ANALYST, APRIL 1989. VOL. 113 429 Investigation by Gas Chromatography - Mass Spectrometry of Potential Contamination Incurred by the Use of Crimp-cap Vial Closures* Stuart J. Pattinson and John P. G. Wilkinst ADAS, Harpenden Laboratory, Harpenden, Hertfordshire AL5 ZBD, UK Nine types of crimp-cap vial closures were examined to determine the propensity of their septa t o discharge contaminants into sample vials under a range of conditions designed to simulate those that could occur in use. Gas chromatography - mass spectrometry was used to identify the solvent-extractable components. The presence of an unpierced polytetrafluoroethylene facing reduced the degree and rate of sample contamination. Sixty components identified at significant concentrations in the solvent extracts of the vial closures were characterised.Gas chromatographic retention times, quantitative and mass spectral data are presented. Of potential importance in the field of pesticide residue analysis was the identification of ethylenethiourea and diphenylamine as possible contaminants associated with certain vial closures. Keywords: Contamination; vial closure; septum; gas chromatography; mass spectrometry The automated analysis of samples by gas (or liquid) chromatography requires the presentation of samples in a uniform manner and in a form that is easily manipulated by the equipment involved. Sampleiextract vials are closed with screw- or crimp-caps that incorporate a septum, allowing aliquots of the sample to be withdrawn through a fine needle with a syringe.Experience in this laboratory has shown that with certain combinations of solvent and septum material, contaminants transferred from the septum into the sample solution may cause interference in subsequent analyses. Interference from septum materials is not a new phe- nomenon. Since 1962 bbghost peaks” obtained in gas chro- matography (GC) , using temperature-programmed analyses, have been attributed to injection port septum contami- nants. 1-3 In such instances contamination of the vial contents does not occur and the interference may be removed by the use of a more suitable septum in the injection port. Compari- sons of various GC septa have been made in an attempt to assess those that produce the least amount of contamina- tion.’ 5 These studies yielded only comparative data and no attempt was made to identify or quantify the compounds causing the contamination, Another problem, that of sample loss encountered in the GC analysis of moisture-sensitive compounds, may arise through the use of GC septa that contain trace amounts of water.6 To produce the physical properties required of a septum, the incorporation of a variety of compounds (initiators, accelerators, fillers, antioxidants, plasticisers, extending oils, lubricants, etc.) in the elastomer is necessary.The leaching of such additives from products fabricated from elastomers has resulted in contamination problems in bacteriological ,7 phar- maceutical ,X agricultural9 and medicall‘k 13 applications. In trace analysis, the occurrence of contaminant peaks in chromatograms can be misinterpreted as indicating the presence of components of interest or importance. The subsequent identification of these peaks, and/or the repeated analysis of the sample, wastes resources and reduces the efficiency and cost-effectiveness of the automated equipment.To prevent the transfer of compounds from the septum into the sample, many vial closure septa have a poly- tetrafluoroethylene (PTFE) layer, which faces the sample when the closure is attached to the sample vial. This impermeable layer minimises the transfer of contaminants from the septum into the vial contents (and the absorption of * Crown Copyright 1988. t To whom correspondence should be addressed. the vial contents into the septum). However, when this layer is punctured by the syringe needle, the seal is destroyed and transfer may occur.Repeated piercing could cause an increase in the rate of contamination. Handling and storage of extracts in vials before