Environmental Health PerspectivesVol. 46, pp. 87-99, 1982

Nonvolatile Organic Compoundsin Treated Watersby C. D. Watts,* B. Crathorne,* M. Fielding* andS. D. Killopst

Over the past decade much information has been published on the analysis of organicsextracted from treated water. Certain of these organics have been shown to be by-products of thechlorination disinfection process and to possess harmful effects at high concentrations. This hasresulted in increased interest in alternative disinfection processes, particularly ozonation. Thedata on organics had been largely obtained by using gas chromatography-mass spectrometry,which is only capable of analyzing, at best, 20% of the organics present in treated water.Research in key areas such as mutagenicity testing of water and characterization of chlorinationand ozonation by-products has emphasized the need for techniques suitable for analysis of theremaining nonvolatile organics.

Several methods for the isolation of nonvolatile organics have been evaluated and, of these,freeze-drying followed by methanol extraction appears the most suitable. Reverse-phase HPLCwas used for separation of the methanol extract, but increased resolution for separation of thecomplex mixtures present is desirable. In this context, high resolution size exclusion chromatog-raphy shows promise. Characterization of separated nonvolatiles is possible by the application ofstate-of-the-art mass spectrometric techniques.Results obtained by these techniques have shown that the nonvolatile organic fraction of

chlorinated drinking water consists of many discrete compounds. Among these, some of thechlorinated compounds are almost certainly by-products of disinfection. Studies of the by-products of ozonation of fulvic and humic acids isolated from river waters have indicated asimilar proportion of nonvolatile organics. Further, ozonation can result in the release ofcompounds that are trapped in the macromolecules.

IntroductionThis paper presents a short review of research

into nonvolatile organics in water, with particularemphasis on analytical methodology. In order tolimit the size of the paper, the review is necessarilyselective in the citation ofresearch from the scientificliterature. Results of recent work at the WaterResearch Centre (WRC) will be used to highlightimportant areas of research into nonvolatile organiccompounds in treated water.The need for methods of analysis of nonvolatile

organics stems from the rapid growth of interest inorganics in treated water over the past ten years

*Water Research Centre, Medmenham Laboratory, P. 0. Box16, Henley Road, Medmenham, Marlow, Bucks, SL7 2HD,England.

tOrganic Geochemistry Unit, University of Bristol, School ofChemistry, Cantock's Close, Bristol, BS8 1TS, England.

(1, 2). This interest arises from concern over thepossible long-term health effects from ingestion oforganics in treated waters, particularly haloge-nated species (1). The vast majority of identificationsof organic material in treated waters have beengenerated by the use of gas chromatography-massspectrometry (GC-MS) techniques, which provide arelatively rapid and straightforward method ofanalysis. It is now generally accepted that 80-90%of the organic matter present in both raw andtreated waters is not amenable to this type ofanalysis due to insufficient volatility. Further,there is some evidence that a similar proportion ofthe halogenated organic matter produced by chlori-nation is also not amenable to GC-MS analysis (3).Methods for the characterization of the nonvolatileorganics in treated waters are thus urgently neededto give information to complement that obtainedfrom GC-MS. Undoubtedly, concern over possiblehealth effects of the nonvolatile organics, which

88

represent the bulk of the organic matter in water,is heightened by the paucity of information con-cerning them. The rapidly expanding application ofshort-term bioassay tests, such as the Ames test, todrinking waters is beginning to emphasize this needfor qualitative techniques for nonvolatile compounds.Many of the mutagens detected do not seem to beidentifiable by GC-MS methods (4, 5).

There are many diverse problems associatedwith the characterization of nonvolative organics intreated waters, but most of these are a result of thelow levels often present, the wide range of com-pound types and extreme range ofmolecular weights.Currently there is no analytical method availablefor nonvolatiles that is comparable in sensitivityand ease of application to the GC-MS approach usedfor volatiles. It is unlikely that a single method willbe sufficient and thus the various steps in theanalysis of nonvolatiles are reviewed below.

Isolation MethodsIsolation of a specific compound from a very

dilute aqueous solution of a complex mixture ofcomponents is a relatively easy process, since theextraction method can be tailored to the chemicaland physical properties of that compound. Howev-er, isolation of the broadest possible fraction of acomplex mixture of unknowns is a considerablymore difficult problem. A large variety of methodshave been examined for isolation of organics fromwater, and these have been recently reviewed (6).Most of these have been developed specifically forvolatile compounds and are, therefore, outside thescope of this paper. Relatively few methods havebeen developed for nonvolatiles, but several of themethods used for volatiles have been applied, withor without modification.Research specific to the isolation of nonvolatile

organics from water has been mainly concernedwith the humic substances. Methods used includeadsorption, precipitation and liquid-liquid extrac-tion, and these have been reviewed in some detail(7, 8). Recent research has focused on the use ofXAD resin adsorption which has been shown toprovide high recoveries (>80%) of humics from rawwaters (9, 10). However, XAD is not recommendedfor the concentration of humic substances afterchlorination, as these are very poorly adsorbed (3).Other techniques which have been applied to non-volatile organics in water are carbon-chloroformextraction (11), reverse osmosis (12) and freeze-drying (13, 14) which suffers form the limitation ofbeing a batch process. Result from these methodssuggest that freeze-drying and reverse osmosisprovide extracts containing the widest range ofcompounds. With freeze-drying, the recoveries

WATTS ET AL.

obtained for particular compound types was foundto be solvent dependent, but methanol provided thehighest overall recoveries (14).

Separation MethodsSeparation ofnonvolatile organics ideally requires

a method that avoids derivatization, does not chem-ically alter the organics during separation, and hashigh resolving power and high sensitivity of detec-tion for separated compounds. While no methodmeets all of these criteria, high performance liquidchromatography (HPLC) approaches this ideal andis the method of choice for the separation of nonvol-atile organics.A variety of fractionation procedures have been

used prior to HPLC, particularly where fractionsare being prepared for mutagenicity testing. Thus,thin-layer chromatography, column liquid chroma-tography and chemical fractionation have all beenapplied for preseparation of nonvolatile organics,and these approaches have been recently reviewed(5,15). Modern HPLC often involves the use ofseveral chromategraphic methods and most of thesehave been used- for separation of nonvolatile organ-ics from treated waters. Size exclusion chromatog-raphy (SEC) has shown that the molecular weightsof nonvolatiles in water range from a few hundredto >15,000 (3,16,17). One drawback in the character-ization of nonvolatile organics according to molecu-lar size by use of SEC is the competition betweenexclusion processes and nonexclusion processes suchas adsorption. This is a particular problem for anunknown mixture of compounds of widely differentpolarity and can invalidate a molecular weight cali-bration produced using a mixture of known com-pounds.

