Organic micropollutants in air and water

Organic micropollutants in air and water

Organic r&ropoIhrtants are sampled by dynamic errrichment on a porous polymer coiumn and subsequently thermally eluted in a flow of helium. The eluted...

2MB Sizes 7 Downloads 44 Views

Organic r&ropoIhrtants are sampled by dynamic errrichment on a porous polymer coiumn and subsequently thermally eluted in a flow of helium. The eluted compounds enter gas chromato_&aph inlet and then pass into two parallel gIass capillary cohrmns with different stationary bhases, in which they are separated. The ‘+pa+ed compounds are detected by gas_chromatographic. (Cc) detectors and by ~a mass spectrometer that is corrected to one or other of the two capillary coiumns by “heart cutting”~ systems.,The GC and mass spectrometric (MS) signafs are fed vfa inierfslces into & minicompllter which controls the TvXSscan.and performs data acquisitio& red+ion and treatment on-line and off-line. The resultirig GC and MS data are displayed on a line primer or a visual display unit. The minicomputer is connected by a telephone line to an IBM 370/X5 computer, where a library search system has been implemented. So&e ditheulties encountered, data on the sampling recovery of model compounds and the ide&%ation of compounds in air and water samples by GC-MS data and library searches are discussed.

W&n the framework of the Environmental Programme of the Joint Research e&e of the Euro&an~ Eeor;omic_Comqmity, we are developing’ an-analytical --urGt for the a+ysis and idemification of oiganic micropollutants in air and water -_: =samPi&._ .. -_ - ..The commonest earlier a&roach to the analysis of organic pollutants irr en-viromr.t~ti s&$& was the determin&on of some @e-sekcted compounds that h&Q&~o r$ I-y,:ofIen-by chance, bee3 recognized as harinful (e.g., pesticides and carginogem+ but~negkcting r&any other-compounds %rbitrari&; simply -use of their ~~~~id&#.&_~ Ip E&X&



. . --




years; tke gradual discovery of a wide variety of compo*ur& in the __;-


-; . . . .. .~ j&g&&. etd’,- _ -, ._ :*--.-*- ..


--.-:.-=-. e&kna;eiit (Lg.; gz+irie zind .ctiorinated [email protected] iu .valio& Q+ of .&&) c&d -fof.e&nkw appr&& iitiir&e .&ih .&e ass-&p&on: that .&e natuiii:~fm.ost -if&- t&b&-$o&&~ is’-&&.&&&&$and &at && sholild~~~:I.~:identi~~ .-IS ixxtaid~ as. pgssi%le hefore &ce&ar& being shown [email protected]$ g &&ialfy~ i&%uL Both the l&boi;&tO+$ of the Envi&&x& [email protected]~ iti ihe X3.&.&2~ and the [email protected] iaboratories invoi&d in the EJX%CCWI’ qoj&t &b3, fdr inst&ce, ~’ -. -. are workirig afong these tin&.. The .method of cboiCe is to couple different tedhniqtiesl &mp$sing a s&p&g procedure, gas &ro.i&ogCaphic (cc) separation iand mks+e&r+metri~ f&Z?) detection. The large +&nt.ofd&& produced by such ati analytitil tinit @is to be handI-. ed &d evahated by 2 comptiter, which, iti add&tin to &educing &d .cqmbining :ee data, is .able to identify (by Il%rary searches) -o£s for. which reference data exist-or to-give struc&-c gpproximations for thos&compoutids for which no reference data are available. .~ At present, these analyticai units are suitable o-dy -for &g&c compounds that are amenable to GC separation, e.g., compounds witiCnot too.high in boiling point and good thermal stability. This factor, of course, limits the range of comp&mds that can be anatyzed &a sample and, in some respects, simplifies the analytic&pr&ile of the sample itself, which may contain compounds that do. not pass the GC separatioti. step. As stated by Grob et nL4, Cc Present ~kriowledgeabout organic substances, mostly iii -water, is biased as analytical not directed by what is in the water or by what we want to find out but by what we are abie to find out by using the available methods”. However, GC stiu remains the method of choice because of its out&n&g separating power. High-pressure liquid chromtitography, when its. sepmtion efliciency has &en increased, could ‘become an al&native separation method for $oiar, high-boiling and thermally unstable compounds, and work is in progress to assess the possi%iIities of Iinking such a technique with MS. DESCRIF’TION