and/or between analyses may result in direct contact of the sample solvent with the septum, providing an opportunity for liquid-phase transfer of contaminants into the sample. In the most extreme instance, cores of the septum material may be introduced into the vial by the sampling needle, particularly when wide-bore syringe needles are used. Sample concentration may also occur because of loss of solvent through the punctured septum.14 A selection of nine types of aluminium crimp-cap vial closures from four suppliers was examined (see Table 1). The vial closures incorporated four types of septum material: natural, silicone, chlorobutyl and fluoro rubbers.The major extractable compounds were identified or characterised by mass spectrometry (MS). The responses produced by these compounds on various element-sensitive detectors were also evaluated. To investigate the effectiveness of PTFE facings in reducing contamination problems, the rates of extraction of selected component compounds from septa without and with PTFE facing (both intact and punctured) were determined under a range of conditions. The compounds identified included: bis(3-tert-butyl-5-ethyl-2-hydroxyphenoxy)methane (A0425) and its oxidation product (A04250); butyl hexade- can o a t e (B H) ; 2,6- di - terf- b u t y l-4- m e t h y 1 p h e no 1 (B HT) ; but y 1 octadecanoate (BO); butyl heptadecanoate (BP); bemothiaz- ole (BZ); 2-(3H)-benzothiazolone (BZO); dibutyl phthalate (DBP); 9,10-dihydro-9,9-dirnethylacridine (DDA); di(2- ethylhexyl) phthalate (DEHP); dimethyl phthalate (DMP); diphenylamine (DPA); 2,6-di-tert-butyl-4-ethylphenol (EDBP); ethylenethiourea (2-imidazolidinethione) (ETU) ; alkanes (C,Hbt2, x = 19-32); 1,2,3,4,4a,9,10,10a-octahydro- 1,4a-dimethyl-7( 1-methylethy1)-I-phenanthrenecarboxylic acid, methyl ester (methyl dehydroabietate, MD); 14- methoxy-3.6,9,12-tetraoxatetradecan-l-o1 (PEG); octadecyl acetate (OA); phenol (P); cyclic polydimethylsiloxanes [SiO(CH3)2], (Si-n, n = 4-22); triphenylphosphine oxide (TPO); and triphenyl phosphate (TPP).Several compounds were characterised but only partly identified (F-I and H-IIIII IIIiIV).1 Experimental Reagents All reagents were of analytical-reagent grade unless stated otherwise. Methanol, HPLC grade (May and Baker); ethyl acetate and cyclohexane, “Distol” pesticide grade (Fisons430 ANALYST, APRIL 1989. VOL. 113 Table 1. Vial closure details (all 1 I-mm diameter). Details of suppliers may be obtained from the authors Code Septum material (A) . . . . . . . . . . (B) . . . . . . . . . . ( C ) . . . . . . . . . . (D) . . . . . . . . . . ( E ) . . . . . . . . . . (F) . . . . . . . . . . (G) . . . . . . . . . . ( H ) . . . . . . . . . . (I) . . . . . . . . . . Natural rubberY Natural rubber* Natural rubber* Natural rubber' Natural rubber^ Natural rubber" Silicone rubber* Chlorobutyl rubber* Fluoro rubber * Septum incorporates PTFE facing.Scientific Equipment); acetone, "Pronalys" grade (May and Baker); BHT (PolyScience Corporation); TPP, gold label (Aldrich); DPA (BDH); and BZ (Fluka) were used as received. The compounds BH and BO were synthesised from hexadecanoic and octadecanoic acids (both from Hopkin and Williams), respectively, and butan-1-01 (BDH). All solvents and reagents were checked for purity before use by gas chromatography - mass spectrometry (GC - MS). Equipment All glassware (including sample vials) was cleaned in an ultrasonic bath, with five solvent changes, and dried at 80°C for 30 min before use. Mass spectrometric detection Capillary GC - MS. A Hewlett-Packard 5790A gas chromat- ograph with a Chrompack CP-Sil 19CB column, 25 m X 0.22 mm i.d., was used under the following conditions: injector temperature, 230 "G; carrier gas (helium) flow-rate.1 ml min-1; temperature programme, 40°C for 2 min then increased at 20°C min-1 to 100"C, held for 1 min, then increased at 10 "C niin-1 to 270 "C and held at this ternperaturc for 22 min. The capillary column was coupled directly to a JEOL DX300 double-focusing mass spectrometer operated under the