Ion-exchange chromatography has provided highresolution separation of nonvolatiles from chlori-nated and ozonated waters prior to identificationand mutagenicity testing (18,19). Reverse-phaseHPLC has been applied to the separation of non-volatiles from treated water to provide fractions formutagenicity testing (20) and mass spectometriccharacterization (13,14). Finally, adsorption HPLCon silica gel G was used to separate neutral organicsfrom a drinking water concentrate prior to massspectrometry (21).Although HPLC shows inherent advantages for

the separation of nonvolatile organics, it suffersfrom two limitations. There is no universal detectorfor HPLC which is equivalent in sensitivity to theflame ionization detector used in GC. Several highlysensitive detectors are available, but these are alsohighly selective and, therefore, inappropriate tothe detection of a wide range of compounds. Detec-tion of separated components is usually achieved by

NONVOLATILE ORGANICS IN WATER

ultraviolet absorption and is therefore limited tothose compounds containing an ultraviolet chromo-phore. Secondly, since nonvolatile organic fractionsisolated from water usually contain a large numberof compounds of diverse chemical type, even highefficiency HPLC columns cannot provide completeseparation into individual components. Thus, meth-ods for characterization of HPLC fractions arerequired that are capable of handling a mixture ofcomponents, which is discussed below.

CharacterizationTechniques for the characterization of nonvolatile

organics can be divided into two broad categorieson the basis of approximate molecular weight, eachrequiring different procedures.High Molecular Weight Material (>2000). This

category consists mostly ofhumic and fulvic acidspolydisperse polymeric materials, which are notamenable to mass spectral characterization as intactmolecules. General methods for their characteriza-tion have been extensively reviewed (8,22) and onlyrecent work which centers around the production ofhaloforms and other halogenated organics from chlo-rination of humic acid fractions (23-25) will be dis-cussed.

Halogenated nonvolatile reaction products ofhumicacids have been characterized by total organohalogenmeasurement and size exclusion chromatography(3,26). The majority of structural studies on aquatichumic materials involve some form of degradationstep to produce smaller, volatile compounds foranalysis by GC-MS. Included in this approach areattempts to identify the volatile by-products fromchlorination (27) and ozonation (28) of humic mate-rials. The substituted aromatic compounds identifiedin degradation studies have similar structures tothose obtained from soil humic material, suggestingthat the two types of humics have similar composi-tions. Recently, attempts have been made to obtainmass spectrometric data on larger fragments ofhumic acid (29). This has involved field desorptionmass spectrometric (FD-MS) analysis of fragmentsformed after chlorination or permethylation ofhumicmaterial, but as yet no structural information hasbeen reported.Low Molecular Weight Compounds (<2000). As

mentioned previously, any method used for charac-terization of nonvolatile organics in HPLC fractionsmust, in most cases, be capable of working with lowlevels of individual compounds in complex mixtures.Thus, most of the classical methods of characteriza-tion, e.g., nuclear magnetic resonance, infrared andultraviolet spectrometry, are unsuitable, and nec-essarily some type of mass spectrometry must beemployed. Conventional electron impact mass spec-

89

trometry using a direct insertion probe offers aconsiderable extension to the types of compoundthat can be handled by GC-MS. However, electronimpact suffers from two limitations: the compoundof interest has to be volatilized prior to ionizationand the molecular ion is frequently of low intensitywith respect to fragment ions. The first limitationcan be overcome to some extent by the use of rapidheating and "in-beam" techniques, which have beenrecently reviewed (30). The second limitation, how-ever, is difficult to overcome and results in extremelycomplex, overlapping mass spectra in the case ofmixtures.

Ideally then, for mass spectrometry of mixturesof nonvolatiles, a method is required that can ionizecompounds in the solid state with transfer of verylittle excess energy. This would provide a massspectrum consisting of intense molecular ions withfew, if any, interfering fragments ions. Severalionization methods which can achieve this are avail-able, for example: direct chemical ionization (DCI)(30), plasma desorption (PD) (31), laser desorption(LD) (32), field desorption (FD) (33) and, mostrecently, fast atom bombardment (FAB) (34). All ofthese techniques provide intense molecular ionsand only limited fragmentation for certain types ofnonvolatile organics and are, to some extent, com-plementary. At the present time, FD-MS has receivedmore extensive application than any of the othertechniques and appears to be applicable to thewidest range of compound types (35). Surprisingly,the use of FD-MS for the characterization of non-volatile organics separated by HPLC has receivedonly limited study. However, it has been used foranalysis of herbicides and their by-products in sur-face waters (36,37) and organics in drinking water(14,21). While it has contributed to the identificationof only few compounds so far (13), it has shown thata complex mixture of nonvolatile organics can bepresent in potable water but it is impossible toresolve this into individual components even bymodern HPLC.A limitation of FD-MS for the identification of

nonvolatiles is the lack of structural informationprovided. Accurate mass measurement of molecu-lar ions can provide empirical formulae, but a num-ber of structures will still be possible. Recently,methods for obtaining structural data from selectedions, namely, collisional activation (CA) (38) cou-pled with either (B/E) linked scanning (39) or mass-analyzed ion kinetic energy spectrometry (MIKES)(40,41) have become more widely available. Usefulfragmentation information similar to that obtainedfrom electron impact mass spectrometry can beobtained for any ion (of sufficient intensity) in amass spectrum, which makes the method advanta-geous for the analysis of mixtures. Reports of the

90

combined use of FD-MS and CA have only recentlyappeared, but the technique offers great promisefor the identification of nonvolatile organics (42).