OF l-I&


Fig. 1 giv& 2 flow scheme of the unit. Or&c~micropoIIutants are sampled by dynamic enrichment on a porous polymer g&s adsorption column and &erwzrds thermally eluted &th a fiow of helium. The duted compounds enter, via ti CiC inlet, two par2lIel glass capillary cohmms (0.3 &.1-D.; 60 m length; Vtian~,lvCdet 2700)

with differentstationary [email protected], in which thecompounds are separated. The separated _compounds are detected by’&me-ionization detectors (FfDs) and a mass spectrometer (quadrupole type, Riber Model QML 51; mass range 4-600; resolution m/&z = 600; deteetabihty 104-10-” g/;jecj,which is connected to one or other of the two cap& lary columns by ‘%eart cutting” systems. The GC and MS signals are fed via interfaces into a minicomputer (General Automation SPC-16/65; S-bit words; l6K core memory; CPU cycle time 960 nsec) which controls the MS sean and performs data acquisition, reduction and treatment on-line and OX-line, in this instance using a supplementary disk memory of 2.5 megawords. The minicomputer is controiled by a teletype or a visual display unit (VDU) (Tektronix 4010-l) and the GC-MS data are displayed either on the VDU or on a line printer (CentronicslOl, 132 columns at 165 charaeters/sec). The minicomputer is linked by a telephone Iine to an IBM 370/165 computer, where a library search based on the method developed by Naegeli and Clerti has been implemented. A data base of 3O$.KlO mass spectra based on the MSDC and Wiley collections of spectra is available. SAMPLING


In a previous papef,

we described the sampling of organic air pollutants


dynamic enrichment on an adsorption column. This technique was judged to be sound by many other workers7-z6 because: (a) the adsorbed compounds can be thermally eluted, thus avoiding the need for ultra-pure soIvents; (b) the adsorption and desorption procedures can be easily automated; (c) the dynamic enrichment allows both complete adsorption of substances that have a retention volume lower than the sample volume passing through the adsorption column and, for more vo!atile compounds, au equilibrium concentration proportional to the concentration in the original _wple; and (d) “&e sampling can be performed in siti.’ From the various adsorbents available, a porous polymer, Tenax CC” (poly-p-2,6-&phenylphenyleneoxide, surface area ca. 30 mz/g, 60-80 mesh, manufactured by AKZQ, Arnhem, The Netherlands), was chosen because it is hydrophobic, has a low adsorption strength, allowing the thermal elution of relatively high-boiling compounds (e.g., terphenyls), has a good thermal stability and does not react with most organic pollutants. A glass column (CQ. 15 cm length, 1 cm LD.) filled with 2.5 g of Tenax wds adopted. The column is preconditioned at 350” for 3 h and subsequently at 200” overnight under a flow of high-purity helium. The air sample (240 1,according to the contamination level) is aspirated in sifu into the porous polymer adsorption column (Fig. 2) by a pump at a speed -of 0.5 l/min. After sampling, the column is closed at both ends and taken to the laboratory for elution and analysis_ water ln contrast to a&, for which adsorption seems to be the most promising sampling techniquei many other sampling procedures have been envisaged for water, e.g., distillation, fmeze&ying, liquid-solid adsorption, headspace analysis, gas-phase strippin& andbatchwise and continuous liquid-liquid extraction. The relative merits and drawbacks of.many of thm techniques have recently been discussed by several

lary ‘He

tube (ghxss1inedSS.l

Fig_ 2. Flow scheme of sampling and dution system.

workers, particularly Mieure and Dietrich”, GrobLg and Bertsch et al.“‘, who indicated that liquid-solid adsorption on porous polymers or gas stripping followed by absorption on charcoal or porous polymers, with subsequent thermal or solvent elution, is very suitable for the sampling of organic substances in waters. Thegasstripping technique has been applied by several workers’0-*3 and a closed loop system has been used by Grob et aL4. Liquid-solid adsorption OQpolymers has been adopted by several others”-27. For water, we use the same system as for air (Fig. 2). After several experiments, we increased the dimensions of the sampling column to contain ca. 4 g of Tenax mixed with 20% of 6Q-80-mesh glass beads in order to facilitate the flow of water and to prevent cracks in the column bed during the elution (heating) &cess. These cracks must not be present, otherwise preferential routes are formed and the water sampled is not able to come into contact with the entire polymer bed, resulting in poor recoveries. Normally, 0.5-1.5 1 of water are sampled, depending on the c~atamination level. Water is aspirated in situ into the porous polymer adsorption columu by a peristaltic pump at 5-7 ml/n&. After water sampling, the column must be dried so as to prevent the residual interstitial water from disturbing the GC-MS operation. The best procedure, which does not cause severe losses of compounds, is to leave the column overnight in a de. siccator containing phosphorus pentoxide at 10 mmIIg pressure. Efution

