terms of specificity and sensitivity.

Based on the literature, Reverse Osmosis/Membrane ~iltration/Dialysis, Tenax-GC Adsorption/Thermal Desorption, XAD Adsorption/Solvent Elution, Liquid-Liquid Solvent Extraction, Bellar Purge and Trap, and Grob Closed Loop Stripping Analysis were considered as possible analytical techniques for evaluation. The two methods selected for evaluation were Closed Loop Stripping ~nalysis (CLSA) and Continuous Liquid-Liquid Extraction (CLLE) with methylene chloride as the solvent, followed by Kuderna-Danish concentration. CLSA is an excellent method for the analysis of volatile, non-polar synthetic organic chemicals. This method is complemented by CLLE, which is suitable for the analysis of non-volatile, moderately polar aqueous organic chemicals. These two methods can provide a powerful combination for the analysis of a broad spectrum of aqueous synthetic chemicals. CLSA and CLLE were the methods selected for removing and concentrating the SOCs from aqueous samples; both these

x i i

ou c,O

cn 'd a d d cd F: PC3 0 3 a-rl 0 -4 c, U.3 ; ca4J a , E : O A -4 a c,

w U a, 0-rl k k a, F:c, 3 0 0

R mc, o a d & 0 kc, G d U c, Pca, aaJPc J i m m

trichloropropene, C3-benzene, hexachlorocyclopentadiene, and 1,1,2,2-tetrachloroethane. Organic analysis of samples taken from the Smith River revealed the presence of N(ethy1)-N(pheny1)- acetamide, 1,3,3-trimethyl-2-indolinone, a xylene, a C3-benzene, a triazine, and a compound detected and designated Unknown E. The concentrations of these Smith River SOCs were in the 1-3 ug/l range.

The CLSA and CLLE methods developed in this research were compared to EPA Method 624 for volatile organic chemicals and EPA Method 625 for extractable organic chemicals. Duplicate samples of SOCs were analyzed; the CLSA and CLLE analyses were performed at the Department of Environmental Sciences and Engineering at UNC, while the EPA procedures were performed by a qualified laboratory. The qualified laboratory routinely performs these EPA analyses on a state-wide basis. The CLSA and CLLE methods demonstrated improved sensitivity and an ability to detect a greater number of SOCs.

Note: Trichloroethylene is also referred to as trichloroethene; tetrachloroethylene is also referred to as tetrachloroethene.

x i v

CONCLUSIONS AND RECOMMENDATIONS

Reliable analytical methods for the analysis of SOCs in environmental water samples are essential in order to safeguard the public and improve environmental quality. The goal of this research project was primarily to evaluate and develop applicable separation and concentration procedures which can be used in such analytical methods for broad range SOC analysis.

The CLSA and CLLE methods are a powerful combination of procedures which complement each other well for the analysis of a broad spectrum of SOCs. A large number of SOC pollutants, many at ng/l levels, were identified in natural water samples by these methods. In a comparison to the currently recommended EPA methods for broad spectrum SOC analysis, using natural water samples, our use of the CLSA/CLLE methods identified and quantified more SOCs than the EPA methods. However, the unfortunate lack of QA/QC information from the laboratory performing the EPA analytical methods for this comparison study does adversely affect the validity of this comparison.

The CLSA and CLLE combination of methods complement each other very well for the analysis of a broad range of SOCs, as shown by our SOC standard and natural water studies. However, there are undoubtedly SOCs which may not be analyzed for by these two methods alone but which may be detected by the Master Analytical Scheme or by the EPAfs 600-series of methods. The combination of CLSA and CLLE is not the final and complete answer to SOC analysis, but considering the impressive broad and general capabilities of CLSA/CLLE and the relative speed and ease of analysis, this combination of procedures appears to offer obvious advantages over other currently available methods. The procedures developed in this work are not complete analytical methods at this point. They are, however, very promising procedures which can be further evaluated and developed into complete protocols such as the EPA methods or the Master - . Analytical Scheme.

~dditional improvement of these methods is definitely possible. For the CLLE analysis the artifact interferences might '

be avoided by distilling the solvent before use. A device for solvent dispersion in the CLLE apparatus might possibly increase the extraction efficiency, and extracting the sample at an acidic pH before the neutral or base extraction has been suggested as a better sequence for optimum recovery (Henderson et al., 1977, pp. 105-112). For both CLLE and CLSA analysis, salting out may decrease the water solubility and increase the recovery of the SOCs. Some highly volatile SOCs that coelute with the solvent are not detected byoCLSA, but lowering the starting GC ovgn .

tegperature from 40 C, which was used in this work, to 25 C or 30 C should allow more of the most volatile SOCs to be detected.

Other improvements on CLSA could include heating of the glass circuit all the way from the water bath to the charcoal filter so as to prevent condensation and associated loss of SOCs before they reach the filter. Larger charcoal filters (This study used 1.5 mg filters, but 5 mg filters are available.) can be evaluated. The CLSA method as used was applicable to trace organic analysis and inappropriate for heavily contaminated water samples. The larger filter might extend the method's applicability to more heavily contaminated waters. The larger filter might also decrease competitive binding effects which were observed even at relatively low concentrations in fairly clean samples. Competitive binding effects can significantly reduce the recovery of SOCs present and need to be better understood. The CLLE/CLSA methods were applied to a variety of water matrices, with the result that the water matrix itself affected the ability of the analytical methods to detect and quantitate aqueous SOCs. These water matrix effects could be more thoroughly investigated for the purpose of gaining understanding of the matrix components that affect recoveries and their relative importance.

This research has been preliminary work in the development of methods for the analysis of a broad range of SOCs in source and finished drinking waters. Further work and improvements are possible, but the CLLE/CLSA combination of procedures appears very promising as a tool to be used for the analysis of SOCs present in the environment.

CHAPTER I. INTRODUCTION

1-1 . Rationale for Determining Synthetic Organic Chemicals in Water.

Many thousands of man-made organic chemicals (synthetic organic chemicals) have been added since World War 11 to those organic chemicals naturally present in the environment. Historically, most synthetic organic chemicals ( S O C s ) are therefore relatively new in the environment and are of major concern since they often have high persistence and are poorly removed by conventional water and wastewater treatment.

Investigations in the United States and throughout the world have shown that a high proportion of existing surface waters and groundwaters are contaminated with industrial and agricultural

- organic wastes from a variety of point and nonpoint sources. In the U.S. alone, about 700 different chemical pollutants, the majority of which are SOCs, have been found in public drinking water systems (Hanson, 1984, pp. 26-27). The effects of most SOCs on the environment and human health is unknown, but low levels of several SOCs have been shown to have serious environmental impacts and adverse health effects in animal studies, including acute and chronic toxicity and carcinogenesis. Current estimates for cancer, for example, indicate that in the U.S., as much as 75-80 percent of the cancer incidence is of man- made environmental origin and mainly due to SOCs, according to some sources (Commoner, 1977, pp. 49-72).

of bec to var The acc the

Since Rachel Carson documented the long-ten adverse effects DDT in her book Silent S~rinq (1963), the American public has ome increasingly aware of the risk of SOCs to human health and the environment as a whole. These SOCs include pesticides, a iety of industrial chemicals, and disinfection by-products. re are regular reports of improper handling and disposal and . . .

idents involving hazardous chemicals at locations all across country.

Concerns about the risks associated with SOCs have prompted legislative actions such as the Clean Water Act, the Clean Air Act, the Safe Drinking Water Act, the Toxic Substances Control Act, the Resource Conservation and Recovery Act ( R C M ) , and the Comprehensive Environmental, Response, Compensation, and Liability Act (Superfund). The development of enforceable regulations and standards is a lengthy process, and to date, there are only seven maximum contaminant level (MCL) standards for SOCs in the National Interim Primary Drinking Water Regulations (NIPDWR) (Table 1-1). The first six SOCs listed in Table 1-1 are pesticides which are either banned or seldom used, thus they seldom occur in drinking water supplies. The standard for total trihalomethanes is primarily intended to regulate

TABLE 1-1

S y n t h e t i c O r g a n i c C h e m i c a l Maximum C o n t a m i n a n t L e v e l s

SOC c o n t a m i n a n t

E n d r i n

L i n d e n e

N e t h o x y c h l o r

T o x a p h e n e

2,4-D

2,4,5-T

T o t a l T r i h a l o r n e t h a n e s

NIPDVR MCL, mg/l

levels of disinfection by-products. Despite these legislative actions, it is evident that current regulations are insufficient to protect humans against the hazards resulting from the SOCs in our environment since the majority of SOCs of concern are not presently regulated. Steps have been taken to evaluate the extent of SOC contamination in our water supplies and to regulate additional SOCs which may be of major concern. Nine commonly found volatile organic chemicals (VOCs) have been proposed for regulation and recommended maximum contaminant levels (RMCLs) set (Table 1-2). VOCs are low molecular weight SOCs with relatively high vapor pressure which therefore evaporate readily. Many more SOCs are currently being considered for regulation in the near future .

RMCLs are nonenforceable health goals set at levels that would result in no known or anticipated adverse health effects with an adequate margin of safety. The recommended limits for noncarcinogens are based on chronic toxicity data, whereas those for proven-animal carcinogens (and suspected human carcinogens) are set at zero. MCLs, on the other hand, are enforceable standards to be set as close to the RMCLs as technically and economically feasible. MCLs are based on health considerations, treatment technologies, cost of treatment, and other factors. There is controversy as to what levels should be set as MCLs for carcinogens. Should the MCL be set at zero, and if so, how can this be tested, or should a certain risk, or number of deaths in a population which are SOC, be allowed? Shou determining factor, or

predict Id cost

* should

ed to be attributable to a specific of treatment technology be the the MCL be set at the minimum

detection limit possible by modern analytical methods? The MCLs for carcinogens will include consideration of all of these options, and no set guidelines have been developed at this time.

It is clear that accurate and sensitive analytical methods for the determination of SOCs in waters are essential in order to solve a large number of environmental problems and to protect human health.

1-2 . Analytical Methods Available

Much of the concern over SOCs in our waters stems from improvements in our analytical capabilities over the last few decades, particularly is possible to detect and even picogram per

during the past 5 to 10 years. Today, it many SOCs at the nanogram per liter (ng/l) liter (pg/l) levels.

Detection of relatively high concentrations of organic pollutants is possible by the direct injection of the sample onto a chromatograph with an appropriate detector, but there is a variety of potential problems associated with this procedure. A more common method is to remove the SOCs of interest from the

TABLE 1-2

Volatile Organic Chemical Recommended Maximum Contaminant Levels

VOC Contaminant

Trichloroethylene

Tetrachloroethylene

Carbon Tetrachloride

l,l,l-Trichloroethane

Vinyl Chloride

1,2-Dichloroethane

Benzene

1,l-Dichloroethylene

1,4-Dichlorobenzene

RMCL, ug/l Proposed MCL, uq/l

Note: Trichloroethylene is also referred to as trichloroethene; tetrachloroezhylene is also referred to as tetrachloroethene.

original sample matrix and to concentrate them for final analysis and identification. There is a varietv of these ~removal/concentration~ methods availagle.

For the more volatile SOCs, a ffheadspacelt methodology is generally applied. This involves the purging and evaporation of the SOCs out of the water, followed by trapping. The trap containing the SOCs can then be extracted with a solvent, or the SOCs can be thermally desorbed from the trap directly into the chromatograph used for separation and identification. The flheadspacelv methods include Bellarfs Purge and Trap method and Grobis Closed Loop Stripping Analysis (CLSA) . Less volatile SOCs are generally either directly extracted from the water wit;? a solvent, or the water is passed over a trap which adsorbs the SOCs, which are then extracted with a solvent or thermally desorbed onto the chromatograph. Methods for extracting and concentrating less volatile SOCs include a variety of liquid- liquid solvent extraction methods, XAD resin adsorption, Tenax-GC adsorption/thennal desorption, and reverse osmosis/membrane filtration methods.

These removal and concentration methods are usually followed by separation and identification by gas chromatograph (GC) with an appropriate detector. A mass spectrometer (MS) is often employed for identification of unknowns, though it may be sufficient to use a flame ionization detector (FID) for gezeral SOcs or electron capture detector (EC) for halogenated organics. High performance liquid chromatography (HPLC) with an apprspriate detector is being increasingly used in trace organic analysis, in particular for analysis of thermally labile organics and for organics that are poorly volatilized.

The Environmental Protection Agency (EPA) has developed a series of methods (601-613) for the analysis of specific classes of SOCs (such as phenols, phthalate esters, purgeable aromatics, etc.). The EPA has also developed methods 624/625 and 1621/1625, which are combinations of methods that will analyze for a 3roader range of SOCs in water. Methods 624 and 1624 are purge ane trap methods that are designed for the analysis of volatile SOcs, while 625 and 1625 are liquid-liquid solvent extraction mezhods for nonvolatile SOCs. These EPA methods have been specifically designed to analyze for the 111 SOCs on the EPAfs Priority Pollutants list, which was developed earlier as a part of a consent decree resulting from litigation.

The number of SOCs which are of environmental and reqAatory concern has increased well beyond the list of 111, and the newer discoveries do not necessarily fall into EPAfs original classification scheme. The additional SOCs have a variety of physical and chemical characteristics; it is desirable to kave analytical methods that will analyze for a broad range of SOCs.

Broad spectrum analysis is essential to detect as many SOCs as possible, even if they are present in very low concentrations. This is the goal of the development of analytical methods in this research project. Broad spectrum analysis is essential in order to Itget the whole picturett and fully understand the degree of SOC pollution, to be able to evaluate possible adverse environmental and health effects, and to take appropriate actions to protect against SOC pollution.

1-3. Research Objectives

The research project presented here attempts to develop analytical procedures which will be suitable for broad spectrum analysis of SOCs in source and finished drinking waters, using commercially available supplies and instrumentation. Most of this work consists of the evaluation and development of applicable separation and concentration procedures to be used in this analytical scheme. Specific objectives include:

- The review of the scientific literature for the purpose of selecting appropriate analytical procedure(s).

- The selection of a representative set of SOCs which will be used to develop, evaluate, and optimize the chosen analytical procedure(s).

- The development, evaluation, and optimization of the analytical procedure(s) chosen, using the SOCs standards in ttcleanw laboratory solutions, as well as in ttnaturaltt water matrices.

- The application of the procedure(s) to broad spectrum analysis of SOCs in source and finished drinking waters.

- The comparison of the procedure(s) to EPA's broad range methods in terms of specificity and sensitivity.

CHAPTER 11. LITERATURE REVIEW

2-1. Introduction.

Reliab only recent relatively compounds ( recognized pollution h increased a available.

le and sensitive method ly become available, so brief history. The nee SOCs) present in water in the past few decades .as increased. Concern s sensitive techniques

s for the analysis of SO trace organic analysis d to identify synthetic supplies has only been , as the degree of SOC over SOCs in our waters for identification have

Cs have has a organic

has become

2-2. Brief Overview of SOC Removal in Water Treatment.

SOCs penetrate groundwater and surface water supplies, both from nonpoint (such as agricultural and other land runoff) and point (such as industrial waste effluent and leaking hazardous waste dumpsites) sources. Unfortunately, conventional water and wastewater treatments are often ineffective for removal of SOCs, and the potential for adverse health effects is therefore a matter of increased concern (Semmens and Ayers, 1985, pp. 79-84; Sorrel1 et al., 1985, pp. 72-78; Thrun and Oberholtzer, 1981, pp. 253-266). Wastewater treatment beyond secondary treatment is capable of removing SOCs at a 90 percent efficiency rate, which may or may not be sufficient (Reinhard et al., 1979, pp. 675- 693). Alum or ferric sulfate coagulation, followed by flocculation, settling, and filtration were found to remove 3-20 percent of four model SOCs in Mississippi River water during drinking water treatment, which is a typical removal efficiency for natural organics with molecular weights of 1000 or less (Semmens and Ayers, 1985, pp. 79-84). Removal efficiency was relatively insensitive to coagulant dose.

Granular activated carbon (GAC) adsorbs SOCs strongly and is probably the most effective and feasible method available today for the removal of a broad range of SOCs from water but adds a considerable expense to water treatment (DeMarco et al., 1981, pp. 907-940; Roberts and Levy, 1985, pp. 138-145). The United States Congress has considered bills which would require the use of GAC in water treatment (Hanson, 1985, p. 19), although EPA withdrew a proposed regulation for carbon treatment under pressure from the water works industry. GAC filtration has been proven efficient in SOC removal when used in a slow sand filtration bed, an option which may be feasible for smaller water utilities (Fox et al., 1984, pp. 62-68). GAC has also been shown to be efficient if used as an additional step in traditional water treatment (DeMarco et al., 1981, pp. 907-940). The efficiency of GAC was also evaluated in an extensive study in which Cincinnati tapwater was analyzed for SOCs before and after

GAC treatment (McGuire et ale, 1981, 530-537). However, GAC is not efficient for the removal of all SOCs, and it mav need to be complemented by another method, such as air stripping, for more complete and efficient SOC removal (McCarty et al., 1979, pp. 683-689). Biological growth on the GAC and early saturation are other problems which hamper the GAC control of SOCs in water.

Of the other methods used for SOC removal durincr treatment, air stripping considered qu organics is a as THMs) but (Roberts and boiling or el to be the on1

is the most commonly used. Though generally ite energy inefficient, air stripping of volatile . promising method for removal of volatile SOCs (such is inefficient for removal of nonvolatile SOCs Levy, 1985, pp. 138-145). For emergency situations, ectric mixing (a form of air stripping) were found y effective in-home treatment methods for removal of

volatile ~~Cs-(~orrell et al., 1985, pp. 72-78), but these are steps that should not be necessary with responsible environmental protection and adequate treatment.

-

2-3. History of SOC Analysis Techniques.

As the potential for adverse human health effects from SOCs in our water-supplies has become increasingly recognized in the past decades, the amount of research invested in analytical methodology has increased. This is particularly true for the past 10 years, as legislative actions have been taken to protect our environment against harmful SOCs.

In the early 1950s, the petroleum industry recognized the need for SOC analysis in order to control water pollution (Headington, 1953, pp. 1681-1685). This industry developed a variety of waste treatment techniques and complemented it with SOC monitoring for pollution abatement. Analysis was done primarily for hydrocarbons and phenols. Methods used included a variety of gravimetric, volumetric, spectroscopic (infrared, , ultraviolet, and colorimetric), and solvent extraction methods. Reflux distillation extraction was shown to improve the recovery and lower the detection limit over the previously used benzene extraction methods. The SOCs of interest were detected down to 100 ug/l - 1 mg/l levels. Preliminary work with a relatively new instrument, the mass spectrometer (MS), was also done, and its powerful capabilities were recognized. Analysis for a variety of petroleum industry products and by-products in water down to 10 ug/l levels was possible with gas stripping with a stream of hydrogen and subsequent MS identification.

During the same period, Braus and coworkersattempted to concentrate SOCs by passing large volumes of water through a column filled with activated carbon, which was then eluted with ether or chloroform (Rosen, 1977, pp. 3-14). The researchers found that even when using 5000 gallons of water, it was

difficult to yield a sample of SOCs that was sufficient to allow extensive analysis.

2-4. Identification Methods in SOC Analysis.

As the gas chromatograph (GC) with packed, and more recently, capillary columns was developed and refined, so was its use in the analysis of SOCs (Schwarzenbach et al., 1984, pp. 167- 253). The general flame ionization detector (FID), more specific electron capture detector (ECD) for halogenated SOCs, and mass spectrometer (MS) were coupled to the GC and greatly improved the ability to analyze for SOCs. For SOCs that are thermally labile and poorly volatilized, there has been an increasing amount of work done using liquid chromatography and high performance liquid chromatography (HPLC) .

Modern GCs, with appropriate detectors, are able to separate and detect most SOCs at the nanogram (10-9 g) and even picogram (10-12 g) levels under the proper conditions (Trussell et al., 1981, pp. 171-186). The improvement in resolution and separation obtained by capillary columns in SOC analysis over the traditionally used packed columns is very significant (Figure 2 - 1) (Grob and Grob, 1977, pp. 75-86; Lin et al., 1981, pp. 861- 906). Sample size requirements, sensitivity, and selectivity are all greatly improved by the use of capillary columns.

The development and improvement of detectors have also helped establish the GC as the major instrument in trace organic separation and analysis (Schwarzenbach et al., 1984, pp. 167-253; Trussell et al., 1981, pp. 171-186). These detectors include the general flame ionization detector, the electron capture detector, the Hall electroconductivity detector, the flame photometric detector, and the mass spectrometer, among others. The mass spectrometer, interfaced with a computer, provides a most powerful detector for the analysis of unknown SOCs.

The majority of SOCs which are of environmental concern today can be analyzed with GC methods. There are, however, a number of potentially harmful SOCs for which GC is inappropriate.

