ISSN:0003-2654    

DOI:10.1039/AN98914FX013

出版商:RSC    

年代:1989

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98914BX015

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANAL'YST. APRIL 1989. VOL. 114 413 Fluorogenic Reagent for Thiols: 4-( N'N-Dimethylaminosulphonyl)- 7 -f I u or 0-2,1,3 - ben zoxad i azo I e Toshimasa Toyo'oka, Takashi Suzuki and Yukio Saito Division of Foods, National Institute of Hygienic Sciences, I - 18- I Kamiyoga, Setaga ya-ku, Tokyo 158, Japan Sonoko Uzu and Kazuhiro lmai Branch Hospital Pharmacy, University of Tokyo, 3-28-6 Mejirodai, Bunkyo-ku, Tokyo I 72, Japan 4-(N,N-Dimethylaminosulphonyl)-7-fluoro-2,1,3-benzoxadiazole (DBD-F) was synthesised for use as a more reactive, thiol-specific fluorogenic reagent than 4-(aminosulphonyl)-7-fluoro-2,1,3-benzoxadiazole (ABD-F). The former had negligible fluorescence whereas its thiol derivatives fluoresced intensely at about 510 nm (excitation occurred at about 380 nm). The DBD-F reacted quantitatively with thiols after 10 min at 50 "C and pH 8.0 and the reaction rates were several times higher than those with ABD-F; it is suggested that the electron withdrawing effect of the dimethylsulphonamide group (S02NMe2) is larger than that of the sulphonamide group (S02NH2).No reaction occurred with alanine, proline, cystine or cysteic acid under the same conditions. The fluorescence intensities of the derivatives were found to be higher in neutral and acidic media than in alkaline solutions. The thiol derivatives of DBD-F were separated by high-performance liquid chromatography and detected fluorimetrically, the detection limits being 0.92, 0.16, 0.13, 0.16 and 0.32 pmol for cysteine, glutathione, homocysteine, N-acetylcysteine and a-mercaptopropionylglycine, respectively.The method was applied to the determination of thiols in rat tissues. Keywords: Fluorescence detection; thiol determination; 4-(N , N - dimethylaminosulphonyl)-7-fluoro-2,7,3- benzoxadiazole; hig h-performance liquid chromatography High - pe r f o r 111 an ce 1 i qui d chromatography ( H P LC) combined with fluorescence detection has been one of the most effective tools for the sensitive and specific determination of substances such as amines and thiols in biological fluids. However, there have been only a few reagents, such as dansylaziridine,IJ bimanes3-6 and maleirnides,7-10 available for the pre- and/or post-column fluorogenic labelling of thiols in HPLC. Recently. two fluorogenic reagents for thiols having the benzofurazan structure, ammonium 7-fluoro-2,1.3-benzoxa- diazole-3-sulphonate (SBD-F) 11-15 and 4-( aminosulphony1)-7- fluoro-2,1,3-benzoxadiazole (ABD-F),l"17 have been repor- ted as pre-column labelling reagents for HPLC.Both SBD-F and ABD-F are highly soluble in water and were suitable for the selective and sensitive detection of low relative molecular mass thiols such as cysteine and glutathione. Both SBD- and ABD-labelled biological thiols were separated by reversed- phase HPLC and detected fluorimetrically in the range 50 fmol-1 pmol. 18 Drastic conditions were required for quanti- tative derivatisation of SBD-F (pH 9.5, 60 "C, 1 h) owing to the weaker electron withdrawing effect of the sulphonic acid moiety located at the para position to the fluorine atom. To overcome this disadvantage, ABD-F, with an aminosulphonyl group at the para position, was synthesised as a more reactive reagent for thiols.The ABD-F did not only react with low relative molecular mass thiols under milder conditions ( ~ € 1 8.0, SO "C. 5 min) but also with the thiol groups in large molecules such as proteins.17 In the course of our studies to develop fluorogenic reagents with the benzofurazan structure, the following reagent was synthesised: 4-(N, N-dimethylaminosulphonyl)-7-fluoro- 2,1,3-benzoxadiazole (DBD-F). As the substituent constant (0) for dimethylsulphonamide (S02NMe2) in the Hammett equation19,'O has not been established, it was decided to compare the reactivities of ABD-F and DBD-F towards thiols. This paper describes the synthesis of DBD-F, its reactivity towards thiols compared with ABD-F and SBD-F and the characteristics of its thiol derivatives.The chromatographic determination of DBD-labelled thiols and the application of the method to rat tissues are also discussed. Experimental Materials 4-Fluoro-2, I ,3-benzoxadiazole was prepared by the method of Nunno et al. 21 and 4-(chlorosulphonyl)-7-fluoro-2,1,3-benz- oxadiazole (CBD-F) was synthesised and purified as described by Toyo'oka and Imai. 16 Homocysteine, glutathione (reduced form), hi-acetylcysteine and cystearnine hydrochloride were purchased from Sigma (St. Louis, MU, USA). Cysteine hydrochloride, alanine, proline and cystine were obtained from Ajinomoto (Tokyo, Japan). 2-Mercaptoethanol (Tokyo Kasei, Tokyo, Japan) , a-mercaptopropionylglycine (Fluka, Buchs, Switzerland), coenzyme A (Boehringer Mannheim, Mannheim, FRG) and disodium ethylenediaminetetraacetate (Na,EDTA) (Kanto Chemicals, Tokyo, Japan) were also used. l-(~-3-Mercapto-2-methyl-l-propionyl)-~-proline (cap- topril) was donated by Sankyo (Tokyo, Japan).Trifluoro- acetic acid (TFA), trichloroacetic acid (TCA) and cysteic acid were from Wako Pure Chemicals (Osaka, Japan). Aceto- nitrile and water were of HPLC grade (Wako Pure Chemicals) and all other chemicals were of analytical-reagent grade and were used as received. Apparatus Proton nuclear magnetic resonance spectra were recorded on a JEOL Model FX-100 spectrometer at 100 MHz using tetramethylsilane as internal standard. Fluorine-19 NMR spectra were recorded on a JEOL Model FX-90Q spec- trometer equipped with a fluorine-19 accessory at 84.3 MHz using TFA as reference (-76.5 p.p.m.)22 and infrared spectra were obtained with a JASCO Model DS 701G spectrometer using potassium bromide discs.Electron ionisation mass spectra (EI-MS) were obtained with an LKB-9000 (Shimadzu, Kyoto. Japan) mass spectrometer and UV - visible absorption spectra were measured with a UVTDEC SO5 instrument * For describing IH NMR characteristics the following abbrevi- ations are used: s = singlet; d = doublct; dd = doublet of doublets; and m = multiplct.414 ANALYST, APRIL 1989, VOL. 114 (JASCO, Tokyo, Japan). For manual measurements a Hitachi 650-60 fluorescence spectrometer with a 1-cm quartz cell was used. Ultraviolet - visible and fluorescence wavelengths were obtained without spectral correction. A Physcotron NS-50 homogeniser (Niti-On, Tokyo, Japan), a Hitachi 05P-21 centrifuge (Tokyo, Japan) and a Yamato MT-31 mixer (Tokyo.Japan) were employed. Reaction temperatures for the DBD-F derivatisation were controlled by a BM-41 water-bath (Yamato). High-performance Liquid Chromatography A Model LC-6A high-performance liquid chromatograph (Shimadzu) equipped with a Rheodyne injector (Model 7125, Cotati. CA, USA) was used; gradient elution was controlled by an SCL-6A system controller (Shimadzu) and the data were processed by a C-R6A Chromatopac (Shimadzu). An Inertsil octadecylsilane (ODS) column (150 X 4.6 mm i.d., 5 pm; Gasukuro Kogyo, Tokyo, Japan) was maintained at 40 "C in a column oven (Gasukuro Kogyo). A Hitachi F-1000 fluores- cence spectrophotometer equipped with a 12-pl flow cell and a Shimadzu SPD-6A UV spectrophotometric detector equipped with an 8-pl flow cell were used to monitor eluates from the column.The UV detector was placed between the analytical column and the fluorescence detector. All solutions used as mobile phases were degassed with an ERC-3320 on-line degasser (Erma Optical Works, Tokyo, Japan) placed before the pump unit. The eluent flow-rate was 1.0 ml min-1. Reaction of Dimethylamine With CBD-F and Synthesis of A solution of dimethylamine hydrochloride (1 g, 12.2 mmol) in 10 ml of 0.1 M borax solution (pH 10) was added dropwise to 70 ml of CBD-F in CH3CN (0.76 g, 3.2 mmol) cooled in ice - water (0-10 "C). Then the reaction solution (pH 8.5) was adjusted to pH 5 with 1 M HC1 and concentrated to about 10 ml under reduced pressure.The acidic solution was extracted with diethyl ether (3 x 200 ml) and one drop was spotted on silica gel (60 FIT4, Merck, Article 5714; solvent, CHC13) for thin-layer chromatography. Three spots were obtained: RF values 0.38, 0.32 and 0.14. The combined diethyl ether extracts were washed with water, dried over anhydrous MgS04 and evaporated in vucuo. The residue was chromato- graphed on a silica gel column (100-200 mesh, 50 X 2 cm; eluent, CHCl?). The 4-(N, N-dimethylamino)-7 ( N , N- dimethylaminosulphonyl)-2,1,3-benzoxadiazole (DDB; RF = 0.14) was separated from the mixture of 4-(N,N-dimethylami- no)-7-(fluorosulphony1)-2,1,3-benzoxadiazole (DBD-S02F; RF = 0.38) and 4-(N, hr-dimethylaminosulphonyl)-7-fluoro- 2,1,3-benzoxadiazole (DBD-F; RF = 0.32), which was then re-chromatographed on the same column with ethyl acetate - benzene (1 + 2) as eluent.The three isolated compounds were recrystallised from hexane - benzene. DDB (orange needles, yield 5%): m.p. 142-144 "C; bH[(CD7)2SOJ, 2.70 (6 H, s, Mec). 3.47 (6 H. s, Med), 6.29 (1 H , d,Jdb = 8.6Hz, Hb), 7.82 (1 H, d, Jab = 8.6 Hz, HJ.); EI-MS, miz 270 ( M + ) ; fluorescence A,,,, (in H,O), excitation 465 nm, emission 560 nm; calculated for C10HL4N403S, C 44.43, H 5.22, N 20.73% ; found, C 44.43, H 5.13, N 20.79%. DBD-S02F (yellow needles, yield 2%): m.p. 161-162 "C; [(CD,),SO], 3.58 (6 H , s, Med), 6.38 (1 [(CD3)2SO] (TFA reference), +69.2 p.p.m.22; EI-MS, mlz 245 ( M + ) ; UV A,,, (in CH3CN), 276 nm (E = 13900 1 mol-1 cm-I), 438 nm (E = 14300); fluorescence A,,, (in H20), excitation 450 nm, emission 560 nm; calculated for C8H8N303FS, C 39.18, H 3.29, N 17.14%; found, C 39.37, H 3.23, N 17.29%.DBD-F (white needles, yield 1%): m.p. 124-125 "C: 8H [(CD3),SOJ, 2.80 (6 H , s, Mec), 7.62 (1 H , dd, J,,F = 4.4 Hz, Ha); EI-MS, miz 245 (M+); UV A,,, (in CH3CN), 321 nm ( E = 4900 1 mol-1 cm-1); calculated for DBD-F H. d , J a b = 8.8 Hz, H'), 8.15 (1 H , d, J a b = 8.8 Hz, Ha); b~ Jdb=7.8Hz,JbF= 10.0Hz,Hb),8.10(1H,dd,J,,b=7.sH~, N Med2 NMed2 F I S02NMeC2 SOZF SO2 N M ec2 DDB DBD-SO2F DBD-F C8H8N303FS, C 39.18. H 3.29, N 17.14%; found, C 39.04, H 3.14, N 16.89%. Synthesis of the DBD Derivative of Homocysteine (DBD- homocysteine) The DBD-F (0.20 g, 0.82 mmol) in 15 ml of CH3CN was added to homocysteine (0.11 g, 0.81 mmol) in 0.1 M borax solution (pH 8.5).The mixture was stirred for 20 min at room temperature, 20 ml of water were added and the excess of DBD-F was removed by extraction with ethyl acetate. The aqueous layer was lyophilised, 50 ml of CH30H were added to the lyophilised residue and the mixture was shaken vigorously then centrifuged at 3000 rev min-1 for 5 min. The extraction procedures were repeated three times with 50 ml of CH3OH. The combined CH30H solution containing DBD-homo- cysteine was poured into a 300-ml round-bottomed flask, evaporated and the residue obtained chromatographed on a Bio-Gel P-2 column (200-400 mesh, 55 x 1.5 cm; eluent, H20). The fluorescent fractions corresponding to DBD- homocysteine were collected and lyophilised (yellow powder; yield 15%; decomposed above 248 "C); aH [(CD,),SO], 7.92 (6 H, s, MeL), 2.0-2.2 (2 H, m, He), 3.2-3.6 (3 H, m, Hd and Hf).Calculated for CI2Hl5N4O5S2Na.H20: C, 36.00; H, 4.28; N, 13.99%. Found: C. 36.44; H , 4.09; N, 13.30%. Fluores- cence in H20: A,,, (excitation), 395 nm; h,,, (emission), 521 nm. (1H,d,J,h=7.6Hz,H'),7.57(lH,d,Jab=7.6H~,H'),2.79 SCHd2CHe2CH'N H2 