Chemical StudiesThe broad spectrum approach to determining the

nature and behavior of nonvolatile organics in waterhas been supplemented by laboratory studies on

specific nonvolatile organics known or believed tobe present in water. Much of this type of workrelates to investigations of the organic by-productsformed during chemical disinfection processes suchas chlorination and ozonation (3,18,27,29). Moststudies have been confined, due to a lack of suitabletechniques, to products identifiable by GC-MS, i.e.,volatile organics. Indirect evidence from such stud-ies, such as results from TOC and TOCl measure-ment, suggests that a considerable proportion ofthe by-products are nonvolatile in nature. Howev-er, its characterization awaits the development andapplication of appropriate analytical methods.The following sections of the paper are devoted

to some relevant features of WRC research. Inparticular, an overall analytical procedure for iden-tification of nonvolatile organics in water, which hasbeen developed recently, is outlined in some detail.Two additional features related to "chemical stud-ies" are included. One involves the laboratory chlo-rination of uracil. 5-Chlorouracil and other haloge-nated compounds were detected in treated water inearlier work (13). The work reported there wascarried out to test the hypothesis that such a com-pound is formed from uracil (a constituent of sew-age effluent) during water treatment chlorination.The other aspect, very much related to the subjectof nonvolatiles in water, is the identification of by-products of aqueous ozonation of humic and fulvicacids and some preliminary findings from this workare presented.

Experimental

IsolationThe procedures used for isolation of nonvolatile

organics from treated water have been reported indetail elsewhere (13). Basically, these involve thefreeze-drying of water samples followed by metha-nol extraction of the recovered freeze-dried solids.Initially, adsorption by XAD was considered as a

potential isolation method and in order to provide acomparison between freeze-drying and resin extrac-tion, the following experiments were carried out.Small samples (5 1.) of a treated water were collect-ed; one was acidified (to pH 3 by using -4N HCI),

WArTS ET AL.

one basified (to pH 10 by using -AN NaOH) and theother left at neutral pH. Each of these samples anda blank consisting of double-distilled water werepassed through a resin column (XAD-2), which wassubsequently eluted with methanol (-400 ml). Themethanol eluates were then concentrated and exam-ined by HPLC as detailed below.

Separation: High Performance LiquidChromatographyThe equipment used consisted of two solvent

delivery systems (Waters Associates, Model 6000A),a gradient former (Waters Associates, Model 660)and two ultraviolet absorption detectors operatedin series (Cecil Instruments, Model CE 212; LDC,Model 1203), at 280 nm and 254 nm. Syringe injec-tions were made through a stop-flow septumlessinjection port for analytical separations (up to 10 RIinjected) and via an injection valve (Rheodyne Inc.,Model 7125) for preparative separations (up to 200RI injected). Columns (20 cm x 7 mm ID) werepacked in our laboratory (13) with 5 ,um particlesize Spherisorb-ODS (Phase Separations Ltd.) andgenerally had an efficiency of about 18,000 theoreti-cal plates (HETP, 0.01 mm; reduced plate height,2.2) as measured from a standard mixture of phe-nols (43).A linear gradient was established from two sol-

vent mixtures consisting of 1% methanol in 0.1%aqeous acetic acid (A) and 90% methanol in 0.1%aqueous acetic acid (B). The gradient was run over30 min from 0% to 100% B. The flow rate wasmaintained at 2.0 m/min.

Size exclusion chromatography (SEC) was car-ried out on a TSK 3000 SW (Toyo Soda Co.) column(30 cm x 7 mm I.D.) with the use of either water oran aqueous solution ofpotassium chloride and sodiumdihydrogen orthophosphate (0. 1M) as eluent (1 ml/min). Injection and detection were performed as forreverse-phase chromatography.

CharacterizationPreparative HPLC fractions were selected on

the basis of a relatively high ultraviolet absorptionfor mass spectrometric analysis on a VG MicromassZAB IF instrument. Electron impact was carriedout at 70 eV electron voltage, 200 ,uA trap current,8 kV accelerating voltage, source temperature of200°C and a resolution of -4000. Samples wereintroduced into the source via an independentlyheated direct insertion probe. Field desorption massspectra were obtained on the same instrument, alsoat 8 kV accelerating voltage and with a total extrac-tion voltage of 10 kV. The FD mass spectra wereobtained at a resolution of4500, using 10 ,m

NONVOLATILE ORGANICS IN WATER

tungsten wires with carbon microneedles (averagelength -40 ,m) grown in the usual manner (44). Allmass spectra were acquired and processed on a VGDatasystem computer.Accurate mass measurements were made on

selected peaks using either data system acquisitionor peak matching. Collisional activation (CA) stud-ies were performed by admitting the collision gas(helium) into a collision cell (in the first field freeregion) to a pressure sufficient to reduce the inten-sity of the parent ion beam by 60%. The CA massspectrum was then obtained by scanning the massspectrometer at a fixed ratio of B/E and recordedoscillographically.

Chemical StudiesChlorination of Uracil. A series of laboratory

experiments was carried out on the aqueous chlori-nation of uracil. Distilled water (2 1.), buffered (pH7.5) with borate, was spiked with uracil (1 and 10mg/l.) and aqueous hypochlorite added to produce achlorine residual (12 and 1.2 mg/I.). The solutionswere stirred continuously and maintained at ambi-ent temperature (20°C) in a water bath. The exper-iments were carried out in the dark, and sampleswere withdrawn at regular intervals for determina-tion of pH, free and total chlorine residuals (DPDmethod) and 5-chlorouracil production. The latterwas monitored by freeze-drying the sample (10 ml),extracting the residue with methanol (3 x 5 ml) andsubmitting the extract to HPLC. One sample fromthe experiment carried out at high chlorine residualwas extracted with n-pentane (2 x 5 ml) and ana-lyzed for haloforms by GC-MS. System blanks werecarried out in an identical manner to that above,but without addition of uracil.Ozonation ofHumic and Fulvic Acids. Ozone

was prepared from dry oxygen (typically -15 mg/min at a flow rate of 180 ml/min), using a GallenkampGE-150 generator. The output was monitored beforeeach experiment by determining the amount of freeiodine liberated from KI solution (45). Reactionswere performed in a twin-necked round-bottomedflask (500 ml) fitted with a bubbler.