,

High performance liquid chromatography (HPLC) is being used more and more for the analysis of SOCs which are thermally labile, poorly volatilized, or for other reasons, poorly identified using GC (Bombaugh, 1984, pp. 317-381; Glaze et al., 1981, pp. 371-382; Haeberer and Scott, 1981, pp. 359-370; Walton and Eiceman, 1978; Watts et al., 1981, pp. 383-398). The Priority Pollutant List includes 11 phenols and nitrophenols which are difficult to analyze for by GC methods. HPLC has proven to be effective in the analysis of these SOCs (Haeberer and Scott, 1981, pp. 359- 370). The great number of large molecular weight or otherwise. poorly volatilized organics can be analyzed by HPLC, coupled to the appropriate detector (which may also be an MS). The

Pocked Column

3 m / 2 m m

ov- l

Gloss Capillary

Column

35 m / 0.28 mm

ov- 1

_-- - ,id.

Figure 2-1

Comparison of Typical Separation

from Packed and Capillary Columns

environmental importance of these organic compounds is even less well known than it is for those compounds that have been identified by GC for some time.

Infrared (IR) spectroscopy, one of the techniques used in the early years of SOC analysis, is today being successfully used in the analysis of an increasing number of groups of including pesticides, phenols, and petroleum related (Kawahara, 1984, pp. 382-444).

At this time, GC methods are still the dominant sot separation and identification. Isotope dilution the past few years been investigated and proposed as

sots, organics

methods for GC/MS has in an

improvement over other GC or GC/MS methods because of increased accuracy and precis 1981, pp. 1907-1911 In isotope dilution determined from the composition by the the isotopic compos (and is known) . Qu

on for quantification (Colby and Rosecrance, Colby and Rosecrance, 1981, pp. 221-230). the concentration of the SOC in a matrix is change produced in its natural isotopic ddition of a known quantity of the same SOC, tion of which has been artificially altered ntitative determinations can be made with a

high degree of accuracy and precision.

A step several o al., 1977 6 This

rarely used t f the SOC ana , pp. 113-134 involves the

oday which may lysis methods ; Thrun and Ob addition of a

, increase the effic is "salting out1' (M erholtzer, 1981, pp considerable amount

iency ieure 0 253- of

clean, inorganic salt (such as NaCl or Na SO ) to the water to be analyzed. The salt will generally decreage the solubility of the SOC in the water, making it more easily extractable or purgeable. The salt added must be inert to the analysis, other than decreasing the water solubility of the SOC.

SOC Extraction and Concentration Methods.

When considering the injection and analysis methods . * - .

available, most researchers have recognized the limited capabilities of the direct injection method of. a contaminated water sample (Dressler 1979, pp. 167-206). A variety of chromatographic separation and detection problems is associated with this method, and high analyte concentrations are generally needed for detection of the SOCs. There is therefore a need to combine sensitive detection with efficient sample preconcentration. Methods for extracting the SOCs from the water matrix and concentrating them before chromatographic separation and analysis have been developed and greatly improve the analysis of low levels of SOCs. The more commonly used methods have been developed for a variety of SOCs and are well documented in the scientific literature and are discussed in the rest of this chapter. These methods are:

Reverse Osmosis/Membrane Filtration/Dialysis

XAD Resin ~dsorption/Elution

Tenax-GC Adsorption/Thermal Desorption

~iquid-Liquid Solvent Extraction

Bellarfs Purge and Trap

Grobrs Closed Loop Stripping Analysis

Membrane Methods in SOC Analysis.

The reverse osmosis/membrane filtration/dialysis group of concentration and extraction methods employs a semipermeable membrane through which water and/or the SOCs are selectively transmitted. In reverse osmosis/membrane filtration, external pressure is applied, and "pure1' water passes through the membrane, concentrating the retained SOCs (Rutanen, 1982). In dialysis, either tvpurett water or the SOCs can pass through the membrane, depending on other solutions and solvents in the system which may or may not attract the SOCs. Direct injection, evaporation as a concentration step, or removal and further concentration by one of the other available methods can then be performed after initial concentration by the membrane method.

Membrane methods for concentration have been used in trace organic analysis of drinking and waste water (Deinzer et al., 1975, pp. 799-805; Keith, 1981, pp. 156-162; Mieure et al., 1977, pp. 113-134). Large numbers of organics were isolated and identified, but it was found that the method is applicable to certain classes of SOCs and inappropriate to others and thus does not give a representative picture of the total SOC content of the water. In the membrane methods, the characteristics of the Sops present, solvents in the system, and the membrane itself are all- determining factors in what is passed through the membrane and what is not. Pore size of the membrane, sorption onto the membrane, and the polarity of the membrane are examples of membrane characteristics that need to be considered. Low molecular weight and polar molecules are poorly concentrated by the polymeric membranes (for example, cellulose acetate) which preferentially adsorb and transmit polar and smaller molecules (Leenheer, 1984, pp. 84-166). Other groups of SOCs such as aromatics are also adsorbed onto the membrane and are lost to analysis.

2-5-B. Resin Methods - 4 A D and Tenax-GC.

In WLD resin adsorption/elution methods, the water sample containing SOCs is passed through a column packed with XAD resin. The SOCs will adsorb onto the XAD resin and can later be extracted with a solvent. Additional concentration through evaporation several var physical ch of certain

is then general ieties, each var aracteristics an types of SOCs.

ly perfo ,iety hav .d being

rmed. The XAD r ing its own chem appropriate for

mesins come ical and the adsorp

The resins are made of either polystyrenedivinyl benzene (XAD 1-5) or polynethacrylate (XAD 7 and 8) copolymers crosslinked to form beads of varying mesh size with a very large surface area (Junk et al., 1977, pp. 135-154; Rutanen, 1982; Stepan and smith, 1977, pp. 339-342). XAD 2, 4, 7, and 8 are the most commonly used, though these macroreticular resins come in forms from XAD 1 to XAD 12 (Musty and Nickless, 1974, pp. 185- 190). The lower the number of the resin, the lower its polarity, indicating the polarity of the SOCs for which it is appropriate. The type of resin and the flow rate of the water passing through the resin column are the main factors affecting adsorption. Even at optimum conditions, the XAD resins have been shown to be applicable mainly to nonpolar SOCs (DeMarco et al., 1981, pp. 907-940; Junk et al., 1977, pp. 135-154; Leenheer, 1984, pp. 84- 166). Highly and even moderately polar SOCs are poorly adsorbed and retained by the resins. Other major problems with the use of XAD resins in the analysis of SOCs are the loss of volatiles during resin drying before solvent elution and the need to clean the resin before use. The cleaning of XAD resins is a very time consuming process, and obtaining truly clean resins is extremely difficult. Dirty "blanksu due to resin contamination have been seen to be the norm in many analyses (Dietrich et al., 1983). These problems make XAD resins inappropriate for the analysis of minute quantities of SOCs over a wide range of polarities and volatilities. An advantage of XAD resin adsorption/concentraf$.on is that large volumes of water can be passed through the resin - -

column, thus allowing for great concentration of the SOCs that are retained. These resins may be very suitable for nonpolar, nonvolatile SOCs if clean resins are used and/or the dirty blanks accounted for.

In spite of their shortcomings, XAD resins have been extensively used in the analysis of organic compounds in water (Burnham et al., 1972, pp. 139-142; Caton et al., 1981, pp. 329- 344; Chriswell et al., 1975, pp. 132501329: Dunlap et al., 1977, pp. 453-478; Junk et al., 1977, pp. 135-154; Leenheer, 1984, pp. 84-166; Lin et al., 1981, pp. 861-906; Musty andNickless, 1974, pp. 185-190; Osterroht, 1974, pp. 289-298; Stepan and Smith, 1977, pp. 339-342; Van Rossum, 1978, pp. 381-392). XAD 2, 4, and 7 resins have been successfully used in the analysis of herbicides, insecticides, and polychlorinated biphenyls (PCBs) in

tap and sea water down to around 1 ug/l levels (Junk and Richard, 1981, pp. 295-316; Musty and Nickless, 1974, pp. 185-190; Osterroht, 1974, pp. 289-298). Inorganic salts are not retained by XAD 2 and 7 resins, making them useful in the analysis of certain SOCs in seawater (Osterroht, 1974, pp. 289-298). However, in another study, the amount of dissolved minerals and naturally occurring organics has been shown to drastically reduce the trapping efficiency and recovery of SOCs on XAD 4 resin (Landrum and Giesy, 1981, pp. 345-355).

Several polycyclic aromatic hydrocarbons (PAHs) and certain organics in wastewater have also been analyzed for with XAD resins (Caton et al., 1981, pp. 329-344; Glaze et al., 1977, pp. 247-254; Keith, 1981, pp. 156-162). Analysis of taste and odor causing SOCs in drinking water has also been performed, using XAD resins (Burnham et ale, 1972, pp. 139-142). Poor recovery of polar compounds, even at as high as 1 mg/l levels, was obserred, while other SOCs were well retained. By increasing the pH to around 12, certain polar compounds such as phenols are retained significantly better on the XAD resin, and analysis at high ug/l levels has been done (~hriswell et al., 1975, pp. 1325-1329). combining resin types in the column has also been tried as a way to improve recoveries. An equal weight combination of XAD 4 and 8 was found to be the most effective in recovering a mixture of 13 commonly found SOC pollutants, but incomplete retention was still observed for certain types of SOCs, including some phthalates, polar, and volatile SOCs (Van Rossum, 1978, pp. 381- 392).

Adsorption/thermal desorption has been tried with XAD resins. Though relatively successful results have been reported for a narrow group of compounds (Ryan and Fritz, 1981, pp. 317- 328), XAD resins generally result in dirty backgrounds upon thermal desorption. Tenax-GC adsorption/themal desorption has been developed for SOC analysis and is not associated with the background problems of XAD resins. In the Tenax-GC method, as, - developed by Pankow, sample water is passed through a column packed with Tenax-GC for SOC adsorption (Pankow and Isabelle, 1982, pp. 25-39; Pankow et al., 1982, pp. 1815-1819; Pankow et al., 1982, pp. 31-43). The Tenax-GC column is dried by centrifugal action, and the SOCs can then be thermally desorbed directly onto the GC. This method, which is possible due to the thermal stability of Tenax-GC, bypasses the need for solvent elution such as is necessary with XAD resins. Other than its increased thermal stability, the adsorptive characteristics and applicabilities of Tenax-GC are very similar to that of XAD resins, which may explain the limited use the method has seen. This method is very flow-rate sensitive and is most applicable to nonpolar and intermediate-to-low volatility SOCs (Pankow et al., 1982, pp. 31-43). Polar SOCs are not well retained by the Tenax- GC, and the more volatile SOCs are lost during centrifugation. Good recovery of a group of seven chemically similar nonpolar,

nonvolatile Priority Pollutants has been shown by Pankow and his coworkers at low ug/l levels (Pankow and Isabelle, 1982, pp. 25- 39; Pankow et al., 1982, pp. 31-43).

2-5-C. ~iquid-Liquid ~xtraction Methods in SOC Analysis.

Liquid-liquid extraction (LLE) is widely used organic substances from water. With highly volati additional concentration is obtained by evaporatio 1984, pp. 84-166; Rutanen, 1982). Several variati available. The batch extraction method is the mos continuous LLE methods, steam distillation/reflux methods, and microextraction are also used.

for extract le solvents, n (Leenheer, ons of LLE a t common, bu extraction

Liquid-liquid extraction extracting most SOCs from the nonpolar and volatile to nonv are, however, not extracted a volatile SOCs are often lost manipulation of the extract (

can generally be efficient in water matrix, including polar to olatile SOCs. Very polar compoun s well as nonpolars, and very during further concentration and Leenheer, 1984, pp. 84-166).

ing

EPA's method for the analysis of nonvolatile SOCs, Method 625, is a batch LLE method (with a continuous LLE method option, briefly mentioned). Method 625 was first introduced when published in the Federal Register on December 3, 1979 (USEPA, 1979, pp. 69464-69575). It was promulgated on October 26, 1984 in its original form (USEPA, 1984, pp. 43233043442). The majority of the nonvolatile Priority Pollutants can be determined by this method, most of them at levels down to 10 ug/l (Boland et al., 1981, pp. 831-838; USEPA, 1979, pp. 69464069575; Weston, 1984, pp. 54-60). The water sample and solvent are combined in a separatory funnel with shaking, then the organic phase (with the SOCs) is separated from the water. High turbidity, or high sediment content is not a significant problem in LIE, as compared to other methods (Krasner et ale, 1981, pp. 689-711; Lopez-Avi,la et al., 1981, pp. 793-828). LLE is a fast, inexpensive, simple,' and straightforward method that allows for easy alteration and modification of such parameters as volume and pH, as the need arises. LLE has proven to be very efficient, even in today's laboratory with more %ophisticatedu instruments and methods.

LLE has been shown to be highly efficient by numerous public agencies and monitoring authorities throughout the country. In independent research the method has often been the method of choice for SOC analysis as well (Colgrove and Svec, 1981, pp. 173701742; Cooper and Wheatstone, 1973, pp. 1375-1384; Lopez- Avila et al., 1981, pp. 793-828). Multiple fractionation beyond the two base/neutral and acid fractions in Method 625 has been shown to be a good approach for the separation and analysis of a large number of different types of SOCs at low ug/l levels and generally results in very good recoveries except for very polar

SOCs (such as some phenols) (Colgrove and Svec, 1981, pp. 1737- 1742). Problems with contaminants in solvents can often be avoided by distillation of the solvent before use. The traditional methods involve a concentration step which increases the relative amount of impurities from the solvent and the time needed for the analysis. Microextraction techniques attempt to minimize these problems. The water to solvent ratio can be 1000:l or greater, as opposed to the 10:l ratio such as in EPA Method 625. This reduces and may even eliminate the need for further concentration. Microextraction has been applied to the analysis of Priority Pollutants in wastewater (Rhoades and Nulton, 1981, pp. 241-252) and evaluated on Priority Pollutant standards (Thrun and Oberholtzer, 1981, pp. 253-266) with varying results and success. Fairly good results were observed on polar SOCs, but standard deviations were generally greater than with other extraction methods. For one study, 100-1000 ug/l levels were essential for good recovery and reproducibility of the SOCs (Junk et al., 1981, pp. 281-294).

By choosing an appropriate solvent, certain groups of SOCs can be selectively extracted. Highly polar phenols and chemically similar SOCs have been successfully extracted and identified in wastewater by such a selective solvent method (Cooper and Wheatstone, 1973, pp. 1375-1384). Many volatile SOCs, including trihalomethanes (THMs), can be extracted by choosing pentane as the solvent (Henderson et al., 1977, pp. 105- 112). In a very simple and rapid process, THMs have been identified at ng/l levels, using the modified pentane extraction method (Glaze et al., 1981, pp. 267-280).

Another less commonly used variety of LLE is the steam distillation/reflux method. Steam distillation has mainly been used to concentrate volatile organic compounds in water, but over 85 percent recovery of pesticides has also been achieved by this method (Rutanen, 1982; Veith and Kiwus, 1977, pp. 631-636). In this procedure, both the aqueous liquid and the solvent are heated to reflux and the condensed vapors are brought into ' -

contact in a capillary tube which is cooled. Both phases are then returned to their original flasks and recycled. The SOC solutes transfer, in time, to the solvent phase (Veith and Kiwus, 1977, pp. 631-636).

Other solvent recycling LLE methods include the continuous LLE methods. Solvent is recycled through the water by an evaporation/distillation-condensation cycle, and the SOCs are extracted and concentrated in the solvent flask. The extraction efficiency is generally increased by continuous extraction (Rutanen, 1982), and it has been used for the determination of a variety of nonpolar (Kahn and Wayman, 1964, pp. 1340-1343; Stachel et al., 1981, pp. 1469-1472) and other SOCs (Ahnoff and Josefsson, 1974, pp. 658-663) in water, including a variety of pesticides (Bruchet et al., 1984, pp. 1401-1409; Xahn and Wayman,

1964, pp. 1340-1343; Wu and Suffet, 1977, pp. 231-237). The continuous LLE has been developed in a variety of styles, including one which continuously extracts large volumes of water passing through the system, which permits great concentration effects and detection down to the ng/l level for many SoCs (Ahnoff and Josefsson, 1974, pp. 658-663; Bruchet et al., 1984, pp. 1401-1409; Stachel et al., 1981, pp. 1469-1472; Wu and Suffet, 1977, pp. 231-237). Extraction efficiencies at a given, recoverable level are comparable for the continuous and batch LLE, but the larger concentration effects possible with the larger volumes of water in the continuous LLE results in improved detection at low levels (Bruchet et al., 1984, pp. 1401-1409). The problems associated with foaming and emulsion formation in other LLE methods are not present in continuous LLE, giving it a substantial edge when such problems are a factor.

2-5-D. Headspace Methods - Purge and Trap and CLSA. Though the pentane extraction method has been used

successfully for certain volatile SOCs, the methods discussed thus far are primarily limited to moderately-to-nonvolatile SOCs. For the volatile SOCs, the so called "headspace" methods are generally employed. These methods include Bellarts purge and trap and Grobts closed loop stripping analysis (CLSA).

EPAts method for the analysis of volatile SOCs in water is Method 624, which was introduced and promulgated along with Method 625 (USEPA, 1979, pp. 69464069575; USEPA, 1984, pp. 43233- 43442). It is a purge and trap method. In this method, the SOCs are purged and evaporated out of the water with an inert gas and trapped on a combination of Tenax-GC, activated carbon, silica gel and/or XAD resin adsorbents (Figure 2-2). The SOCs trapped are thermally desorbed and driven onto the GC by the carrier gas. The purge and trap method is, like Method 625, well tested and used extensively throughout the country for the analysis of volatile SOCs in water. In relatively clean source and drinkinge% waters, the nonpolar, volatile Priority Pollutants (including THMs), for which the method was intended, are generally well identified with purge and trap. Bellar reported detection limits down to 0.5 ug/l for a group of volatile SOCs (Bellar and Lichtenberg, 1974, pp. 739-744; Weston, 1984, pp. 54-60). However, the volatile Priority Pollutants are not the only volatile SOCs of environmental importance. A modified purge and trap method was used for the analysis of SOCs in wastewater sludge (Haile et al., 1981, pp. 763-792). The method performed well for a number of volatile Priority Pollutants but was unsatisfactory for some other environmentally important volatile SOCs. Inconsistent results were seen in many sludge samples tested, including urecoveriesu well in excess of 100 percent for several of the SOCs. A large number of SOCs were identified, but equally inconsistent recoveries of SOCs were seen in another

Pur~fng Device Trap Pockinss and Consfructfon to

Include Desorb Ccpability

Schemctic of Purse and Trcp Schematic of Purge and Trap Device-Purge Mode Device-Desorb Mode

Figure 2-2

Purge % Trap Apparatus

18

s w pup ( , ames aqq A T T ~ T ~ T U ? ~~Aqqam xo

-T) T T m s e uo paddexq n 5 - r ~ ) dmnd e Aq a ~ d m e s - - . ~ o n do01 pue aaedspeaq

*qqoq xo sayxoqexoqq buyqedyayqxed aqq buoue a a u e m s s e Aqy~enb/~oxquoo Aqy~enb xood l o

s y s d p u e dexq pue a5xnd 30 A q y ~ y q e o y ~ d a x xood xaqqya 5uyqeaypuy ' sayxoqexoqe~ Ijuyqedyayqxed aqq uaamqaq palvasqo axan suoyqeyxeh

able1 & a ~ ' ( ~ 8 6 7 : ' x a u x e ~ pue eyxn3) yd3 aqq Aq panuojxad Apnqs A ~ o q e ~ o q e ~ x a q u r xoCeu e uy papnpuy o s p sem d e q pue abxnd

'dd '1861 ' T e qa suy6belds) aqys dmnp -1eoymaqo e uolg saoyxqeu ~ y o s ur s30s j o s y s d p u e aqq x o j Apnqs dexq pue abxnd payaypom

T crnoeraturt Controller -

Heater

Trap

Figure 2-3

CLSA - Purge Mode

+

4-10 ul o f CS, rrcorert-3 tn - extramian

Figure 2-4

CLSA - Solvent Extraction Mode

F: m t P 0 tr 0 m r d a , m F : b , F : -d U 0 , G 0 -d k -d

d ( d U OkE-1k ma, 0 0 - d P c r l k k E Q r d t r

k rd 3 . 2 3 ka, a m - 3 ~ o + , a , m n ~ + ,

a+,rd rd 3 s 3 - t : a , m o m - 3 + , ~ ~ + , a o m a , m k h

k m a , l i r d U a o r l a , a + J a , h 4 + , 3 a o w d . 2 3 a, *F :+ , m u a, ~ : a M o a+, k r n a , r d h k r d a , r d t J a F : r d a 3 32; a $ & C o w Q O d G.23 r d - d 0 a d o a 4 k ha+ , U+, 4J O - l J t n G U 0 -d-d-dal-dUl u a o f d hL:k .233ka+ ,

k ' d d + , O U ? 3 a U h U t n a , d .23U rl wrto m rn -+, c m +,-d fi F: 0 U a 0-d rd P c k w O m a , r l - d a , Q ) F : M O a , * a , 0 ?I& % 0+,+, cd U PC0 & d o - d f d m - + , m e a x - a a , ~ h m ~ c k a m - c r a G a a , ~ o m - 4 a a, F:+J k k

rd rd m + , u 5-i F:-4 m 4 ~ a, k 9 4 R - d rd > U U G F: rdo ] k O 4 J E f d o -dH k4aa Q) 0 F: .23tncd 04JU F: 0.d k O h

c m k rd -d a 0 4 r n k r d u * r b a a , w O+J a,a O- ~ ~ a - l - r k r d m a

4 . 2 3 0 3 W k a+) m k A a, a, a - v m - a , k g m a w 3 0 k W t - c , a, F: 3 P c o a k F: 4J 3 0 I d t r - I D , m G m a , m d PI 0-d w d 3 Q ) 3

G h N W ga F: 0 ia o a l i 3 r d ~ ~ ~ P , W ~ PC (44 a, a,k a, m t n m g o ~ : w = 3u * .23= ga a,-d o m - d

a,omPc+,.c 0 g U - u ~ c m m m 3 4 +,m h r d 4 r d L : $ I r d g l i o ] cordm darn - m o m O d ri a, a , r n ~ ~ r l = f k a o ~ ~ m r l +,aU-da> 0 . 2 3 w E cd rd

a m k + , O r a , k F : O

G a c c r a d . 2 3 d D a 0 3 k 9~ o a, +, a, a, o h a w o rd +,= +, - * ~ C X W ' Z $ & ~ O a , a , O O ~ O l i a , ( d u o 0 W d Q ) = 0 Dk F: I= U C a f d 0 m k 3 k O - d a , % DRJRJW~N m w ~ - d - - c l

m r d U C\T: -d fd O d 4 J 4 J I - i a+,-40.23 r d 3 k d f d r d a , a , G W U > a a , ~ . 2 3 ~ f i . 2 3 a , o w m a , (d fd 3-u-4 m4J F: U a, d 4 J

such as turbidity, high organic, or inorganic content than by nonpurging techniques.