S02N MeC* DBD-homocysteine Time Course of the Reaction of Homocysteine With DBD-F One millilitre of 1 mM DBD-F in 0.1 M buffer (phosphate, pH 6.0 and 7.0; borax, pH 8) containing 8% CH3CN and 1 ml of 10 VM homocysteine in 0.1 M buffer (pH 6.0, 7.0 or 8.0) containing 2 mM Na,EDTA were mixed in a 5-ml glass tube. The tube was capped and heated at the required temperatures (40-60 "C) in a water-bath in the dark.At fixed reaction time intervals, the tube was removed and cooled in ice - water. The fluorescence intensities of the reaction solution were measured at ambient temperature with excitation at 395 nm and emission at 521 nm. The reagent blank without homocy- steine was treated similarly. Calculations of the reaction yields were based on the fluorescence intensity of the standard DBD-homocysteine while pseudo-first-order rate constants were based on the difference between the net fluorescence intensities (sample fluorescence minus blank fluorescence) and that of the standard DBD-homocysteine. Fluorescence Intensities and Wavelength Maxima of DBD- homocysteine at Various pH The relative fluorescence intensities (RFIs) of standard 2.5 p~ DBD-homocysteine, dissolved in 0.05 M Britton - Robinson buffer for a solution pH in the range 2-12 and 0.1 M HCI for a solution pH of 1, were measured at 520 nm (excitation at 395 nm).One litre of Britton - Robinson buffer solution A consistsANALYST, APRIL 1989. VOL. 114 415 -7 / / / d / / / F 1 1 / F F r NMe2 S02CI 1 CBD-F \ I \ \ \ Me2NH \ \ ! \ \ v @) ---I Me2NH S02NMe2 DBD-F Fig. 1. Reaction pathway of dimethylamine with CBD-F I 2 50 300 350 400 Wave1 en g t hin m Fig. 2. Absorption spectra of fluorogenic reagents in acetonitrile. A , DBD-F (0.19 mM); B, ABD-F (0.25 mM): and C, SBD-F (0.15 mM) of 4.5 g of 85% orthophosphoric acid (H,P04), 2.4 g of acetic acid and 2.47 g of boric acid (H3BOJ). Solution B is 0.2 M NaOH (8.0 g 1-1).Buffers of different pH were prepared by mixing suitable volumes of solutions A and B. Use of the Manual Method for the Determination of Thiols With DBD-F To 1.0 ml of thiol (10 p ~ ) in 0.1 M borax solution (pH 8.0) containing 2 mM Na2EDTA was added an equal volume of DBD-F (1 mM) in 0.1 M borax solution (pH 8.0) containing 8% CH3CN. After mixing immediately the solution was heated at 50 "C for 10 min. The solution was then cooled in ice - water. adjusted to pH 2 with 1.5 ml of 0.1 M HC1 and the fluorescence intensity was measured at each maximum wavelength. Cysteic acid, cystine, alanine and proline were also reacted with DBD-F. The reagent blank without thiol was treated in the \ 1 NMe2 DBD-S02F Me2NH 1 N Me2 +lN;o N S02NMe2 DDB same manner as the thiol samples.The detection limit was calculated as that concentration at which the fluorescence intensity of the sample was equal to twice the fluorescence intensity of the reagent blank. High-performance Liquid Chromatographic Separation of DBD-thiols and Detection Limits To a 1-ml brown vial was added 0.4 ml of DBD-F (1.0 mM) in 0.1 M borax solution (pH 8.0) containing 8% CH3CN and 0.4 ml of mixed thiols (cysteine, 4.54 p ~ ; homocysteine, 4.66 VM; glutathione, 5.00 p ~ ; N-acetylcysteine, 5.39 VM; and a-mer- captopropionylglycine, 4.72 p ~ ) in 0.1 M borax solution (pH 8.0, 2 mM Na,EDTA). The vial contents were mixed immediately and the vial was capped tightly and heated at 50 "C for 10 min. After cooling in ice - water, 0.9 ml of water was added to 0.1 ml of the solution and a 4-pl aliquot of the dilute solution was injected on to the HPLC column.A linear gradient elution from 80% eluent A (0.15 M H3P04) to 50% eluent B (CH3CN) over 30 min was adopted for the separation of DBD-thiols. The eluate was monitored with both UV (at 320 nm) and fluorescence (excitation at 390 nm, emission at 520 nm) detection in series. High-performance Liquid Chromatographic Determination of Thiols in Rat Tissues Derivatised With DBD-F Tissues (liver, kidney, spleen, testis, lung and pancreas) were exenterated from an etherised rat (F344, male, 10 weeks old, body mass 300 g), washed with physiological saline solution and stored in a freezer at -80 "C prior to analysis. Aliquots (4 ml g-1 of tissue) of 5% TCA containing 5 mM Na2EDTA were added to the liver (11.0 g), kidney (0.92 g), spleen (0.57 g), testis (1.48 g), lung (1.08 g) and pancreas (0.53 g) tissues with cooling in ice - water.The samples were homo- genised, centrifuged at 3500 rev min-1 for 10 min and to each 10 ~1 of supernatant were added 150 pl of DBD-F (27 mM) in CH&N and 440 pl of 0.1 M borax solution (pH 8.0) containing416 ANALYST, APRIL 1989, VOL. 114 1 mM Na2EDTA. The solution was heated at 50 "C for 10 min, then cooled in ice - water and a 2 0 4 aliquot was subjected to HPLC. The eluate was monitored at 520 nm (excitation at 390 nm). The supernatant from the homogenised tissues and the reagent blank without supernatant were treated and chro- matographed in the same manner. Results and Discussion Structural Assignments of Benzofurazan Derivatives Two positions in the 4-(chlorosulphonyl)-7-fluoro-2,1 ,3-ben- zoxadiazole (CBD-F) skeleton are susceptible to nucleophilic attack, one of these is the sulphonylchloride group (-SO$J) (Fig.1). By nucleophilic substitution, it is possible to sq'nthesise various derivatives having an -S02R group. For example, ABD-F was synthesised by the reaction of CBD-F with ammonia solution, as described by Toyo'oka and Imai.lh With dimethylamine as nucleophile, three compounds in addition to the hydrolysate of CBD-F were obtained from the reaction with CBD-F, although the yields were very low. One of the compounds was identified as 4-(N, N-dimethylamino)-7- (N, N-dimethylaminosulphonyl)-2,1,3-benzoxadiazole (DDB) from mass spectral data (miz = 270, M + ) , the 1H NMR spectrum and elemental analysis (calculated for CIoHI4N1O3S).The other two compounds were isomers ha\.ing the same molecular ion (miz = 245) (mass spectral analysis) and the same composition (elemental analysis calculated for C8H,N303FS). The 'I3 NMK spectrum of one of the isomers showed the aromatic H-F couplings, the coupling constants of which (IL,,, = 7.8 Hz, JhF = 10.0 Hz and JaF = 4.4 Hz) were similar to those of ABD-F (J:,[, = 7.6 Hz, JbF = 10.0 Hz and JaF = 4.3 Hz) and SBD-F (Jah = 7.6 Hz, JhF 1 I00 75 s 50 .- >- 25 V I 0 15 > 30 45 60 Ti me/min Fig. 3. Time courses 01 the reaction of homocysteine with DBD-F 0. pH 8 0. 50 "C, A, p1-I 7 0, 60 "C, 0, pH 7 0, 50 " C , 0. pH 7 0 . 40 "C. m d A. pH 6 0, 50 "C Homocysteine ( 5 WM) and DBD-F (500 LIM) nere reacted in 4% CH,CN - 0 1 M buffer (pH 6 (!-8 0) contdning 1 n i h i Ya'EDTA Med\ured at 521 nm (excitation at 395 nm).dt fixed time interkdls Table 1. Rate constants for the reaction of homocysteine with DBD-F, ABD-F and SBD-F in 0.1 M phosphate buffer (pH 7.0). Fluorescence ~vas measured at 521 nm (excitation at 395 nni). Honiocysteine (5 LLM) and DRD-F (500 p h t ) were reacted in 0 . 1 M phosphate solution (pH 7.0. 1 mhi Na,EDTA) containing 4% CH3CN kimin 1 Reagent 40 "C 50 "C 60 "C DBD-F . . . . 6.76 x 10 2 1.93 x 10 1 3.07 x 10 - 1 ABD-F . . . 2.54 x 10 '* 7.40 x 10k" 1.16 X 10 ' SBD-F . . . ND' ND' 3.55 x 1 0 - 3 ' * From reference 16 (ND = not determined). = 10.5 HZ and Jilt. = 3.9 Hz).ll.l" The UV - visible absorption spectrum was almost identical with those of ABD-F and SBD-F (Fig.2). From these results, the structure was identified as that of DBD-F, the desired compound. The third compound was expected to be 4-(chlorosulphonyl)-7-(N, A'- dimethylamino)-2,1,3-benzoxadiazole. However, both ele- mental analysis and mass spectral data showed the existence of fluorine. The IYF NMR chemical shift was +69.2 p.p.m.,2' suggesting that the fluorine was not covalently bound to an aromatic carbon, but to the sulphur in the sulphonyl group. Therefore, the third compound was identified as 4-(N, N-dimethylamino)-7-(fluorosulphonyl)-2,1,3-benzoxa- diazole (DBD-S02F). When CBD-F was reacted with 40% aqueous dimethyl- amine instead of dimethylamine hydrochloride, DBD-F was not formed in the reaction mixture and the yields of DDB and DBD-SO'F were 15 and 4%, respectively.The yield of DBD-F was improved by using anhydrous dimethylamine (yield, 10%) instead of dimethylamine hydrochloride (yield, 1 Yo). Considering the structures and yields of the reaction products, the nucleophilic attack of CBD-F by amine seems to occur mainly at the aromatic carbon bonded to F. The chlorine atom of the S02Cl group in the reaction intermediate (Fig. 1, I) is then displaced by the fluorine atom of the €IF generated in the previous step to give DBD-S02F (Fig. 1 ) . An alternative route to DBD-SO'F via 4-fluoro-7-(fluorosulphonyl)-2,1,3- 100 I 1 2 4 6 8 10 12 PH Fig. 4. Effect of pH on the fluorescence intensity of DBD-hoino- cysteinc. Detection at: A , (cxcitation) 395 nm and (emission) 520 nm; and B, (excitation) 265 nni and (emission) 530 nm.Standard DBD-hoinocysteine (2.5 p~ each) was dissolved in (1.05 xi Britton - Robinson buffer (pH 2-10) o r 0.1 hi HCI (pH 1) Table 2. Detection limits and wavelength maxima for various thiols. To 1.0 ml of 10 LLM thiol in 0.1 hi borax (pH 8.0) solution containing 2 miv NaZEDIA was addcd an equal volume of 1 m h i DBD-F in 0.1 51 borax (pH 8.0) solution containing 8%) CH,CN. After heating at 50cC for 10 min, the reaction solution was adjusted to pH 2 by the addition of 1 .5 ml of 0.1 hi HCl and the fluorescence intensity was measured a t the wavelength maximum L,, of fluorescenceinm Thiol Excitation Emission Cysteamine . . . . . . 402 537 CoenzymeA . . . . 398 527 Homocysteine . . . . 396 52 1 Glutathione .. . . . . 393 Captopril . . . . . . 398 524 A;-Acetylcysteine . . . . 394 Cysteine . . . . . . 387 5 10 52 1 522 propionylglycine . . 386 516 Cystine . . . . . . ND* ND Cysteicacid . . . . . . ND ND ND Alanine . . . . . . ND Proline . . . . . . ND ND a-Mercapto- ' ND = not detected Detection limit1 pmolml I 34 49 63 72 77 92 100 299 ND ND ND NDANALYST. APRIL 1989. VOL. 113 417 I F Timeimin Fig. 5. Chromatograms o f biological thiols derivatised with DBD-F. ( 0 ) Fluorescence detection at (excitation) 390 nm and (emission) 520 nm: and ( h ) UV detection at 320 nm. A, DBD-cystcine; B, DBD-glutathione; C . DBD-homocysteine; D, DBD-N-acetyl- cyst e i n e ; E . D B D - a- me rc a p t o p r o pi o n y I g l y ci n e : and F . D R D - F , Amount of thiol injected, 2 pmol; column, Inertsil ODS (150 x 3.6 mm i.d..5 urn) at 40 "C; eluent A, 0.15 M W,PO,; eluent B, CH,CN; linear gradient elution. 