For both a!- and water, the elution of the adsorption columns is carried out thermally in a sense op_msi’te to that of the sampling (Fig. 2, b-ta) with a flow of helium of ca. 15 ml/n&r. During the elution at 270” (20 min for air and 40 min for water), the eluted compounds are trapped in a glass-lined stainless-steel U-tube (I/16 in. for air, l/8 in. for water) at -180”. When the elrrtion is completed, they are flash-heated into the two GC glass capillary columns_ The flash temperature of 300” is reached in 15-20 sec. Fig. 3 shows the actual configuration of the elution system. Recovery

Tables I and II give some recoveries for model compounds in air and water.




Fig. 3. Actualconfiguration ofelutionsystem_1 = Tenaxcolumn;2 = oven; 3 = U-tubecondenser; 4 = flashheatercables; 5 = dual-columngas chromatograph. TABLE I RECOVERY OF MODEL ORGANIC POLLUTANT COMPOUNDS FROM AIR Compounds&xxbed on Tenax at 20” and desorbedat 250-270” undera fiow of helium. Compound


(%) n-Hexane LightpP,troIeum (b-p. XU-100”) Lightpetroleum(b.p. 100420”) Diphenyi Acetone Diphenytaraine -01


100 90 92 96 97 93 90

e Mean relativestandarddeviation:5 % at the ppb level.

In air, the recovery for compounds as volatile as n-hexane is 100% provided that the retention volume of this compound at 20” (abont 30 I) is not exceeded. Compounds of different polarity can easily be adsorbed and eluted from Tenax with a mean reIative standard deviation of about 5% at the ppb’ level. In water, the recoveries of compounds that arc very soluble, such as pyridine and phenol, are poor. Tests were made with spiked water, prepared by dissolving the model compounds in methanol and adding them to water that had been purified by passing it through a Tenax column. The recovery tests were normally performed at a pH value near to neutrality;


Throughoutthisarticlethe Americanbillion(104 is meant.

.~ i _- .’ .: ~. :. _.(..,‘J... -::-; : :,,:: &; y1-:..I -..: .: -:.:.-. -.f: : .-:1; _.-; -1. -y.:-y y; : ‘-.fp -.;;;..;.: -.. -.-.. -. -..‘. .. --:. -. Y. : ~. ;y.-; -1.. -. -: ~-,--.;.,.;, .y ---.: ._. .:~-;. ,‘f;tiB’g$II..: :,;~;.~.I .:.-.. . ..-.: _-~. -. :..&&+&- .:. ;&$.pA&EL&&&p*L~L&~ ~Mg.$&&&*M +&& -;:.- .’ &m&&& z&i&b& at 206 by &s&g uadtiati~+qfheIitlm.~ .-.

Cotipad :

;A&M (Pm



Pyridiae n-c,. Aniline Phenol P+-~-

48.7 -14.2 .- 30s 19.8 19.9.

wa&r d&r%& through Tee - _-

Meen recovery -("/of

-g&&f&g starwkrd fievwJf Phi


-30 3 s8 14.

4: 70 89 45 89


-&d d&bed I


zk.25&& -....

.:,.. . ..- :

_ .-._ .

_- '. -:










.- _-



.. -

other experimentscar&d out at acidic or basic &I +alu& did not substa&ahy irkprove th&-recoveries. In orderto increasethe recoveryfor water-soIublecomp.ounds; one could envisagethe use of ttio Tenax cohnnnsin series.Benzene,in spiteof a good meatirecovery,has a comparativelyhigh relativestandarddeviation; dnringthe tong drying of the~cohrmnin the desiccator,some losses of the more volatile compounds might occor. S!s?AR.ATION