Solutions of fulvic and humic acids (200 mg)extracted from river water as previously described(46) were prepared in aqueous 1M NaOH (25 ml)and diluted with double-distilled water (175 ml).Solution pH was adjusted to -6 by addition ofaqueous HCl prior to ozonation (1 hr, typically 5 mg03/mg fulvic or humic acid). By means of a tech-nique based upon adsorption onto XAD-2 resin (47),unozonized and ozonized solutions were extractedafter acidification to pH 1 with aqueous HCI. Afterconcentration under reduced pressure ( - 2 ml),the aqueous eluate was examined by HPLC.

91

The XAD-2 resin column was sequentially elutedwith ether (25 ml) and methanol (30 ml). The etherfraction was concentrated (to -1 ml) by evapora-tion prior to GC and GC-MS analysis, and themethanol fraction was concentrated (to -2 ml)under reduced pressure prior to HPLC analysis.The XAD-2 resin was regenerated using a modified

procedure to reduce the amount of organic contam-inants released from the resin (48). Blank experi-ments were carried out for all of the ozonations andresulted in chromatograms which were essentiallydevoid of peaks.GC was performed on a chromatograph equipped

with flame ionization detector (Carlo Erba Fractovap2150), Grob splitless injector, and a wall-coatedOV-1 Pyrex capillary column (25 m). GC-MS wascarried out in the electron impact mode (Finnigan4000) with computerized data acquisition and pro-cessing (INCOS 2300). A wall-coated silicone-fluidfused silica capillary column (20 m) was used with aGrob splitless injector.HPLC equipment used was similar to that out-

lined earlier in the experimental section consistingof two-solvent delivery systems (Waters Associ-ates, Model 6000 A), a gradient programmer (WaterAssociates, Model 660), and an ultraviolet absorp-tion detector (LDC Spectromonitor II) operating at240 nm. A reverse-phase column (25 cm x 4.6 mmI.D.) of C18-bonded silica (DuPont Zorbax ODS) wasused with a linear gradient of water/methanol (0%to 100%) containing an ion-pairing reagent (tetrabutylammonium bromide: 0.01M).

Results and DiscussionIsolationAs mentioned above, our studies with model

compounds have indicated that freeze-drying andmethanol extraction is a very efficient method forisolation of a wide range of nonvolatile organicsfrom drinking water. However, it does suffer fromtwo limitations: (1) only a fixed sample size can beprocessed in one batch and (2) a proportion of theinorganics in the sample is dissolved in the metha-nol extract. These problems led us to explore thepossibility of using XAD-2 resin as a method forobtaining a nonvolatile organic concentrate thatwas relatively low in inorganics from a large vol-ume of water sample.

Figure 1 compares the reverse phase HPLC chro-matograms derived from a drinking water sampleat neutral pH using the freeze-drying/methanolextraction and XAD-2 resin/methanol extractiontechniques. The XAD-2 extract exhibits a morepronounced "hump" of unresolved compounds andfewer discrete peaks than the freeze-dried extract.

92

Furthermore, a larger proportion of the XAD-2extract (eight times) had to be used to obtain asimilar ultraviolet response on the chromatograms.The XAD-2 extracts obtained at acidic and basic pHgave chromatograms similar to the neutral pH extractbut with fewer discrete peaks. From this compari-son of XAD-2 extraction with freeze-drying, it ispossible to make the following point with respect toultraviolet-absorbing compounds present in treatedwater. XAD-2 resin is less efficient at concentrat-ing nonvolatile organics than freeze-drying and resultsin an HPLC chromatogram with fewer discretepeaks.

It is probable that this lower efficiency of XAD-2is to some extent sample-dependent, since the watersample used to compare the different extractionmethods was a chlorinated drinking water. It hasbeen suggested (3) that the adsorption efficiency ofXAD-2 resin for humics is significantly lower forwater that has been chlorinated, as compared to thesource water.

Freeze-drying and methanol extraction appearsto offer the best overall recoveries for a wide range

COLUMN: SPHERISORB-ODS

MOBILE PHASE: LINEAR GRADIENT FROM 1% MeOH IN 0.1% AD. ACETIC ACIDTO 90% MeOH IN 0.1% AD. ACETIC ACID

DETECTION: UV (254 nm)

a) FREEZE DRIED EXTRACT:

315 20TIME (min)

b) XAD-2 EXTRACT: 4 pI from 100 pl of MeOH extract

SAMPLE VOLUME 301

FIGURE 1. HPLC proffles of methanol extracts of treatedwater samples concentrated by (a) freeze-drying and (b)resin adsorption. Eight times more sample was neededto produce chromatogram (b).

WATTS ET AL.

of organics. The major problem with this methodremains the significant quantity of inorganics whichis dissolved by the methanol. These elute as broadpeaks on reverse-phase HPLC and SEC and caninterfere with mass spectrometric characterization,which is discussed below. We have evaluated sev-eral methods for desalting methanol extracts (14),but none of these have been successful in reducingthe level of inorganics without adversely affectingthe organics. It is possible that the problem will bealleviated by the use of SEC as a further HPLCseparation step.

SeparationReverse-phase HPLC is the most suitable method

for the separation of complex mixtures of nonvola-tile organics isolated from treated water. HPLCchromatograms of treated water extracts consist ofa large early eluting peak and many discrete peakssuperimposed on a "hump", presumably consistingofunresolved ultraviolet-absorbing compounds (e.g.,Fig. la). Using preparative reverse-phase HPLC,it is possible to separate an organic extract into alarge number of fractions for characterization bymass spectrometry. Unfortunately, it is not practi-cable to examine all of these using the newer meth-ods of MS characterization and it is essential todevelop criteria for selection of particular HPLCfractions. Previously, we have selected fractionsaccording to either relatively high ultraviolet absorp-tion and/or relatively high total organohalogen (TOX)values (14). However, recent MS evidence (seebelow) has indicated that the use ofTOX levels as acriterion is severely handicapped by the presence ofinorganic chlorides. Preparative reverse-phase HPLCfractions that are selected on the basis of highultraviolet absorption are always found to consist ofmixtures of compounds by subsequent MS analysisand only rarely consist of one major compound.This complexity has encouraged us to seek a fur-ther separation step in an attempt to reduce thecomplexity of the mixtures prior to MS character-ization. Further separation according to molecularsize was an attractive possibility and a high per-formance size exclusion column was evaluated forthis purpose.The retention volumes and molecular weights of

a series of proteins, used to provide a molecularweight calibration, and some model compounds areshown in Table 1. These were obtained on an SECcolumn using both water (very low ionic strength)and 0. 1M KCl/NaH2PO4 (high ionic strength) as themobile phase. With water as the mobile phase, anelution order based on molecular size was notobtained. One of the proteins, albumin, was irre-versibly adsorbed on the column and the largest

Table 1. Retention volumes on SEC under different elution conditions.