The most extensive comparison of analytical methods for SOC analysis in water was one carried out jointly by researchers from EPAts Health Effects Research Laboratory (HERL) and Battelle Columbus Laboratories (Melton et al., 1981, pp. 597-674). The methods compared were purge and trap, CLSA, LLE, and XAD-2 resin, with the first two methods presumed most applicable to volatiles and the latter two to nonvolatiles. The CLSA method included the use of a capillary column, while the purge and trap method employed a packed column, according to recommended EPA methodology. Overall, 183 different SOCs were identified in the study. Six were detected by purge and trap, 107 by CLSA, 90 by LLE, and 58 by the XAD-2 resin method. The six VOCs detected by purge and trap were also detected by CLSA, but the two most volatile of these VOCs were detected more accurately by purge and trap. This indicates that the advantage of purge and trap over CLSA is in the analysis of the few very volatile VOCs where the solvent peak present in CLSA will interfere. Since purge and trap is a thermal desorption method, it does not use any solvent for introducing the VOCs on the GC, and there is therefore no solvent peak. The researchers who had in earlier work identified about 500 unique purgeable organics in water by CLSA, found that the average SOC concentration of SOCs identified by CLSA in this study was 9.2 ng/l and that for 215 SOCs studied, the minimum detection limit was in the 1-10 ng/l range. The researchers agree about the state of the art, highly sensitive, and comprehensive nature of CLSA and speculate that it will replace purge and trap as the method of choice in VOC analysis. It should be pointed out that the purge and trap method was EPA Method 601, which is a method used primarily for halogenated purgeable SOCs. Nevertheless, purge and trap appeared inferior, as several halogenated volatile to semivolatile SOCs that were identified by CLSA, and ideally, should have been by purge and trap, were missed by the purge and trap method. Other than f o y t h e t w o most volatile VOCs, CLSA was superior in all respects and is complemented well by LLE or XAD-2 for the less volatile organics. LLE analysis was efficient for the groups of SOCs that the XAD-2 method analyzed for and even more, covering a broader range of SOCs and complementing CLSA the best.

Summary of Literature and Choice of Methods.

The EPAfs rules and guidelines for the analysis of SOCs in water were first published in the Federal Register on December 3, 1979 (USEPA, 1979, pp. 69464-69575) . These included several specific methods, aimed at narrow groups of SOCs but also two methods for broad spectrum analysis of SOCs. These methods are:

- Method 624, Bellarts Purge and Trap

- Method 625, Liquid-Liquid Extraction

These EPA methods involve different extraction and concentration techniques. Method 624 is for volatile SOCs and Method 625 for nonvolatile SOCs. Both methods involve the use of packed column GC/MS for separation and identification.

The EPA methods have since been promulgated (USEPA, 1984, pp. 43233-43442) and are the methods to be used for monitoring and determination of SOCs in water by regulatory authorities. Since these methods are more than five years old, there is reason to believe that they may not necessarily be the best methods for broad spectrum analysis, in particular when considering the rapid development and improvements that have been taking place in the field of SOC analysis in the past few years.

The Master Analytical Scheme (MAS) is a collection of analytical techniques (most of which have been described earlier in this chapter) that have been organized into a comprehensive protocol (Garrison et al., 1981, pp. 17-30; Gebhart et al., 1981, pp. 31-48). It is broad in scope and includes methods for preanalysis (scouting) to determine optimum sample volume and amount of internal standard needed, as well as methods for the analysis of ionic interactable compounds. The MAS has strong QA/QC and can be used as a model of good protocol validation. However, the MAS is unnecessarily large and complex for the purposes of most SOC analysis, involving up to 9 fractions from each sample for final GC/MS analysis.

For this study, the EPA8s methods and the other scientific literature available on SOC analysis in water were carefully reviewed in order to select the most appropriate method or methods for broad spectrum analysis of a range of potentially harmful SOCs in water. This includes SOCs of diverse physical. and chemical characteristics. The most commonly used methods of- concentration and analysis were identified and the scientific literature studied in more depth.

It is useful to classify the methods into degrees of volatility and polarity of the SOCs that they will analyze for (Table 2-1). For this study, the object was to analyze for the greatest range of SOCs, using as few methods as possible. There is a significant degree of overlap between the methods, and these classifications are generalizations but still quite useful.

The reverse osmosis/membrane filtration/dialysis methods were eliminated as potential methods at an early stage. The membrane methods are, in general, too selective for concentration of SOCs, not concentrating low molecular weight (volatile) or polar SOCs well. There is also unpredictable adsorption of SOCs

T a b l e 2 - 1

A P P L I C A B I L I T Y OF SOC COXCEXTRATIOS/AHALYSIS METEQDS

M e t h o d

R e v e r s e O s m o s i s / Y e a b r a n e F i l t r a t i o n / D i a l y s i s

Tenax-GC A d s o r p t i o n / ; T h e r m a l D e s o r p t i o n

X A D - R e s i n A d s o r p t i o n / E l u t i o n

L i q u i d - L i q u i d S o l v e n t E x t r a c t i o n

B e l l a r ' s P u r g e a n d T r a p

G r o S ' s C l o s e d L o o p S t r i p p i n g A n a l y s i s

V o l a t i l i t y

Low

L o w - H e d i u m

N e d i u m - H i g h

M e d i u m - H i g h

P o l a r i t v

Low

L o w - X e d i u m / H i g h

L o w - H i g h

Low-)1edi urn

L o w - X e d i u m

onto the membrane. These methods require an additional extraction/concentration step, necessitating the use of one of the other five methods as well.

The Tenax-GC method requires very clean Tenax-GC, and its strongly organic adsorbent nature makes this difficult. A substantial modification to the GC, which is not commercially available, is necessary for the thermal desorption step. Finally, this method is not applicable to the less volatile and polar SOCs, making it limited in range as compared to the other methods.

The XAD resin method can analyz if a combination of different resins be used very well for the polar and resins are even harder to clean and as a result, dirty blanks are very c Department of Environmental Sciences XAD resins and liquid-liquid extract in Haw River water showed XAD resins comparison to liquid-liquid extracti

e for a broad range of is used, but it still nonvolatile SOCs. The keep clean than Tenax- omrnon. Work in the and Engineering at UN ion for the analysis o to have many problems on (Dietrich et al., 1

' SOCs cannot

.se GC, and

'C with tf SOCs in

983).

For this study, liquid-liquid extraction was chosen for the analysis of nonvolatile SOCs. This is a simple, straightforward method that can easily be modified as the need arises. It is also used in EPAVs Method 625 for nonvolatile SOCs. Of the methods for nonvolatiles, liquid-liquid extraction was judged to cover the broadest range of applicability, and it extracts in particular, the more polar SOCs better than the other methods. Of the variety of liquid-liquid extraction methods available, a continuous liquid-liquid extractor (CLLE) was chosen. The CLLE was expected to have several advantages over the traditionally used separatory funnel, including:

- less person-hours needed for the extraction,

less worker contact with extraction solvents,

- less water matrix interferences due to foaming and emulsion formation during the extraction, which in turn would improve recoveries, reproducibility, and detection limits,

- possibly better recovery of the more volatile and polar SOCs.

For the analysis of the volatile SOCs, the purge and trap method was judged to be less applicable than the CLSA method. Purge and trap is used in EPAk Method 624 for volatile SOCs. Purge and Trap does not cover as broad a range as the CLSA, bei limited mainly to highly volatile, nonpolar SOCs. CLSA analyze better for the less (medium) volatile SOCs in addition to the

highly volatile ones and covers a broader range of polarity as well. CLSA is also superior to purge and trap in detection limits, having a two to three orders of magnitude edge.

CLLE and CLSA were thus judged to be the most methods for use in broad spectrum analysis of SOCs findings in the literature study justified the inc third method or the substitution of any other comb methods for CLSA and LLE. These two methods compl other better than any other combination of methods in this research.

applicable in waters. No lusion of a ination of ement each and are used

CHAPTER 111. MATERIALS AND METHODS

3-1. Selection of Analytical Methods.

Based on the literature review, CLLE and CLSA were found to be the most applicable methods to broad spectrum analysis of SOCs and the ones that will be used in this research. These methods complement each other better than other combination of methods available, in terms of the SOCs they are applicable to. The materials and instruments used in CLLE and CLSA are commercially available and well documented, an important consideration for the development of methods which might be used by a variety of public and private organizations concerned with water quality.

3-2. Standards Selection.

In order to evaluate, optimize, and develop the chosen methods, CLSA and CLLE, a comprehensive set of standard SOCs to be used in the experimental process needed to be selected. These standards must cover a broad range of chemical and physical characteristics in order to be representative of the wide range of SoCs one may encounter in environmental water analysis. In order to compile such a representative list of standard SOCs, a large amount of pertinent literature, reports, and permits were studied (Table 3-1). From these valuable sources of information, along with chemical and physical data on the SOCs, a representative set of standard SOCs was chosen. Relative importance in terms of environmental concern was another major factor that needed to be taken into account. It was important to choose a variety of SOCs, which necessitated exclusion of several SOCs of environmental concern that were too closely related in chemical and physical properties. Molecular weight, molecular structure, volatility (vapor pressure) , and polarity (water solubility), as well as relative retention time in gas chromatography, were the major chemical and physical + .

characteristics considered. Vapor pressure is a good indicator of volatility, and though water solubility is not a direct measure of polarity, it is an indicator and a valuable measure in water analysis for SOCs.

A set of 16 SOCs was chosen, 12 of which would be used as primary internal standards (added to the original samples before any workup has begun) and 4 of which would be used as secondary internal standards (added to the sample before final injection onto the GC or GC/MS, commonly called external standard) (Table 3-2). The matching up of primary and secondary internal standards was based on relative retention times (matching those which elute at about the same times) and degrees of physical and chemical similarity (Table 3-3).

T a b l e 3-1

SOURCES OF IXFOR3ATION FOR S E L E C T I S G STXNDARDe'SOCs

A . R e p o r t s o n VOCs a n d S O C s c o m m o n l y f o u n d i n d r i n k i n g wa te r ( f r o m A m e r i c a n W a t e r W o r k s A s s o c i a t i o n K o r k s h o p s o n VOCs a n d S O C s ) ( 1 , 2 ) .

B . R e p o r t s o n t h e u s e a n d d i s c h a r g e o f i n d u s t r i a l a n d a g r i c u l t u r a l S O C s i n K o r t h C a r o l i n a (XPDES p e r m i t s a n d a s t u d y o n p e s t i c i d e u s e ) ( 2 3 ) .

C . EPA's l i s t s o f :

i . P r i o r i t y P o l l u t a n t s . i i . C u r r e n t l y r e g u l a t e d S O C s ( w i t h N C L ' s ) .

i i i . VOCs p r o p o s e d f o r r e g u l a t i o n ( w i t h R X L ' s ) . i v . SOCs d i s c u s s e d f o r f u t u r e r e g u l a t i o n .

( 1 , 2 , 8 8 , 8 9 , 9 0 )

D . P r e v i o u s r e s e a r c h w o r k o n S O C s i n N o r t h C a r o l i n a w a t e r s ( i n c l u d i n g Haw R i v e r r e s e a r c h p e r f o r n e d i n t h e D e p t . o f E n v i r o n m e n t a l S c i e n c e s a n d E n g i n e e r i n g , U n i v e r i s t y o f N o r t h C a r o l i n a ) ( 2 O , 2 4 ) .

Table 3-2

LIST OF SOCs CHOSEN AS STANDARD C O Y P O U X DS

Primary Standards:

l,l,l-Trichloroethane Trichloroethene 1,2-Dibromoethane (EDB) Tetrachloroethene 1 , 1 , 2 , 2 - T e t r a c h l o r o e t h a n e Phenol 2-Chlorophenol Naphthalene 1,4-Dichlorobenzene Atrazine Tris(chloropropyl)phosphate Nethoxychlor

Secondary Standards:

ILlol. Idt. Sol.

500 1100 4310 150

2900 93000 28500

30 79 70

no data no data

4500 17000

145 no data

Vap. Pres.

96 58 11 1 A 5 0.2 2.2 1 1.2 low low low

19 1 1.5 * - .

low

Mol. Wt. = Molecular veight, grams/mole Sol. = Approximate Solubility in H20, mg/l at 200C Vap. Pres. = Approximate Vapor pressure, torr, at 200C

Table 3-3

HATCEUP OF P R L X A R Y A N D SECONDARY I N T E R N A L STXXDARDS

Primary Internal Standards

l,l,l-Trichloroethane

Trichloroethene . . 1,2-Dibronoethane

Tetrachloroethene

1,1,2,2-Tetrachloroethane

Phenol

2-Chlorophenol

Naphthalene

Atrazine

Tris(chloropropyl)phosphate

Methoxychlor

Secondary Internal Standards

Arnetryne

These 16 SOCs are generally of environmental concern and are representative of numerous other SOCs of environmental importance. Methoxychlor, a pesticide, is the only one on the list that is currently regulated with an MCL of 100 ug/l. l,l,l- trichloroethane, l,4-dichlorobenzene, trichloroethylene, and tetrachloroethylene are among the 9 VOCs that have been proposed for future regulation, with RMCLs set at 200, 750, 0, and 0 ug/l respectively. Trichloroethylene and tetrachloroethylene are proven animal carcinogens, therefore the RMCL is set at 0. Atrazine, ametryne, and 1,2-dibromoethane are widely used pesticides. 1,2-dibromoethane and the rest of the SOCs on the -

list are widely used in a variety of industries. Other SOCs considered which were included in preliminary GC studies were: carbon tetrachloride, toluene, 1,2-dichloropropane, 1,2,4- trimethylbenzene, 2,4-D, lindane, endrin, and pentachlorophenol. These were eliminated for a variety of reasons, including frequent solvent or laboratory contaminants coeluting with another SOC standard or otherwise being poorly chromatographable. Equally good and representative substitutes were found for these, eliminated SOCs. The SOCs chosen are frequently occurring water pollutants and, more importantly, are good representatives of the broad range of SOC contaminants one may encounter in source and finished drinking water.

Preliminary GC Work - Relative Response Factor Determinations.

Standards of high and known purity were obtained from the EPA, and solutions of these SOCs were made up. Stock solutions were made up by dissolving each of the standards individually in acetone and calculating the concentration, according to the weight of SOC added to the volumetric flask containing the acetone.

- . A solution of primary internal standard was made at the desired concentration by adding the required amount of each of the 12 SOC stock solutionsto a volumetric flask containing . acetone. Acetone was used as a solvent because it facilitates the rapid equilibration of the SOCs with the water matrix. For the secondary internal standards solutions, these 4 SOCs were combined by dissolving the appropriate amount of each in a volumetric flask containing the solvent in which the samples would be at the point of injection onto the GC, therefore, in methylene chloride for CLLE and in carbon disulfide for CLSA.

To evaluate and optimize the methods, the relative response factors of the primary to secondary internal standards needed to be established, using the GC. This was essential in order to be able to determine the recovery of the primary internal standard in the later experiments by knowing its relative response to an

equal amount of secondary internal standard. The responses were determined at a broad range of primary:secondary internal standard ratios in order to evaluate the degree of linearity of the relationship. Primary and secondary internal standards were mixed in methylene chloride, and later in carbon disu1fi.de for comparison, at 1:1, 1:10, and 10:l concentration ratios. The concentrations were 4 mg/l:4 mg/l, 4 mg/l:40 mg/l, and 40 mg/l:4 mg/l, resulting in a two orders of magnitude difference in the relative concentrations. These concentrations were chosen since a concentration of 4 mg/l was a desirable goal for a minimum concentration of the final, concentrated samples from both CLLE and CLSA. A 4 mg/l concentration in the final sample would be equivalent to a starting concentration in the water of approximately 1 ug/ l in CLLE and 50 ng/l in CLSA. Burdick and Jackson brand methylene chloride and J.T. Baker reagent grade carbon disulfide were used in this work.

A Varian 3700 GC with a hydrogen flame ionization detector was used as the main analytical instrument. Fused silica capillary columns were used since they are superior to packed columns in the separation of trace organics. Two stationary phases (column coatings) and two column lengths were studied. The 30 meter column had better separation power than the 15 meter column, in particular for the SOCs eluting earliest. The stationary phase chosen was DB-5. The other phase studied, DB-1, is slightly less polar, but both DB-1 columns tried exhibited severe peak splitting and poorer separation for SOCs dissolved in both the methylene chloride and carbon disulfide. In the end, 30 meter long, 0.256 mm inner diameter, 1 urn film thickness DB-5 columns (available from J & W Scientific, Rancho Cordova, CA) were used, one for SOCs dissolved in methylene chloride and another for the SOCs dissolved in carbon disulfide.

Chromatograms were recorded and peak areas determined by a Shimadzu C-R2AX integrator.

A variety of chromatographic conditions was tried to obta'in' optimum separation and resolution and a reasonable length of time for the run. Tbe injector temperature was 325 C, detector temperature 350 C , flow rate about 1 ml/min., and column pressure 16 psi. Helium was used as the carrier gas. ;She following oven temperature psogram was chosen to be used: 40 C isgthermal for 10 minutes, 6 C/minute temperature program, and 280 C isothermal for 10 minutes. For the CLSA runs, it was gound that the temgerature program could be stopped at 180 C rather than at 280 C since SOCs that do not elute by 180 C are not purged and analyzed by CLSA. These programs resulted in GC runs of 60 minutes for the CLLE and 43 minutes For the CLSA samples. Samples of 1.0-1.5 ul were on-column injected into the GC, though up to 2.0 ul could be injected if the sample was sufficiently %lean.

The Department's mass spectrometer (MS) was used when its capabilities were needed for SOC identification and verification. The same capillary column used in the routine GC analysis was transferred to a Hewlett-Packard 5710-A GC, which is coupled to the VG-Micromass Model 7070F Double Focusing Mass Spectrometer, equipped with a combined electron ionization/chemical ionization source. Data reduction was accomplished with a VG 2035 Datasystem.

The relative response factors and their respective standard deviation were determined by multiple injections on the GC. Five to ten injections of each of the different concentrations were generally-required to reach the desired standard deviation of less than 5 percent of the calculated value. Carbon disulfide was briefly compared to methylene chloride in this respect, and the solvent used did not influence the relative responses. Each SOC solution was prepared twice and the relative response factors calculated. These response factors were combined since the values were virtually identical, resulting in one relative response factor with a standard deviation less than 5 percent of the-value. Relative response factors were calculated by dividing the response (peak area) of the primary internal standard to that of the secondary internal standard (Table 3-4, Figure 3-1). Trichloroethylene posed a slight problem, as it is a contaminant of methylene chloride, the solvent. This was corrected for by calculating the relative response factor of trichl~roeth~lene- to other regularly present solvent contaminants (chloroform and cyclohexene) in pure solvent injections and then subtracting the appropriate amount of the trichloroethylene peak (as determined by the pure solvent run) in the runs with the standards.