80% A to 50% B over 30 min; flow-ratc. 1 .O ml min - 1 . The retention time (min) is indicated on each peak benzoxadiazole (Fig. 1 , I I ) is not possible as the fluorine atom is not present in the initial stage of the reaction. The low yield of DBD-F (101') suggests that the routes depicted by the broken lines (Fig. 1) may also be minor. Both of the production routes to the bis(dimethy1amino) compound (DDB) are possible via DBD-S02F and DBD-F. The path- ways are similar to those described for the reactions of amines Lvi t h c h 1 ori d e (DNS-Cl) and 4-fluoro-7-nitro-2,1,3-benzoxadiazole (NBD- F), respectively.2-i.24 Comparison of the Reactivities of DBD-F, ABD-F and SBD-F Towards Homocysteine The reactivity of 4-fluorobenzofurazan towards nucleophiles ( e .g . , thiols and amines) was dependent on the electron withdrawing effect of the substituent at the 7-position of 4-fluorobenzofuran, the order of which was as follows: nitro (B = 0.78)> sulphonamide (a = 0.62)> sulphonic acid (B = 0.38). Thus, NBD-F was the most reactive followed by ABD-F and, finally, by SRD-F. As the reactivity of DBD-F appeared to be similar to those of ABD-F and SBD-F. the time course of the reaction of DBD-F with homocysteine, as a representative thiol, was determined at various conditions (pH 6.0-8.0, 4@60 "C). As shown in Fig. 3, the reaction yields increased with increasing time, under all conditions.The reaction was slowest at pH 6.0 and SO "C and the yield after heating for 60 min was about 45% (Fig. 3). Although the time courses of the reaction at SO and 60 "C in pH 7.0 solution were similar, degradation at 60 "C was higher than at SO "C. On the other hand, the reaction at SO "C in pH 8.0 solution reached completion in 10 min. However, a slight decay of fluorescence intensity was observed with longer heating periods and, therefore, for thiols, the reaction was carried out at SO "C in 0.1 M borax solution (pH 8.0) containing Na2EDTA for 10 min. Pseudo-first-order rate constants at pH 7.0 were measured using solutions of 0.5 mM DBD-F and S VM homocysteine at various temperatures. From Table 1 it can be seen that the rate constants were several times higher than those with ABD-F.16 5 -dim e t h y 1 amino - 1 -naphtha len e su 1 phony 1 5 10 15 20 25 30 35 1 I I I I O 5 10 15 20 25 30 35 5 10 15 20 25 30 35 B L 5 10 15 20 25 30 : 5 I 5 10 15 20 25 30 35 B I I I I I I 5 10 15 20 25 30 35 Ti meimi n Fig.6. Chromato rams obtained from rat tissues. ( a ) Liver; ( b ) spleen; (c) testis; (a!! lung; ( e ) kidney; and U, pancreas. A, Cysteine; and B, glutathione. DBD-F (150 pl, 27 mM) in CH3CN and a pH 8.0 solution of borax (440 pl, 0.1 M) containing 1 mM Na2EDTA was added to the supernatant liquid (10 pl) obtained from the homo- genised tissue. After reaction at SO "C for 10 min, a 20-ul aliquot was injected on to the HPLC column. HPLC conditions were as in Fig. 5 Thus, it was demonstrated that in this instance the electron withdrawing effect of the dimethylsulphonamide group (S02NMe2) is larger than that of the sulphonamide group The DBD-F itself had negligible fluorescence and the DBD-F derivative of homocysteine had a long fluorescence wavelength (excitation at 395 nm, emission at 521 nm) similar to the SBD and ABD derivatives.From these results, DBD-F appeared to be a suitable alternative reagent to ABD-F and SBD-F for the determination of thiols. (S0,NHZ).418 ANALYST, APRIL 1989, VOL. 114 Effect of pH on the Fluorescence Intensity of DBD-homo- cysteine The fluorescence intensities of SBD- and ABD-thiol are dependent on the pH of the medium, as described pre- viously.13 16 To obtain the optimum pH for the determination of thiols the correlation of pH with fluorescence intensity was studied.As shown in Fig. 4, relatively high fluorescence intensities were obtained in the pH range 2-6, which gradually decreased with increasing pH. The maximum fluorescence intensity was obtained at pH 2. The low intensity obtained in 0.1 M HC1 might be caused by the presence of C1- ion in the medium, becauve halide ions such as C1- and Br- can reduce the fluorescence." For the pH range 1-9 fluorescence wavelengths were in the range 52&523 nm (excitation, 393-396 nm). Blue shifts of the excitation maxima (from 395 to 267 nm) were also observed in alkaline medium at pH 10 and 12. However, the emission maxima were similar to those obtained at other pH values. Determination of Various Thiols Using the Manual Method Considering the time courses of the reaction of homocysteine with DBD-F (Fig.3), derivatisation conditions were selected for the determination of thiol compounds [SO OC, 10 min, 0.1 M borax solution (pH 8.0) containing 1 mM NazEDTA and excess of DBD-F]. High fluorescence intensities were ob- se rve d for so 1 u t i o n s con t ai n i n g bio 1 ogi call y import ant t h i o 1 s and for drugs having an SH group. As can be seen from Table 2, the lowest detection limit was obtained for cysteamine, while the highest was obtained for cysteine. The detection limits (the fluorescence intensity minus twice the fluorescence intensity of the reagent blank) with the exception of cysteine, were in the range 34-100 pmol ml-l and the difference for the thiols tested was less than a factor of three. The reason for the higher detection limit for cysteine compared with those for other thiols was not clear.The same trend was also observed using both SBD-F and ABD-F.13.16 In contrast, amino acids (alanine or proline) and oxidised forms of thiols (cystine or cysteic acid) did not react with DBD-F under these conditions. However, considering the reactivity of DBD-F towards thiols compared with that of ABD-F as described above, DBD-F may react with amines and phenols under relatively drastic conditions. These are now under investigation and will be reported elsewhere. The excitation and emission maxima of the thiols tested were in the range 386402 and 510-537 nm, respectively (Table 2). The fluorescence intensity of DBD-F itself in the detection wavelength range for DBD-thiols was significantly lower than that of ABD-F and similar to that of SBD-F. This feature of DBD-F makes it more suitable than ABD-F for the determination of thiols using the manual method.Moreover, its higher reactivity compared with SBD-F also makes it more suitable for the labelling of thiols. High-performance Liquid Chromatographic Separation and Detection of DBD-thiols The separation of DBD-thiols was studied using a reversed- phase HPLC column (Inertsil ODs). From the results obtained with a binary eluent system, the retention times of DBD derivatives were, as expected, decreased with increased CH3CN concentration and at higher pH of the eluent. When a solution of 0.15 M H3P04 - CH3CN (8 + 2) was used as isocratic eluent, two derivatives, DBD-N-acetylcysteine (1 8 min) and DBD-ck-mercaptopropionylglycine (23 rnin), had longer elution times than the other thiols (DBD-cysteine, 5 min; DBD-glutathione, 6.5 min; and DBD-homocysteine, 8 min).Therefore, a linear gradient elution was adopted in order to shorten the run time. As shown in Fig. 5(a), DBD-cysteine, DBD-glutathione, DBD-homocysteine, DBD-N-acetylcysteine and DBD-a-mercaptopropionylgly- cine were separated completely with elution in this order within 15 min by a linear gradient from 80"/0 eluent A (0.15 M H7PO4) to 50% eluent B (CH&N) over 30 min. Under the chromatographic conditions, unreacted DBD-F was eluted at 17 min (monitored by UV detection at 320 nm) [Fig. 5(b)]. The peak corresponding to the hydrolysate of DBD-F was not observed on the chromatogram obtained using fluorescence detection (Acx, 390 nm; A,,, , 520 nm).The chromatogram shows that neither DBD-F nor its hydrolysate fluoresce at 520 nm with corresponding excitation at 390 nm. Under the selected conditions, the detection limits (signal to noise ratio of three) for DBD-cysteine, DBD-glutathione, DBD- homocysteine, DBD-N-acetylcysteine and DBD-a-mercapto- propionylglycine were 0.92, 0.16, 0.13, 0.16 and 0.32 pmol, respectively. High-performance Liquid Chromatographic Determination of Thiols in Rat Tissues When the proposed method was applied to the determination of thiols in various tissues, a large peak corresponding to reduced glutathione appeared on the chromatograms for liver, spleen, testis and lung [Fig. 6(a)-(d)]. On the other hand, both cysteine and reduced glutathione were the main thiols found in kidney and pancreas.The small peaks eluted after 5 min seem to be relatively low molecular mass thiols such as short-chain peptides containing a half-cystine residue, as the peaks of endogenous compounds derived from the reaction without DBD-F were eluted within 5 min (data not shown). From the chromatograms thus obtained, it is clear that fluorescence detection at 520 nm (excitation at 390 nm) is appropriate for the specific detection of DBD-thiol derivatives and for avoiding interferences from endogenous compounds. This is an obvious advantage of using DBD-F for the determination of thiols in biological samples. Further, the simultaneous determination of thiols and disulphides reported previously26 may be possible with the combined use of both DBD-F and SBD-F or ABD-F.Among the synthesised compounds, DBD-S02F would be expected to react with nucleophiles such as amines and thiols, as it has an active site in the molecule. When DBD-S02F was reacted with dimethylamine, however, it afforded no adduct in the solution owing to its instantaneous hydrolysis. In conclusion, DBD-F is more reactive than ABD-F and may be an alternative to ABD-F and SBD-F for the sensitive and specific detection of thiols using both the manual and HPLC methods. It could also be used as a differentiation reagent for thiols located in large molecules, as suggested for ABD-F. '7 The authors thank Y. Watanabe of Chugai Pharmaceutical for performing mass spectral measurements and N.Fujii of JEOL for performing NMR spectral measurements. Thanks are also due to C. K. Lim, Clinical Research Centre, for his review of and comments on the manuscript. References 1. 2. 3. 4. 5 . 6. 7. Scouten, W. H., Lubcher, R . , and Baughman, W., Biochim. Biophys. Actu, 1974, 336, 421. Lankmayr, E. P., Budna, K. W., Muller, K., and Nachtmann, F., Fresenius Z . Anal. Chem., 1979, 295, 371. Kosower, N. S . , Kosower, E. M., Newton, G. L., and Ranney, H. M., Proc. Nutl. Acad. Sci. USA, 1979, 76, 3382. Kosower, N. S . , Newton, G. L., Kosower, E. M., andRanney, H. M., Biochim. Biophys. Acta, 1980, 622, 201. Newton, G. L., Dorian, R . , and Fahey, R. C., Anal. Biochem., 1981, 114. 383. Baeyens, W. R. G., van der Weken, G., and de Moerloose, P., Anal. Chim. Acta, 1988, 205, 43. Kanaoka, Y., Yakuguku Zasshi, 1980, 100, 973.ANALYST, APRIL 1989, VOL. 114 419 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. Weltman, J . K., Szaro. R. P., Frackelton, A. R . , Bunting, J . R . , and Cathou, R. E., J . Biol. Chem., 1973, 248, 3173. Machida, M.. Machida, M., Sekine, T., and Kanaoka, Y., Chem. Pharm. Bull., 1977, 25, 1678. Takahashi, H., Nara, Y., and Tuzimura, K., Agric. Biol. 22. Chem., 1978,42, 769. Imai, K., Toyo’oka, T . , and Watanabe, Y., Anal. Brochem., 1983, 128, 471. 23. Toyo’oka, T.. and Imai, K., J . Chrornatogr., 1983, 282, 495. 24. Toyo’oka, T., and Imai, K., Analyst, 1984. 109, 1003. 25. Toyo’oka, T., Imai, K., and Kawahara, Y.. J . Phurm. Biomed. 26. Anal., 1984, 2, 473. Araki, A . , and Sako, Y., J . Chromatogr., 1987. 422, 43. Toyo’oka, T . , and Imai, K., Anal. Chem., 1984, 56. 2461. Toyo’oka, T., and Imai, K., Anal. Chem., 1985, 57, 1931. Imai, K.. and Toyo’oka, T . , Methods Enzymol., 1987, 143,67. Hanimett, L. P . , “Physical Organic Chemistry,” Sccond Edition, McGraw-Hill, New York. 1970. 20. 21. Ritchie, C . D., and Sager, W. F., “Progress in Physical Organic Chemistry,” Volume 2, Wiley, New York, 1964. Nunno, L. D., Florio, S . , and Todcsco, P. E., J . Chem. SOC. C, 1970, 1433. Dungan, C. H., and van Wazer, J. R . , “Compilation of Reported F19 NMR Chemical Shifts,” Wiley-Interscience, New York, 1970. Gray, W. R . , Methods Enzymol., 1967, 11, 139. Imai, K., and Watanabe. Y., Anal. Chim. Acta, 1981,130,377. McClure, D. S., J . Chem. Phys., 1949, 17, 905. Toyo’oka, T., Uchiyama, S . , Saito, Y., and Imai, K., Anal. Chim. Acta, 1988, 205, 29. Paper 8104012E Received October loth, 1988 Accepted December 2nd, 1988