following conditions: ion source temperature, 200°C; ionisation potential, 70 eV; mass range, m / z 20-500; and scan rate, 1 scan s-1. Packed column GC - MS. A Dani 3800 gas chromatograph with a 0.5 m x 2.5 mm i.d. column of 6.9% SE30 on Chromosorb W HP, 10&120 mesh, was used with an injector temperature of 230°C and a carrier gas (helium) flow-rate of 30 ml min-1. The column was coupled via a jet-separator to a VG 7035 double-focusing mass spectrometer operated under the following conditions: ion source temperature, 200 "C; ionisation potential, 30 eV; low-resolution selected ion monitoring (SIM); isothermal packed column GC.The compounds studied and the ions monitored are given in Table 2. Procedures Experiment I. Determination of solvent-extractable septum rziaterials Five of each type of vial closure were immersed in 20 ml of each of the solvents, viz., methanol, ethyl acetate and cyclohexane, in stoppered glass tubes and stored at room temperature. After 24 h the samples were examined for any visible deformation of the septum material. After 30 d the closures were removed from the solvents. The solutions were analysed by capillary GC - MS and the major components were characterised and quantified [either by comparison with standards, when available, or calculated from their total ion current (TIC) response].Table 2. Compounds studied and ions monitored by packed column GC - MS Oven Ketention temperature! time/ Compound Ions monitored "C S BH . . . . . . mi2 312 ( M + ) , 56 240 41 BHT . . . . . . miz 220 (A!+). 205 170 31 BZ . . . . . . m/z 135 ( M I ) , 108,69 130 47 TPP . . . . . . miz 326 (M' ), 325 240 69 BO . . . . . . m/z340(Mt).56 240 72 Table 3. Effects on \epta o f immersion for 24 h in various solvents Solvent Vial closure Methanol Ethyl acetate (A) . . . . (B) . . . . (D) . . . . (E) . . . . (F) . . . . (G) . . . . (H) . , . . (C) . . . . (I) . . . . No change No change No change No change No change No change No change No change Highly swollen Swollen * Slightly swollen: Slightly swollen Highly swollen Highly swollen Highly swollen No change No change Highly swollen * Swollen = 50% increase in diameter.t Highly swollen = 100% increase in diameter. $ Slightly swollen = cu. 10% increase in diametc Cyclohexane Highly swollent Slightly swollen Highly swollen Highly swollen Highly swollen Highly swollen Slightly swollen Highly swollen No change x-. Experiment II. Determination of the rate of extraction o j selected compounds from the inner surface of unpierced PTFE-faced vial closures into ethyl acetate Four vials were prepared for each of the eight vial closure types that incorporated PTFE facings [ (A)-(H)]; ethyl acetate (1 ml) was placed in a vial and sealed with a closure using a Wheaton hand-operated crimping tool.The sealed vials were stored in a horizontal position 5 0 that the solvent was in contact with the inner face of the closure. A sample of the solvent from two of the vials of each type was analysed by GC - MS, with SIM, after solvent contact periods of 1 and 70 h. The compounds determined were: DDA and BH/BO for (A), BH/BO for (B), BHT for (C), BH/BO for (D), BHT for (E), BZ for (F), BH/BO for (G) and TPP for (H). Experiment I l l . Determination of the rate of extraction of selected compounds from the inner surface of thrice-pierced PTFE-faced vial closures into ethyl acetate Thirty-six vials were prepared for each of four vial closure types, viz., (D), (E), (F) and (H), following the procedure used in Experiment I1 except that each septum was pierced three times (in different positions) with a standard 22 gauge (0.71 mm 0.d.) syringe needle with a 22 degree bevel. The vials were stored as in Experiment 11.The vials were sampled (in triplicate) after periods ranging from 4 to 70 h and the solvent was analysed for the selected cornpounds as described in Experiment 11. Experiment IV. Determination of the rate of extraction of selected compounds from the inner surface of unpierced vial closures (without a PTFE seal) into ethyl acetate Twenty-two vials were