The analysis of compIex mixtums of organic pollutauts in air and water.remeet this requires a high-resolutionseparationsystem4-W.Glass capillaryc~ohuuris quirementfor severalreasons: (a) they have a good separation.power,(b) ‘+ey -allow t&e elution of &lativeiy high-boilingcompounds at relativtiiylow temperatnres,(c) the amount of materialthey can handleis in &e range obtaiuedby most of the actual sampling-enrichmentsystems and (d) the carrier gas fiowyrates(24 ml/h) are compatiblewith the pumping capacityof most of the presentlyavailablemass spectrometersfor a directGC-MS connection. Two ~hnnn.s in parallelare employed, with differentstationaryphsses (Norm mally OV-101 and OV-225 or SolarIOC) in orderto ev&hratithe Kovgts retentionindices([email protected]) for the GC [email protected] of compound class&. The d1 values are norr%a$f comparedwithreferencevalues obtaikdon the same stationaryliquidson a separate -GC system”. The-gfasse~pihary&hrmm~weredrawnfro&-s&a-glass tub&of 8 mm 0-D and LD. using -A glass--d&wingmachjne: Mornrallycolumns of 60 rn-fe&tb andi0.3mm I.D._are&d. ;-,. ’ &vek&ipr~dures. and mater& & used for &lumn pm-tr&tment5 etchjng withhydio&loric ahdtg21, [email protected]~~,~surf~tants~~~, de&sition~ofpa&ct+sueh and itat+ : s- Siianox 101” and Carbowax=-=. SiinihrIy,f& &hnnn fiklitig~dynami& Gethods tie..used.-ye- obtainedi&ebest:[email protected] nsingthe foffo&ng conditiok.’ -ForOV-1Oi cohunns,+e [email protected]&ies [email protected] [email protected]@&h.ane [email protected]@d:three timeswith a .l”/, ‘y-(&Fl%?C)* di$do~oinethtie. After drymg,;[email protected]?$ so+iop of GV;;lOl...rnm jr-oefane-is ::.. ..:..:-

~3k~&&0~~~ &~a._&cn&b;

A c&u& pq#e&2x this maker is &od t^oretutti’lbl&dkacaFfxx~s, basr’c-co&ou~& such aS amlike anb d5netiyiamZne anoTweakfy &%ib i=arn&ti& sticli as dikrctbyfphenofs and afcbhofs. However, strongfy aci&c

compounds such as phenols and &orophenois aIre ehzted with some taihng. The use of Carbowax ZJM; instead ofBTppC &d nof improve the~elutfon ofthese fatter com=-xxeB?? i I? For GV-225 or Sifar ICK! cofumns,~ the soda-g&s capiEa&s are etched at 500” for 2% h with gaseow hydrogen ehioride. After [email protected], the residual hydrogen chforideis removed with drynitrogen and the columns are se&d until they are to be co&e& A IOx s&ion ofthe stationary phase in [email protected] or chiorofWm iIs &ssed through the cohunn (40 cm/mm). These simple procedures are reproducible, and- although they do not give optimal cohnnns, they represent a good compromise for our purposes. A 60-m column normafly exhibits a separation number* of 30-40 relative to GL3-CId hydrocarbons. GC-MS


As the carrier gas flow-rate in the g&s capilhzry columns used (ca. 4 ml/mm) can be managed- easily by the pumping system of the MS, direct GC-IMS coupling was decided upon. MOreOveF,it was considered desirable to have (a) at the end of each c&umn, a splitter to an FID in order to cuntrol the eluted compounds with a sensitivity higher than that of a normal total ion current monitor and to facilitate quantitative measurements, and (b) the two columns co;mnected alternatively and on-line to the mass spectrometer. The scheme adopted is shown in Fig. 4. id A1


: _:__ ---: __&+__&,_ mv B


L_,__ vent

Fii 4. Gc-MS “k3rt CUttiIlg~ COnnection. The mass spectrometer is connected to the GC columns by a platinum capiilary tube (0.15 mm E-D., ~(1.50 cm Iength), allowing a Flow-rate of?-4 &/tin. [email protected] leaving columns A and B always enter the GC detectors, whereas whether 0; tiot they enter the mass spectrometer depends on tie “‘heart cutting’“~conditions, Le., & helium flow-rates in a and b. These flow conditions are regulated by two ekctromagnetic v&es (_tiv> bmv) and the _restrietors r. With an appropriate helium