ISSN:0003-2654    

DOI:10.1039/AN9891400413

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST. APRIL l%Y. VOL. I13 42 1 Determination of Chlortetracycline Residues in Tissues Using High-performance Liquid Chromatography With Fluorescence Detection W. John Blanchflower, Robert J. McCracken and Desmond A. Rice Department of Agriculture, Veterinary Research Laboratories, Stormont, Belfast BT4 3SD, UK A method is described for the determination of chlortetracycline residues in tissue samples. The samples were extracted into a hydrochloric acid - glycine solution and the extracts concentrated and purified on cyclohexyl-bonded reversed-phase cartridges. Any chlortetracycline present was converted t o iso-chlor- tetracycline at pH 12, which was then separated from interfering compounds on a reversed-phase polymeric column using high-performance liquid chromatography with fluorescence detection.The detection and determination limits of the assay were 20 and 50 ng g-l, respectively, making it suitable for statutory residue testing purposes. Keywords: Chlortetracycline; iso-chiortetracycline; high-performance liquid chromatography; fluorescence detection; tissues Chlortetracycline (CTC) is a broad spectrum antibiotic with a high degree of activity against a wide range of Gram-positive and Gram-negative bacteria. It is frequently added to swine feeds for the prophylaxis and treatment of bacterial diseases, often in combination with other antibacterial agents such as sulphonamides or penicillins. This had led to the need for methods to monitor residual levels of CTC in meat products, to ensure that they fulfil the relevant government require- ments for tolerance limits.Microbiological assays can be used as an initial screening method for CTC, but chemical methods are normally required to confirm and quantify any residues found. Several chemical methods have been described for the determination of CTC in biological samples. 1-3 These methods have been based mainly on high-performance liquid chromatography (HPLC) in order to achieve the sensitivity and specificity required to detect CTC at residual levels. We found, however, that in practice these methods suffered from a lack of reproducibility; this disadvantage has also been discussed by Moats.5 In addition, they lacked the sensitivity required to detect residual levels of CTC, i.e., less than 100 ng g-1. Some of the problems were due to interactions between CTC and the residual silanol groups present in the silica-based support materials used in the HPLC columns.This resulted in peak tailing and poor sensitivity.6 Sharma and Bevilll described a lengthy column conditioning procedure using ethylenediaminetetraacetic acid and excess of CTC, and Knox and Jurand7 showed that improvements could be obtained by removing the residual silanol groups from the silica support material. Further improvements were obtained by the use of polymer-based packing materials.8-") Other problems were due to procedures involving liquid - liquid extraction steps. Sharma and Bevilll extracted CTC in sample homogenates from an aqueous phase into ethyl acetate using ion-pairing agents and then re-extracted the CTC from ethyl acetate into orthophosphoric acid prior to injection on to the HPLC column.We found, however, that with this method only 20-30% of the CTC added to blank tissue samples could be recovered. The authors attempted to overcome this problem by using spiked tissue samples as standards, but with such low values for recoveries, this technique is obviously subject to inaccuracies. Because of the difficulty in partitioning CTC into organic solvents. other workers have mainly used solid-phase extrac- tion and clean-up. Many of the solid-phase cartridges however caused problems similar to those found with reversed-phase columns, i.e., binding of CTC to the silica support material. Oka et al. 3 found that treatment of the cartridges with oxalic acid gave improved recoveries.The use of aqueous-based eluents with solid-phase cartridges, however, generally resul- ted in poor sensitivity due to the relatively large volumes of eluent required and the difficulty in reducing their volume. Moats5 overcame thi5 problem by injecting large volumes of extract (200-800 pl) on to the column and using a gradient to elute the CTC. All the reported HPLC methods also use UV detection of CTC and this limits further the sensitivities that can be achieved. The method that we have developed has overcome many of the problems discussed above by ( a ) using an aqueous-based extraction; ( h ) carrying out the concentration and clean-up steps on a cyclohexyl-bonded solid-phase cartridge; (c) converting CTC to the highly fluorescent iso-chlortetracycline (iso-CTC) derivative; (d) performing the separation on a polymeric reversed-phase column: and (e) employing fluor- escence detection.Experimental Apparatus The HPLC system consisted of a Vl'aters Model 501 pump (MilliporeiWaters, Harrow, Middlesex, UK), a Waters Model WISP 710A autosampler and a Perkin-Elmer Model LS5 fluorescence spectrophotometer fitted with an LC flow cell (Perkin-Elmer, Beaconsfield, Buckinghamshire, UK). The excitation wavelength was set at 340 nm and the emission wavelength at 420 nm. Results were recorded on a chart recorder (J. J . Lloyd Instruments, Southampton, UK) oper- ated at 4 mm min-1. The reversed-phase polymeric column was a PLRP-S type, lOOA, 5-pm, 150 x 4.6 mm i.d. column (Polymer Laboratories. Church Stretton. UK). The solid- phase extraction cartridges were Analytichem International type Bond-Elut cyclohexyl cartridges, 3-ml capacity (Jones Chromatography, Hengoed, Glamorgan, UK) fitted with syringe adapters. Reagents Acetonitrile was of HPLC grade; all other reagents were of analytical-reagent grade.Chlortetracycline was obtained from Sigma (Poole, Dorset, UK). A stock standard solution (1 mg ml-1) was prepared in methanol and was stable for 1 month if stored in an amber glass bottle at 4°C. A dilute422 ANALYST, APRIL 1989. VOL. 114 standard (10 pg ml-1) was prepared in methanol immediately before use. The working standard (100 ng ml-1) was prepared in the mobile phase (see under Method). The mobile phase consisted of a mixture of 875 ml of 0. I M glycine (adjusted to pH 12 using a 10 M sodium hydroxide solution) and 125 ml of acetonitrile.The solution was de-gassed using helium. The mobile phase was delivered at a rate of 1 ml min-1. Method Cut the tissue samples into small cubes and store at -20°C until frozen or until required. Place the frozen cubes in a domestic food blender and pulverise until the tissue forms a fine powder. Weigh 5-g aliquots into 100-mI wide-mouthed polyethylene centrifuge bottles, add 45 ml of 0.1 M glycine in 1 M hydrochloric acid and homogenise for 1 min. Centrifuge at 2500 g for 15 min and filter the supernatant through a plug of glass wool into a 150-ml conical flask. Re-homogenise the precipitate using a further 50 ml of glycine - hydrochloric acid, centrifuge, filter as before and combine the two filtrates.Prepare a Bond-Elut cyclohexyl cartridge by washing it with 10 ml of methanol followed by 10 ml of water. Using a syringe, pass 20 ml of the tissue extract through the cartridge. Wash with 7 ml of water, remove the excess of water by passing a small plug of air through the cartridge and elute the CTC using 7 ml of methanol. Collect the eluate in a 60 x 16 mm amber glass vial. Evaporate to dryness at 60 "C under nitrogen. Allow the vials to cool to room temperature and dissolve the residue in 2 ml of the mobile phase. Cap the vials and leave them at ambient temperature (25°C) for a minimum of 2.5 h before HPLC analysis, to facilitate conversion of CTC to iso-CTC. Simultaneously prepare a 100 ng ml-1 standard by diluting 20 p1 of the 10 pg ml-1 standard with 1.98 ml of the mobile phase in an amber glass vial, and also leave this at ambient temperature for a minimum of 2.5 h before HPLC analysis.Transfer the sample extracts in mobile phase into auto- sampler vials and set the autosampler to inject 25p1 aliquots, with a run time of 10 min. Alternatively, manual injection can be used. The vials must be protected from light because iso-CTC is light-sensitive. Record the peaks on a chart recorder and calculate the results from peak-height measure- ments. Results A typical chromatogram of a 100 ng ml-1 standard, a blank muscle extract and an extract from a muscle sample containing about 180 ng g-1 of CTC is shown in Fig. 1. The iso-CTC peak elutes at 7.2 min. The linearity of the assay was checked by running a set of standards containing between SO and 500 ng ml-1 of CTC through the procedure in duplicate.Allowing for dilution factors, the 500 ng ml-1 standard is equivalent to 1 pg g-1 in tissue. Least-squares linear regression analysis of the data gave the equation J = 0 . 4 7 ~ + 3.2, where y was the peak height (mm) and .x the concentration of CTC (ng ml-I). The correlation coefficient was 0.9988. The assay is therefore linear up to at least 1 pg g-1 of CTC in tissue. The recovery of CTC in the assay was determined by spiking tissue samples with CTC and carrying them through the procedure. The results are shown in Table 1; the recoveries ranged from 87 to 94%. The precision of the assay was checked by performing six replicate analyses of two muscle samplcs containing CTC using the described method.The results for the mean, standard deviation and coefficient of variation (CV) were 230 ng g-1.9.7 ng g-1 and 4.2%. respectively, for muscle 1 and 86 ng g-1, 11.4 ng g-1 and 13.3%. respectively, for muscle 2. An experiment was carried out to determine the length of time required for the CTC - iso-CTC reaction to reach completion at pH 12 and 25 "C. The effect of natural light on t > cn C a C a C a U .I- .- + .- E 3 U - a) I I 8 4 0 8 4 0 8 4 0 Timeim in Fig. 1. Chromatograms of the fluorescence produced by injecting 25 pl of ( a ) the iso-CTC derivative o f a 100 ng ml-1 CTC standard; ( h ) a blank tissue extract; and (c) a tissue extract containing about 180 ng g 1 of CTC. The iso-CTC derivative elutes at 7.2 min Table 1. Recovery of CTC' added to swine muscle containing low levels o f CTC CTC added/ CTC found 1 Recovery.n!? g- ngg-l 70 - Muscle A . . . . . . . . 0 58 200 233 87 400 403 86 200 200 94 400 379 92 - Muscle B . . . . . . . . 0 11 * Results are the mean of duplicate analyses. 100. 1 2ot I' 1 2 3 4 24 48 72 96 120 I I 0 Time/h Fig. 2. Rate o f formation of the fluorescent iso-CTC derivative from a 100 ng ml 1 CTC standard in the mobile phase at pH 12. Reaction carried out in (A) amber glass vials or (B) clear vials exposed to natural light at 25 "C the reaction was also studied. Two sets of 100 ng g-1 standards were prepared in the mobile phase. One set was protected from the light in amber glass vials and the other was prepared in clear vials and exposed to natural daylight.The fluores- cence intensity was measured initially at 30-min intervals for 4 h and then every 24 h for 5 d; the results are shown in Fig. 2. It can be seen that the reaction is complete after 2.5 h at 25 "C and that the iso-CTC derivative is stable for several days if protected from light. In the presence of light the iso-CTC fluorescence reaches only 50% of that produced by the protected samples and is destroyed over a period of several hours.ANALYST, APRIL 1989. VOL. 114 Discussion As discussed above, we found that previously described methods for the assay of CTC in tissues were unsuitable for statutory residue testing because of poor sensitivity and lack of reproducibility. It has been known for some time that CTC forms the highly fluorescent iso-CTC derivative under alkal- ine conditions.In 1949 Levine et al. I * reported a spectroflu- orimetric method for the assay of CTC in pharmaceutical preparations using this property, but there have been no reports of its use with HPLC. The method that we have developed couples fluorescence detection of the iso-CTC derivative with HPLC to give improved sensitivity and specificity over methods using UV detection of CTC. The detection and determination limits of the assay are 20 and SO ng g-1, respectively. Oxytetracycline and tetracycline do not interfere with the assay. To the best of our knowledge CTC is the only tetracycline to form a fluorescent derivative under the alkaline conditions described. The others require chelation with metal ions such as aluminium or magnesium.The use of a polymeric reversed-phase column eliminates the problems of CTC - silica interactions. It also allows the use of an alkaline mobile phase for separation of iso-CTC, in contrast to silica-based columns where dissolution of silica takes place at high pH. Several types of' solid-phase cartridge were investigated for use in the concentration and clean-up steps of the assay. As reported previously,3 conventional CIK reversed-phase cart- ridges gave low recoveries of CTC unless some form of pre-treatment was applied. We found, however, that either cyclohexyl- or phenyl-bonded cartridges gave good recoveries without any pre-treatment. The former were selected for the assay as they appeared to give a more efficient clean-up of the extracts. The use of solid-phase cartridges also eliminated the need for any liquid - liquid extraction steps, which in previously reported methods also led to low recoveries.A pH of 12 was selected for the mobile phase because it was found that this gave good separation of iso-CTC on the 423 polymeric column and was also a suitable pH for the CTC - iso-CTC reaction. At this pH, conversion was complete in 2.5 h at room temperature (Fig. 2), but in practice the extracts were normally left overnight for the reaction to take place. The use of a higher temperature would have increased the rate of reaction, but it also appeared to decrease the stability of the iso-CTC derivative. The described method is now used routinely in this laboratory for the statutory chemical confirmation of CTC residues in swine muscle. We have found it to be equally suitable. however, for liver or kidney samples. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. References Sharma, J . P., and Bevill, R. F., J. Chromatogr., 1978. 166, 213, Sharrna, J . P., Perkins, E. G., and Bevill, R. F . , J . Chro- rnatogr., 1977, 134, 441. Oka, I-I., Matsumoto. H., and Uno. K., J . Chromarogr., 1985, 325, 265. Onji, Y., Uno, M., and Tanigawa, K.. 1. Assoc. OJf. Anal. Chem., 1984: 67, 1135. Moats. W. A . . J . Chromatogr., 1986, 358, 253. Moats. W. A . . J . Chromatogr., 1986. 366, 69. Knox, J . E l . , and Jurand, J . , J. Chromatogr., 1975, 110, 103. Wolfs, K., Roets, E.. Hoogmartens, J . , and Vanderhaeghe, I<., J. Chromatogr., 1986, 358. 444. Reeuwijk, H. J . E. M.. and Tjaden, U. K., J . Chromutogr., 1986. 353. 330. Tsuji. K., Robertson, J . H., and Beyer, W. F.. Anal. Chem.. 1974, 46, 539. Levine, J . , Garlock, J . R.. and Fisbach, H., J . A m . Pharm. Assoc., 1949, 38, 473. Paper 8103773F Received September 26th, 1988 Accepted December 12t12, 1988