Retention volume, mla

Solute Mol. wt. H20 0. 1M KCl/NaH2PO4

Ferritin 444000 7.0 6.8Catalase 232000 5.0 8.0Albumin 67000 NE 8.8Ribonuclease 13700 ND 10.5Adenosine diphosphate 550 4.9 11.1Chlorouridine 278 5.8 11.5Chlorouracil 146 6.0 11.7Uracil 112 7.7 12.0Anisole 108 17.0 ND

'ND = not determined; NE = not eluted.

protein, ferritin, eluted after a smaller one, cata-lase. Further evidence of the occurrence of ioniceffects was provided by the small retention vol-umes of four of the model compounds and theexceptionally high retention volume for the small-est model compound, anisole. Changing the mobilephase to one of higher ionic strength resulted in theexpected elution order for the proteins and modelcompounds based purely on molecular size. Theinherently low peak capacity of SEC columns is also

.0 DECREASING MOLECULAR WEIGHT

15 10 5

RETENTION VOLUME (ml)

FIGURE 2. Analysis of a reverse-phase HPLC fraction froma dring water extract using size exclusion chromatography.

demonstrated in that a retention volume of only -5ml is available for separation of compounds with amolecular weight spread of -400 to 500,000. Thus,these columns are appropriate for a crude separa-tion of nonvolatile organics, and are best applied tofractions obtained from reverse-phase HPLC ratherthan to a total drinking water extract. In order togive some insight into the resolving power of SECcolumns for nonvolatile organics of drinking water, abroad reverse-phase HPLC fraction was chromato-graphed on a SEC column. The chromatogram (Fig.2) shows two major peaks with some fine structuresuperimposed suggesting that the extract containsnonvolatile organics of two broad molecular sizeranges. It is not possible to determine the molecu-lar size range from this chromatogram, since this isdependent on the nature of the compounds used forcalibration of the column which cannot be extrapo-lated to a complex mixture of unknown composi-tion.

CharacterizationThe results obtained using electron impact (EI)

and field desorption (FD) mass spectrometry bothshow that reverse-phase HPLC fractions of drink-ing water extracts contain mixtures of compoundswith a wide range of molecular weights in agree-ment with other reports (21). An example of aFD-MS of a reverse-phase HPLC fraction of atreated water extract is shown in Figure 3. Thisrepresents one of the more complex fractions obtainedbut illustrates the wide range of compounds pres-ent, since the majority of ions can be considered tobe molecular ions under the FD-MS conditions used.Identification of these compounds presents manyproblems and the identifications reported to date(13) generally constitute cases where HPLC frac-tions have contained only one or two major com-pounds. Halogenated compounds are present, asillustrated by the FD-MS of another HPLC fractionof a treated water extract (Fig. 4). This shows

NONVOLATILE ORGANICS IN WATER 93

94 WATTS ET AL.

I* 9 1X

8 0

6 0 241 556

40_

20 37

60

xi40

639

2011

/88 o_e 115.3

on_1 I9122

FIGURE 3. FDMS of a reverse-phase HPLC fraction from a treated water extract.

I00

13 0 ~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~~0

4 09 8

20 60 100 140

60 1 99 226

xi

40

20 2 15

167~~~~~~~~~~~~~~~~~~-10 220 60

FIGURE 4. FDMS of a reverse-phase HPLC fraction from a treated water extract indicating halogen-containing compounds.

NONVOLATILE ORGANICS IN WATER

several ions with isotope ratios indicative of com-pounds which contain chlorine and/or bromine. Themethodology we are applying to attempt identificationof the nonvolatile organics is shown schematically(Fig. 5).

First, low resolution FD-MS is obtained on afraction to provide nominal mass data for the organ-ics present and information on the emitter currentat which they desorb. Then accurate mass data andhence empirical formulae are obtained on each FD-MSpeak of interest by using suitable reference com-pounds. Either a calibration table is created by theDatasystem which then automatically performs theaccurate mass determination or the peak is man-ually matched against a reference ion and its accu-rate mass calculated. Structural information is thenobtained for each compound by the use of collisionalactivation and B/E linked scanning. This techniqueproduces the equivalent of a normal electron impactmass spectrum for any molecular ion of sufficientintensity generated by FD-MS.An example of a FD-CA mass spectrum obtained

by linked scanning is shown in Figure 6. Thecompound used to produce this spectrum was tetra-butylammonium chloride and the FD-CA spectrumwas obtained on the tetrabutyl ammonium ion, m/z242. The structure of this ion is clearly indicated bythe elimination of methane (m/z 226), ethane (m/z212), propane (m/z 198) and butane (m/z 184). Theformation of the tributylammonium ion (m/z 185)and similar eliminations from this (e.g., m/z 169,elimination of methane) provide further structuralinformation. The combination of this data with theempirical formula information would allow relativelyeasy identification of this compound.

Application of this approach is still in its earlystages, but it is possible to comment on some of theproblems that have arisen. FD-MS is a relativelysensitive technique, requiring only a few nanogramsof compound, but produces low intensity fluctuatingion beams as compared to EI-MS. In order to carry

95

out collisional activation and linked scanning on aFD ion beam, a minimum of 0.5,u/g of each compoundmust be deposited on the emitter. Not only doesthis mean that there is a relatively high (for massspectrometry) minimum sample requirement, butalso if there is a very complex mixture, it may onlybe possible to examine one or two of the compoundspresent before exhausting all of the sample. Thepresence of large amounts of inorganics in methanolextracts of treated water and in some derived HPLCfractions can interfere with FD-MS in several ways;(1) they limit the absolute amount of a fraction thatcan be deposited on a FD emitter, (2) they can giverise to high molecular weight cluster ions for polarmolecules (49) and (3) they can yield, by formationof cluster ions, FD-MS spectra with ions at similarmolecular weights to compounds of interest, whichcan complicate interpretation of the mass spectrum(33). However, it is likely that the problems due toinorganic constituents will be overcome by eitherimproved HPLC resolution and/or FD-MS handlingtechniques. Even ifthese problems cannot be solved,FD-MS with collisional activation and linked scan-ning is by far the most promising technique avail-able for the identification of nonvolatile organicsisolated from water samples.