prepared, using (I) closures, and stored as in Experiment 11. The vial contents were analysed for BI-I and BO using GC - MS, with SIM, in duplicate at 5-min intervals. Results and Discussion The purpose of Experiment I was to investigate the nature and concentration of the organic components present in the septaANALYST, APRIL 1989.VOL. 114 43 1 Table 4. Calculated concentrations and capillary GC retention times (relative to chlorpyrifos-methyl = 1 .000) of compounds identified in the 20-ml ethyl acetate extracts of five vial closures [Experiment (I)] Nitrogen-containing compounds- BZ* . . . . . . . . . . . . BZO . . . . . . . . . . . . DDA . . . . . . . . . . . . DPA* . . . . . . . . . . . . ETU . . . . . . . . . . . . Phosphor us -containing compounds- TPO . . . . . . . . . . . TPP* . . . . . . . . . . Sulphur-containing compounds- BZ* . . . . . . . . BZO . . . . . . . . ETU * . . . . . . . . Other organic compounds- A0425 . . A04250 BH* . . BHT* .. BO* . . BP . . DBP* . . DEHP" DMP* . . EDBP . . F-I . . H-I . . H-I1 . . H-I11 . . H-IV . . Alkanes MD . . OA . . P" . . PEG . . Si-n . . . . . . . . . . . . . I . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . I . . . . . . . . . . . . . . . , . . . . . . I . . . , , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . * . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . , . . . . . . . . . . . . . . . . . . . . . Relative retention time 0.485 0.988 0.945 0.767 0.884i 1.452 1.153 0.485 0.988 0.884t 1.288 1.364 0.975 0.632 1.061 1.018 0.926 1.168 0.665 0.691 0.336 0.363 0.546 0.557 0.700 I 1.178 0.990 0.393 0.977 I Concentrationipg nil- I Vial closure type 10 (F) 10 1 20 5 1 0 1 50 10 5 10 10 10 2 5-10 5 2 5 50 50 SO 5 10 10 5 25 25 20 10 10 10 10 1 5 5 10 20 10 10 5 20 10 10 1 5 1 5 1 50 5 2 5 50 5-10 2-10 2-10 2- 10 1 5 5 0 5- 100 * These compounds were identified by their mass spectra and their identity was confirmed by comparison with an analytical standard. The 1- The GC peak obtained for ETU tailed badly. $ The relative retention times of members of the alkane and polydimethylsiloxane series were as follows.C,9-C3,: 0.797, 0.847, 0.895, 0.944, 0.989, 1.031, 1.072, 1.112, 1,150, 1.190, 1.237. 1.293, 1.359 and 1.439, respectively. Si-4-Si-21: 0.245, 0.331. 0.446, 0.558, 0.659, 0.747. 0.823. 0.893, 0.956, 1.015, 1.069, 1.120, 1.170, 1.226. 1.380, 1.491, 1.638 and 1.833, respectively. remaining compounds were identified solely by their mass spectra. that could cause interference on GC analysis. The choice of a 30-d immersion period was based on the observation that this produced an equilibrated solution. Only one of the nine types of septa examined appeared to be physically affected by immersion for 24 h in methanol; however, immersion for a similar period in ethyl acetate and/or cyclohexane caused swelling of all the septa examined (see Table 3). Extractable contaminants were obtained from all the vial closure types studied. The compounds identified and their concentrations (measured from their GC - MS responses obtained by analysis of the ethyl acetate extracts) are presented in Table 4.Results for ethyl acetate are given because, of the three solvents used, ethyl acetate was found to extract the widest range of compounds. Capillary GC reten- tion times, relative to chlorpyrifos-methyl (1216 s), are also given. Hydrocarbons and other organic compounds contain- ing oxygen, phosphorus, sulphur, nitrogen and chlorine were detected. All the vial-closure extracts contained BH and BO in almost equal amounts. The extracts were also analysed using a variety of element- sensitive detectors; all the extracts produced responses on an electron-capture detector (ECD).In all instances the com- plexity of the results obtained made direct correlation with the MS data impossible, although a series of negative peaks observed intermittently on electron-capture detection analysis of the (G) extract appeared to be due to the series of silicones identified by GC - MS. The lack of reproducibility of this result is still not fully understood