ISSN:0003-2654    

DOI:10.1039/AN9891400421

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST. APRIL 1989, VOL. 114 425 Simultaneous Determination of Fenamiphos, its Sulphoxide and Sulphone in Water by High-performance Liquid Chromatography Rai Singh* Soil Science and Plant Nutrition, University of Western Australia, Nedlands, Western Australia 6009, Australia A method for the simultaneous determination of fenamiphos and its two metabolites, fenamiphos sulphoxide and fenamiphos sulphone, in water is described. The proposed method is based on the separation of the compounds on a silica gel column with ultraviolet detection and overcomes the problems of inadequate separation and possible decomposition associated with the gas-chromatographic determination of these compounds. The application of the method to groundwaters showed that a minimum detection level of the order of 10 pg 1-1 could be achieved easily with pre-concentration of the samples on Sep-Pak C18 disposable cartridges.Keywords: Fenamiphos analysis; fenamiphos metabolites; organophosphorus pesticides; high-perfor- mance liquid chromatography Fenamiphos [ethyl 3-methyl-4-(methylthio)phenyl isopropyl- phosporamidate, Nemacur] is a broad-spectrum. non-volatile, neniaticide insecticide and is commonly used for nematode control in horticultural crops. In soils fenamiphos is oxidised, both microbiologically and chemically, to fenamiphos sul- phoxide (f. sulphoxide) and then to fenamiphos sulphone (f. sulphone). The structural formulae of the three compounds are illustrated in Fig. 1. Both metabolites have nematicidal activity' and are toxic to man.2 The oral LDSo (rats) of the sum of fenamiphos, its sulphoxide and sulphone, expressed as fenamiphos, is 15-19 mg kg-1 as reported by Krause el a / .; The oxidation products are more polar and have more mobility in soil than the parent compound.4 With some gas-chromatographic methods reported in the literature. either the parent compound only,5 or the sulphone as the total residue after permanganate oxidation of both the parent compound and the sulphoxide,6 are determined. Brown' reported a gas-chromatographic method for the determination of fenamiphos and its sulphoxide and sulphone using column chromatography with silica gel to separate the sulphoxide and sulphone. Satisfactory separation of the sulphoxide and sulphone has not been possible on both packed' and capillary8 GC columns. Some aspects of the gas - liquid chromatographic analysis of the sulphone have been discussed by Krause.9 Sulphoxide and sulphone species of organophosphorus pesticides are difficult to determine because of their low volatility, decomposition at elevated temperatures10 and difficult separation.8 High-performance liquid chromatography (HPLC) would appear to be a more suitable alternative to gas - liquid chromatography (GLC).To my knowledge, no HPLC method for the simultaneous determination of fenamiphos and its metabolites has yet been reported. This paper describes the simultaneous determination of fenamiphos, f. sulphoxide and f. sulphone in water by HPLC. Experimental Reagents Fenamiphos, f . sulphoxide and f . sulphone (99% pure) were obtained from Bayer Australia (Botany, NSW, Australia).Liquid-chromatographic grade methanol, acetonitrile and phosphoric acid were obtained from BDH (Sydney, Aus- tralia). Water used to prepare HPLC eluents was distilled twice using a silica still (TQS, Wiesbaden. FRG). On leave from Haryana Agricultural University. Hisar. India. 0 CH3 II / - P - N H -CH I \ O-CC2H5 CH3 CH, Fig. 1. f. sulphone (111) Structural formulae of fenamiphos (I); f. sulphoxide (11) and Sample Pre-concentration Water samples were passed through a 0.45-pm membrane filter (type SM 11306, Sartorius, Gottingen, FRG) and spiked with fenamiphos, f. sulphoxide and f. sulphone to concentra- tions ranging from 2 to 100 pg 1-1. The samples were then passed through Sep-Pak CI8 disposable cartridges (Millipore, NSW, Australia) at a rate of 3-5 ml min-1.The cartridges were pre-conditioned with 5 ml of methanol followed by 5 ml of water. The samples thus concentrated were extracted with 1 ml of acetonitrile and were ready for analysis. Liquid Chromatography Samples were analysed using a high-performance liquid chromatograph (ETP-Kortec, Sydney, Australia) consisting of an HPLC pump (K 35M), an automatic sampler (K 65) and a variable wavelength UV detector (K 95). Data were processed with an LDC/Milton Roy integrator. The compounds were separated on a Spherisorb silica gel column (Phase Separation, Clwyd, UK) (250 X 4.6 mm i.d., particle size 5 pm). The following chromatographic conditionsANALYST. APRIL 1989, VOL. 114 were used: eluent, acetonitrile - water (20 + 80, V / V ) ; pH, 2.2 (adjucted with phosphoric acid): flow-rate, 0.5 ml min-1; detector wavelength, 200 mm: injection volume, 200 p1; temperature.ambient = 20 & 2°C. A Sphericorb ODS (C,,) column (250 X 4.6 mm i.d., particle size 10 pm) (Phase Separation) was also used to separate the compounds. The following conditions were used: eluent, methanol - water (65 + 35, WV); flow-rate, 1.5 ml min-I. Other conditions were as used for the silica gel column. Gas Chromatography - Mass Spectrometry The identities of peaks of fenamiphos, t. sulphoxide and f. sulphone were verified on a gas chromatograph equipped with a mass-selective detector and an H-P 5 column (25 m x 0.31 mm i.d.)? by collecting the individual compounds as they eluted from the HPLC column. The conditions were as follows: injection port temperature, 250 "C: initial column temperature, 175 "C (programmed to 250 "C at 25 "C min-1); carrier gas, helium (44 cm s-1).Detection was carried out with a single-ion monitoring mass spectrometer (Hewlett-Packard 5970) at rti'z 303.2, 303.2 and 320.2 for fenamiphos, f. sulphoxide and f. sulphone, respectively, on the basis of maximum abundance of ions. Results and Discussion A spectrophotometric scan of fenamiphos, f. sulphoxide and f. sulphone at wavelengths between 190 and 400 nm showed absorption maxima for all three compounds at 200 nm. Fenamiphos and f. sulphone gave a further peak of much lower intensity at 248 and 226 nm. respectively. For the simultaneous determination of the three compounds and considering the peak maxima, a wavelength of 200 nm was selected. Liquid Chromatography The three compounds could not be separated simultaneously on an ODS (Clx) column under isocratic conditions because f.sulphone and f. sulphoxide showed similar retention times. Fig. 2 shows the chromatograms of ( a ) f. sulphone and tenamiphos and of ( h ) f. sulphoxide. The sulphone and sulphoxide could be separated by changing the composition of the mobile phase and the flow-rate but the parent compound did not elute from the column. Gradient elution could not be used because of the problem of base-line drift during analysis. al 0 C m +! 2 1) a a) A 0.01 A T B I I I I 0 5 10 15 20 +- a, .- i\l 0.01 A ! -c I 1 I I 0 5 10 15 tim in Fig. 2. Chromatograms of ( u ) A , f. sulphone and B. fenamiphos.t:ach 5 big nil-' i n water; ( h ) C, f . sulphoxide, 5 b~g mlV1 in water. Conditions: eluent, 65% MeOH in water; column, ClS ODS (particle size. 10 pin); flow-rate. 1.5 ml min I ; detector wavelength. 200 nm; detector sensitivity, 0.08 a.u.f.s.; injection volume, 200 1.11 All three compounds could be separated, isocratically, on a silica gel column. A typical chromatogram is shown in Fig. 3. Base-line resolution between the three compounds was achieved in less than 15 min. Calibration graphs with good linearity were obtained for all compounds in the concentration range studied (0.2-4 ug ml-1). Reproducibility The reproducibility of the proposed method was determined by estimating the variation in four replicate runs of the three compounds at a concentration of 1 ug ml-1.The relative standard deviations ("/o) obtained for f. sulphone, f. sulphox- ide and fenamiphos were 3.0, 4.3 and 3.9, respectively. Long-term reproducibility was also determined by comparing the relative response in peak heights of the three compounds at a concentration of 1 pg ml-1 for a number of assays (12 = 5 ) , over a 4 week period. The relative standard deviations (Yo) obtained were 4.0, 4.3 and 5.6 for f. sulphone, f. sulphoxide and fenamiphos, respectively. No significant change in the retention times of the compounds was observed during this period. Based on these observations, no unfavourable effect of the mobile phase on the long-term performance of the column is expected. The low pH (2.2) of the mobile phase should not affect the performance of the chromatographic column as silica is only soluble significantly at a pH above 9.11 The column, connective tubing and pump in contact with the eluent were made of marine-grade stainless steel (SS 316) and were considered to be resistant to phosphoric acid at pH 2.2.However, it must be noted that a satisfactory separation of the three compounds is possible even in the absence of phosphoric acid but the detector response for the compounds is affected? particularly during analysis of groundwater samples. Application to Groundwater Samples The environmental applicability of the normal phase HPLC method was tested with two groundwater samples from horticultural areas of Western Australia. Groundwater from Carnarvon represents a banana plantation area and Wanne- roo groundwater a vegetable growing area around Perth.In the Carnarvon area fenamiphos is used for bananas at 10 1 of active ingredient (a.i.) per hectare twice a year with irrigation water. In the Wanneroo area the recommended amount for various vegetable crops varies from 5 to 10 1 of a.i. per hectare, once or twice per crop growing season (2-3 crops per year). Total organic carbon contents of Carnarvon and Wanneroo waters were found to be 13.9 and 21.1 pg ml-1, respectively. During analysis the water sample with the lower content of organic matter (Carnarvon) did not give any interfering peaks. al V c m e 2 a n 10.002 A l,i, 0 4 8 12 16 [0.002 A L I I L I I 0 4 8 12 16 timin Fig. 3. Chromatograms o f (a) A. f . sulphone; B. f . sulphoxide; and C. fenamiphos, each 1 pg ml-1 in water; and ( h ) unspiked distilled water.Conditions: eluent, 20% MeCN in water (pH 2.2); column. silica gel (particle size, 5 pin); flow-rate, 0.5 ml min-I; detector wavelength, 200 nm; detector sensitivity, 0.02 a.u.f.s.; injection volume, 200 plANALYST, APRIL 1989. VOL. 113 427 1 I I I I 0 4 8 12 16 0 4 8 '12 16 timin Fig. 4. Chromatograms of Wanneroo groundwater ( a ) spiked with A. f . sulphone: B, f . sulphoxide; and C. fenamiphos, each at a level of 1.0 big 1-1; ( h ) spiked with A, 1 pg 1-1 olf. sulphone; B, 1 pg 1 1 off. sulphoxide: and C, 2 big 1 - 1 of fenamiphos. Peak D is unidentified. Dashed line indicates the chromatogram of an unspiked groundwater sample. Conditions: as for Fig. 3, except detector sensitivity ( a ) 0.04 o r ( h ) 0.08 i1.u.f.s.The other sample, however, produced a significant peak, tailing up to the retention time of fenamiphos. Sample clean-up Cleaning of the samples was attempted with Sep-Pak cart- ridges of florisil and silica with limited success. The extraction of the compounds retained on the cartridges by acetone (as mentioned by Peterson and Winterlins) resulted in a highlq turbid extract due to precipitation. and low recoveries after filtration. Clean-up of the compounds is difficult and there is also the possibility of breakdown of fenamiphos during the clean-up procedure .s Hence camples were processed without clean up. In Fig. 3 , chromatograms of water with a high organic carbon content with and without spiking the three cornpounds after pre-concentrating ( a ) 100 and ( b ) 1000 times are shown.The higher level of pre-concentration resulted in increased interference [Fig. 4(6)]. No fenamiphos was detec- ted in unspiked samples. Recoveries The recoveries of the three compounds from three different surface u aters and groundwaters were quantified using a silica gel column. The average percentage recoveries from the water samples spiked at various levels with the three compounds are given in Table 1. Better than 73% recoveries for all three ;ompounds were obtained with different quality waters. Generally, low recovery values with higher variability were obtained in groundwaters at the lowest spike level (2 pg I-'). The recoirery was sufficiently reproducible and for ground- water samples a correction factor for 100% recovery can be applied.The detection limit of the method will depend upon the interference by organic matter. Based on these results however, detection levels of at least 10 pg 1-1 can be achieved from groundwaters. Table 1. Recovery of fenamiphos, f . sulphoxide and f . sulphone from water samples spiked at three levels Recovery, " '% Spike/ Distilled Carnarvon Wanneroo Compound Fig 1 water water water F.sulphone . . 100 87.2(2.3) 82.6(3.7) 94.4(3.5) 10 99.0 (2.7) 107.9 (5.3) X2.0 (3.7) F.sulphoxide . . 100 8X.5(2.9) 96.0(5.6) 92.6(2.1) 10 (17.0 (3.9) 79.2 (3.6) 86.8 (4.2) Fenamiphos . . 100 90.4 (2.4) 92.9 (5.X) 102.6 (8.9) 2 87.9 (4.4) x3.3 (9.9) 72.9 (6.0) 2 X1.2(3.5) 81.8(11.4) XX.S(2.0) 2 84.0 (3.3) 74.7 (2.0) 76.1 ( 1 3) 10 85.8 (3.0) 91.9 (6.2) 101.1 (3.4) * Values in parentheses are the standard errors ((YO) based on three replicate measurements. Conclusion With high - pe r f o r m an ce 1 i q ui d chromatograph y ten am i p h o s and its two major metabolites can be determined, simul- taneoucly. in water samples at low levels.Detection level\ of the order of 10 ug 1-1 can be achieved easily by pre- concentrating the samples on Sep-Pak Cls cartridges. All experiments were conducted in the laboratories of the CSIRO Division of Water Resources, Wembley , Western Australia. The author is grateful to Robert Gerritse and John Adeney (CSIRO) and Graham Aylmore (University of Western Australia) for their assistance and advice. The Australian International Development Assistance Bureau has supported the research by granting a fellowship to the author. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. Waggoner, T. B.. and Khasawinah, A. M.. XesidiieRev., 1974, 53, 79. Waggoner, '1'. B., J . Agric. Food Chenz., 1972, 20. 157. Krause, M.. Loubser, J . T.. and de Beer, P. R . , J . Agvic. Food Chem., 1986, 34, 717. Bilkert. J . N., and Rao. P. S. C.. J . Eizviron. Sci. Health.. 19x5, H20, 1. Sagredos, A . N., and Eckert, W. R . , Bcitr. T(ihukfbr.wh.. 1977, 9, 107. Thornton, J . S., J . Agric. Food Chenz., 1971, 19, 890. Brown. M. J . , J. Agric. Food Chem., 1981. 29. 1129. Peterson, D., and Winterlin. W., J . Agric. Food Cheni.. I9X6, 34, 153. Krause, M.. Analysr, 1985. 110, 673. Hill, A . R . C.. Wilkins. J . P. G.. Findlay, N. R. I.. and Lontay, K. E. M., Analyst, 1984, 109, 48.3. Unger, K. K . . "Porous Silica. Its Properties and Use as Support in Column Liquid Chromatography." Elsevier, Amsterdam. 1979. p. 13. Pciper 81032 70J RectGiwi Aiigirst lOth, 1988 Accepred December Sth, 1988