GC and MS data acquisition, aspwell as the MS scan, are o&r&led by the GA SPC 16/fi5minicomputer”. A commencement dialogue permits one of three operation modes (GC, MS, GC-MS) to be selected aad to change or_save previously established experimental parameters. The MS scan takes co. 27,13 or 9 samples per mass_unit instead of jumping from peak maximum to peakmaximum; The Iattet. method tends to introduce major intensity variations owirrg to changes in peak shape at the relatively high&B working pressure during GC!:MS operation. j . IQ order to save minicomputer time, an interface (PEGASO), constructed by o-urElectronics Division39,digitahy smoothes the MS ion &rre~t signals-and detects peak maxima. Thus, sufficient computer time is available to reduce on-lice data from up to 4 GC! channels (sampling rate 20 Hz per channel) during MS operation; hilassspectra are continuously taken at a scan rate of 600 a.m.u./secduring GC peak elation, and at variable intervals otherwise (every i-1 5 see). This can be done without loss of information, as the mass spectrometer is less sensitive than the-FID, unless it is operated in the mass chromatogram mode, when it is run ?&ninuousiy. The GC data are reduced on-line. For each peak, its number, the ehttion time and height of the start, inflexion, maximum and end-points, the first derivative at the infiexion points, and the front and back half-areas are calculated on-line and stored OP disk. The- separation of fused peaks, the choice of the separation algorithm @e$endicufar drop, ILange+), the baseline corm&or!, the determination of corrected p&k area and the calculation of Kov+_tsretention indices are performed off-Jiine.An anaIysis report is output 0% the line printer. s&d.m ti stored directly on a disk and processed off-line. Mo& than 3(wo speCtracan be stored for a single GC-MS run. Processing is display-oriented. With simple comm~ands,reconstruct&i gas or mass chromato_mamsare generated. By tieans of the cursor OQ t&e YDU, they &I be expanded and spectra at interestingpoints, together with related background spectra, en be ‘selected;Single -or averaged background spectra can be subtracted._ &I data ac+&itiosi and pro&s&g software h-&been [email protected] developed, _ is very aexible and can easily~beadapted to hew problems,


n air Sample poUuted with car exhaust

1g_ %33%s*Q .i!t af.


t&611 [email protected]~g a GC-MS run. A re_port is~givenon the finep&ter &ztainingGk pe&: numbersin increasingorder, the numb&rsof those associatedmass speqtr&fqr-which one (or several)hit(s) has been found, the identifiedcompound names, tkkespective Kovde retention indices and the similari~ factors-S; This job- is run efitirelyon theminkomputer~ Fig. 5 shows an F chromatogra&(above)and the correspondingkconkucted chromatogram(below) of a car e_xhaust pollutedair sample.The names indicated were Figned by use of the reversesear& library. The interpretation of&e mass spectraof GC pea.¬ identifiedby the reverse search is effectedin a second step, for which the Clerc library-searcha!go&.hnP~ was adopted. This algorithm was designedparticularlyto IZndsimi!arly~ structured compounds, even if the spectrumof an unknown compound itself is not available in the hi. Mass spectralfeaturesratherthan orig$al mass spectraare compared with a featurelibraryderivedfrom the MSDP atid, recently,the Wiley1 c&k&ions of spectra.The sum of the featuresof a spectrumis called a signature.Featuresare coded in one bit, Le., they are eitherpresentor absent. Signatur+ of unl&~own qectra are generatedby the minicomputerand, up to 20 at a time, transferredto our calculation centre, wheret&e searchalgorithmand the library are implementedon an IBM 370/165 computer. The resuitinghit lists of thenten best f&s are output on the line printer. Fig. 6 shows two starch results,which were obtainedwith the MSDC library only. The upperset (a) is the hit list of a compound that is obviously a C+ubstituted