The linearity of the relative response factor over the two orders of magnitude studies was evaluated at this point. When the relative response factors (Table 3 - 4 , Figure 3-2) were adjusted for dilution and plotted (Figure 3-3), the response factors were found to be fairly linear over the range of

+ . interest, with the possible exception of the phenols and naphthalene. The deviation in relative response from the 1:l to the 10:l and 1:10 mixtures (corrected for dilution) is generally less than 5 percent and always less than 10 percent for all SOCs. Since the work in the development and optimization experiments was to be done with roughly 1:l mixtures of the primary to secondary internal standards and at the same concentration levels used in the 1:l relative response factor determinations, the relative response factors (including standard deviations) determined in the 1:l relationship were used when calculating recoveries in these studies. If the relative concentrations or ratios were other than approximately 1:1, the appropriate adjustment in relative response factor was made.

Table 3-4

RELATIVE RESPONSE FACTORS OF PR1XARY:SECONDARY INTERNAL STAKDARDS

Stzndard

Trichloroethene

Tetrachloroethene

Phenol

2-Chlorophenol

1,4-Dichlorobenzene

N a p h t h a l e n e

Atrazine

'Tris(chloropropyl)phosphate .0782 .899 9.09 (isomer I) ( .0018) ( .028) (063)

Tris(chloroprophy)phosphate .0270 .306 3.17 (isomer 11) (.0012) ( .018) ( A 9 )

M e t h o x y c h l o r

(parentheses i n d i c a t e standard de v i a t i o n s )

1.1.1-Trichloroethnne

r 9

Trichloroethene

Phenol 2-Chlorophenol

1,LDichlorobenzene

K a p h t h a l e n e

Hethoxychlor

Primory:Secondory Int. Std. cot&. Rctio Naphthalene 0 Atra2ir.e

v Hethoxychlor 0 Tris(chloropro~yl)~hosphate 1,G-Dichlorobenzene V 2-Chlorophenol

Primory:Secondary Int. Std. Conc. Ratio A Phenol 0 Tetrachloroethene

Trichloroethene 0 1,1,2,2-Tetrachloroethane p l,l,l-Trichloroethane 1.2-Dibromoethane ( D S )

Figure 3-2

Relative Response Factors

Primory:Secondory Int. Std. Conc. Ratio v ?hen01 A 1.1.1-Trichloroeihne - . .r:c-lcrce~hene 0 1.1,2,2-ietrachlcrcefha::e

Te~rachlorcethene 1.2-Dibrcaoethane (EE3)

Prirnory:Secondary Int. Std. Conc. R o t i o 0 Saphthalene Tris(chloropropy1)phosphate

~et.h.oxych\or 0 Atrazine A 1,LDichlorobenzene 2-Chlorophenol

Figure 3-3

Dilution Adjusted Relative Response Factors

3-4 . Evaluation and Optimization of Analytical Methods - CLLE

For the liquid-liquid solvent extraction of SOCs, a continuous liquid-liquid extractor was chosen. Commercially available CLLE apparatuses exist for both the extraction with lighter-than-water solvents such as ether (Figure 3-4) and heavier-than-water solvents such as methylene chloride (Figure 3- 5). As in EPAfs Method 625, methylene chloride was used as the solvent. This is a very efficient general solvent for SOCs and is more stable, easier to keep pure, and in general, less dangerous and easier to handle than the other commonly used solvent, diethylether. The CLLE equipment used was purchased from Continental Glassblowing Lab. Inc. in Richland, NJ. Other glassware and laboratory supplies used are available from most general suppliers, such as Ace Glass and Fisher Scientific. All glassware was chromic acid soaked, followed by exhaustive rinsing with distilled, deionized water and oven drying. Boiling chips, filter paper, and magnesium sulphate (drying agent) were soxhlet extracted with methylene chloride. The CLLF apparatus was generally set up in a hood for safety reasons. Several boiling chips were placed in the round bottom flask, along with about 250 ml of methylene chloride. One hundred and fifty ml of methylene chloride were poured into the CLLE apparatus, and 2.0 liters of the water sample were carefully added and primary internal standard spiked into the water. The condenser was attached, and the cooling water and the heating mantle turned on. The heat was adjusted so that the solvent was kept at a low continuous boil, with 2-4 drops per second of methylene chloride dripping off the condenser and into the water sample. Twelve to twenty-four hours of extraction, a commonly used time for SOC analysis, was used.

Once the extraction was completed, the heat and cooling water was turned off. The 250 ml of methylene chloride/SOC extract was transferred to an Erlenmeyer flask with a ground glass stopper, and the round bottom flask was rinsed once with a few milliliters of methylene chloride which was added to the a . extract. ~agnesium sulphate (or sodium sulphate) was added to the extract as a drying agent and the flask swirled. The presence of fluffy, fknowyu drying agent that slowly settled indicated that enough drying agent had been added. The samples could either be stored at 4 C at this point (with the flask sealed tightly with Teflon tape) or concentrated right away with a ~uderna-~anish (K-D) apparatus (Figure 3-6).

For the K-D concentration, the extract was filtered through filter paper directly into the K-D apparatus. The Erlenmeyer flask was then rinsed with a few milliliters of methylene chloride which was added to the rest through the filter. The condenser was replaced and the Bottom 2-3 inches of the K-D apparatus was submerged in a 70 C water bath. The level of submersion of the K-D apparatus was controlled so as to keep the clearly visible evaporating solvent level only slightly beyond

Figure 3-4

Liquid-Liquid Extractor for Liquids Lighter than Water

Condenser

Water Sample(

- '7) Per forotrd

rnl flask a -

Mantle

Figure 3-5

Liquid-Liquid Extractor for Liquids Heavier than Water

40

Figure 3-7

Figure 3-6 Micro-Snyder Column Evaporator

Kuderna-Danish Evaporator

41

the top of the condenser. This rate resulted in evaporation of the 250-300 ml of methylene chloride to the desired 2-4 ml in about 2 or 3 hours. Further evaporation was possible by either attaching a micro Snyder column to the K-D solvent collection tip (Figure 3-7) and continuing the concentration in the same way as with the K-D or by transferring the solvent to a sample vial and allowing it to evaporate (sometimes aided by slight heat from a hot plate) until the desired final volume of 0.3 - 0.5 ml was reached. At this point, the sample was ready for addition of the secondary internal standard and injection onto the GC for analysis.

Little information is available on the operational aspects of CLLE and therefore a lot of experiments had to be performed in order to determine the optimum extraction and concentration conditions. Having optimized the CLLE conditions, the C U E was evaluated against batch LLE. Batch LLE, using separatory funnels, is the method of choice in EPA1s Method 625, and a direct comparison of CLLE and Method 625 was carried out in a preliminary experiment, once the optimum conditions had been estimated.

The parameters studied and experiments performed with the CLLE were:

Extraction time - SOC recovery was determined using different extraction times. 6, 12, and 24 hour extraction times were evaluated. The solvent volume was 250 ml, and the pH was not adjusted (pH was 6-7).

Extraction volume - SOC recovery was determined using different extraction solvent volume. Methylene chloride volumes of 100, 250, and 400 ml in the round bottom flask were evaluated. The extraction time was 12 hours, and the pH was not adjusted (pH 6-7).

' - Extraction pH - SOC recovery was determined at different water sample pHs. pHs of 2, 7, and 12 were evaluated. The solvent volume was 250 ml and the extraction time 12 hours.

pH variations - SOC recovery was determined at different water sample pHs, changing the pH once or twice during the extraction. pH combinations of 7 and 2, 12 and 2, and 7, 12, and 2 were evaluated. The solvent volume was 250 ml, and the extraction time was 12 hours at each pH, except for the pH 12 and 2 experiment, which was kept at each pH for 18 hours.

Comparison of CLLE to Method 625 - SOC recoveries were determined in a preliminary experiment comparing the optimized CLLE method with EPA Method 625. CLLE

conditions used were 250 ml of solvent and 12 hours extraction at each of pHs 7 and 2.

6. Minimum detection limits - Minimum statistically reliable detection limits were estimated.

All optimization experiments were performed using an initial SOC concentration of 2-5 ug/l in distilled, deionized water solutions. For each of the CLLE and CLSA experiments performed, multiple (generally 4 conditions were done, distilled water. Mu1

or 5) setups using the-same experimental - including one system blank of deionized, tiple injections onto the GC were done for

each of the setups, achieving-a standard deviation of less than 5 percent (generally 1-2 percent). The results from these setups were averaged and the final standard deviations calculated. Due to the propagation of errors, the final standard deviations were sometimes quite large but generally less than 10 percent of the analysis value. Natural water matrix effects and source and finished drinking waters were to be studied later. The results from these experiments with CLLE are discussed in Chapter IV.

Evaluation and Optimization of Analytical Methods - CLSA. For the more volatile SOCs, CLSA was selected as the method

of analysis. A commercially available CLSA apparatus was purchased for this purpose (Figure 3-8). The CLSA apparatus is made by Brechbuhler AG in Switzerland and distributed by Erba Instruments Inc., Peabody, MA. The charcoal filter traps, Teflon sleeves, sample vials, and sample bottles were also purchased from Erba Instruments. Good sources of information about performing CLSA, including possible problems, are Daniel Cutugno's Master's Thesis (UNC, 1984) on the construction and operation of a CLSA (Cutugno, 1984) and the commercial literature accompanying the CLSA apparatus, as well as work done by Grob (Haeberer and Scott, 1981, pp. 359-370).

a -

The experimental process used in CLSA is straightforward. Before each use, the charcoal filters were cleaned by passing 5- 10 volumes of a size equal to the filter holder of each of the solvents acetone, methylene chloride, and carbon disulfide (in that order) through the filter. The filters were dried by either placing them in a glass tube under vacuum for 5-10 minutes (most commonly recommended) or in a drying oven at low heat for several minutes. The filter was then placed in its place in the CLSA. Primary internal standard was added to the one liter water sample in the sample bottle, the sparger put into the bottle, the bottle submerged into the preheated water bath, and the glass loop sealed tightly by clamping the ball and socket joints together. The trap heater was allowed to reach its set temperature, and the purge/circulation pump was turned on. Water bath temperatures commonly used were in the 30-50 degree centigrade range, and the

Charcoal Trap Heater Control Unit

Figure 3-8

CLSA Apparatus

4 4

ws Pump

filter trap heater was kept ~ O O C above the water bath so as to prevent water condensation. A common purge/run time was 2 hours. At the completion of the run, the pump, water bath, and trap heaters were turned off and the loop disconnected.

The charcoal trap/filter holder was removed with the FEP connection tube which connects the filter holder to the sample vial (Figure 2-4). The filters were extracted by placing 10 ul of solvent on the filter, moving the sample vial and filter holder apart slightly and back together about 20 times (which moves the solvent back and forth through the filter and extracts the SOCs). To collect the solvent in the sample vial, a piece of ice was held against the sample vial. his pulled the solvent through the filter and beyond to the sample vial. At this point, the vial/filter holder was swung in the air, forcing the solvent to the bottom of the sample vial. This extraction process was repeated twice more, using 5 ul of solvent each time, and the final solvent volume collected was 10-15 ul (Some solvent is lost due to evaporation and wetting of the filter.). The filter was cleaned by running several filter-holder volumes of extraction solvent through it and was dried in a drying oven before storage. At the completion of the extraction, secondary internal standard was added, the vial capped with a Teflon cap, and Teflon tape wrapped around the top. The samples were generally analyzed by GC or GC/MS within one or two days. Carbon disulfide is the most commonly used solvent in CLSA. Methylene chloride is occasionally mentioned as an elution solvent. Using methylene chloride would be convenient since it is the solvent used in CLLE, and the analyses by CLLE and CLSA would be very compatible and easily compared (use of the same GC columns and GC retention times for instance) .

The CLSA manual and the scientific literature do contain operational instructions, but these were worked out for noncommercial CLSA units which may not necessarily perform the same as commercially produced CLSAs. For this reason, some a .

operational parameters needed to be tested and optimized for this study. process

The developmental experiments performed 50 optimize the were :

Purge temperature - SOC recoveries at various bath/trap temperatures werg determined. The trap temperature was maintained at 10 C above the bath temperature. This has been shown to be the optimum temperature difference. Thg purge temperatures studied were 25, 30, 37, and 50 C. Purge time was 120 minutes.

Purging/stripping time - SOC recoveries after various stripping times were estimated. Purging times of 30, 60, 120, and 180 m.&nutes were studied. Purge temperature was 30 C.

Elution solvent - SOC recoveries using methylene chloride and carbon disulfide as the elution solvents were compared. Samples of 100, 200, and 250 ng/l were run and eluted by methylene chloride and compared to carbon disu1fi.de eluted runs with 200 ng/l SOCs in the original sample, The 200 ng/l methylene chloride samples were eluted a second time with 20 ul solvent to determine the completeness of the first elution. Pugge times were 120 minutes, and purge temperature was 30 C.

Minimum detection limit - The minimum statistically reliable detection limit was estimated.

The optimization experiments were performed with SOC concentrations between 25 and 250 ng/l (generally 100 ng/l) in distilled, deionized water solutions. Natural water matrix effects and source and finished drinking waters were studied later. The results from these experiments with CLSA are discussed in Chapter IV.

3-6. Application of Methods to Source and Finished Drinking Waters.

To examine the practical value of the methods that were developed on lrcleanll SOC solutions and to determine possible water matrix effects, application of the methods to natural water samples was required. These developed methods would also be compared to EPA1s methods by sending samples to a qualified laboratory for complete SOC analysis.

It was decided to use 3 source waters for the water matrix effect studies and preliminary SOC identifications. From these preliminary studies, one site was selected for a more thorough investigation into the SOC contamination of the source water and its finished drinking water. Complete SOC analysis was a - performed, applying CLLE and CLSA to the same samples.

In order to select appropriate water sources to study the SOC analysis methods, the surface water sources in North Carolina were reviewed. Reports on North Carolina industries (DeRosa, 1984) and NPDES discharge permits were studied.

A variety of personnel at the North Carolina Department of Environmental Quality (who deal with general fresh water quality in the state) and the Department of Health and Human Services (who deal with drinking water quality in the state) were consulted. The many options were weighed against each other and points on three different rivers were selected for source waters with a high potential for SOC contamination (Figure 3-9). The three water sources were:

1. Smith River,

at Eden,NC.

,LPhilpott Dam

2. Deep River,

below Asheboro,NC.

3. Yadkin River,

above Sal i sbury,NC.

g i g h Point a

Figure 3-9

' Natural Water Sourses used in Methods Evaluation Work

47

Smith River at Eden, North Carolina - Severe weekly water quality changes, including color, taste, and odor have been observed at the Fieldcrest Water Treatment Plant in Eden. These weekly water quality changes have been correlated to surges of water from an upstream dam which opens once a week. This water passes by heavily industrialized areas before reaching Eden.

Deep River below Asheboro, North Carolina - High Point, Randleman, Asheboro and other communities discharge wastes into the Deep River. Gulf and Goldston are two communities downstream which utilize the Deep River as a drinking water source. Acute Daphnia toxicity has been recorded for Deep River water below the Asheboro wastewater treatment plant.

adk kin River above Salisbury - A great number of industrial and municipal waste sources in the Winston- Salem and other areas upstream from Salisbury discharge into the Yadkin River. Salisbury uses the Yadkin River as a drinking water supply.

These three source waters were used to evaluate the water matrix effects on the recovery of the standard SOCs. Preliminary unknown SOC identification was done on these three sources as well, and one of them was selected for a more in-depth study of SOCs in the source water, as well as in the finished drinking water. Choosing the source for in-depth investigation was based on the number, concentration, and diversity of the SOCs found in the preliminary experiments. This source water was also used in a comparison between the CLLE/CLSA combination and EPAts Methods 624 and 625.

The

1 .

2.

3

4 .

experiments performed were:

Water matrix effects - SOC recoveries and minimum detections would be determined using three different natural waters and compared to those obtained with llcleanu distilled deionized water.

Preliminary SOC identification - preliminary analysis and identification of SOCs in the three natural water sources would be done.

In-depth SOC identification of source and finished drinking water - one of the three natural water sources would be studied more closely, and SOCs would be thoroughly identified in its source and finished drinking water.

Comparison to EPAfs methods - SOC spiked source and finished drinking water from the source under thorough

investigation would be sent to a qualified lab for complete SOC analysis. EPAts Methods 624 and 625 would be used by the qualified lab and results compared to simultaneous analysis of the same water, using the CLLE and CLSA methods.

These studies involved three setups for each water source: one system blank (distilled deionized water), one with the unspiked source water (for unknown SOC determination and so as to be able to account for any presence of the standard SOCs not originating from the spike), and one with standard SOC-spiked source water (for percent recovery and water matrix effect determination). Each of these setups was real, like all experimental runs in this research, performed in duplicate so as to ensure statistical validity.

GC/MS was employed for identification of unknowns in these water analyses. Peak areas were obtained in the GC/FID runs, and once identification was accomplished by GC/MS, the peak areas could'be used for estimating concentrations. The quantification was done by first calculating the relative response of the SOC of interest to the secondary internal standard that is chemically most similar to it. no wing the relative response of the primary:secondary internal standards which are closest to the SOC of interest in characteristics (and making appropriate adjustments if necessary), knowing the percent recovery in the water matrix of the SOC standard to which the SOC of interest is similar, and knowing the concentration of the secondary internal standard allowed for estimates of SOC concentration. This is admittedly an estimation since the assumptions and extrapolations are based on SOCs that are chemically similar but not identical. A greater degree of confidence in the results would be obtained by the use of radiolabelled isotopes or water matrix recovery studies for each of the SOCs identified in the natural waters, which neither time nor resources allowed. The approximate concentrations calculated probably have no more than a 50 percent error margin or standard deviation, which is fairly good for organic analysis at these concentration levels. The approximate concentrations reported give a good picture of the SOCs present, the relative importance of the various contaminants, and an idea of the amount of each of the SOC contaminants. In the comparison study'of EPA's methods to the CLLE/CLSA methods, SOC-spiked source and finished drinking waters were used to enable comparison of the methods in terms of their analysis for the standard SOCs, as well as for the unknowns present.

The results from these experiments with CLLE and CLSA and natural water samples are discussed in Chapter IV.

Water samples were obtained in clean glass bottles. The glass sampling containers were soaked in chromic acid, rinsed with copious amounts of distilled, deionized water, and then

solvent rinsed with either acetone or methylene chloride. sampling containers were allowed to dry prior to obtaining water samples. One gallon amber glass sample bottles were used for CLLE samples, and one or two liter CLSA bottles with ground glass tops were used for CLSA samples.

Samples of influent and effluent water from the Fieldcrest Water Treatment Plant were taken from the influent and effluent sampling taps there. These sampling taps were opened and allowed to run freely for several minutes, prior to obtaining a sample for analysis of SOCs. River water samples were obtained from the sites designated in the text. The procedure for obtaining river water samples involved wading out into the river several feet from the bank and then obtaining the water sample by completely submerging the sampling container below the surface of the water.

3-6-A. In-Depth SOC Analysis of the smith River - Background and Experimental Design.

As mentioned in the last section, an in-depth investigation of one water source was to be performed. The quantity and diversity of SOCs in the Smith River made this the source chosen. This source was also chosen because of an interesting current water quality problem. The background to this water quality problem and the experimental design for applying the analysis methods to the Smith River follow.

The Fieldcrest Water Treatment Plant (which uses the Smith River as its water source) has for a long time provided the city of Eden and Fieldcrest Mills (textile factories) with potable water. As the city of Eden grew, so did the demand for water, and the city built the Eden Water Treatment Plant which uses the Dan River as the water source (the Smith and Dan Rivers converge just south of Eden). In order for Fieldcrest Mills to receive water from their Water Treatment Plant and for the city to utilize the existing distribution system, both the Eden and - Fieldcrest Water Treatment Plants put their finished water into the same distribution system, and both, in effect, supply the Fieldcrest Mills and the private citizens of Eden with their water. The Fieldcrest Water Treatment Plant contributes about one-third of the water in the distribution system.

Operators at the Fieldcrest Water Treatment Plant and the water consumers of Eden have reported water quality problems for several years. The water treatment plant influent is reddish black on regular occasions, which frequently results in taste and odor complaints. The water quality problem seems to be a weekly phenomenon and occurs mainly during the fall, winter, and spring. Sometime between late Monday evening and Wednesday morning is the peak in poor water quality. At this time, the water changes to reddish black, the pH will occasionally increase by up to one pH

water sample was sampled 12 hours after the influent sample, which is the theoretical time it takes the water to pass through the plant. The water treatment plant is relatively old but well maintained. It chlorinates the plant influent, uses conventional alum flocculation, precipitation and settling, and filters through anthracite filter beds.

3-6-B. Comparison of CLLE/CLSA to EPA1s Recommended Methods for Broad Spectrum Analysis.

As mentioned in Section 3-6, a comparison of CLLE/CLSA methods to the EPA recommended methods (Methods 624 and 625) were performed to fully understand the applicability of the methods developed. Four samples were sent to a modern qualified laboratory for broad spectrum analysis by EPA1s methods. These methods are Method 624 (volatile SOCs) and Method 625 (nonvolatile/extractable SOCs). In order to compare the methods, not only for the analysis of the SOCs in the natural water, but also for the SOC standards used in this work, the 12 standard SOCs were spiked into the water samples as well.