ISSN:0003-2654    

DOI:10.1039/AN9891400425

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

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

ISSN:0003-2654    

DOI:10.1039/AN9891400429

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST, APRIL 1989, VOL. 113 435 Synthetic Inorganic Ion-exchange Materials Part XLIX.* Adsorption and Desorption Behaviour of Heavy Metal Ions on Hydrated Titanium Dioxide Mitsuo Abe, Peng Wang,t Ramesh Chitrakar and Masamichi Tsuji Department of Chemistry, Faculty of Science, Tokyo Institute of Technolog y, 2- 12- I , Ooka yama, Meguro-Ku, Tokyo 152, Japan The ion-exchange selectivity of a number of divalent metal ions was studied as a function of pH in nitrate and chloride media. The order of selectivity was Pb2+ > Hg2+ > Cd2+ > Mg2+ in nitrate solution and Pb2+ > Cd2+ > Ca2+ > Mg2+ > Hg2+ in chloride solution. A good linear relationship between the logarithm of the distribution coefficient of the heavy metal ions and their effective ionic radius was found.On the basis of the Kd values, the separation of a mixture of Cd2+, Hg2+ and Pb2+ and the group separation of Hg2+ and Pb2+ from several common metal ions were achieved on an ion-exchange column containing amorphous hydrated titanium dioxide. Keywords ; Hydrated titanium dioxide; ion exchange; chromatographic separation; transition metals; lead separation Many useful inorganic ion-exchange materials have been synthesised during the last two decades and many have found application in the fields of analytical chemistry, radio- chemistry, environmental chemistry and biochemistry. 1 Inor- ganic ion exchangers possessing high selectivities for certain ions or groups of ions*-' can be utilised for the chromato- graphic separation of many elements. Hydrated titanium &oxide (HTDO) has attracted considerable attention as an ion exchanger owing to its high sorption selectivity for certain metal ions.Many reports have described the basic characteris- tics and sorption properties of this material.&-11 Three types of HTDO are known: amorphous, anatase and rutile."Il Several new types of HTDO, e.g.. crystalline HTDO fibres"-13 and crystalline HTDO with a layered structure, have been prepared recently. 14-16 The applications of HTDO have been mainly concerned with the recovery of uranium from sea- water."-1') Little effort has been devoted to the chromato- graphic separation of heavy metal cations, except for the separation of 1 % ~ - YSr and 1 W s - '"Pr 20 and chromate and phosphate ions.21 More recently, separations of Ni2+ - Co2+ and Zn*+ - Cd'+ - Cu2+,11 Cu*+ from Mg2+ - Ca2+ - Sr2+, Cu2+ from Ni2f - Co*+, Mg*+ from Ba2+ - Zn2+ 22 and Cu2+ from Zn?+ - Mn2+ - Co2+ - Ni2+ 23 have been achieved with an HTDO column.Heavy metal ions such as Hg2+ and Pb*+ are hazardous and are frequently discharged in industrial wastewater; much attention has been directed towards studying their effect as environmental pollutants. Hence there is a need to develop a method to separate these biologically toxic elements from common metal ions also present in contaminated natural w a t e r s . This paper describes the adsorption and desorption behav- iour of a number of heavy metal ions on amorphous HTDO (Am-HTDO) at different pH values and in different media. Their selectivities and the applicability of some chromato- graphic separations of these heavy metal ions are also discussed.Part XLVIII of this series has been submitted for publication in i Present address: Harbin Institute of Technology, Department of Solveizt Extr. Ion Exch. Applied Chemistry, Harbin, China. Experiment a1 Reagents The reagents used were supplied by Wako Pure Chemical Industries (Japan). Titanium tetrachloride was of the highest available grade (>99% as metal); all other chemicals were of analytical-reagent grade. Standard solutions of the metal ions were prepared by dissolving the high-purity metals (>99.9%) in the minimum amount of nitric acid. A solution of Hg2+ was prepared by dissolving mercury(I1) nitrate in 5% (VIV) nitric acid in order to prevent hydrolysis of the Hg*+ ion.Preparation of the Am-HTDO Ion Exchanger This was prepared as described previously. I1.Z3 Water (150 ml) was added slowly to TiCI4 (SO ml) and a 2 . 8 ~ sodium hydroxide solution (800 ml) was added to the resulting solution. The precipitate formed was washed with de- mineralised water, left for 2 d and filtered under suction. This procedure was repeated until the pH of the filtrate reached a constant value (about pH 12). The precipitate was then air-dried at room temperature until it had been transformed into a semi-transparent glassy gel. This gel was immersed in dG-mineralised water to break it down into fine particles. The product was ground and sieved to 100-200 mesh size. In order to obtain the hydrogen ion form, the sample was conditioned as follows.About 10 g of the sample were placed in a column (10.0 X 1.0 cm i.d.) and 0.1 M hydrochloric acid was passed through the column until the amount of Na+ ion in the effluent was negligible ("a+] < 1 0 - j ~ ) . The sample was then washed with de-mineralised water until the pW of the supernatant solution was in the range 4-S. The resulting product was air-dried to a constant mass. Apparatus A Varian-Techtron 1100 atomic absorption spectrometer was used for the determination of the metal ion content in the aqueous phase. A Model HM-5B TOA pH meter was used for all pH measurements. X-ray powder diffraction patterns were recorded using a JEOL Model JDX-7E X-ray diffractometer with Ni-filtered Cu K a radiation. A Rigaku Denki Model 8001 Thermoflex was used for the thermal analysis studies at a heating rate of 10°C min-1 by employing a-A1203 as the reference material.Infrared spectra were measured by the436 ANALYST, APRIL 1989. VOL.. 113 KBr disc method with a JASCO Model DS-701G infrared spectrometer. Equilibrium Studies The equilibrations were carried out batchwise. A 0.10-g amount of Am-HTDO in the hydrogen ion form was treated with 10.0 ml of a 1.0 x 1 0 - 4 ~ solution of the metal ions at various pH values with a constant ionic strength of 0.1. The mixtures were shaken intermittently at 30 i 0.5"C. After equilibrium had been attained, the pH and metal ion content of the supernatant solutions were determined. The distribu- tion coefficient ( K d ) was calculated from the following equation: amount of metal ion in exchanger amount of metal ion in solution Kd = X volume of solution (ml) mass of exchanger (g) The concentrations of the metal ions in the exchanger were calculated from the differences between the initial and final concentrations in the aqueous phase.The separation factor, a(AIB), was determined from the equation where &A and Kc,B are the distribution coefficients of ions A and B, respectively. Chromatographic Separation Separations of mixtures of metal ions were carried out on a column ( 5 X 0.5 cm i.d.) of Am-HTDO in the hydrogen ion form. A mixed solution containing 1 pmol of each metal ion was loaded on the top of the column and then eluted with different eluents. The eluents were charged continuously with a high-pressure pump (Nihon Seiinitsu Kagaku, Model NP-DX-2).The effluents were collected by using a drop- counting type fraction collector (Ohtake Works, Model UM-160). The effluent fractions were analysed for their metal ion concentration and the recoveries of the metal ions were calculated from the difference between the initial and equilib- rium concentrations. Results and Discussion Characterisation of Am-HTDO Amorphous HTDO is a semi-transparent glassy gel. The samples were characterised both before and after conditioning with 0.1 M I1C1 by X-ray diffraction, infrared spectra and thermal analysis. The results showed good agreement with those reported previously.11 25 The composition of the gel in the hydrogen ion form can be written as TiO?. 1 .6H20. Rate of Metal Ion Sorption The rate of exchange of different metal ions was measured qualitatively in order to determine the equilibrium di\tribu- tion coefficients ( K d values).Fig. 1 shows the rate of sorption of the metal ions on Am-HTDO from their nitrate or chloride solutions. The ion-exchange reactions of these metal ions are relatively fast. An exchange equilibrium appears to be reached within 4 d; about 4-7 d are required to determine the equilibrium KCi valueu. The K,, values were therefore deter- mined by equilibrating the solid with the metal ion solution for 7 d. 100 ( a ) Final pH I 0 2 4 6 PH 3.20 K+ 5.10 -x Cd2+ Hg2+ 1.75 20 Fig. 2. Values of K,, for metal ions in nitrate media. Exchanger, 0.10 g; total volume, 10.0 ml; concentration of metal ion, 1 x 10 A M ; temperature. 