but may be associated with the state of cleanliness of the detector. All but the (G) extract produced responses on a nitrogen - phosphorus detector (NPD). The seven major responses were identified as being due to the five nitrogen- and two phosphorus-containing compounds given in Table 4. No phosphorus-containing compounds other than the two mentioned above were detected using a flame photometric detector (FPD) in the phosphorus mode. No compounds were detected using the FPD in the sulphur mode, although three sulphur-containing compounds were identified by GC - MS.The most important findings in terms of pesticide residue432 ANALYST. APRIL 1989 . VOL . 114 Table 5 . Mass spectra o f compounds in "Eight Peak Index of Mass Spectra" format Compound A0425 . . . . A04250 . . BH . . . . BHT . . . . BO . . . . BP . . . . BZ . . . . BZO . . . . DBP . . . . DDA . . . . DEHP . . . . DMP . . . . DPA . . . . EDBP . . . . ETU . . . . F-I . . . . H-I . . . . H-I1 . . . . H-I11 . . . . H-IV . . . . HC-19 . . . . MD . . . . OA . . . . P . . . . . . PEG . . . . Si-4 . . . . Si-5 . . . . Si-6 . . . . Si-7 . . . . Si-8 . . . . Si-9 . . .. CI.H..O.Si, Relative Mass t o charge ratios molecular Relative intensities of most abundant ions mass 368 366 312 220 340 298 135 151 278 209 390 194 169 234 102 I00 180 214 214 292 268 314 3 12 94 252 296 370 444 518 592 666 Si-10 . . . . C~J&,~jO~oSil~. 740 TPO . . . . C.HH. SOP 278 TPP . . . . CIHH1504P 326 191 366 56 205 56 56 135 96 149 194 149 163 169 219 102 43 57 102 97 97 57 239 43 94 59 28 1 355 73 73 73 73 73 277 326 178 178 41 57 41 57 108 151 150 195 167 77 168 57 30 72 68 97 57 57 71 240 55 66 58 282 73 34 1 28 1 355 429 147 278 325 175 368 189 176 57 43 221) 206 57 43 41 55 69 82 123 69 31 104 193 209 57 71 76 135 167 51 191 234 73 42 82 39 97 41 57 99 99 41 41 55 85 99 299 314 83 57 39 65 149 104 283 133 267 356 429 342 147 415 147 221 147 221 281 221 77 199 77 65 163 57 163 57 29 55 145 5.5 55 29 73 271 63 91 45 70 40 223 97 192 70 41 92 50 84 77 220 159 72 40 27 53 69 109 41 67 56 109 69 99 55 113 129 41 69 97 51 55 45 76 73 265 357 268 430 343 325 282 356 281 355 430 429 355 51 201 170 233 M' 135 312 100 69 66 51 43 24 20 15 51 I91 175 100 95 80 75 65 52 55 52 100 73 257 100 78 68 66 48 45 24 22 8 177 105 100 30 25 17 12 7 6 6 25 285 73 100 78 76 58 47 42 28 27 17 61 43 I00 57 37 32 25 22 21 18 7 45 54 100 38 23 12 11 9 8 7 100 52 63 100 85 80 25 20 17 15 15 85 56 205 100 11 9 7 7 6 6 5 2 179 165 100 17 16 10 8 8 3 3 10 55 279 100 40 35 31 29 25 23 14 0 164 194 100 23 13 10 10 9 9 8 8 170 66 100 48 27 17 15 14 13 10 100 115 39 100 27 23 20 16 10 7 7 20 60 104 100 45 16 15 14 14 7 6 11)O 29 57 100 32 21 20 20 18 13 12 0 55 82 100 97 80 50 48 43 42 4U 12 104 55 100 83 81 48 37 32 28 23 12 55 69 100 92 75 64 46 38 31 28 7 41 127 100 75 56 27 22 20 18 17 5 157 173 100 21 14 13 11 10 9 9 13 61 41 100 83 78 76 71 66 58 57 0 63 70 100 43 30 30 7 7 6 6 100 207 132 100 63 15 13 13 11 6 3 1 191 207 100 27 17 14 8 7 5 5 0 269 358 100 74 53 37 26 16 10 8 0 431 147 100 77 52 26 23 18 17 17 0 83 123 100 72 27 24 21 21 18 15 7 341 283 100 48 43 18 16 13 13 8 .401 357 100 70 34 29 26 23 17 15 . 281 207 100 43 35 27 26 17 15 13 . 207 341 100 44 37 30 22 19 16 12 . 183 152 100 45 30 19 18 18 16 14 45 94 215 100 70 61 35 30 30 24 24 100 P H-Ill DMP Hc I I H-IV PI BH ETUDYP \ TPP 1 BO I 200 400 600 800 1000 1200 1400 Timeis Fig . 1 . Capillary GC . MS TIC chromatogram of the ethyl acetate extract of vial closure (H)ANALYST.APRIL 1989, VOL. 11.1 433 analysis (our particular interest) were the discoveries of ETU and DPA in the extracts of the (H) and (A) vial closures, respectively. Ethylenethiourea is a toxicologically significant metabolic degradation product of the important and widely used ethylenebisdithiocarbamate fungicides" and DPA is used as a post-harvest treatment agent.16.17 Clearly the contamination of samples with either of these compounds could lead to the production of highly inaccurate information. The presence of TPP in the (H) extract was