Fig. 6_ Printouts

of MS idk~tion








be-e_ -&in& to the similarity of~tke difkqt [email protected]~ t&e assignment of etkylbenzene as the correct answer was made on the basis of Gc retention data. This_examl$e is o&z of our early attempts and the ethylbenzene spectrum is now included in the reverse searck Irbrary. Fig. 6b reports a i&s evident example. A matck oftke unknown compound against the entire MSDC library resulted in a compietely non-significant hit list with very different and complex &mpounds, ar,ar-dickloroanisoIe being in the last position. As a chlorine pattern classiiler gave char evidence for the presence of cklorine in the unknown compound, a second match against the sub-kbrary ot‘ only cklori%-containing compounds was made and gave the result shown Fig.-6b. Tn spite of the relatively bad match factors, four of the first five kits have a dicklorobenzoxymetkyl group in common, wkich was useful evidence in identifying the unknown compound as the pesticide 2,4-D. The last example shows that pre-&ssi&ation of an unknown spectrum45may be useful not only in saving computer time by reducing the library to be matched, but also by increasing the structure elucidation power of a search result, especially when only relatively poor similarity indices are obtained. The application of further classifiers is being explored. The identification of unknown structuresby a library search may be complicated if the names of the kit list members, which are often trivial, do not indicate rcfated structures.Tkis [email protected] can be circumvented if tke reference data Gle contains a structure code, suck as the Wiswesser Line Notation (WLN), which is available for tke Wiley iegistry. Witkin the framework of a data bank for environmental chemicals (a description of this data bank, ECDIN, is given in refs. 52 and 53), the Crossboti’ p,ograrn developed by LCI was implemented on our IBM 370/165computer. Via teleprocessing, it generates, from the WLN, semi-structural formulae that can be displayed either on tke line printer or on the VDU. Using tkis option, sub-libraries of selected compound classes required in connection with tke classification of unknown spectra can easily be ienerated. Tke structuralfragment option also offers an additional criterion, apart from similarity indices, for the assessmentof a search result. As illustrated in Fig. 6b, a structural fragment or a combination of suck fragments that is common to several members of a hit list can more reliably be supposed to be present in the unknown compound. Tkis option also needs furtker exploration. PRACIZCAL


Fig. 7 shows the effectiveness of the sampling of air on Tenax. Ckromatogram 3 is the result of the sampling of 2 1of air from near a car parking area on Tenax, and ckromato_gram C of the sampling of 40 1 of air in a woody area far from direct contamination by car exhaust fumes. Both samples were taken in winter. By comparison of ckromatograms B and C with ckromatogram A, which representsa standard mixture of gasoline plus diesel oil, it follows tkat both air samples are polluted by little except gasolinr$ “ml*. Tke iarger sampling volume of 40 1 for the sample from the woody area accounts for the higher proportion of high-boiling materials in ckromatogram C tkan in A. Tke peaks numbered in ckromato,v A, which were also identified in ckromatogr&s I3 and C, are listed in Table IIf. ~-Fig. 8 representsthe chromatogram of an air sample of 40 1taken in a woody

Fig 7. ?rofSs of &auk vohtiies in air. (-4). Standard rkture of @s&be &is diesel 02; the nurrkred peaks ue those &xT2bk III; (B). 2 iofairfmmacarparking~(~~);(C),401of air fmm 2 woody area (winter).Glass czpilhry column, 60 m x 0.3 mm ID_, OV-lot. Helium flowDetector: FlD, rate cu. [email protected] Temperature progr2mme: 40” for 5 II&l; 40 to 250” at 2”/dn_ 270”.

area duringthe [email protected] peaks are those [email protected] Table III, and again: the air appearsto be polluted tiostly by gasoline. Neverth&!ess,_the @romatow ontains some dZemxes comparedwith that for the wintersample(Fig;-?; chromato_c): a smalls pi&&on of ligh~com~unds ad SO~F r;nknO~n @.mak&dwith ti a&Sk, probablybelongingto -tie hydrocarbonenissfoos‘from the t&s. The ident&ation of &se conqo~ds is in progress. Fig. 9 shows the eEe&venF of water sampling, Chrematagq& 4 is the


40 &rs wmay


(zcea [summerl

-_ 36




gasoline and diesel -oil superimposed on the “fingerprint” of the tip water,The n&nbered peaks refersto those listed in Table I_& Chromat0gram.s B and C show the organic componnds present in 1 I of tap water and in I 1 of Lake Maggiore surfwater. The two chromatograms appear similar and in fact Lake Ma&ore is used as a source for this tap water. Comparison of B and C with A makes it evident that both contain compounds present in ,aofines5 (marked -with asterisks). Contrary to reports by some workers 22,it does not seem that tap water, in spite of the chlorination process, contains more compounds than the precursor. This is probabiy due to the mild chlorination process used, which, performed at a pH vahre of ca. 8, does not cause easy incorporation of cblorin,a’ into the aromatic compounds (e-g., those befonging to the gasoline type)%_ Chromatograms D and E show the organic compounds present in 11 of bottled mineral waters from two different sources. Both samples appear “clean” and contin, surprisingly, similar peaks. A possible explanation could be that these compounds belong to the plastic tubing U,SXIin the bottle fzlling system or to the plastic liners of the stoppers, the last huge peak having in fact been identified as a phthafate.