30 IL 0.5 "C; and ionic strength.0.1 ki (NaNO, + FINO, or NaOH) 104 I 1 I 2 4 6 8 I I I 0 2 4 6 8 0 ' ' Timeid Fig. 1. Rate o f metal ion sorption. Exchanger, 0.25 g; total volume, 25.0 ml; concentration of metal ion, 1 x 10-4 M ; ionic strength, 0.1 M ; and temperature, 30 IL 0.5 "C. (a) Nitrate solution (NaNO? + HNO, or NaOH); and ( b ) chloride solution (NaCI + HCI or NaOH) I I I U" 0 2 4 6 8 PH Fig. 3. 0.1 M (NaC1 + HCI or NaOH). Other conditions as in Fig. 2 Values of K,, for metal ions in chloride media. Ionic strength,ANALYST. APRIL 1089, VOL. 114 437 Table 1. Values of the distribution coefficient (K,/ml 8-1) and separation factor (a) on Am-HTDO and Bio-Rad AG 50W-X8’7-3 Ions Ion exchanger Solution Am-HTDO . . . . 0.IMNaCl (pH 5 ) Am-HTDO . . . . 0.1 M N ~ N O ~ (PH 3.5) Bio-Rad AG50W-X8 .. . . O.2MHCI Bio-Rad AGSOW-X8 . . . . 0 . 2 ~ H N 0 3 Hg2- 33 2.42 Mg’+ 8 1.8 Hg’ 0.9 68.9 ME’+ 295 1.3 Mg’+ 80 Ca’ t 15 5.38 10.7 PW 1 62 1.35 Cd’+ 392 1.2 Ca7+ 430 4.65 Cd’+ 160 43.8 Cd2 + 84 6.3 Ca’ 480 2.3 Cdz+ 2000 >5 Hg?+ 7000 >1.5 Mg2- 530 1.5 Hg2 1 1090 1.3 PW+ > 104 Pb’ + > 103 Ca‘ + 790 Pb’ + 1420 103 r I 0 - E , Ic“ 102 I 0 1 2 3 4 PH Fig. 4. Conditions as in Fig. 2. 0, NO,-; A , CI-; and 0. Br- Values of Kd for Pb’+ on Am-HTDO in different media. Hg2+ - Mn2+ I I 0 0.5 1.0 1.5 EWA Fig. 5 . Relationshi between log Kd and effective ionic radius (EIR) of heavy metal ions ?correlation coefficient, 0.9964). Conditions as in Fig. 2 Distribution Coefficients (Kd Values) A graph of the logarithm of the Kd values of the metal ions against the final pH values (Figs.2 and 3) shows a linear relationship with a slope of about unity (1 J-1.3) for divalent ions and 0.33 for the K+ ion for all the systems studied at an initial concentration cf metal ion of 1 X 10-4 hi. The Kd value for the K+ ion was included as being representative of those alkali metal ions that are usually present in actual water samples. The slopes do not correspond to the valency of the ions exchanged. Similar results were obtained for the adsorp- tion properties of metal ions on amphoteric inorganic ion exchangers. 11 This might be due partly to the uptake of anions (e.g., Cl-) because of the amphoteric behaviour of Am- HTDO and partly because the amount of dissociated H+ responsible for ion exchange was not constant because of the weakly acidic nature of Am-HTD0.l6 For Am-HTDO, the following order of selectivity was observed: Pb*+ > Hg2+ >> Cd*+ > Ca*+ > Mg*+.Of these ions Pb2+ and Hg*+ exhibit very high Kd values in nitrate media (Fig. 2), whereas the selectivity order was Pb2+ >> Cd2+ > Ca2+ > Mg2+ > Hg2+ in chloride media (Fig. 3). In the latter instance a very low Kd value for Hg2+ was observed due to the formation of a chloride complex. The large difference in the values of Kd for Hg2+ in the two media should make it possible to separate Hg2+ from other matrix ions. Fig. 4 shows the Kc1 values for Pb*+ on Am-HTDO in different media. Lower Kd values were found in bromide solution, which may be due to the formation of a bromide complex. The Kd values and separation factors, a(A/B), are sum- marked in Table 1.the values obtained on Bio-Rad AG 50W-X827-29 being included for comparison. Selectivity for Heavy Metal Ions The selectivity order of Am-HTDO for heavy metal ions was found to be Pb*+ > Hg*+ > Cd*+ (Fig. 2). The log Kd values at pH 3.15 for the heavy metal ions correlated very closely to their effective ionic radius (EIR) (correlation coefficient, 0.9964) as shown in Fig. 5 . The EIR values used were those reported by Shannon.”) A similar selectivity order has been reported previously. 11 This selectivity order might be explained as follows. For the heavy metal ions, the observed order of selectivity was the same as that of their increasing ionic radii, i.e., their decreasing hydrated ionic radii.This suggests that the energy required for the dehydration of the metal ions so that they can occupy a site in the exchanger plays an important role in determining the selectivity series for the transition metal ions.22 On the other hand, according to the principle of hard and soft acids and bases (the HSAB principle), hard acids prefer to bind to hard bases and soft acids to soft bases.31 The hydrated titanium dioxide (HTDO) can act as a Lewis base and the heavy metal ions as Lewis acids. The softness of the cations increases as the ionic radius of the heavy metal ions increases. The interaction of transition metal ions with438 ANALYST, APRIL 1989, VOL. 114 E + 16 20 24 28 32 36 Fraction No. Fig. 6. Chromatographic separation of heavy metal ions.Exchanger, 1.25 g; column, 5 x 0.5 cm i.d.; loading, 1 pmol of each metal ion; and flow-rate, 0.15 ml min-1. Volume of each fraction, 7 ml r- 0 . 5 ~ HBr 16 20 24 28 32 36 * 0.002 M HNO, Fraction No. Fig. 7. common metal ions. Conditions as in Fig. 6 Chromatographic separation of Hg2+ and Pb2+ from some HTDO, acting as a soft base, could be expected to increase with an increase in EIK. Ion-exchange Chromatographic Separation It is evident from studies of the Kd values on Am-HTDO that some selective separations should be feasible for these heavy metal ions. Hence a number of separations were attempted on a column of this material by using various eluents. Separation of cadmium, mercury and lead On the basis of the distribution coefficients, the separation of a mixture of Cd*+, Hg2+ and Pb2+ was attempted; their elution curves are shown in Fig.6. A 0.01 M nitric acid medium was used as the eluent for Cd2+, 0.01 M hydrochloric acid for Hg’& and 0 . 5 ~ hydrobromic acid for Pb2+. Complete separation was achieved with recoveries of 100,98, 98.8% for Cd2+, Hg2+ and Pb2+, respectively. Separation of mercury and lead f r o m common metal ions This was performed with 0 . 0 0 2 ~ nitric acid for Na+, K+, Mg2+ and Ca2+, 0.01 M hydrochloric acid for Hg2+ and 0.5 M hydrobromic acid for Pb2+ (Fig. 7). Recovery of the common metal ions was quantitative. The yields of Hg2f and Pb2+ were quantitative (99’/0) with complete separation. The excellent separation achieved on Am-HTDO can be applied to the separation of highly toxic metal ions, e.g., Hg2+ and Pbzf, in drinking water. In conclusion, Am-HTDO is useful for the concentration and chromatographic separation of highly toxic heavy metal ions such as Hg2+ and Pb2+ from common metal ions and can be applied to the determination of very low concentrations of Hg2+ and Pb2+ in natural water.1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23, 24. 25. 26. 27. 28. 29. 30. 31. References Clearfield, A . , Editor, “Inorganic Ion Exchange Materials,” CRC Press, Boca Raton, FL, 1982. Vesely, V., and PakBrek, V., Talanra, 1972, 19, 219. Abe, M., and Ito, T., Bull. Chem. SOC. Jpn., 1967, 40, 1013. Abe, M., and Uno, K., Sep. Sci. Technol., 1979, 14, 355. Abe, M., and Ito, T., Kagaku Zasshi, 1967, 70, 440. Abe, M., and Ito, T., Nippon Kagaku Zasshi, 1965, 86, 814.Amphlett, C. B., McDonald, L. A . , and Redman, M., J . Inorg. Nucl. Chem., 1958, 6 , 236. Babyrenko, Yu. Ya., Domaltov, Yu. D., and Bragina, M. I . , Zh. Prikf. Khim. (Leningrad), 1970, 43, 1152. Heitner-Wirguin, C . , and Albu-Yaron, A., J . Inorg. Nucl. Chem., 1966, 28, 2379. Levi, H. W . , and Schiewer, E., Radiochim. Acta, 1966, 5 , 126. Abe, M., Tsuji, M., Qureshi, S. P., and Uchikoshi, H., Chromatographia, 1980, 13, 626. Fujiki, Y., Komatsu, Y . , Sasaki, T., and Ohta, N., Nippon Kagaku Kaishi, 1981, 1656. Sasaki, T., Kornatsu, Y., and Fujiki, Y., Chem. Lett., 1981, 957. Marchand, R . , Brohan, L., and Tournoux, M., Mater. Res. Bull., 1980, 15, 1129. Ohta, N., and Fujiki, Y., Yogyo Kyokai Shi, 1980, 88, 1. Izawa, H . , Kikkawa, S . , and Koizumi, M., J . Phys. Chem., 1982, 86, 5023. Ogata, N., Nippon Kaisui Gakkai-Shi, 1971, 24, 197. Keen, N. J., J. Br. Nucl. Energy SOC., 1968, 7, 178. Yamashita, H., Ozawa, Y., Nakajirna, F., and Murata, T.. Nippon Kagaku Kaishi, 1978, 1057. Lavrukhina, A. K., Malyshev, V. V., and Rodin, S . S., Zh. Vses. Khim. Ova., 1963, 8, 227. Kraus, K. A., Phillips, H. O., Carlson, T. A., and Johnson, J. S . , “Proceedings of the 2nd International Conference on the Peaceful Uses of Atomic Energy,” IAEA, Geneva, 1958, Volume 28, p. 3. Sasaki, T . , Komatsu. Y., and Fujiki, T . , Solvent Extr. Zon Exch., 1983, 1, 775. Sasaki, T . , Kornatsu, Y., and Fujiki, T., Sep. Sci. Technol.. 1983, 18, 49. Inoue, Y., and Tsuji, M., Bull. Chem. SOC. Jpn., 1976.49. 11 1. Tsuji, M., and Abe, M., J . Radioanal. Nucl. Chem., Articles, 1986, 102, 283. Inoue, Y., andTsuji, M., Bull. Chem. SOC. Jpn., 1978,51,479. Strelow, F. W. E., Rethemeyer, R., and Bothma, C. J. C., Anal. Chem., 1965, 37, 106. Strelow. F. W. E., Rethemeyer, R . , and Bothma, C. J. C., Anal. Chem., 1960, 32, 1185. Strelow, F. W. E . , Rethemeyer, R . , and Bothma, C. J . C . , Anal. Chem., 1971,43, 7. Shannon, R . D., Acta Crystallogr., Sect. A , 1976, 32, 751. Pearson, R. G., J. Chem. Educ., 1968, 45, 581 and 643. Paper 8104367A Received November 2nd, I988 Accepted December 14th, I988