also of specific interest to us, as it is used as an internal standard in this laboratory. The presence of these compounds presumably stems from their use in the formulation of the septum material: ethylenethiourea is used as a vulcanisation accelera- tor in the manufacture of chlorobutyl elastomersix; diphenyl- amine is added to vulcanised rubber as an antioxidant or as a vulcanisation acceleratorlq; and TPP is used as a plasticiser.20 Mass spectral data for the compounds included in Table 4 together with data for seven of the cyclic dimethylsiloxanes are presented in "Eight Peak Index of Mass Spectra21" format in Table 5 .The TIC chromatogram obtained for the ethyl acetate extract of vial closure (H) is shown in Fig. 1. Of the nine vial closures studied, six [(A)-(F)] used natural rubber as the septum material. The results of GC - MS analysis of the extracts of five of these types of vial closure [(B)-(F)] showed similarities. whereas those of (A) were markedly different.The five vial closures (B)-(F) all contained BH, BO and alkanes in the range C19-C32; however, each vial closure also contained additional compounds that helped to distin- guish it from the others. Vial closures (B), (C) and (F) contained BZ, (C) and (E) both contained BHT, (C), (D) and (E) contained A0425, whereas only (B) contained P. The main difference between vial closure (F) and the other natural rubber extracts was the presence of an unidentified compound in high concentration with a short retention time, designated F-I (plus, at a longer retention time, a compound that was identified as a methylated constituent of rosin, MD, which is added to rubbers to improve their adhesive quali- ties).22 Although F-I appears to have a relative molecular mass of 100 (by ammonia chemical ionisation MS), the molecular ion was not observed. The mass spectrum exhibits ions at mlz 72, 82 and 85, which are probably the [M-28]+, [M-18]+ and [M- 15]+ ions, respectively (produced by loss of carbon monoxide, water and a methyl radical, respectively, from the ionised molecule). This suggests that F-I may be an aldehyde or a ketone.In contrast to the other natural rubber vial closures, the (A) extract yielded several nitrogen-containing compounds (DPA and DDA being identified), and OA, together with BH and BO. The alkanes were not detected. The extracts of the vial closure incorporating silicone rubber [ (G)] contained a homologous series of cyclic dimethylsilox- anes (Si-4-Si-22). These silicon compounds resulted either from decomposition of the siloxane elastomer.23 or were present as contaminants introduced during the manufacture of the silicone rubber, as it is unlikely that they arose from room temperature solvolysis.The members of the Si-n series for n > 8 produced similar mass spectra over the mass range studied (mlz 20-500), largely because the diagnostic mol- ecular ion region was not acquired. The mass spectra of compounds Si-4-Si-10 are given in Table 5 . All members of the Si-n series studied produced a characteristic ion at miz 73, due to the rearrangement ion [Si(CH,),]+. The extract of the vial closure incorporating chlorobutyl rubber [(H)] contained ETU and TPP (mentioned above). A polyethylene glycol methyl ether, with the probable structure CH30(CH2CH20)sH (PEG), was found in high concentra- tion.The reconstructed chromatograms of the ions at miz 58 and 59 [due to (CH2CH20CH2)+ and (CH2CH20CH3)+, respectively. characteristic of these compounds] indicated a low concentration of a higher homologue with a longer retention time. Such glycol ethers are employed as surfactants during the emulsion polymerisation process used to produce chlorobutyl rubber. Other compounds whose mass spectra indicated that they were related to the structure of the rubber itself were also found; these appeared to be buteneiisoprene oligomers, some of which were chlorinated. The four major oligomers were designated H-I, H-11, H-I11 and H-IV. The mass spectrum of H-I indicated that it had a structure corresponding to (C5Hx)(C4H,),: H-I1 