Fig. 9. Prof%s of organic compounds in water. (A), 1 I of tap water spiked with 10 ppb of gasoline and IO ppb of dieseloil; numberedpeaks are those in Table HI; (B), 11 of tap water; (Q 11 of Lake Maggiore surface water (winter) (this watter is used to produce tap wafer B); (D), 11 of bottled mined water from utmpany LSCR; (E), L 1 of bottled mineral water from company SB; (F), Tenax column bIank. Chromatographic conditions as in Fig. 7. ACKNOWLEDGEMENTS

The authors are grateful TVMr. De Boni for his valuable technical help in making the glass capillary columns. REFEP.ENcEs I B. Versino, & Kn6ppel, M. de Groat, A. Peil, J. Poehnan, H. Shauenberg, H. Vissersand F. Geiss, Eru0anatysi.sII, Second European Conference on AnaiyticaI Chemistry, Budapest, 25-N August, 197.5. 2 S. R HeJler,J. M.

McGuke and W. L. Budde, Environ. Sci. Technol., 9 (1975) 210. 3 European mnomic Community, Cooperation and Cooniinatioti in the Field of Scientificand Technical R esearcb, Action COST 64b, AnaIysis of Orgamk Micropollutants in Water, EECCOST Reports EUCO/MDV/23/74 (AIUZKQZReport 1973) and EUCO/MDU,!44/75 (AIKK~~ Report 1974). Brussels. 4 K. Grob, % Grab, Jr. and G. Grab, 1. Chromatogr., 106 (1975) 299. 5 P. R. Naegeli and 3. T. clerc, Anal. Chem., 46 (1974) 739A. 6 3. Ver&o, M_ de Groat and F. Geiss, Chromafographia, 7 (1974) 302. 7 R. E. Raiser, in A. ZIatkis (Editor), Advancesin Chromatography 1970, ChromatographySymposium, Houston, Tesas, 1970, p_ 21. 8 A. Drhieks, B. K. Krotoszynski, J. Whitfield, A. O’Donuel and T. Burgwald, Environ-sci. Teehnol., 5 (1971) lZZO_

.- -3*8..-m _ -1~

~. -_


; :

-.&&kS~~&nl.~ .._

-9 i. Gmb an&G. &ob; J. Chrcmiitogr.;62 (197f) I.-. :-‘.I:--I -‘- .-i 10-W. A. Alte and-P. M:Teli, J.~ Chro&zrogr., 62 (1971) @-. -.. .: : _-- -I-.‘- ‘.I ; ,_ 11 $E. -Kaiser,-An& C&n-, 45 (1973) 9655. : ..---. :. : .12 &%&is, K. A. Lichtesteinand [email protected]& C~%o/n&o&zp&, 6 (1973) 67.. -f3..v. Bert&, R C. Chsszgand A. Z&&is, J. Clrromatogr.SC& l;! (1974) 175. ‘-; 14 A. Raymond and G. Guiochon, Environ. Sri- Tec&w& 8. (1974) 143. -. 15 A. R2ymoncj tid G. Gtiochon, J: Chroma~ogr. ,$ci., 13 (1975) 173. 16 E. D. PeUizzari,J. E. By&, B. H. Carpenterand E. !&g&i, &iron. Sci. IT&&& 9 (1975)352. 17 R, vati Wijkj in A. ZIatkis (Editor), rldsancesin [email protected],Chmtito&aphy Symposkmn, Houston, TexaS, 1970, p. 122. IS,.J. P. Mieure and M. W. Dietrich. J. Cfuonxzro~. Sci., li (1973) 559.. : 19 EL Grab, J. Cbromatogr., 84 (1973) 255. 20 W. Bertsch, E. Anderson and G. HoIzer, J. Chromarogr., 112 (1975) 701. -21 T. k BelIaratid J. J. Lichtenberg,E.P.A. &purr, 670/[email protected]& Natiooall&vir~nns&tal R--n& Center, Cincinnati, Ohio, November 1974. 22 B. J. Dowty, D. R. Carlisle and 3. L. Laster, ~fiviron. Sci. Te&wl., 9 (1975) 762. 23 A. K. Bumhun, G. Vi CXder, J. S.-Fritz, G. A. Junk, H. J. S&c and R WiIIis, A&. Cheni.. 44 (1972) 139. 24 .G. A. Junk, J. J..Ric’hard,M. D. Griesq, D. Witiak, J. L. ~itiak, M. D. ArgueIIo, R. Vi&, H. J. Svec, J. S. Fritz and G. V. Calder, J. Cbromarogr., 99 (1974) 745. 25 C. Osterroht, J. Cbromatogr., EC1(1974) 289. 26 P. R. Musty and G. NickIess; J. Chromafogr., 89 (1974) 185. 27 V. Leoni, G. Puccetti and A. GrclIa, J. Chromcrogr., 106 (1975) 119. 28 G. Schomburg, H. Husmann 2nd F. Weeke,-& Chromatogr., 112 (1975) 205. 29 M. Novomy and K-Tsar&, Chromat&rapkta, 1 (1968) 332. 30 G. Alexander, G. Garz5 and G. P&lyi,J. Cbromatogr., 91 (1974) 25. 3 1 G. AIexznder and G. A. F. M. Rutten, J. Chromarogr., 99 (1974) 81. 32 M. Novotny and A. .Zl2tkis, Ch~omatogr.Rev., 14 (1971) 1. 33 E. J. Mak, J. Ch7ol;mto,or.SC& 9 (1971) 318. 34 G. A.- F. M. Rutten and J. A. Luyten, J. Cfirorn~fogr.,74 (1972) 177. 35 A. L. German and E. C. Horning, 1. Chromarog&- Sci., 11 (1973) 76. 36 D. A. Cronin, 3. Ctiromarogr., 97 (1974) 263. 37 G. Schomb-urg,It Husmaon and Fe Week-e, J. ~krotxzogr~. 99 (1974) 63. 38 H: KniippeI, B. V&rsino,W. G. Town, He Schsuenburg,A- PeiI, J. PoeImzn and F. Geis, An.nuaJ Meeting