The samples analyzed were:

1. Fieldcrest Water Treatment Plant influent with 0.5 ug/l SOC standards by CLSA and Method 624.

2 . Fieldcrest Water Treatment Plant influent with 5 ug/l SOC standards by CLLE and Method 625.

3. Fieldcrest Water Treatment Plant finished water with 0.125 ug/l SOC standards by CLSA and Method 624.

4. Fieldcrest Water Treatment Plant finished water with 2 ug/l SOC standards by CLLE and Method 625.

a - one of each of the four samples was delivered to the

qualified laboratory for analysis by Methods 624 and 625, and another identically spiked and treated of each of the four samples was analyzed by CLSA and CLLE within a few days at the Department of Environmental Sciences and Engineering.

After the completion of this project, further work was done by State and Federal officials to address the water quality problems of Eden. The North Carolina Department of Environmental Management, official From Virginia, and personnel from EPA Regions 3 and 4 completed a special study on the problem. The Martinsville Waste Water Treatment Plant was found to be a major source of the contamination. NPDES permits are in the process of being modified, and a new WWTP is being proposed that would have its NPDES permit adjusted to reflect the results of the study. Additionally, the Fieldcrest WTP recently incorporated treatment with polymers.

CHAPTER IV. RESULTS AND DISCUSSION

4-1. CLLE Optimization Experiments.

CLLE was the method chosen for development for the analysis of less volatile SOCs. The results from the evaluation and optimization experiments follow here.

Figures 4-1 and 4-2 show GC runs typical of the CLLE methods optimization experiments. It is important to note that three of the volatile SOCs used as standards (l,l,l-trichloroethane, trichloroethylene, and tetrachloroethylene) appear to be - methylene chloride (extraction solvent) artifacts which are present in blanks present in the or 3) but do appear Danish (Figure 4- explain. For thi

at unpredictable levels. These SOCs are not iginal solvent as are some other SOCs (Figure in methylene chloride concentrated by Kuderna- 4) at levels much higher than concentration ca s reason, the CLLE results for these SOCs are

considerably less reliable than for the other SOCs. These three SOCs can be-analyzed for well with CLSA though, so one does not have to rely on CLLE for their analysis.

4-1-A. CLLE - Extraction Time Optimization. The optimum extraction time at which an increase in

extraction time does not produce a noticeable increase in extraction efficiency was determined in these experiments. The percent recoveries of each of the 12 SOC standards were determined at the 3 extraction times tested (6, 12, and 24 hours) (Table 4-1, Figure 4 - 5 ) .

Three volatile SOCs (l,l,l-trichloroethane, trichloroethylene, tetrachloroethylene) are inconsistent and unpredictable in CLLE analysis due to their presence as solvenJt . . artifacts. The phenols are also poorly recovered, with recoveries of less than 10 percent for phenol and 0 for 2- chlorophenol. The very water soluble phenols are poorly recovered in most methods that involve removal of the SOC from the water matrix, which includes all commonly used methods of analysis except for some HPLC techniques. The phenols are better recovered at lower pHs, which will be evident in the results from other experiments.

The extraction efficiency increases beyond 6 hours but not beyond 12 hours extraction for most SOCs. There is a levellinq out of extraction efficiency in the 10-12 hour extraction rang& In this study, 12 hours will be used as the optimum extraction time .

- c- --2- - -

Trichloroethene

'ietrachloroechene e_

Table 4-1

CL L E - SOC PERCENT RECOVERIES vs. E X T R A C T I O X TIYE

Extraction Time (hrs.)

Standard

l,l,l-Trichloroethane "

Trichloroethene *

Tstrschloroethene X-

1,1,2,2-Tetrachloro2thane

Phenol

2-Chlorophenol

Saphthalene

Atrazine

Nethoxychlor

-x- substantial amount present in blank sample parentheses indicate standard deviations conditions: pH=7, 250 ml extraction volume, 2.5ug/l S O C std.

Extroc!ion T ime (hrg.)

V 1.2-3ibronoethane ( 3 3 ) V Phenol 0 1.1.2.2-Tetr2chloroethane frichloroet5ene

Tetrachloroethene

Figure 4-5

CLLE - % Recovery vs Extraction Time

5 8

4-1-B. CLLE - Extraction Volume Optimization. The percent recoveries for each of the 12 SOC standards were

determined with three different extraction solvent volumes (100, 250, and 400 ml) (Table 4-2, Figure 4-6). Most SOCs show a very moderate to virtually no increase in recovery, with increase in solvent volume. The solvent volume appears not to affect the SOC extraction efficiency as long as a certain minimum amount of solvent is used.

There were some inconsistencies in the results. Tetrachloroethylene is recovered fairly well (36 percent) at 250 ml but not at all at 100 or 400 ml. This is one of the volatile SOCs not reliably analyzed for by CLLE. The phenols do exhibit a slight increase in efficiency with increased solvent volume, but the large deviations in the results for the phenols (probably due to the unfavorable pH for phenol extraction) make these results questionable.

Large solvent volumes resulted in more contamination and dirtier blanks. A solvent volume of 250 ml was most appropriate for the operation of the CLLE apparatus and resulted in an overall extraction efficiency at-least as good, if not better, than the other volumes. It was thus determined that 250 ml of methylene chloride was the optimum solvent volume to be used in the CLLE.

4-1-C. CLLE - pH Optimization. The percent recoveries for each of the SOC standards were

determined at three pHs (2, 7, and 12) to determine the optimum pH (Table 4 - 3 , Figure 4-7). The pH is not a linear scale, and the presentation of extraction efficiency versus pH in Figure 4-7 does not show the whole, true picture. The chemical characteristics (such as extraction efficiency) of an SOC generally change drastically around its pKa and do not follow the. simplified linear relationship depicted in the figure. This plot does, however, give an indication of how well the SOCs are extracted at three commonly used pHs.

The extraction efficiency was highest at unadjusted pH (pH of 6-7) for all SOCs. Problems with recovering the phenols are once again obvious, and the large standard deviations in the recoveries of the phenols are a proof of this problem, indicating the unpredictable and generally low extraction efficiency of phenols. It was surprising that the phenols did not exhibit better recovery at the lower pH. However, the low pH samples were quite dirty, resulting in large standard deviations and low reliability for the SOC data. The recoveries of the SOCs at the higher pH are also poorer than one might expect, all being less than 50 percent recovered. This is surprising, in particular,

T a b l e 4-2

C L L E - S O C P E R C E N T R E C O V E R I E S VS. S O L V E N T VOLU>IE

E x t r a c t i o n V o l u m e ( m l s )

1 0 0 250 400 S t a n d a r d

l,l,l-Trichloroethane ++

T r i c h l o r o e t h e n e *

1 , 2 - D i b r o n o e t h a n e (EDS)

T e t r a c h l o r o e t h e n e

1 , 1 , 2 , 2 - T e t r a c h l o r o e t h a n e

P h e n o l

2-Chlorophenol

N a p h t h a l e n e

1 , 4 - D i c h l o r o b e n z e n e

A t r a z i n e

M e t h o x y c h l o r

* s u b s t a n t i a l a m o u n t p r e s e n t i n b l a n k s a m p l e p a r e n t h e s e s i n d i c a t e s t a n d a r d d e v i a t i o n s c o n d i t i o n s : pH=7, 12 hrs. e x t r a c t i o n t i m e , 2.5ug/1 S O C s t d .

60

Extraction Volume (ml)

0 1,2-Di broroet hane ( D 3 ) A Phenol 0 1.1.2.2-Tetrachloroeth3ne 0 Trichloroechene

B T e t r a ~ h l o r o e t h e n ~

150 2 5 0 550

f x t roc t ion Volume (ml) . . V Methoxvchlor A Naphthalene

~ r i s ( c h l o r o ~ r 0 ~ ~ 1 )phosphate . 1 ,L-Dichlorobenzene 0 ~ t r a z i n e 0 2-Chlorophenol

Figure 4-6

CLLE - % Recovery vs Extraction Solvent Volume

T a b l e 4-3

CLLE - S O C PERCENT RECOVERIES vs. EXTRACTIOX pH

E x t r a c t i o n pH

l , l , l - T r i c h l o r o e t h a n e

T r i c h l o r o e t h e n e

S , 2 - D i b r o n o e t h a n e (EDB)

T e t r a c h l o r o e t h e n e $F

1 , 1 , 2 , 2 - T e t r a c h l o r o e t h a n e

P h e n o l

2 - C h l o r o p h e n o l

S a 2 h t h a l e n e

1 , 4 - D i c h l o r o b e n z e n e

A t r a z i n e

W e t h o x y c h l o r

s u b s t a n t i a l a G o u n t p r e s e n t i n b l a n k s a m p l e p a r e n ~ h e s e s i n d i c a t e s t a n d a r d d e v i a t i o n s c o n d i t i o n s : 250 rnl e x t r a c t i o n vol., 12 hrs., 2.5u9/1 S O C std.

0 1.2-Dibror.oethane ( E D B ) pH

E Phenol -- - - , 1,1,2,2-Tetrackloroethane 0 Trichloroethene Tetrachloroethene

2 4 6

0 Hechoxychlor pH

fl Tris(chloropropy1)phosphate 0 Atrazine

A Naphthalene

A l,l-Dichlorobenzene

Figure 4-7

CLLE - % Recovery vs Extraction pH

6 3

for a slightly basic SOC expected to be extracted of 6-7 appears to be the standards, based on this

such as atrazine, which would be well at a higher pH. The unadjusted pH best pH for the extraction of the SOC experimental data.

4-1-D. C U E - pH Variations and Complementary pHs. The extraction efficiencies for the SOC standards were

evaluated at three pH combinations, 12 and 2, 7 and 2, and 7, 12, and 2, and the results are presented in Table 4-4 and Figure 4-8 .

A pH combination of 7 and 12 proved to have the best extraction efficiency, even though its extraction time was 24 hours, as opposed to 36 hours for the other two pH conditions. The recoveries were 10-30 percent higher for most SOCs using the pH 7 and 2 combination as compared to the other pH combinations. Additionally, these two other pH combinations resulted in more contamination (and dirtier blanks) than the 24 hour extraction, in particular, for the triple pH (7, 12, and 2) runs. Based on these experiments and those presented in the last section, a pH of 12 seems not to benefit the extraction efficiency for any of the SOC standards under any of the conditions tested. An interesting observation in these experiments, as well as in those presented in 4-1-C, is that 1,1,2,2-tetrachloroethane appears to need a neutral or acidic initial pH to be extracted. An initial alkaline pH results in no extraction at all of this SOC, possibly due to base degradation of it. In these experiments, the phenols are far better extracted than in earlier experiments, in particular with the pH 7 and 2 combination. The pH combination of 7 and 2 appears to have the highest extraction efficiency of any single pH or combination of pHs.

It should be pointed out that a pH of 7 and 2 may not necessarily be the best combination for all SOCs, even though it is good for the SOC standards in this study. There are SOCs and groups of SOCs with characteristics that might make another pH'or pH combination better for their extraction, and if their presence is suspected, the method can be modified. For initial, broad .

spectrum screening and analysis of a sample, the extraction pH combination of pH 7 and 2 appears to be the most efficient.

4-1-E. Preliminary Comparison of CLLE and EPA Method 625.

Having optimized the various parameters for CLLE of solutions containing SOCs, a preliminary comparison of CLLE to the traditionally used separatory funnel (using EPA Method 625) could be made. The extraction efficiencies of both methods were evaluated and compared (Table 4 -5 , Figure 4-9).

Table 4-4

C L L E - SOC PERCENT R E C O V E R I E S v s . pH V A 2 I A T I O K S

Extraction p H s

Standard

l,l,l-Trichloroethane <$

Trichloroethene *

1,2-Dibromoethane (EDB)

Tetrachloroethene -%

Phenol

Xaphthalene

1,4-Dichlorobenzene

Atrazine

Tris(chloropropyl)phosphate

-x- substantial amount present in blank sample parentheses indicate standard deviations conditions: 2 5 0 r n l extraction volume, 2.5up/l SOC std.

pH-Voriotiona 1.2-aibrocoetkane (EDB)

Phenol fetrachloroethene 2-Chlorophenol

1.1 .?.i-Tetrachloroethane

Naphthalene

1 ,&-Dic hloro benzene

Figure 4-8

Tris(chloropropy1)phosphate

Hethoxychlor

CLLE - % Recovery vs pH Variations / Complementary pH's

T a b l e 4 -5

P 2 E L I Y I X A R Y CO>IP.4RISO?i OF EXTRACTION ' E F F I C I E 3 C Y OF C L L E TO EPX YETHOD 6 2 5

M e t h o d

S t a n d a r d C L L E 6 2 5 -

T r i c h l o r o e t h e n e *

' T e t r a c h l o r o e t h e n e -X-

P h e n o l

S a p h t h a l e n e

A t r a z i n e

-2 s u b s t a n t i e l a m o u n t p r e s e n t i n b l a n k s a m p l e p a r e n t h e s e s i n d i c a t e s t a n d a r d d e v i a t i o n s c o n d i t i o n s : p H = 7 , 250 m l e x t r a c t i o n v o l u m e , 2 . 5 u g / l SOC s t d . 2 x 1 2 h r s . e s t r a c t i o n t i n e f o r C L L E

1.2-Di b r o a o e t h a n e (EDB) P h e n o l

CLLE N a p h t h a l e n e

Figure 4-9

CLLE - % Recoveries in a Preliminary Comparison

of CLLE to EPA Method 625

m fd m a ,

4J ~ d - a m

7 k U d - m hc, 4J

m k 2 3 2 m c- r id u m a ,

F: c f d 9 3 5 m % F:

H 0 w m O U G - d W O Q ) v 3 O m - W f d k h f d 4J-d E

m - d Id (C(3 j l ,d * a 3 a, +r G W O - d r ncu rn~ l - r - m m a) e a ~ m =I UN+J o O a h a a , " - d o c * m%i H O d G G + , - d d k t4Ua.I

m + , r d a c ~ m 3 m a o U U U . k C orn3-da, o w 2 a a m 0 k m Q ) k U G k -A o w Q l w 0-4 c 0-4 -E au+J Q)4J o a k *

w d a r l w Q) O Q r l f d Q ) Q ?v m ~ z - d c a r n ~ ~ k m m c c c ~ : dh + J U C r d * C f d O 0 7 k O

2 : z g 1.-2sy.:.: 32 k 3 ' d E i c u t n a , d C ) G ad* k c 0 r d o p o o u k u x a c r d - r l o a U a k C , + , k 0 a0F:Kl rn R c m o h a w u rd 3 m x o h - ~ ~ Q ) o ~ A ~ o F : ~ ) a k - r f ~ : U L n a U E , G 7 0 Q ) d f d Q ) k - d

k c u I k u h m C - ~ A U +)+) h a , W N a , a , f d d 0-v a fd f d a k C - a G kua-4 U 0 G G 3 Ei a, a, D r d d I3 -V G d + J ld#-dB u) > - d o a , Q ) g r d k d o m

S C m o a + ) k m rd+,g a + , d m * I U 3 f d x a a * k U @ h a , W-cu-d k a , a , k a 7 h m

- w a m a f d a+,@-d E i P G a o N w c aa m a - d a O & Q 3 0 E: - 0 f d Q fd id Z= 3 G d w mr= m k arc -4 td -4 a) k-4 a a, u - a x a,@ m+J m e~ k a d m > k r c k a m - r i a m r l o - d fd a m ~ ) u o a , o ace a&+,

~ o u P ( ~ R ~ A - ~ , G u ~ ~ P , ~ m m - d w a , 0 0 # O Q ) a,C, c N k b W . . U l - d % l Q) Q)U ld -U U c 0 - d f d k Q ) + J 0 , G B 9 , E : a , 7 a G ::tn a, ~ + k a , d o r m ~ u +J 4J rd+J,G&cU fd-4111 O G 0 r d m - 4 a c r n u w m d ~ s + J - + d 8 0 -'a + J O f d + J + J > k Q

a , h d ' X a , 3 O ! ~ k d k U l Q C Q ) 2 z e g @ 3 d f d O O a , E = o C,XCf + J Q ) O a J 4 J u w e > @ a, a + J F : e k d

d a, 0-data o a o C Q ) Q ) O h G U U a , 4 J a , a , - - d U a a U , G , G e k a a ) - 4 + ) m G r l + ) e 0 3 E C , ~ Q ) O & ~ U O ~ - d f d U O Q ) ~ U l

,G 3 k w f d u =Id k fd U A 0 .c as23 a, rc u f d a ) k +J

a a + J a , m X c e , a , ~ a o

- r ( m a , ~ d c x $ ~ - ~ d a , x ~ ~ d = f k 3 fd k U U - d a a a fd m a Q ) mw

G -4

m aa,a a, d & d m D

w d O f d m a , F: c u 4 J a m c u d (d

O - 2 * & c c u c l k O e m a , d f d O a ,

a - d - d - d G f d + J - d G k d 3 Clctn-IJ E d E m am: > -d & m a, w - d u-4 fd 3 c a c - d U I O 3 3 k -da,3c, ma , * o o

o Q ) Q ) a c a w E-:a, O a , rl a o m m ~ d - r l ~ a - r l u *- d C:-d-d-d4JC) k4J m d 0

0 k r d a fdQ) \+J % o c a , a ucar(i,m -4 +J C o ~ o c o 710 w a o k a a a 4 J d G a m + ) + ) fd m a 0 + J f d = f O a c m m k a , c c 0 4 J l d o o +,34Ja,

U E C x 0 7 , G 4J a d OC)

m X 0 4 a k-d A U k 0 3 0 Q f d G

h4 m s a c w -85: *-r( d 4 E G U rl w ( d 7 m d t n f d a id * a , a " \ $ 5 5 L ) ' d , G r O @ - d m -ti a,+J+J k G a 3 3 c 4J4J c =f+J*d a, 0.- m rd a, a)+r-d.rd 7 c-r lr l

- d B U E r d 3 d O M \ +)-dF:-dk e 3 f d - d U D a * -d k 0 c 0 0 >+J 0 c 4 J m m a 4 J a , - d d rdc , m a , a - r l 3 + , a , m k a , o

- X d W U A - d 4 J a L n a a a a , a, c N G k Q ) a +,a ,+Ja, fd a, UldG U L

Q) 3 7 r d L % % c a 3 a

+J 0 0 0 - d 0 U ~ m m c c a , a w U G A O A + J f d Q ) - d m L = 4J - d f d W - 4 - d 3 R - d C ) a c d a U R a a o a , @ \ - k 3 - d r c o a c - + J U t P a , k O a d d m O f d k 7 G 0 , G 0 O a , = 1 a , - d H a , 4Jrl.U k c # > d G O a0 a,?vAfd+J - 0 a , - d L n a d f d k [OOfdU) E+J ca, arc 422

U + , Q ) O U Q ) o o o f d 0 7 m - 4 m Q ) e U c G + J O Q ) g 0 - d k f d G a , + J k 0 A G 0 0 0 a,G-dG Q) 0 a r d + , ~ ~ : ~ ~ + ~ m a ~ i w

aJisJi l "a"

G d ' a u f d 7 a J r 3 C f O k

a, 4J 3 a,

3 G m m o B c c d

4-2. CLSA Optimization Experiments.

CLSA was the method chosen for the analysis of the more volatile SOCs. Figures 4-10 and 4-11 show GC runs typical of CLSA methods optimization experiments. Preliminary CLSA runs indicated that there would be no problems with the three volatile SOCs that were poorly recovered by CLLE. Their presence as solvent artifacts is not a problem, since carbon disulfide is used instead of methylene chloride as the solvent. In CLSA, the solvent is not concentrated, and benzene (a carbon disulfide contaminant) and diacetone alcohol (from the acetone in which the SOC standards were dissolved) are the only SOCs attributable to the solvent found in CLSA runs (Figures 4-12 and 4-13). The three least volatile of the SOC standards (atrazine, tris(chloropropyl)phosphate, and methoxychlor) were not volatilized and recovered by CLSA. These are, however, recovered very well by CLLE. CLSA was able to analyze for the SOC standards in the volatility range from naphthalene to 1,1,1- trichloroethane, except for the highly polar phenols and 1,2- dibromoethane. Occasionally, 1,2-dibromoethane, the more volatile of these three SOCs, was recovered by CLSA, but this was unpredictable. CLLE complements CLSA well, since it is capable of analyzing for these SOCs. CLLE analyzed for the SOC standards in the range from 1,2-dibromoethane to methoxychlor. The two methods thus appear to complement each other well (Figure 4-14).

4-2-A. CLSA - Purge Temperature Optimization. The effects of purge temperature on efficiency are shown in

Table 4-6 and Figure 4-15. The purge temperature did not significantly affect the efficiency of SOC recovery. The more volatile SOCs are recovered slightly better at lower temperatures and the less volatile somewhat better at higher temperatures. It appears that once good volatilization is achieved for any given SOC, additional temperature increases decrease the trapping a - .

efficiency of the charcgal filter. For the SOC standards, a purge temperature of 30 appears to be best, and this will be-the temperature used in this work.