Chemical StudiesChlorination. The results obtained from labo-

ratory chlorination of uracil are shown in Table 2. Anegligible drop in free chlorine residual of the sys-tem blanks was observed during the experiments.GC-MS analysis of a low chlorine residual experi-ment showed negligible production of haloforms.Several points emerge from study of these results,albeit from a limited number of experiments: (1) theexperiment using a high chlorine residual demon-strates that formation of 5-chlorouracil from chlo-rine and uracil is rapid; (2) 5-chlorouracil is appar-

FD-CA mass spectrum of tetrabutylammoniumFIGURE 5. Scheme for mass spectrometric analysis.

Q.

1 4 2 4

8 2 2

1 5 I 9 2 6

9

FIGURE 6.chloride.

96

ently not the ultimate product of the chlorinationreaction; and (3) although the ultimate product(s)were not identified, they are not haloforms, at leastunder the conditions used.The method used to monitor the course of the

reaction did not indicate the appearance of anyproducts other than 5-chlorouracil, hence the endproduct(s) either does not absorb in the ultravioletregion at 254 or 280 nm or is sufficiently volatile tobe lost during freeze-drying or is not extracted intomethanol.

Ozonation. The compound types identified byGC-MS of the XAD-2/ether extracts before andafter ozonation of humic and fulvic acids are shownin Table 3. Identification is limited to the functionalgroups that can be recognized from the mass spec-tral fragmentations, and does not exclude the pres-ence of further oxygenated functionalities in eachcompound. Full details of the structural assignmentswill be published at a later date. A typical GC-MSchromatogram of an ether extract of ozonized fulvicacid is shown (Fig. 7) to indicate the complexity ofthe mixture of volatile organic products.These results show that the volatile products due

to ozonation of humic acid are mainly aromatic,

WATTS ET AL.

while those of fulvic acid are chiefly aliphatic. Suchfindings are in agreement with previously publishedresults (8). Ozonation of humic acid does not seemto liberate any volatile compounds resulting fromfragmentation of the polymer skeleton, althoughchanges are apparent among some of the compo-nents detectable after XAD-2 resin extraction ofunozonized material. With fulvic acid, ozonationleads to the appearance of certain aromatic acidsand normal, branched and possibly cyclic alkanes.While the aromatic acids found in the humic acidextract only after ozonation may result from oxida-tive cleavage of the macromolecular structure, thealkanes found in the fulvic acid extract only afterozonation were presumably trapped within the poly-meric matrix. However, the release of alkanes andDDT and related compounds which are extractedby XAD from humic acid prior to ozonation may bedue to a function of the macromolecular structure,which is known to be more "open" at the low pHinvolved in the XAD extraction procedure (22).Estimation of the proportion of volatile organicsextracted before and after ozonation suggests thatthere is little difference for humic acid (<1%) andonly a slight increase (-8% before and -12%

Table 2. Laboratory chlorination of uracil at high and low chlorine residual.

Time, Uracil, Cl-uracil Chlorine conen, mg/lmin concn, pg/l concn, pg/l Free Total

High Cl 0 1000 0 120.5 5 50 11

30 5 20 1060 0 0 10

Low Cl 0 10 0 1.2 1.20.5 5 0.5 1.1 1.115 5 0.5 0.75 0.7530 4 0.5 0.55 0.5545 5 0.5 0.4 0.450 4 1.0 0.35 0.35

Table 3. Compound types present in XAD-2 ether extracts of ozonized and unozonized humic and fulvic acid.

Humic acid Fulvic acid

Alkyl benzenes (chiefly methyl and ethyl)a Hydroxybenzoic acidsNaphthalene and methylnaphthalenesa Aliphatic alcohols (up to C20)Hydroindenesa Aliphatic diols (up to -C10)'Aromatic acids (1,2,3-ring) Aliphatic acids (mainly branched up to - C20)Aromatic carbonyl compoundsbSubstituted phenols (CHO, OMe, COMe, COOH groups) Aliphatic diacids (C_10)Heterocyclics (chiefly furans and benzofurans) Possibly ethers and ketones (aliphatic and mixed aliphatic-aromatic)Aliphatic acids (n-C,18)DDT and derived compounds Alkanes (normal, branched, -C2o30 and possibly cyclic)bn-Alkanes (n-C23 32)

a Present only in unozonized ether extracts.b Present only in ozonized ether extracts.e Probably also includes hydroxy acids, more predominant in ozonized sample.

NONVOLATILE ORGANICS IN WATER

100 -11

RIC_

500 1000 1500 208:20 16:40 25:00 33

FIGURE 7. GC-MS chromatogram of an XAD-2 ether extract from ozonized fulvic acid.

after) for fulvic acid. Thus, despite the severe con-ditions of ozonation (-10:1 w/w ozone to humicacid), only a relatively small amount of the humicand fulvic acids was converted to volatile organiccompounds.

Ozonation of both humic and fulvic acid resultedin the disappearance of most of the color of thestarting solutions, which could suggest that a majorstructural change occurred. Since no major changewas detected in the volatile components, it seems

that such changes should be detectable in the non-volatile components. The reverse-phase ion pairHPLC chromatograms ofthe XAD-2/methanol extractboth before and after ozonation (Fig. 8) of humicacid show similar features, i.e., three major peaks.A method is, therefore, required to provide infor-mation on the structural changes undergone by thepolymeric matrix of the humic and fulvic acids. Oneappropriate method is FD-MS (see above), whichwill be applied to the nonvolatile ozonation productsin the near future.The results ofTOC measurements of the aqueous

solutions which are applied to the XAD column andthe aqueous eluate indicate that the majority of theorganic matter is not adsorbed onto the resin col-

)00 25003:20 41:40

SCANTIME

umn. These water-soluble organics are also nonvol-atile in nature and exhibit similar chromatographicbehavior to those in the methanol extract of theXAD column. Investigations of these water-solubleorganics will be reported at a later date.