and H-IT1 were monochlorinated analogues of H-I; and H-IV corresponded to (C5HX)(C3H&.Several other minor GC peaks exhibited the miz 97 [(C7H13)+] ion characteristic of these compounds, but their mass spectra were too weak to report. The fluoro rubber (Viton) septum, (I), produced the least number of contaminants on extraction. Compounds that were identified in the extract included the two butyl esters, BH and BO, and TPO in low concentration. The compounds identified from the vial closure extracts include a range of elastomer additives. The DDA, EDBP, BHT and A0425 are all added as antioxidants in the formulation of elastomers. Several oxidation products were also found: A04250 (the oxidised form of A0425), BZO (the oxidised form of BZ) and TPO (the oxidised form of triphenylphosphine, which was not found itself).Phthalate plasticisers (DMP, DBP and DEHP) were identified in some of the extracts. The butyl esters, BH and BO, found in all the extracts examined are also used as plasticisers. The objective of Experiments 11-TV, in which the vials were stored horizontally, was to simulate in an easily controlled and reproducible fashion the effect of vial contents coming into contact with the septum material, e.g., by splashing or accidental inversion. The irreproducibility of such contact in a practical situation will hinder the identification of any sample contamination thus produced. In Experiment 11, analysis of the vial contents with vial closures having an intact PTFE facing showed that, after 1 h, measurable extraction had occurred with only three of the eight vial closure types examined (see Table 6).Comparison with the results from Experiment I, which gave an indication of the amount of solvent-extractable material in the septa, shows the effectiveness of the PTFE facing in minimising the transfer of septum contaminants into a solvent stored in the vial. Why the ratio of BH to BO observed in this experiment (1 : 3) differs from that found in Experiment I (1 : 1) is unclear, although it must be remembered that the concentrations of these compounds were close to their limits of detection. In Experiment 111, the rates of extraction of five selected compounds (chosen using the results from Experiment I) from four types of thrice-pierced vial closure were determined.The compounds were detected in the contents of all the vials examined. The rates of extraction varied widely according to the vial closure type and compound determined. Benzothia- zole was extracted very slowly compared with the other four contaminants studied. As the extraction time increased, the variation in the measured concentrations of the extractives also increased. The average concentrations detected after 70 h are given in Table 7. (It should be noted that individual values Table 6. Average concentrations of selected cornpounds in 1 ml of ethyl acetate sealed in vials with intact PTFE faces after horizontal storage for 1 and 70 h (Experiment TI) Concentrationipg ml- I Vial closure Compound After 1 h After 70 h (B) . . . . . . . . . . BH 0.1 0.2 BO 0.3 0.6 BH 0.1 0.1 I30 0.3 0.3 (E) ., . . . . . . . . BHT 0.05 0.05 (H) . . . . . . . . . . TPP 0.1 0.1434 ANALYST, APRIL 1989, VOL. 114 Table 7. Average concentrations of selected compounds in 1 ml of ethyl acetate sealed in vials with thrice-pierced PTFE-faced septa after horizontal storage for 70 h (Experiment 111) Concentration/ Vial closure Compound pg ml-1 (D) . . . . . . . . . . . . BH 20 (D) . . . . . . . . . . . . BO 20 (E) . . . . . . . . . . . . BHT 1s (F) . . . . . . . . . . . . BZ 0.1 (H) . . . . . . . . . . . . TPP 1 0 varied by a factor of as much as 3.) These results emphasise the importance of the integrity of the PTFE face in preventing contamination of the sample with compounds present in the septum. From Experiment I it is apparent that the septa of (I) vial closures, which do not have a PTFE facing, swell rapidly when in contact with ethyl acetate.In Experiment IV, when determining the rate of extraction of BH and BO from the inner surface of