4-2-B. CLSA - Purge Time Optimization. Table 4-7 and Figure 4-16 show the effects of varying purge

times. As with purge temperature, the purge time does not appear to significantly affect the recovery of the SOCs. The less volatile SOCs are recovered slightly better with longer extraction times, whereas the highly volatile SOCs are recovered well at shorter extraction times (possibly due to losses through minute leaks in the system). Once a minimum required purging time for acceptable recovery is met, there appears to be little reason to extend it, since the amount of contamination clearly

Trichloroethene

1.1.2-Trichloroethane

Tetrachloroethene

T a b l e 4 - 6

CLS.4 - S O C P E R C E S T RECOVERIES v s . PVZGE TE?IPEXXTURE

P u r g e T e m p e r a t u r e ( o C )

S t a n d a r d

T r i c h l o r o e t h e n e

I

T e t r a c n l o r o e t h e n e

P h e n o l

N a p h t h a l e n e

A t r a z i n e

p a r e n t h e s e s i n d i c a t e s t a n d a r d d e v i a t i o n s c o n d i t i o n s : t r a p t e r n p . = l C o c a b o v e b a t h , 2 0 0 n g / l S O C s t d .

Purge T e m p e r a t u r e (degrees .C) 1.4-Dichlorobenzene o Naphthalene

v 1,1,2,2-Tetrachloroethane B Trichloroethene o Tetrachloroethene v l,l,l-Trichloroethane

Figure 4- 1 5

CLSA - 5% Recovery vs Purge Temperature

7 6

Table 4-7

Standard

CLSX - SOC PERCENT RECOVERIES vs. PURGE T I H E

Purse Time (minutes

l,l,l-Trichloroethane

Trichloroethene

l,2-Dibronoethane ( E D 3 )

a Tetrachloroethene

1,1,2,2-Tetrachloroethane

Phenol

2-Chlorophenol

Naphthalene

1,4-Dichlorobenzene

Atrazine

~ T r i s ( c h l o r o p r o p y l ) p h o s p h a t e

Me t h o x y c h l o r

parentheses indicate standard deviations c o n d i t i o n s : purge temp.=30oC, 200 ng/l SOC std.

77

P u r g e T ime (min . )

1,4-Dichlorobenzene o Naphthalene o Tetrachloroethene A Trichloroethene

A 1,1,2,2-Tetrachloroethane 6 1,1,l-Trichloroethane

Figure 4-16

CLSA - 5% Recovery vs Purge Time

Table 4-8

CLSA - SOC PEZCENT RECOVERIES, COXPARING RETHYLESE CHLORIDE TO CARBOS DISULFIDE AS ELUTIOX SOLVENT

Solvent

Standard MeCl 1 >IeCl 2 XeC1 3 YeC1 3/2 CS2 -

Trichloroethene

1,2-Dibromoethane (EDB) 0 5.4 0 0 0 ( 4 . 7 )

Tetrachloroethene

Phenol 0 0 0 0 0

Saphthalene

parentheses indicate standard deviations conditions: 30oC purge temp., 120 min. purge time. NeCl 1=100 ng/l SOC std., MeCl 2=250ng/l SOC std., ?kc1 3= 200 ng/l std., NeC1 3/2=;bIeCl 3 extracted a second tine uith 20 ul solvent, CS2= 200 ng/l SOC std.

Table 4-9

CLSA - SOC PERCERT RECOVERIES v s . SOC C O X C E N T R A T I O N

SOC Concentration (ne/l)

25 - 5 0 - 100 200 - Standard

Trichloroethene

1,2-Dibromoethane (EDB)

Tetrachlcroethene

Phenol

2-Chlorophenoi

Naphthalene

1,4-Dichlorobenzene

Atrazine

T r i s ( c h l o r o p r o p y l ) p h o s p h a t e

Nethoxychlor

parent'heses indicate standard deviations conditions: Bath temp=30oC, Purge t i n e = 1 2 0 min., Soluent=CS7

' . SOC Standard Concentrat ion (ng/l)

0 Tetrachloroethene r Trichloroethene A 1,4-Di~hlorobenzen~ v I , 1,l-~richlo'roethane D l,192.2-Tetrachloroethane a 1,2-Dibronoethane (EDB) O Naphthalene

Figure 4-1 7

CLSA - % Recovery vs SOC Conc, entra tion

is the minimum amount required for adequate ng/l sample would be the minimum detection possible, even when taking the lower than 1 into account, by increasing the injection v

GC re limit. 00 per olume

spons Thi

cent to 1.

'e, a .s sho recov 5-2.0

10 uld be eries ul.

The one order of magnitude greater sensitivity possible with the GC-FID should make it possible to further lower the minimum detection limit (in spite of the increased noise that would result) below the 10 ng/l level estimated for the CLSA under the conditions used in this work. A 2 liter sample bottle can also be used instead of the one liter used in this work, which ideally would double the amount of SOC available for analysis, The volume of solvent (carbon disulfide) could probably also be decreased somewhat without decreasing the recoveries proportionately. These steps should bring the minimum detection limit of CLSA to low ng/l levels. EPA1s recommended method for the analysis of a broaa range of volatile SOCs, Method 624 (a purge and detection which it

rap method) imit in the applicable

, is generally 500 ng/l to 1

cons 0 ug/ and

idered to have a min 1 range for the SOCs 625 use an MS as the

imum for

detector, which may have a slightly lower sensitivity than the FID used in the CLSA work. This difference does not, however, account for the 2-3 order of magnitude difference invdetector' limit.

Application of CLLE and CLSA Methods to Source and Finished ~rinking Waters.

As explained in Section 3-6, the Deep River, Yadkin River, and smith River were chosen as water sources for an initial survey, each being sampled and analyzed once at about the same time. The Smith ~iver was chosen for a more in-depth study because of an interesting current water quality problem with this water source and clearly evident SOC contamination.

4-3-A. Water Matrix Effects.

The diverse and numerous constituents that make up a natural water matrix were expected to affect the analysis of SOCs. The SOC standards were again used and their recoveries from natural waters determined and compared to recoveries obtained in the %leanu water studies in Sections 4-1 and 4-2 to study matrix effects.

The results from the water matrix experiments with CLLE are presented in Tables 4-10 and 4-11 and percent recoveries compared to those obtained with clean SOC solutions (Table 4-5). Similarly, the CLSA water matrix results are presented in Tables 4-12 and 4-13 and compared to the data in Table 4-9.

T a b l e 4-10

S t a n d a r d

C L L E WATER MATRIX EFFECTS. S O C P E R C E N T RECOVERY FOR SATURAL VATEXS

W a t e r S o u r c e

T r i c h l o r o e t h e n e

1,2:Dibronoethane (EDB)

T e t r a c h l o r o e t h e n e * . 1 , 1 , 2 , 2 - T e t r 2 c h l o r o e ~ h 2 n e

P h e n o l

N a p h t h a l e n e

A t r a z i n e

M e t h o x y c h l o r

* substantial * * 2.5 ug/l spi

*** not chromato parentheses indi -Smith Aa = Smith Smith Ab = Smith

Deep<$-X- Y a d k i n S m i t h Aa S m i t h Ab .

amount present ke rather than graphically res cate standard d River, extract River, extract

in blank 5 ug/l olved eviation ion pH = ion pH =

sample

T a b l e 4 - 1 1

CLLE -?;'ATE% ) IATRIX EFFECTS - SOC PERCENT R E C O V E Z Y FOR NATURAL WATERS

Water S o u r c e

Standard S m i t h B A m i t h Ca S m i t h C3

T r i c h l o r o e t h e n e 3+

Phenol

S a p h t h a l e n e

A t r a z i n e

* substantial amount present in blank sample * * * not chromatographically resolved parentheses indicate standard deviations Ca = Fieldcrest WTP influent, Cb = Fieldcrest WTP effluent ( 6 / 1 0 / 8 5 ) B = Fieldcrest WTP influent ( 5 / 1 5 / 8 5 )

T a b l e 4-12

CLS.4 - WATER MATRIX EFFECTS. SOC P E R C E N T RECOYEZIES FOR NATUR.4L WATERS

W a t e r S o u r c e

D e e o ahi in S n i t h A - S n i t h E S n i t h C

T r i c h l o r o e t h e n e

T e t r a c h l o r o e t h e n e

A t r a z i n e

T r i s ( c h l o r o p r o p y 1 ) p h o s p h a t e

N e t h o x y c h l o r - -

* indicates 500 ng/l spike rather than 100 ng/l * * * not chromatographically resolved parentheses indicate standard deviations

Table 4 - 1 3

CLS-4 - \+'.ATER > l A T R I X EFFECTS. SOC PERCEXT RECOVERIES FOR NATURAL WATERS

Smith River Water P-

Standard Source Upstream Influent Finished

Trichloroethene

1,2-Dibromoethane (EDB)

Tetrachloroethene

1,1,2,2-Tetrachloroethane

P h e n o l

2-Chlorophenol

Naphthalene

Atrazine

Methoxychlor

* * * not chromatographically resolved parentheses indicate standard deviations conditions: 100, 200, 500 and 125 ng/l spike respectively

For the CLLE samples, various water matrix effects could be observed this gen variabil solution sample a organic For the

. T era1 ity S.

he Deep ~iver sample ly resulted in decrea (standard deviation) The Fieldcrest Water

was sed as Tre

heavily contaminated, and recovery and increased compared to the clean water atment Plant finished water

lso contain compounds, other less

ed large numbers and high and SOC recoveries were s contaminated sources, the

concentrations of ignificantly reduced. Yadkin River and

Smith River waters (other than finished-water from the treatment plant), matrix effects were less evident, and reduced recoveries were observed mostly for the large molecular weight SOCs such as atrazine and tris(chloropropyl)phosphate. The phenols exhibited unpredictability again. In an extraction performed at pH 12 and 2 (source: Smith A b ) , recovery of 1,1,2,2-tetrachloroethane and the phenols was significantly poorer than at pH 7 and 2 (source: Smith Aa), supporting the previously mentioned hypothesis of base degradation of 1,1,2,2-tetrachloroethane.

For the CLSA samples, the matrix effects are comparable to those seen with the CLLE. The greater concentration factor in CLSA appears to make it somewhat more sensitive to the various types of water matrix interferences (Tables 4-12, 4-13, 4 - 9 ) . The two most contaminated samples (the Deep River and the finished water from the Fieldcrest Water Treatment Plant) show severe matrix interference. The Deep River sample was so heavily contaminated that none of the SOCs could be chromatographically separated and identified. Due to poor chromatography, it is difficult to identify the type of contamination, but it appeared to be hydrocarbons. The Fieldcrest Water Treatment Plant finished water had such a great number of intermediate volatility compounds at high concentrations that the SOC standards in the elution range from tetrachloroethylene through the phenols could not be identified. The high concentrations of these SOCs make it more difficult for the SOC standards to compete for the binding sites on the small charcoal filter of the CLSA, partially excluding them from analysis at an early stage. For the other cleaner samples, there was a 10-40 percent reduction in the ' - recovery of the SOC standards and slightly increased standard deviations (decreased reproducibility) as compared to the "pureu SOC solution results discussed in Section 4-2 (Table 4 - 9 ) . The more volatile SOCs appear to be only slightly affected by the matrix, just as in CLLE. An interesting deviation from this reduced recovery trend was 1,2-dibromoethane, which showed better recovery in some of the natural waters than in pure solutions. This could possibly have been due to water matrix constituents which decreased the water solubility of 1,2-dibromoethane, but the variability in this S O C V s recovery were so great that conclusions were difficult to make.

In the different sources studies, the water matrix contributed to reduced SOC recovery from 0 to 50 percent. Overall, the natural water matrices decreased the predictability

and reproducibility of SOC analysis to a varying degree. Even the same water source had different water matrix effects from one day to the next, as seen in the data for Smith A, B, and C. The components that create these water matrix effects are numerous and-cannot easily be identified, and their relative importance altering the SOC recovery is unknown. For this reason, each source or sample has to be studied for its particular water matrix effects (using a number of representative internal standards) in order to determine the applicable SOC recoveries be used when identifying and quantifying SOCs in that particul water.

4-3-Be Preliminary SOC Survey.

The preliminary survey of the Deep, Yadkin, and Smith Rivers included an attempt to identify SOC contaminants in those water sources. The smith River was chosen for an in-depth study (section 4-3-D). Smith A is the first sample taken and used for the preliminary SOC survey, and the samples referred to as Smith B and Smith C are from later sampling dates of the Smith River - -

for the more in-depth study.

The results from preliminary SOC analysis and quantification (according to the method explained in Section 3-6) of the three natural water samples by CL-u and CLSA are presented in Tables 4- 14, 4-15, and 4-16.

The Deep ~iver water was so heavily contaminated (probably due to the sampling site's proximity to the Asheboro Wastewater Treatment Plant discharge site) that satisfactory analysis was not possible. The amount and number of components in the sample was so great that separation and analysis was possible for very few of the compounds present. For this reason, MS analysis was not performed on these samples, and the SOCs reported in Table 4- 14 are only a fraction of the compounds actually present. Many . compounds were present at concentrations greater than those of the few reported. Substantial dilutions of the Deep River samples might have made analysis possible but probably only for the dominant contaminants at the highest concentrations. The SOC contaminant present at the highest concentration of those that could be identified in the Deep River was toluene, which was present at 5-10 ug/l levels. An unidentified SOC referred to as Unidentified A was observed in the CLLE sample of the Deep River, also at 5-10 ug/l levels. This unidentified SOC was seen in other samples analyzed by MS. Other SOCs were identified in the Deep River samples at lower concentrations, as reported in Table 4-14. BHT and a series of phthalates are among the contaminants seen in all CLLE samples, including the blanks (They are relatively large, nonvolatile molecules, not analyzed for by CLSA.) at greatly varying and inconsistent concentrations, generally in the 2-20 ug/l range. For this reason, it was

T a b l e 4 - 1 4

soc

P R E L I > I I N A R Y SOC ANALYSIS OF THE DEEP AXD YADKIN RIVERS

l , l , l - T r i c h l o r o e t h a n e T r i c h l o r o e t h e n e 1 , 2 - D i b r o n o e t h a n e T e t r a c h l o r o e t h e n e 1,1,2,2-Tetrachloroethane N a p h t h a l e n e

T ~ l u e n e X y l e n e ( i s o m e r s ) C 3 - b e n z e n e ( t r i o r e t h y l -

m e t h y l i s o m e r s ) T r i c h l o r o b e n z e n e ( i s o n e r )

U n i d e n t i f i e d A U n i d e n t i f i e d B

4 - m e t h y l - 3 - p e n t e n - 2 - o n e D i a c e t o n e a l c o h o l $$*

A p p r o x i m a t e C o n c e n t r a t i o n ( u g / l )

: - D e e ~ R i v e r Yr:':-X- Y a d k i n R i v e r

CLSA CLLE

B u t y l a t e d h y d r o x y t o l u e n e ( B H T ) -%

D i m e t h y l p h t h a l a t e * D i e t h y l p h t h a l a t e * D i b u t y l p h t h a l a t e -%

D i e t h y l h e x y l p h t h a l a t e O t h e r u n i d e n t i f i e d p h t h a l a t e *

CLSX C L L E

y e s 11

11

11

I I

11

"SOCs a l s o p r e s e n t i n b l a n k s a t v a r y i n g l e v e l s +t*SOCs o r i g i n a t i n g f r o m t h e a c e t o n e (SOC s p i k e s o l v e n t ) * + V l e e p R i v e r s a m p l e s were t o o h e a v i l y c o n t a m i n a t e d f o r g o o d

a n a l y s i s

Table 4 -15

SOC XXALYSIS OF SXITH RIVER WATER BY CLLE - FIELDCXEST WATER PLANT INFLUENT

Apprpxinate Concentration (ug/l)

Snith A ' Snith B Snith C

Phenoi Tris(chloropropyl)phcsphate Naphthalene

Xylene (dimethyl benzene, isomer) C3-benzene (tri or e t h y l - n e ~ h y l

isorner) . Unidentified E Unidentified F

H-(ethyl)-N-(phenyl)acetzmid 1 , 3 , 3 - T r i m e t h y l - ? - i n d o h o n e Atrazine Triazine pesticide (not Atrazine)

4-nethyl-3-penten-2-one *??

Diacetone alcohol ** Butylated hydroxy toluene (BHT) * Dimethyl phthalate * Diethyl phthalate ?$

Dibutyl phthalate K-

Diethylhexyl phthaltate " Other (unidentified phthalate) a

y e s 11

*SOCs also in blanks in varying levels * * S O C s originating from the acetone (SOC spike solvent) *s5SOCs that are extraction solvent artifacts Fieldcrest Gater Treatment Plant Influent, 5 / 3 , 5/15, and 6/10/55.

Table 4-16

soc

soc

Tetrachloroethene Naphthalene

4N.ALYSIS OF S?l ITH R I V E R WATER BY CLS.1 - F I E L D C X E S T WATER PLAXT I X F L U E N T

Xylene (isomers) C3-benzene (tri or ethyl-

dethyl isomer) Trichlorobenzene Benzonitrile

Unidentified A Unidentified C Unidentified D

Approximate

S n i t h A

Concentration ( u g / l )

Snith B Smith C

* . Fieldcrest !dater Treatnent Plant Influent, 5 / 3 , 5/15,.and 6 / 1 0 / 5 5

difficult to estimate the amount of these SOCs present in the different water sources.

The Yadkin River water was very clean, with few SOCs present. One of the xylene isomers was identified at 2-4 ug/l, and an SOC referred to as Unidentified B was present at 1-2 ug/l. Other than those two SOCs, ng/l levels of four other SOCs were found. A larger number of SOCs was identified in the preliminary study of the Smith River water than in the Deep and Yadkin River studies (Tables 4-15 and 4-16). A total of 15 different SOCs originating in the water sample were identified by CLSA and CLLE. The concentrations ranged from 40 ng/ to 2 ug/l. The main contaminants included phenol, a xylene isomer, a C -benzene isomer, a triazine (not atrazine) , N- (ethyl) ON- @h2nyl) - acetamide, 1,3,3-trimethyl-2-indolinone, and a compound referred to as Unidentified E. Unidentified E is an SOC which had been frequently seen in studies of the Haw River, North Carolina but has not been identified (Dietrich et al., 1983). All of these SOCs were present in the 1-2 ug/l range.

4-3-C. Solvent Artifacts.

Two SOCs identified at 10-20 ug/l levels in the SOC spiked Yadkin River CLLE samples were acetone dialcohol and 4-methyl-3- penten-2-one. These two compounds were present only in other CLLE samples spiked with SOC standards and are attributable to the acetone in which the SOC standards were dissolved. Both can be formed from acetone by the base or acid catalyzed aldol condensation (Lin et al., 1981, pp. 861-906):

Diacetone alcohol

These acetone artifacts were also seen in other CLLE samples. This could be due to the pH change done in the CLLE extraction which may have catalyzed the reaction, or it could be related to the I1reaction chamber1! characteristics of the Kuderna- Danish concentration step.

In the Smith A CLLE samples and other CLLE samples, additional solvent artifacts were identified. These artifacts were related to the extraction solvent (methylene chloride) and not acetone. The artifacts are a cyclohexenol isomer, 2- cyclohexenone and cyclohexanone. These SOCs are present at significant levels, dominating the chromatograms at concentrations that would be equivalent to 10-20 ug/l in the original water samples. These artifacts originate from cyclohexene, which is a preservative in methylene chloride. In the 250 ml of methylene chloride used for extraction there is about 4000 ug of cyclohexene preservative. During extraction and concentration, some of the cyclohexene artifacts are formed. For chlorinated water samples, chlorinated artifacts from the cyclohexene are often formed. These observations led to some experiments in which the artifact formation theories were verified. From chlorinated water samples, for which methylene chloride was used as the extraction solvent, large amounts of chlorinated cyclohexanol and chlorinated cyclohexane were seen in the final analysis. When the same water was dechlorinated before extraction, the chlorinated cyclohexene artifacts were not seen. When the same water was extracted with diethyl ether (which has no cyclohexene preservative), none of the chlorinated or nonchlorinated cyclohexene related artifacts were found.

4-3-D. An In-Depth Survey of the Smith River.

The quantity and diversity of the SCCs found in the Smith River during the preliminary study made it the most suitable for further study. The Deep River site was too heavily contaminated. by wastewater related constituents, and the Yadkin River site was too clean. The results of the SOC analysis of the Water Treatment Plant influent for each of the three sampling dates are presented in Tables 4-15 and 4-16. The results of the SOC analysis of the four sources analyzed on the last sampling date are presented in Tables 4-17 and 4-18.