10 15TIME (min)

FIGURE 8. Reverse-phase ion pair HPLC chromatogram of anXAD-2/methanol extract of humic acid after ozonation.

97

co

98 WATTS ET AL.

ConclusionsNonvolatile organic compounds account for most

of the organic matter present in both raw andtreated water. Many of the industrial, agriculturaland pharmaceutical compounds in widespread useare nonvolatile and undoubtedly some proportion ofthem find their way into the aquatic environment.Studies of the organic by-products of chlorinationand ozonation ofwater have shown that a significantproportion of these are nonvolatile. The increasingapplication of short-term bioassay tests to drinkingwater has shown that nonvolatile organics could beresponsible for some of the mutagenic response.

Until recently there have been no really suitabletechniques available for the isolation, separationand characterization of nonvolatile organics fromwater. Consequently, few data are available on thetypes and amounts of nonvolatile organics presentin drinking water.

Recently there have been significant analyticaldevelopments, particularly in mass spectrometry,of direct relevance to nonvolatile organics. Newtechniques have emerged for the characterizationof nonvolatile organics and the approach usingfreeze-drying, HPLC separation and FD-CA massspectrometry appears to offer great promise in thisrespect. Without doubt, the identification of non-volatile organics in water would benefit considera-bly from a more widespread application of newtechniques such as these.Much of the work discussed in this paper was

funded by the Department of the Environment.Permission from the Department of the Environ-ment and the Water Research Centre to publishthis work is gratefully acknowledged.

Much of the work discussed in this paper was funded by theDepartment of the Environment. Permission from the Departmentof the Environment and the Water Research Centre to publishthis work is gratefully acknowledged.

REFERENCES

1. Jolley, R. L., Brungs, W. A., and Cumming, R. B., Eds.Water Chlorination: Environmental Impact and HealthEffects, Vol. 3, Ann Arbor Science, Ann Arbor, Mich.,1980.

2. Keith, L. H. Identification and Analysis of Organic Pollu-tants in Water. Ann Arbor Science, Ann Arbor, Mich.,1976.

3. Glaze, W. H., Saleh, F. Y., and Kinstley, W. Characteriza-tion of nonvolatile halogenated compounds formed duringwater chlorination. In: Water Chlorination: EnvironmentalImpact and Health Effects, Vol. 3, Ann Arbor Science, AnnArbor, Mich., 1980, pp. 99-108.

4. Fallon, R. D. and Fliermans, C. B. Formation of non-volatile mutagens by water chlorination: Persistence andrelationship to molecular weight of organic material inwater. Chemosphere, 9: 385-391 (1980).

5. Loper, J. C. Mutagenic effects of organic compounds indrinking water. Mutat. Res., 76: 241-268 (1980).

6. Jolley, R. L. Concentrating organics in water for biologicaltesting. Environ. Sci. Technol. 15: 874-880 (1981).

7. Gjessing, E. T. Physical and Chemical Characteristics ofAquatic Humus, Ann Arbor Science, Ann Arbor, Mich.,1976.

8. Povoledo, D., and Golterman, H. L. Humic Substances:Their Structure and Function in the Biosphere. Centre forAgric. Publ. and Documentation, 1975.

9. Leenheer, J. A. Comprehensive approach to preparativeisolation and fractionation of dissolved organic carbon fromnatural waters and wastewaters. Environ. Sci. Technol. 15:578-580 (1981).

10. Thurman, E. M., and Malcolm, R. L. Preparative isolationof aquatic humic substances. Environ. Sci. Technol. 15:463-466 (1981).

11. Tabor, M. W., Loper, J. C., and Barone, K. Analyticalprocedures for fractionating nonvolatile mutagenic compo-nents from drinking water concentrates. In: Water Chlori-nation: Environmental Impact and Health Effects, Vol. 3,Ann Arbor Science, Ann Arbor, Mich., 1980, pp. 899-912.

12. Deinzer, M., Melton, R. and Mitchell, D. Trace organiccontaminants in drinking water; their concentration byreverse osmosis. Water Res. 9: 799-805 (1975).

13. Crathome, B., Watts, C. D., and Fielding, M. The analysisof nonvolatile organic compounds in water by high-performance liquid chromatography. J. Chromatogr. 185:671-690 (1979).

14. Watts, C. D., Crathorne, B., Crane, R. I., and Fielding, M.Development oftechniques for the isolation and identificationof non-volatile organics from drinking water. In: Advancesin the Analysis and Identification of Organic Pollutants fromDrinking Water, Vol. I, L. H. Keith (Ed.), Ann ArborScience, Ann Arbor, Mich., 1981, Chapt. 25.

15. Waters, M. D., Nesnow, S., Huisingh, J. L., Sandhu, S. S.,and Claxton, L. Environmental Science Research: Applica-tion of Short-term Bioassays in the Fractionation andAnalysis of Complex Environmental Mixtures, PlenumPress, New York, 1978.

16. Glaze, W. H. and Peyton, G. R. Soluble organic constituentsof natural waters and wastewaters before and after chlori-nation. In: Water Chlorination: Environmental Impact andHealth Effects, Vol. 2, Ann Arbor Science, Ann Arbor,Mich., 1978, pp. 3-14.

17. Gloor, R., and Leidner, H. Universal detector for monitor-ing organic carbon in liquid chromatography. Anal. Chem.51: 645-647 (1979).

18. Jolley, R. L. Identification of organic halogen products.Chesapeake Sci. 18: 122-125 (1977).

19. Jolley, R. L., and Cumming, R. B. Oxidant effects oncomplex mixtures of non-volatile organics in polluted waters:examination by HPLC and bioscreening. Ozone Sci. Eng. 1:31-37 (1979).

20. Cumming, R. B., Lee, N. E., Lewis, L. R., Thompson, J.E., and Jolley, R. L. Relationship of disinfection to mutage-nicity in wastewater effluents. In: Water Chlorination:Environmental Impact and Health Effects, Vol. 3, AnnArbor Science, Ann Arbor, Mich., 1980, pp. 881-898.