39 40 41 42 43 44 45 46 47 48

49 50

of the ~sArbeirsgemeinschaft Massenspektrotnetrie”,

Schioss Resensbarg,

G. F. R, .29.9.-

1.10.1915. S. Amic, F. SoreI and A. Termanini, AnnzralReport Z97#/75, Commission, of the European Communities, Joint Research.Centre, lspra, in preparation. & S; Hert& g. A Hites znd K. Biemann, And Chem, 43 (1971) 681. S. L. Gmtch, A&. @sn., 43 (1971) 1362. S. R. He&r, Anai. Chem., 44 (1972) i951. K. S. Kwok, R. Venkataraghavanand F. W. M&Lafferty,J. Amer. C/&z. Sot., 95 (1973) 4183. T. L. Isenhour and P. C. Jurs-Anal. Chetn:, 43 (1971) MA. J. Franz+, Chromatogrcphiq 7 (1974) 518. G. Si Zander, A. J. Stupcr and P. C. Jurs, AJZZ$.Ciiem...47 (1975)‘1085. A. Bucti, A. M. DuflieId, G. SchroIl, C. Djerasi, A. B. Delfino, B. G. Buchanan, .G. L. Sutherland, E. A. Fei&neaum and J. Lederberg,J. A;ner. C&n. Sot., 92 (1970) 6831. F. P. -4bramson, Anal. .Ch,m+ 47 (1975) 45. J. T. Clerc, E Emi, C. Jo&, J. Mei& P. NZgeli and R Sc&wz&&?c& z. Anrii c&g-, 264 (1973) 192 Mim Spectrai Lib&~, amass Spectmme& Data Centre, A&, .AIdermaston;.Great &i&in, 1972. .’

51. Mms Spectrai Data C&ection, Wiley, New Y&k, 1975. 52 F. Geis and Ph. Bourdezu, ti F.. Cm&ton and .F.. Korte (Ed2ors);B&romne~ZaZ -Safety, Vol. >;-1976, p. 15.

[email protected]


53 M. Bogi, F. Geiss, J. IX Petrie2nd W- G.-Tok P~&e&z&&&&&f2 C&f&e&; Amrterikan, .23-25 Mzrch, Z976, +SLIB, London, tti be p&Ii&e&~ ‘1. : .‘54D.R.~ E. Hy$& and G. Palmer, Pestic SZ,. 5 (f97ij 319. .._. : .. ~. ~. 55 K. Gmb a;?d G. Grab, J. Cbro&z+~&+ 90 (1974) 303; _ -~ 56 R. M. tZar+on, R;E. Car&on, H. L. Kop.w.and+ C.apIe~E+~n;‘~~~ T;cF;M&l97j).67~.[ :. ..’ ... ..-