The source water (collected about 1500 feet downstream of Philpott dam) was very clean, with no SOCs identified by CLLE and only 6 identified by CLSA at very low concentrations. The upstream sample (collected about 3 miles downstream from Martinsville, near the Virginia-North Carolina border) was unfortunately sampled 12 hours after the supposed peak contamination concentration had passed that point, but nevertheless, a fairly large number of SOCs were identified at elevated concentrations. The CLSA analysis showed 7 SOCs, all at

T a b l e 4 - 1 7

S O C ANALYSIS OF S\ . l ITH R I V E R WATER BY CLLE - F O U R SAXPLING POINTS

A p p r o x i a a t e C o n c e n t r a t i o n ( u g / l ) e:

S o u r c e U p s t r e e n I n f l u e n t F i n i s h e d S O C

T e t r x h l o r o e t h e n e

S y l e n e ( i s o m e r s ) C 3 - b e n z e n e ( t r i o r e t h y l -

m e t h y l i s o m e r )

S-(ethyl)-8-(pheny1)-acetamid 1,3,3-Trimethyl-2-indolinone T r i a z i n e p e s t i c i d e ( n o t a t r a z i n e ) 3 , 3 , 3 - T r i c h l o r o p r o p e n e

&-!I2 t h y l - 3 - p e n t e a - 2 - o n e X-G

D i x e t o n e a l c o h o l +x-

E u t y l a t e d h y d r o x y t o l u e n e ( 3 H T ) )(

D i m e t h y l p h t h a l a t e * D i e t h y l p h t h a l a t e E-

y e s 11

I t

t 1

1 t

t t

y e s ' 11

1 t

1 I

1 t

t 1

D i b u t y l p h t h a l a t e %- 11

D i e t h y l h e x y l p h t h a l a t e * 1 1

O t h e r u n i d e n t i f i e d p h t h a l a t e * I I

*S .OCs a l s o p r e s e n t i n b l a n k s i n v a r y i n g l e v e l s " * S O C s o r i g i n a t i n g f r o m t h e a c e t o n e ( s p i k e s o l v e n t ) "j:-j%OCs t h a t a r e e x t r a c t i o n s o l v e n t a r t i f a c t s S z n p l i n g d a t e : 6 / 1 0 / 8 5

Table 4-18 soc AXILYSIS OF SNTH RIVEX WATER BY CLSA -

FOUR SAHPLING POINTS

Approximate Concentration ( u g / l )

Source U ~ s t r e a m Influent Finished

Bromodichloromethane Dibromochloromethane Dichloroiodonethane Tetrachloroethene Napthalene Toluene Xyle2.e (isomers) C3-benzene (tri or ethyl-

methyl isomer) Trichlorobenzene (Lscner) Benzonitrile Benzaldehyde Chlorodimethyl benzece

(isomers) Dichlorobenzene (isomer) Hexachloropentadiene

Unidentified X Unidentified C Unidentified D Unidentified G Unidentified H

Sampling date: 6/10/85

fairly low concentration levels. The CLLE analysis of this upstream sa ug/l) of a atrazine) , indolinone

.mple sho xylene i N< ethyl) and Unkn

wed relatively somer, a C3-ben -N (phenyl) -acet own Eo The lat

I1highV1 conc zene, a tri amide, 1,3, ter three S

entrations azine (not 3-trimethy OCs turned

1-29 out to

significant pollutants in all the Smith River analyses, except the sample from the river source. These concentration levels may not necessarily be considered high and are substantially below

-

the levels thus far discussed as possible future maximum contaminant levels for the non-cakinoqenic SOCs. Concentrations of 2 1 ug/l were higher than the levels of most other SOCs, and therefore, this concentration is designated as llhigh.l'

In the plant influent (Tables 4-15 and 4-16) , a total of 10 SOCs were identified by CLSA and 11 SOCs by CLLE (disregarding the acetone and methylene chloride artifacts). Unidentified C, Unidentified D, and xylene were the most significant pollutants identified by CLSA at concentrations up to 1 ug/l. By CLLE, the most significant pollutants were the three reoccurring major Smith River pollutants (with 1,3,3-trimethyl-2-indolinone having the highest concentration at up to 8 ug/l), and xylene and Cg- benzene were in the 1.5-8.0 ug/l range. Another unknown, Unidentified F, was present at 1.5-2.0 ug/l on two of the three sampling dates in the plant influent.

The finished treated water showed significant and unexpectedly high levels of SOC pollution. Figures 4-18 and 4-19 show typical GC chromatograms of CLLE and CLSA analysis of Fieldcrest Water Treatment Plant finished water. The CLSA analysis identified 18 SOCs, mostly at low concentrations. Three trihalomethanes were identified at 0.2-0.3 ug/l levels, and 8 other SOCs were found at low levels (0.05-0.15 ug/l). At higher concentrations, toluene (0.4 ug/l), Unidentified G (0.5 ug/l), Unidentified H (1.0 ug/l), and a xylene (1 ug/l) were found. Two other xylene isomers were found at high levels by CLSA (4.0 and 20 ug.1 respectively), which incidentally, is a 40 fold increaae for all three xylenes over their concentration in the plant influent. By CLLE, 7 SOCs were identified in the finished water (other than the solvent artifacts). This included the three major Smith River contaminants seen by CLLE (at 4-6 ug/l), 3,3,3- trichloropropene (12 ug/l), a C3-benzene (1.5 ug/l), 1,1,2,2- tetrachloroethane (1 ug/l), and two xylenes (1 and 4 ug/l).

When comparing the plant influent and finished water, it is evident that the treatment process does not adequately remove the SOCs present. Additionally, trihalomethanes (THMs) were present in the finished waste and not in the influent. These THMs are most likely disinfection byproducts formed during the water treatment process. For a few SOCs and for xylene in particular, the concentration is higher in the effluent than in the influent. 3,3,3-trichloropropene is present at a high concentration (12 ug/l) in the finished water but not at all in any other sample.

These observations are not easy to explain. It may be due to grease or chemicals used in the water treatment process or to some step in the treatment process which for some reason caused a surge of these SOCs to be released from the sludge or carbonaceous filter bed which coincided with the sampling time. Another reason could be that the influent was poorly mixed, resulting in sampling error.

The major recurring S O C s of concern at the Fieldcrest Water Plant appear to be Unidentified E, N (ethyl) -N (phenyl) acetamide, 1,3,3-trimethyl-2-indolinone, and xylenes. These were identified in all samples and each at concentrations over 2.5 ug/l. To a lesser degree, Cj-benzenes and Unidentified F are contaminants of concern, and the significant amount of 3,3,3-trichloropentene which was found in the finished water may be of importance. A large number of other SOCs were found less frequently and at lower concentrations.

In summary, the water from just below the Philpott Dam had very little SOC contamination. The sample taken downstream of the dam and below the majority of the industrial outfalls (but above Eden) had more S O C s and higher concentrations than at the dam. It is important to remember that this sample was not taken at the supposed peak of contamination as were the treatment plant samples, which may explain its lower SOC content. The water treatment plant influent which is below the industrial outfalls had more S O C contamination than the upstream samples, and the plant finished water had the highest SOC contamination of all.

It is not possible at this point to conclude what contaminant causes the water quality problems at Eden. Many dyes and pigments, which could cause the discoloration of the water, are nonorganic, and others are organic but not chromatographable by GC/FID. The problem may not be due to a dye or pigment, and it could easily be a combination of compounds that is responsible for the problems in Eden. Some significant SOC contamination was. observed, and this problem warrants further investigation (See footnote, Section 3-6-A. ) . 4-3-E. Comparison of CLLE/CLSA to EPA Recommended Methods.

In order to compare the methods developed to those methods recommended by the EPA, waters containing SOCs were analyzed with both sets of methods. Analysis by the E P A methods was performed by a qualified laboratory, and the CLLE/CLSA analyses were performed in the Department of Environmental Sciences and Engineering on the same water samples.

Based on the percent recoveries previously calculated from the water matrix studies and the relative responses obtained. the spiked SOCs and other identified SOCs were identified and

quantified by CLLE/CLSA, as if all were originally present in the sample. The results in Tables 4-19 and 4-20 show the SOCs found and-their approximate amounts for both the CLSA/CLLE and Methods 624 and 625. The SOCs identified and quantified by CLSA/CLLE are the same as reported in 4-3-D, except for the spiked SOC standards.

For the water treatment plant influent, CLSA underestimated the amount of SOC standards present by 10-20 percent. CLLE underestimated some and overestimated other SOC standards, generally within 20 percent of the amount actually spiked. Atrazine, however, was underestimated by about 33 percent, and the phenols were not detected at all.

The finished water was more of a problem for the analysis of the SOC standards. The large numbers and high concentrations of SOCs present appeared to have affected the identification and recovery of the's~c interferences and p matrix effects. Wi

standard robably t th CLSA,

S, due o compe two of

in part to chroma titive binding an the senerally we1

tographic d other 1 recover

SOCs (tetrachloroethylene and 1,1,2,2-tetrachloroethane) had interference from other SOCs in the sample, and 1,1,1- trichloroethane and trichloroethylene were not recovered in the analysis. 1,2-dibromoethane, naphthalene, and 1,4- dichlorobenzene were identified and quantified well. By CLLE, 1,2,2-tetrachloroethane and the phenols had interference from other pollutants, but the rest of the SOC standards that could be expected to be analyzed for by CLLE were. Underestimation of the amount of these S O C ~ was the pule, generally by 10-30 percent of the amount actually there.

For the Fieldcrest Water Treatment Plant influent analyzed for volatiles, the CLSA method identified and quantified 10 SOCs in addition to all 7 of the SOCs in the SOC spike to which it was applicable. EPA's Method, Method 624, did not detect the spikes and only identified acetone (the solvent in which the spikes were. dissolved). For the same water source, the CLLE method identified and quantified 6 SOCs (other than solvent artifacts) in addition to 8 of the 10 in the SOC sp ike to which it was applicable. The EPA's recommended method, Method 625, did not detect any SOCs in the same water sample.

For the Fieldcrest Water Treatment Plant finished water, the CLSA method identified and quantified 18 SOCs in addition to 3 out of the 7 SOC standards it is generally applicable to. EPA1s method identified only acetone and chloroform and did not quantify them. For the same water source, the CLLE method identified and quantified 8 SOCs (other than solvent artifacts and general lab contaminants) in addition to 7 of the 10 SOC standards to which the method is applicable. The EPAis recommended method did not detect any SOCs in the same water sample.

Table 4 -19

SOC ANALYSIS O F S N I T H R I V E R WATER - A CO>lPARISON O F CLSA TO EPA METHOD 6 2 4

Approximate Concentration (ug/l) -

SOC Standards

l,l,l-Trichloroethane Trichloroethene 1,2-Dibromoethane (EDE) Tetrachloroethene 1,1,2,LTetrachloroethane phenol 2-Chlorophenol Saphthalene

Other SOCs Identified

In£ luent Finished CLSA - 624 CLSA 6 2 4 -

Acetone Trichloronethane (chlorofora) Tetrachloroethene Bromodichloromethane Dibromochloromethane D i c h l o r o i o d o ~ e t h a n e Toluene 0.2 Xylene .05,0.1,0.5 C3-benzene (tri or ethyl-methyl

isomers) Trichloro benzene (isomer) Benzonitrile Benzaldehyde ~ h l o r o d i m e t h ~ l b e n z e n e Dichloro benzene (isomers) Hexachloropentadiene Unidentified A Unidentified C Unidentified G Unidentified H

Influent was spiked with 5 0 0 ng/l SOC std. Finished water was spiked with 1 2 5 ng/l SOC std. --- indicates too much noise o n chromatogram for identification S D = Xot detected Fieldcrest Vater Treacnent Plant, 6/10/35

Table 4 - 2 0

S O C A N A L Y S I S OF S > f I T E RIVER WATER - A COMTARISOX OF C L L E TO EPA METHOD 6 2 5

Approximate Concentration (ug/l)

SCC Standards

l,l,l-Trichloroechane * Trichloroethene 1,2-Dibromoeth~ine (E2B) Tetrachloroethene * 1,1,2,2-Tetrachloroethane Phenol 2-Chlorophenol Naphthalene 1,L-Dichlorobenzene Atrazine T r . i s ( c h l o r o p r o p y l ) p h o s p h a t e

Other SOCs Identified

Influent CLLE 625 -. -

Tetrachloroethene Sylene 2.5 C3-benzene (tri or ethyl-methyl

isomer) 1.6 Unidentified E 2.5 Unidentified F 1.5 N-ethyl-N-zethyl-acetnid 4 . 0 1,?,3-Trimethyl-2-indolinone 6 . 0 3,3,3-Trichlorpropene

Dichlorocyclohexane %*

Cyclohexenol (isomer) -E?:

2-Cyclohexenone ** Cyclohexanone *-x 4-methyl-3-penten-2-one *** Diacetone alcohol i'i'3:- .

Finished C L L E 625 -

-- 1.7 1.4

noisy I I

S O C std. spikes: 5 u g / l in influent, 2 ug/l in finished uater. -2SOCs also present in blank to a large extent *<-SOCs that are extraction solvent artifacts

.*-3QSOCs originating from the acetone (SOC spike solvent) <- 2'" '< - - --xLaboratory contaminants also present in blanks

. Fieldcrest Iv'ater Treatment Plant, 6/10/55

It was chloroform. CLSA method chlorinated the GC run.

disappo One re (it was water)

inting that the CLSA failed to analyze for ason why chloroform was not detected by the undoubEedly present in this highly

-

could be that it coeluted with the solvent EPA'S method has an advantage over CLSA for a few

very volatile SOCs, such as chloroform, in that it used direct thermal desorption onto the GC and no solvent which can interfere in the GC separation and identification. However, a lower initial GC temperature might separate the solventTand chloroform in the CLSA runs. Another reason for not detectina chloroform could be that the low molecular weiaht volatiles (Such as chlorof o m ) compete sites on the carbon temperature for the

poorly with larger molecules *or binding trap or that the trap was kept at too high highly volatile chloroform to adsorb well.

It must be pointed out that the Quality Assurance/Quality Control (QA/QC) procedures of the laboratory which performed the analysis using the EPA methods were not made available. Without such information, the com~arison of the CLSA/CLLE and EPA methods cannot be validated satis?actorily. For instance, it may be that the laboratory automatically reports and quantifies only findings above a certain concentration or looks only for certain SOCs or simply performed poor analysis. The latter is most likely not the case, but without QA/QC information, one cannot conclude whether the lack of SOC detection is due to the EPA method itself or the manner in which the laboratory carried out the analysis.

REFERENCES

Ahnoff, M., Josefsson, B., I1Simple Apparatus for On-Site Continuous Liquid-Liquid Extraction of Organic Compounds from Natural Waters,I1 Analytical Chemistry, 46: 658-663, 1974.

American Water Works Association, "Participant Materials for AWWA Research Foundation Workshop on VOCS,'~ AWWA Research Foundation, Denver, CO, 1982.

~merican Water Works Association, Itparticipant Materials for Workshops on Revised Regulations,ll AWWA Research Foundation, Denver, CO, 1983.

Bellar, T.A., Lichtenberg, U.J., llDetermining Volatile Organics at Microgram-per-Litre Levels by Gas Chromatography,I1 Journal of the American Water Works Association, 66:. 739- 744, 1974.

Boland, P.A., Kingsley, B.A., Stivers, D.F., Pomerantz, I.H., lvProtocol for the Analysis of a Broad Range of Specific Organic Compounds in Drinking Water," in Advances in the Identification and Analysis of Orqanic Pollutants Water, Vol. 2, Keith, L.H., Ed., Ann Arbor Science, Ann Arbor, Michigan, pp. 831-838, 1981.

Bornbaugh, K.J., "The Use of HPLC for Water ~nalysis,~~ in Water ~nalysis, Volume 111, Orqanic Species, Minear, R.A., and ~eith, L.H., Eds., Academic Press, Inc., New York, pp. 317- 381, 1984.

Boren, Ha, Grimvall, A., Savenhed, R., nModified stripping ~echnique for the Analysis of Trace Organics in Water," Journal of Chromatoqraphy, 252: 139-146, 1982. 4

\

Bruchet, A., Cognet, L., Mallevialle, J., I1Continuous Composite sampling and Analysis of Pesticides in Water," Water Research, 18: 1401-1409, 1984.

Burnham, A.K., Calder, G.V., Fritz, J.S., Junk, G.A., Svec, H.J., Willis, R., Vdentification and Estimation of Neutral organic contaminants in Potable Water," Analytical chemistry, 44 : 139-142, 1972.

Cato n, J. E. , Barnes, M.P., Vomparis Quantification Aqueous Media. ~nalysis L.H., Ed Arbor, M

PP. chigan

Z.K., Kubota, H., ~riest, W.H., Maskarin on of Several Methods for Collection and of Polycyclic Aromatic Hydrocarbons from in Advances in the Identification and

anic Pollutants in Water, Vol. I, Keith, 329-344, Ann Arbor Science Publishers, An , 1981.

~hriswell, C.D., Chang, R.C., Fritz, J.S., Vhromatographic ~etermination of Phenols in Water," Analytical Chemistry, 47: 1325-1329, 1975

Colby, B.N., Rosecrance, A.E., I1Determination of Priority Pollutants in Industrial Wastewaters by Stable Isotope Dilution Gas Chromatography/Mass spectrometry," in ~avances in the ~dentification and Analysis of Organic Pollutants -- Water, Vol. I, Keith, L.H., Ed., pp. 221-230, Ann Arbor Science Publishers, Ann Arbor, Michigan, 1981.

Colby, B.N., Rosecrance. A.E.. 'IMeasurement Parameter Selection -for ~uantitative soto ope-~ilution Gas Chromatography/Mass Spectrometry, Analytical Chemistry, 53 : l907-l9ll, 1981.

Coleman, W.E., Allen, W.K., Slater, R.W., Voto, S.J., Melton, R. G. , Kopf ler, F. C., Aurand, T.A. , vlAutomatic Quantification and Statistical Evaluation of Organic Contaminants Using a computerized Glass Capillary Gas Chromatography/Mass Spectrometry System and rob Closed-Loop Stripping," in - Advances in the Identification and Analysis of Orqanic Pollutants Water, Vol. 2, pp. 675-688, Keith, L.H., Ed., Ann Arbor Science, Ann Arbor, Michigan, 1981.

Coleman, W.E., Melton, R.G., Slater, R.W., Kopfler, F.C., Voto, J., Allen, W.K., Aurand, T.A., "Determination of Organic contaminants by the c rob-Loop-Stripping Technique," Journal of the American Water Works Association, 73: 119-121, 1981. --

Coleman, W.E., Munch, J.W., Slater, R.W., Melton, R.W., Kopfler, F.C., vlOptirnization of Purging Efficiency and Quantification of Organic Contaminants from Water Using a 1-L Closed-Loop- Stripping Apparatus and Computerized Capillary Column GC/MS,I1 Environmental Science and Technoloqy, 17: 571-576, 1983.

a .

Colgrove, S.G. Svec, H.J., "Liquid-Liquid Fractionation of Complex Mixtures of Organic Cornp~nents,!~ Analytical Chemistry, 53: 17374742, 1981.

Commoner, B., Vhernical Carcinogens in the ~nvironment,~~ in Identification and Analysis of Oraanic Pollutants Water, Keith, L.E., Ed., pp. 49-72, Ann Arbor Science Publishers, Ann Arbor, MI, 1977.

Cooper, R.L., Wheatstone, LC., "The Determination of Phenols in Aqueous effluent^,^^ Water Research, 7: 137501384, 1973.

Cutugno, D., llEvaluation of the Closed-Loop Stripping Technique for Concentration of Trace Volatile Organics in Water and its Application to the Haw River," Master's Thesis, University of North Carolina, Chapel Hill, NC, 1984.

~einzer, M., Melton, R., Mitchell, D., "Trace Organic contaminants in Drinking Water: Their Concentration by Reverse ~smosis," Water Research, 9: 799-805, 1975.

DeMarco, M.L., Stevens, A.A., Hartman, D.J., lfApplication of Organic Analysis for Evaluation of Granular Activated Carbon Performance in Drinking Water Treatment," in Advances in the ~dentification and Analysis of 0rsanic Pollutants Water, Vol. 2, Keith, L.H., Ed., Ann Arbor Science, Ann Arbor, ~ichigan, pp. 907-940, 1981.

DeRosa, P.M., lVPotential for Chemical Contamination of Community Surface Water Supplies in North Carolina," Master's Technical Report, University of North Carolina, Chapel Hill, NC, 1984.

Dietrich, A.M., Millington, D.S., Christman, R.F., "Specific Identification of Organic Pollutants in Haw River Water Using Gas Chromatography/Mass Spectrometry," Water Resources Research Institute, Publication No. UNC-WRRI-83-206, North Carolina, 1983.

Dressler, M., "Extraction of Organic Contaminants from Water with Porous Organic PolymersfN Sournal of Chromatographv, 165: 167-206, 1979.

Dunlap, W.J., Shew, D.C., Scalf, M.R., Cosby, R.L., Robertson, J.M., "Isolation and Identification of Organic Contaminants in Ground Water," in Identification and ~nalysis of Orsanic Pollutants Water, Keith, L.E., Ed., pp. 453-478, Ann Arbor Science Publishers, Ann Arbor, MI, 1977.