these vial closures, a more rapid sampling scheme was therefore found to be necessary. Both compounds were identified in the samples at levels of ca. 2 pg ml-1 after only 10 min. After 1 h their levels had increased to ca. 3 pg ml-1. This demonstrates the speed of transfer and approach to equilibrium that may occur with these compounds when a PTFE facing is absent. After 30 min the septa has swollen to such an extent that they were buckling either into, or out of. the sample vials. It must be remembered that some of the solvents used in this work are not recommended for use with all of the septa studied.For example, fluoro rubber septa are not recommen- ded for use with methanol or ethyl acetate24 (the reason being clear from Tzble 3). In view of the fact that BH and BO were present in all of the vial closures examined, it should be noted that all the vial closures were supplied in plastic packaging, which could be a source of plasticiser contamination. Conclusions To produce the properties required for septum materials the incorporation of modifying compounds is necessary. The results from the experiments carried out in this work have indicated both the number and variety of those compounds that may be leached out of commercial vial caps under conditions similar to those encountered in normal use. This serves to illustrate the care needed both in the selection of septa and during the planning and execution of analytical experiments.The effectiveness of an intact PTFE facing in minimising the contamination of vial contents with septum-related materials 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 1s. 16. 17. 18. 19. 20. 21. 22. 23. 24. has been demonstrated. However, as soon as this seal is punctured, transfer may occur. The rates of transfer of several compounds through punctured PTFE-faced septa have been calculated and shown to vary widely according to compound type. It is therefore apparent that for minimum sample contamination, vials should not be sampled repeatedly over long periods of time. It should, however, be possible to store sample vials unpierced for extended periods without signifi- cant contamination occurring. References Kollhoff, R. H., Anal. Chem., 1962, 34, 1840. Croll, B. T., Chem. I n d . , 1971, 789. Smith, E. D., and Sorells, K. E., J . Chromatogr. Sci., 1971, 9, 15. Olsavicky, V. M., J . Chromatogr. Sci., 1978, 16, 197. Callery, I. M., J . Chromatogr. Sci., 1970, 8, 408. Smith. E. D., Sorrells, K. E., and Swinea, R. G., J . Chromatogr. Sci., 1974, 12, 101. Quinn, P. A., J . Chromatogr. Sci., 1974, 12, 796. Hopkins, J . L., Cohen, K. A., Hatch, F. W., Pitner, T. P., Stevenson, J . M., and Hess, F. K., Anal. Chem., 1987, 59, 784A. Kurtz, F. E., J. Dairy Sci., 1962, 45, 1573. Baylocq, D., Majcherczyk, P., and Pellerin, F., Ann. Pharm. Fr., 1986, 43, 329. Salmona, G., Assaf, A . , Gayte-Sorbier, A., and Airaudo, Ch. B., Biomed. Mass Spectrom., 1984, 11, 450. Airaudo, Ch. B., Gayte-Sorbier, A., Momburg, R., and Laurent, P., J . Chromutogr., 1986, 354, 341. Peterson, M. C.. Vine, J., Ashely, J . J . , and Natio, R. L., J . Pharm. Sci., 1981, 70, 1139. Hodgson, D. W., and Watts, R. R., J . Assoc. Off. Anal. Chem., 1982,65, 89. Vettorazzi, G., Residue Rev., 1977, 66, 137. Smock, R. M., J . Am. Soc. Hortic. Sci., 1957, 69. 91. Johnson, D. S . , Chem. Znd., 1985,77. Roff, W. J . , and Scott, J . R., “Fibres, Films, Plastics and Rubbcrs,” Butterworth, London, 1971. Vimalasiri, P. A. D. T., Haken, J . K., and Burford, R. P., J . Chromatogr., 1984, 300, 303. Bennett, H., Editor, “Concise Chemical and Technical Dic- tionary,” Fourth Edition, Gower Technical Press. London, 1987. Mass Spectrometry Data Centre, “Eight Peak Index of Mass Spectra,” Third Edition, Royal Society of Chemistry, London, 1983. “The Merck Index,” Seventh Edition, Merck, New Jersey, 1968. Watt, J . A. C., Chem. Br., 1970, 12, 519. “Pierce and Warriner Catalogue 1987-88,” Pierce Chemical Company, Netherlands, 1986, p. 218. Paper 8102505C Received June 24th, 1988 Accepted December 5th, 1988