21. Shinohara, R., Koga, M., Kido, A., Eto, S., Hart, T., andAkiyama, T. Analysis of low volatile neutral organics inwater by HPLC/FD-MS. Paper presented at A.C.S. Divi-sion of Environmental Chemistry Symposium; reprints ofpapers, Vol. 19, No. 1, 151 (1979).

22. Schnitzer, M. Chemistry of humic substances. Environ.Geochem. 2nd Int. Symp. 7: 89-107 (1975).

23. Stevens, A. A., and Symons, J. M. Measurement oftrihalomethane and precursor concentration changes. J.Am. Water Works Assoc. 69: 546-554 (1977).

NONVOLATILE ORGANICS IN WATER 99

24. Rook, J. J. Chlorination reactions of fulvic acids in naturalwaters. Environ. Sci. Technol. 11: 478-482 (1977).

25. Oliver, B. G., and Lawrence, J. Haloforms in drinkingwater. A study of precursors and precursor removal. J. Am.Water Works Assoc. 71: 161-163 (1979).

26. Oliver, B. G. Chlorinated non-volatile organics produced bythe reaction of chlorine with humic materials. Can. Res. J.11: 21-22 (1978).

27. Christman, R. F., Johnson, J. D., Pfaender, F. K., Nor-wood, D. L., Webb, M. R., Hass, J. R. and Bobenrieth, M.J. Chemical identification of aquatic humic chlorinationproducts. In: Water Chlorination: Environmental Impactand Health Effects, Vol. 3, Ann Arbor Science, Ann Arbor,Mich., 1980, pp. 75-98.

28. Lawrence, J., Tosine, H., Onuska, F. I., and Comba, M. E.The ozonation of natural waters: product identification.Ozone Sci. Eng. 2: 55-64 (1980).

29. Barofsky, D. F., Barofsky, E., and Wershaw, R. L. FD-MSof Humic Acid. Paper presented at 26th Annual Conferenceon Mass Spectrometry and Allied Topics, 1978.

30. Cotter, R. Mass spectrometry of non-volatile compounds.Anal. Chem. 52: 1589A-1606A (1980).

31. McNeal, C. J., MacFarlane, R. D. and Thurston, E. L. Thinfilm deposition by the electrospray method for Californium-252 plasma desorption studies of involatile molecules. Anal.Chem. 51: 2036-2039 (1979).

32. Posthumus, M. A., Kistemaker, P. G., Meuzelaar, H. L. C.,Ten Noever De Brawn, M. C. Laser desorption-mass spec-trometry of polar, non-volatile bioorganic molecules. Anal.Chem. 50: 985-991 (1978).

33. Beckey, H. D. Principles of Field Ionization and FieldDesorption Mass Spectrometry, Pergamon Press, London,1977.

34. Barber, M., Bordoli, R. S., Sedgwick, R. D., and Pyler, A.N. Fast atom bombardment of solids as an ion source inmass spectrometry. Nature 270-275 (1981).

35. Schulten, H. R. Biochemical and environmental applicationsof field ionisation and field desorption mass spectrometry.Int. J. Mass Spectrom. Ion Phys. 32: 97-283 (1979).

36. Schuiten, H. R. and Stober, I. Determination andidentification of biocides of the phenylurea, carbamateand thiocarbamate type in surface waters by highpressure liquid chromatography and field desorptionmass spectrometry. Fresen. Z. Analyt. Chem. 293:370-376 (1978).

37. Pankow, J. F., Isabell, L. M., and Barofsky, D. F. Theidentification of chlorophenoxyphenols in soil and watersamples by solvent extraction and field desorption massspectrometry. Anal. Chim Acta 124: 357-364 (1981).

38. McLafferty, F. W. Collisional activation mass spectra. Phil.Trans. Roy. Soc. (London) 293A: 93-102 (1979).

39. Beynon, J. H., and Boyd, R. K. Scanning of sector massspectrometers to observe fragmentation of metastable ions.Org. Mass Spectrom. 12: 163-165 (1977).

40. Bozorgzadeh, M. H., Morgan, R. P., and Beynon, J. H.Application of mass-analysed ion kinetic energy spectrome-try (MIKES) to the determination of the structures ofunknown compounds. Analyst 103: 613-622 (1978).

41. Kondrat, R. W. and Cooks, R. G. Direct analysis ofmixtures by mass spectrometry. Anal. Chem. 50: 81A-92A(1978).

42. Straub, K. M., and Burlingame, A. L. Use of fielddesorption/collision induced dissociation (FD/CID) massspectrometry in sequencing modified polynucleotides. Ad-vances in Mass Spectrometry, Vol. 8 (1980), in press.

43. Bristow, P. A., and Knox, J. H. Standardization of testconditions for high performance liquid chromatographycolumns. Chromatographia 10: 279-289 (1977).

44. Beckey, H. D., Hilt, E., and Schulten, H. R. High tempera-ture activation of emitters for field ionization and fielddesorption. J. Phys. E, 6: 1043-1044 (1973).

45. Lanting, R. W. Modification of the potassium iodide proce-dure for improved stoichiometry. Atmos. Environ. 13:553-554 (1979).

46. Packham, R. F. Studies of organic colour in water. Proc.Soc. Wat. Treat. Exam. 13: 316-334 (1964).

47. Junk, G. A., Richard, J. J., Grieser, M. D., Witiak, D.,Witiak, J. L., Arguello, M. D., Vick, R., Svec, H. J., Fritz,J. S., and Calder, G. V. Use of macroreticular resins in theanalysis of water for trace organic constituents. J. Chromatogr.99: 745-762 (1974).

48. James, H. A., Steel, C. P., and Wilson, I. Impurities arisingfrom the use of XAD-2 resin for the extraction of organicpollutants in drinking water. J. Chromatogr. 208: 89-95(1981).

49. Rechsteiner, C. E., Jr., Youngless, T. L., Bursey, M. M.,and Buck, R. P. Field desorption of cluster ions: a method oftesting the mechanism of ion formation. Int. J. MassSpectrom. Ion. Phys. 28: 401-407 (1978).