~llington, J., "Analysis of Volatile Organics on Sediments and in Associated Water," in Advances in the Identification and Analvsis of Orsanic Pollutants Water, Vol. 2, pp. 729- 746, Keith, L.H., Ed., Ann Arbor Science, Ann Arbor, a - .

Michigan, 1981.

Fox, K.R., Miltner, R.J., Logsdon, G.S., Dicks, D.L., Drolet, L.F., "Pilot Plant Studies of Slow Rate Filtrationftl Journal of the American Water Works Association, 76: 62-68, 1984. --

Garrison, A.W., Alford, A.L., Craig, J.S., Ellington, J.J., Haeberer, J.M., McGuire, J.M., Pope, J.D., Shackelford, W.M., T h e Master Analytical Scheme: An Overview of Interim Procedures," in Advances in the Identification and Analvsis of Orsanic Pollutants & Water, Vol. I, Keith, L.H., Ed., - pp. 17-30, Ann Arbor Science Publishers, Ann Arbor, Michigan, 1981.

Gebhart, J.E., Ryan, J.F., Cox, R.D., Pellizzari, E.D., Michael, L.C., Sheldon, L.S., "The Master Analytical Scheme: Development of Effective Techniques for Isolation and concentration of Organics in Water," in Advances in the ~dentification and Analysis of Orqanic Pollutants b Water, Vol. I, Keith, L.H., Ed., pp. 31-48, Ann Arbor Science Publishers, Ann Arbor, Michigan, 1981.

Glaze, W.H., Henderson, J.E., Smith, G., nAnalysis of New Chlorinated Organics Compounds in Municipal Wastewaters after Terminal ChlorinationfN in Identification and Analysis of Orsanic Pollutants Water, L.E., Ed., pp. 247-254, Ann z b o r Science Publishers, Ann Arbor, Michigan, 1977.

Glaze, W.H., Jones, P.C., Saleh, F.W., Vize Exclusion, Reverse Phase and Weak Anion Exchange Chromatography of Natural Organics in Waterfff in Advances in the Identification and ~nalvsis of Orqanic Pollutants Water, Vol. I, Keith, L.H., Ed., pp. 371-382, Ann Arbor Science Publishers, Ann Arbor, Michigan, 1981.

Glaze, W.H., Rawley, R., Burleson, J.L., Mapel, D., Scott, D.R., "Further Optimization of the Pentane Liquid-Liquid Extraction Method for the Analysis of Trace Organic Compounds in Water," in Advances in the Identification and Analysis of Orsanic Pollutants Water, vol. I, Keith, L.H., Ed., pp. 267-280, Ann Arbor Science Publishers, Ann Arbor, Michigan, 1981.

Grob, K., nOrganic Substances in Potable Water and in its Precursor, Part I., Methods for Their Determination by Gas- Liquid Chromatography,ll Journal of Chromatosra~hv, 84: 255- 273, 1973.

Grob, K., "Organic Substances in Potable Water and in its Precursor, Part 11, ~pplications in the Area of Zurich," - Journal of Chromatosraphv, 90: 303-313, 1974.

Grob, K., "Organic Substances in Potable Water and in its Precursor, Part 111, The Closed-Loop Stripping Procedure Compared with Rapid Liquid Extraction,li Journal of Chromatoqraphy, 106: 297-315, 1975.

Grob, K., Grob, G., "Glass Capillary Gas Chromatography in Water Analysis: How to Initiate Use of the Method," in Identification and Analvsis Orqanic Pollutants i~ Water, Keith, L.E., ~d.,~~. 75-86, Ann Arbor Science Publishers, Ann Arbor, Michigan, 1977.

Grob, K., Zurcher, F., %tripping of Trace Organic Substances from Water Equipment and Procedureftf Journal of Chromatosraphv, 117: 285-294, 1976.

Gurka, D.F., Warner, J.S., Ynterlaboratory Comparison Study: Methods for Volatile and Semivolatile CompoundsfW U.S. E.P.A. Grant Report No. 68-03-3098, 1983.

Haeberer, A.F., Scott, T.A., "High Performance Liquid Chromatography Analysis of ~itrophenols,~ in~dvances in the Identification and Analysis of Orqanic Pollutants Water, Vol. I, Keith, L.H., Ed., pp. 359-370, Ann Arbor Science Publishers, Ann Arbor, Michigan, 1981.

Haile, C.L., Shan, Y.A., Malone, L.S., Northcutt, R.V., "Development of Methods for the Analysis of Purgeable organic Priority Pollutants in Municipal and Industrial Wastewater Treatment sludge^,'^ in Advances in the Identification and Analysis of Orsanic Pollutants Water, Vol. 2, Keith, L.H., Ed., Ann Arbor Science, Ann Arbor, Michigan, pp. 763-792, 1981.

Hanson, D., f'Multimedia Approach to Pollution Control Urged," Chemical and Engineerinq News, pp. 26-27, Nov. 26, 1984.

Hanson, D., "Stricter Drinking Water Bills Passed,' Chemical and Enqineerinq News, p. 19, July 1, 1985.

Headington, C.E., ItMethods for Determining Hydrocarbons and Phenols in Water," Analytical Chemistrv, 25: 1681-1685, 1953.

Henderson, J.E., Peyton, G.R., Glaze, W.H., "A Convenient Liquid- Liquid Extraction Method for the Determination of Halomethanes in Water at the Parts-Per-Billion Level/' in Identification and Analysis of Organic Pollutants Water, Keith, L.E., Ed., pp. 105-112, Ann Arbor Science Publishers. - - - Ann Arbor, Michigan, 1977.

Jolley, R.L., Jones, Jr., Guy, Pitt, Jr., W.W., Thompson, J.E., - . tf~etermination of Chlorination Effects on Organic Constituents in Natural and Processed Waters Using High Pressure ~iquid Chromatographyfl' in Identification and Analysis of Orqanic Pollutants Water, Keith, L.E., Ed., pp. 233-246, Ann Arbor Science Publishers, Ann Arbor, Michigan, 1977.

Junk, G.A., Ogawa, I., Svec, H.J., "Extraction of Organic Compounds from Water Using Small Amounts of Solvent, in Advances in the Identification and Analysis of Orsanic Pollutants Water, Vol. I, Keith, L.H., Ed.! pp. 281-294, Ann Arbor Science Publishers, Ann Arbor, Michigan, 1981.

Junk, G.A., Richard, J.J., "Anionic and Neutral Organic Components in Water by Anion Exchange," in Advances in the identification and Analysis af Orqanic Pollutants h Water, Vol. I, Keith, L.H., Ed., pp. 295-316, Ann Arbor Science Publishers, Ann Arbor, Michigan, 1981.

Junk, G.A., Richard, J.J., Fritz, J.S., Svec, H.J., IfResin sorption Methods for Monitoring Selected Contaminants in Water," in Identification and Analysis a Organic Pollutants in Water, Keith, L.E., Ed., pp. 135-154, Ann Arbor Science - Publishers, Ann Arbor, Michigan, 1977.

Kahn, L., Wayman, C.H., IfApparatus for Continuous Extraction of Nonpolar Compounds from Water Applied to Determination of chlorinated Pesticides and IntermediatesfVt Analvtical chemistry, 36: 1340-1343, 1964.

Kawahara, F.K., ItInirared Spectrophotometry of Pollutants in Water Systems," in Water Analysis, Volume 111, Orqanic Species, Minear, R.A., and Keith, L.H., Eds., Academic Press, Inc., New York, pp. 382-444, 1984.

Keith, L.H., nOrganic Pollutants in Water: Identification and Analy~is,~~ Environmental Science and Technolosy, 15: 156- 162, 1981.

Krasner, S.W., Hwang, C.J., McGuire, M.J., "Development of a Closed-Loop Stripping Technique for the Analysis of Taste- and Odor-causing Substances in Drinking Waterftl in Advances in the Identification and Analysis of Orqanic Pollutants a -- Water, Vol. 2, pp. 689-711, Keith, L.H., Ed., Ann Arbor Science, Ann Arbor, Michigan, 1981.

Landrum, P.F., Giesy, J.P., wAnomalous Breakthrough of Benzopyrene During Concentration with Amberlite XAD-4 Resin from Aqueous Soluti~n,~~ in Advances in the Identification, and Analysis a Orqanic Pollutants a Water, Vol. I, Keith, - - L.H., Ed., pp. 345-355, Ann Arbor Science, Ann Arbor, Michigan, 19 8 1.

Leenheer, J.A., Voncentration, Partitioning, and Isolation Techniquesftt in Water Analysis, Volume 111, Orqanic Species, Minear, R.A., and Keith, L.H., Eds., Academic Press, New York, pp. 84-166, 1984.

Lin, D.C.K., Melton, R.G., Kopfler, F.C., Lucas, S.V., flGlass capillary Gas ~hromatographic/~ass Spectrometric ~nalysis of Organic Concentrates from Drinking and Advanced Waste Treatment Waters,I9 in Advances in the Identification and Analysis a Orqanic Pollutants h Water, Vol. 2, Keith, L.H., Ed., pp. 861-906, Ann Arbor Science, Ann Arbor, Michigan, 1981.

~opez-Avila, V., Haile, C.L., Goddard, P.R., Malone, L.S., Northcutt, R.V., Rose, D.R., Robson, R.L., "Development of Methods for the Analysis of Extractable Organic Priority Pollutants in Municipal and Industrial Wastewater Treatment Sludgesfft in Advances in the Identification and Analysis of Orsanic Pollutants Water, Vol. 2, Keith, L.H., Ed., pp. 793-828, Ann Arbor Science, Ann Arbor, Michigan, 1981.

March, J., Advanced Orsanic Chemistry, Reactions, Mechanisms and Structures, McGraw Hill Publishing, New York, p. 692, 1968.

McCarty, P.L., Argo, D., Reinhard, M., ffOperational Experience with Activated Carbon Adsorbers at Water Factory 21fVf Journal of the American Water Works Association, 71: 683- 689, 1979.

McGuire, M.J., Krasner, S.W., Hwang, C.J., Izaguirre, G., vfClosed-Loop Stripping Analysis as a Tool for Solving Taste and Odor Problemsfff Journal of the American Water Works Association, 73: 530-537, 1981.

Melton, R.G., Coleman, W.E., Slater, R.W., Kopfler, F.C., Allen, W.K., Aurand, T.A., Mitchell, D.E., Voto, S.J., Lucas, S.V., Watson, S.C., Vomparison of Grob Closed-Loop Stripping Analysis with Other Trace Organic Methods,If in Advances the ~dentification and ~nalysis of ~rsanic Pollutants &J - Water, Vol. 2, pp. 597-674, Keith, L.H., Ed., Ann Arbor science, Ann Arbor, Michigan, 1981.

Mieure, J.P., Mappes, G.W., Tucker, E.S., %eparation of Trace Organic Compounds from WaterffV in Identification and Analysis of Orsanic Pollutants Water, Keith, L.E., Ed., pp. 113-134, Ann Arbor science Publishers, Ann Arbor, Michigan, 1977.

Munch, D.J., Munch, J.W., Feige, M.A., lick, E.M., Brass, H.J,., IfA Scheme for the Routine Analysis of Purgeable Compounds by. Gas Chromatography/Mass SpectrometryfW in Identification and Analysis of Orsanic Pollutants Water, Keith, L.E., Ed., pp. 113-134, Ann Arbor Science Publishers, Ann Arbor, Michigan, 1977.

Musty, P.R., Nickless, G., "Use of Amberlite XAD-4 for Extraction and Recovery of Chlorinated Insecticides and Polycarbonated Biphenyls from Water," Journal of Chromatosraphv, 89: 185- 190, 1974.

Osterroht, C., ffDevelopment of a Method for the ~xtraction and Determination of Non-polar, Dissolved Organic Substances in Sea Waterfff Journal of Chromatoara~hy, 101: 289-298, 1974.

rl G ofdid U G U id k-d k 3 k E O Q , = h2 -=

G 4 J na) E i w m

-CI Olp. or6 cn m a C d k k O d B - 4 4

m m -a)-dm 4 J 3 Q

C m - r l I a a ~n -5 r-

44 . 0 r i d C d r l -a 4J-4 Q ) 3 C 3 U 0 a)-d

xEU aa: o w

k 01

. m CO

k or O d d ccr d

U 4

rn-4 u ' a G Z 0 A G U - C, O r l O H r l

c - 0 k d

-rl Q) Q) 4J4J PI r d r n r d

p e a = 4

0 4111

Schwarzenbach, R.P., Bromund, R.H., Gschwend, P.M., Zafirou, O.C., Volatile Organic Compounds in Coastal Sea~ater,~~ Aquatic Geochemistrv, 1:93-107, 1978.

Schwarzenbach, R.P., Giger, W., Grob, K., "Gas Chromatography," in Water Analysis, Volume 111, Orsanic Species, Minear, RoAe, and Keith, L.H., Eds., Academic Press, Inc., New York, pp. 167-253, 1984.

Schwarzenbach, R.P., Molnar-Xubica, Em, Giger, W. Wakeham, S.G., lVDistribution, Residence Time, and Fluxes of Tetrachloroethylene and 1,4=Dichlorobenzene in Lake Zurich, Switzerland," Environmental Science and Technoloqy, 13: 1367-1373, 1979.

Semmens, M.J., Ayers, K., "Removal by Coagulation of Trace Organics from Mississippi River Water," Journal of the American Water Works Association, 77: 79-84, 1985.

Sorrell, R.K., Daly, E.M., Weisner, MOJO, Brass, H O G . , "In-Home Treatment Methods for Removing Volatile Organic Chemicals," Journal of the American Water Works Association, 77: 72-78, 1985.

~praggins, R.L., Oldham, R.G., Prescott, C.L., Baughman, K.J., "Organic Analyses Using High-Temperature Purge and Trap ~echniques,~~ in Advances in the Identification and Analysis of 0rsanic Pollutants Water, Vol. 2, pp. 747-762, Keith, - L.H., Ed., Ann Arbor Science, Ann Arbor, Michigan, 1981.

Stachel, B e , Baetjer, KO, Cetinkaya, M., ~ueszein, J., "On site continuous ~iquid-Liquid Extraction of Nonpolar Organic Compounds in Water," Analytical Chemistry, 53: 1469-1472, 1981.

Stepan, S.F., Smith, J.F., %ome Conditions for the Use of a .

Macroreticular Resins in the Quantitative Analysis of organic Pollutants in Water,I1 Water Research, 11: 339-342, 1977.

Suffet, I.H., Radziul, J.V., fiGuidelines for the Quantitative and Qualitative Screening of Organic Pollutants in Water Supplies, "Journal of the American Water Works Association, 68: 520-524, 1976.

Thomas, Q.V., Stork, J.R., Lammert, S.L., "The Chromatographic and GC/MS Analysis of Organic Priority Pollutants in Water," Journal of Chromatoqra~hy Science, 18: 583-593, 1980.

Thrun, K.E., Oberholtzer, J.E., I1Evaluation of the ~icroextraction Technique to Analyze Organics in Water," in Advances in the Identification and Analysis of Orsanic Pollutants Water, Vol. I, Keith, L.H., Ed., pp. 253-266, Ann Arbor Science Publishers, Ann Arbor, Michigan, 1981.

Trussell, A.R., Lieu, F.Y., Moncur, J.C., "Part Per Trillion ~nalysis of Volatile, Base/Neutral and Acidic Water contaminants on a Single Fused Silica Capillary column,^ in Advances in the Identification and Analysis of Organic Pollutants Water, Vol. I, Keith, L.H., Ed., pp. 171-186, Ann Arbor Science publishers, Ann Arbor, Michigan, 1981.

U.S. E.P.A.! aGuidelines Establishing Test Procedures for the Analysls of Pollutants; Proposed regulation^,^^ Federal Resister, Vol. 44: 69464-69575, Dec. 3, 1979.

U.S. E.P.A., nGuidelines Establishing Test Procedures for the Analysis of Pollutants Under the Clean Water Act; Final Rule,I1 Federal Reqister, Vol. 49: 43233943442, Oct. 26, 1984.

U.S. E.P.A., I1National Interim Primary Drinking Water Regulationsf1I Publication No. EPA-570/9-76-003, 1976.

Van Rossum, P., I1Isolation of Organic Water Pollutants by XAD ~esins and Carbon, Journal @ Chromatoqraphy, 150: 381-392, 1978.

Veith, G.D., Kiwus, L.M., I1An Exhaustive Steam-Distillation and Solvent-Extraction Unit for Pesticides and Industrial chemical^,^^ Bulletin Environmental Contamination and Toxicoloc~, 17: 631-636, 1977.

Verschueren, K., Handbook of Environmental Data on Orqanic chemicals, Van Nostrand Reinhold Co., New York, 1977.

Walton, H.F., Eiceman, G.A., aTrace Organic Analysis of Wastewater by Liquid Chr~matography,~~ Proceedings of the 9th ater rials Research Symposium, April 10-14, Maryland, 1978.

Watts, C.D., Crathorne, B., Crane, R.I., Fielding, M., llDevelopment of Techniques for the Isolation and ~dentification of Nonvolatile Organics in Drinking Water," in Advances in the Identification and Analysis of Orsanic Pollutants Water, Vol. I, Keith, L.H., Ed., pp. 383-398, Ann Arbor Science Publishers, Ann Arbor, Michigan, 1981.

Weston, A.F., "Obtaining Reliable Priority-Pollutant Analyses," Chemical Ensineerinq, April 30: 54-60, 1984.

Wever, Jr., W.J., Jodellah, A.M., I1Removing Humic Substances by Chemical Treatment and Adsorpti~n,~~ Journal of the American Water Works Association, 77:132-137, 1985.

Wu, C., Suffet, I.H., "Design and Optimization of a Teflon Helix continuous Liquid-Liquid Extraction Apparatus and its ~pplication for the Analysis of Organophosphate Pesticides in Water,!' Analytical Chemistrv, 49: 231-237, 1977.

LIST OF PUBLICATIONS

~ietrich, A.M., G.S. Durell, R.F. Christman, "GC/MS ~dentification of Cyclohexene Artifacts Formed During Analysis of Chlorinated Water Samples," Proceedings of the 33rd Annual Conference on Mass Spectrometry and Allied Topics, 1985.'

APPENDICES

Paqe

Analysis Report of Broad Spectrum Analysis by EPA Methods 6 2 4 and 625 - ~ieldcrest WTP ~inikhed water . .

. . . . . Mass Spectra of Primary Internal Standards

. . . . Mass Spectra of Secondary Internal Standards

Mass Spectra of Some Major Identified SOC Pollutants in the Natural Waters Studied . . . . . . . . .

Mass Spectra of Some Major Identified SOC Pollutants in the Natural Waters Studied Which Were Also Major Laboratory or Analysis Method Contaminants

Mass Spectra of Some Major Unidentified SOC Pollutants in the Natural Waters Studied . . . . . . . ' . . .

Mass Spectra of Solvent Artifacts Identified . . . . . .

AVALYSIS REPORT

DHS Form Ida0 ( R e v . 2 - 7 5 ]

,X3h;iTORY

Analysis Report of Broad Spectrum SOC Analysis

StL'PLE HUHSE.? DESCAIPTION

by €PA Methods 624 and 625 - Fieldcrest WTP Finished Water

(performed by a certified laboratory)

1 1 8

NU.YBER I n ; ,. L , , :

5 !;'79.4.4 +&LI

A I [?cT) 5 ~ 2 4 2 1 I /&T) 5?2243! c I Lq&qrJ:,-

ALU!RKS

S : ~ Z Q ~ i 3 1 I

RESULTS 13 1

5qi ?-,,- L . wy

I I I ! I

i a / I - Acetone probably T r m b o t t l e I o n t a r m a t i c ? . EPA ('ethcd 5 2 4 .

1 ( *Sase/neutra l m d 3cld ex t rac t ib r . , SPA r e t h s o 625 .) i I !

Acetcne , ch lo rofc rn I sresent- 3 /

acetoce present- 9 /

!:a coz?oucds ~ d e c t f isa+

KO coc?ounds l d s n t f i e d a

a - I

I

CMYE!'TS: 0 C 9 #CUT ~e

I I

!

Trichloroethene

Tetrachloroethene

P h e n o l

&I,:. 1 ' ' I " I ' I " .dl 1 '

Atraz ine

s

1 0 0-

e 0,

a 0,

1 1 . r 0,

2 0,

0

2 7 0 3 1 0 3 5 0 3 9 0 4 3 0

2 2 7

1 ' 4

Methoxychlor

B u t y l a t e d h y d r o x y t o luene ( Y H ' i ' )

D i e t h y l p h t 5 I a 1 a t e

D i e t h y 1 h e x y l p h t h a l a t e

p h t h a l a t e

7

Unidentified A

Unidentified B

Unidentified C

Unidentified E

Unidentified F

( p o s s i b l y a C7H140)

Diacetone alcohol

'I'

Cyclohexanol

Dich lorocyc lohexane

(isozer )