ISSN:0003-2654    

DOI:10.1039/AN9891400435

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST, APRIL 1989, VOL. 114 439 Synthesis, Characterisation and Metal Sorption Studies of a Chelating Agent-loaded Anion-exchange Resin Lucy Joseph and Vadasseril N. Sivasankara Pillai Department of Applied Chemistry, Cochin University of Science and Technology, Cochin-682 022, India A selective chelating resin was prepared by loading pyrogallolsulphonic acid (PGS) on a conventional anion-exchange resin, Seralite SRA-400. The loaded resin was stable in 0.1 M HCI. Its equilibrium sorption capacity as a function of pH and the exchange kinetics for various metal ions were studied. The loaded resin showed high selectivity towards MoVl, V V and Fell'. The sorption capacities of the PGS-loaded resin for Felll, V V and MoV1 were 0.69, 0.75 and 0.4 mmol g-1, respectively. The loaded resin was used for the separation and enrichment of these metals.Keywords: Metal sorption; chelating sorbent; p yrogallolsulphonic acid; ion-exchange resins Ion exchange has evolved as a major technique for the separation and enrichment of metal ions at trace levels. The selectivity of chelating exchangers is often superior to that of conventional ion exchangers. Investigations into the synthesis and application of chelating exchangers are being actively pursued by a number of workers. The use of these resins in the field of analytical chemistry is limited by the complexity of their synthesis. Recentiy, the potential of anionic chelating agents loaded on anion-exchange resins was discussed by Brajter.1 A series of papers have been published by Brajter and co-workersz-6 describing the synthesis and application of aromatic complexing agents containing sulphonic acid groups attached to anion-exchange resins.Going et u1.7 have reported that 2-(3'-sulphobenzoyl)pyridine-2-p~ridylhydrazone can be used for the pre-concentration and separation of Fe, Co, Ni, Cu, Zn, Cd, Hg and Pd. which can then be determined by atomic absorption spectrometry (AAS). Chromotropic acid, 5-sulphosalicylic acid and 7-iodo-8-hydroxyquinoline-5-sul- phonic acid have been employed by Lee and Lees for the pre-concentration of metal ions prior to their determination by AAS. The pre-concentration of Cu and Hg has been studied by Chikuma et al.9 using the suiphonic acid derivative of dithizone. Based on this principle we have synthesised a selective resin by incorporating pyrogallolsulphonic acid (PGS) in a commer- cially available anion-exchange resin, Seralite SRA-400.When attached to the anion-exchange resin, PGS exhibits some interesting chelating properties, similar to those of TironlO (disodium 4,5-dihydroxybenzene-1,3-disulphonate), and provides efficient separations of a number of metal ions. This paper describes the synthesis, characterisatipn and metal sorption characteristics of PGS-loaded resin (R-PGS). Experimental Apparatus and Chemicals An ES Research Model 1400 pH meter with a combination electrode system, a Hitachi Model 200-20 UV - visible spectrophotometer and a Plasmascan (Labtam) Model 8410 inductively coupled plasma atomic emission spectrometer were used. All chemicals used were of analytical-reagent grade.Metal ion solutions were prepared by dissolving the salts in doubly distilled water and concentrations were checked by standard methods. The chloride form of a commercially available strongly basic anion-exchange resin, Seralite SRA-400, with 8% cross-link- ing (SISCO Research Laboratories, Bombay, India) was used for preparing the chelating agent-loaded resin. Pyrogallolsul- phonic acid was prepared using the procedure described by Pollak and Fulnegg.11 Preparation of R-PGS Seralite SRA-400 (chloride form) was powdered and sieved and particles of 10&200 mesh size were soaked in an aqueous solution of PGS for 8 h with occasional stirring. The resin was filtered, washed with 0.1 M HC1 and then with water to render it neutral and dried at 60 "C for 24 h.The resin was allowed to swell and then conditioned in the appropriate buffers before sorbing the metal ions. Measurement of PGS by Spectrophotometry A sample of purified PGS was dissolved in distilled water and the solution was diluted to 100 ml in a calibrated flask to give a concentration of 1% mlV. Aliquots of this solution were diluted to give concentrations ranging from 0.002 to 0.02% mlV and the absorbance of the solutions was measured at 438 nm. A calibration graph of absorbance versus concentration was constructed. Kinetics of the Sorption of PGS on Anionic Resin A weighed portion of the anion-exchange resin was shaken with 20 ml of reagent solution of known concentration and the residual concentration of the reagent was measured at various intervals.Sorption Studies The sorption characteristics of R-PGS for Fe"', CU", MI1, Mn", MoVI, VV and Be" were studied by batch and column methods. The pH of the solutions was adjusted using 0.1 M NaOH solution or 0.1 M HC1 unless specified otherwise. The concentrations of all metal ions except MoVI were determined by UV - visible spectrophotometry. For MoVI, inductively coupled plasma atomic emission spectrometry (ICP-AES) was used. Kinetics of the Sorption of Metal Ions on R-PGS Weighed portions of the loaded resin were shaken with 20 ml of buffered solutions of the metal ions (400 pg ml-1). Aliquots of the solutions were withdrawn at various intervals using a filter stick fitted with a sintered-glass disc and the residual coricentrations of the metal ions were measured.440 ANALYST, APRIL 1989, VOL.114 Effect of pH on the Sorption of Metal Ions A sample (0.1 g) of the loaded resin was shaken with 20 ml of the metal ion solution (400 pg ml-1) at different pH. After 24 h, the residual concentrations of the metal ions were measured. Breakthrough Studies For the column technique, the metal ion solutions (after adusting their pH) were passed through a glass column (10 x 0.5 cm i.d.), packed with the resin and pre-conditioned with the appropriate buffer, at a flow-rate of 0.5 ml min-1. The effluent fractions were collected in 10-ml portions and analysed for the presence of the metal. Separation of Metal Ion Mixtures on R-PGS Synthetic mixtures of metal ions were used in this study. The solutions were buffered appropriately and 1-ml aliquots were passed through a column of R-PGS.The metal ions were 0.6 0.4 01 C m e a a a 0.2 C 350 400 450 500 Wavelengthinm Fig. 1. pH dependence of the absorption spectrum of PGS. Concentration of PGS, 0.01% m/V. A. pH 1; B, pH 2; C, pH 3; D, pH 5-9; and E, pH 11 eluted at a flow-rate of 0.5 ml min-1. The eluate was collected in 10-ml portions and the metal ion concentrations were determined. Results and Discussion Nature of PGS The sodium salt of PGS is soluble in water. The compound is sensitive to air oxidation and can be partially converted to the quinone form as indicated by the presence of carbonyl absorption bands in the infrared spectrum. The band assign- ments are 3441 cm-1 (phenolic OH), 1624 cm-1 (C=O) and 1168 cm- (SO,-).Pyrogallolsulphonic acid absorbs in the visible region with an absorption maximum at 438 nm. The absorption spectrum in the visible region is not affected by changes in pH in the range 5-9 (Fig. 1) and a calibration graph of absorbance versus concentration gives a straight line with a slope of 5.9 1 g-1 cm-1, which passes through the origin when extrapolated. The resin used in this work had a reagent loading capacity of 0.8 mmol g-1. Studies on the sorption behaviour of PGS have shown that the rate of uptake on Seralite SRA-400 is very fast; half-saturation occurs within 5 min (Fig. 2). The tenacity of the loading was tested by treating the loaded resin with HCl of 0 40 80 120 Timeimin Fig. 2. Kinetics of the sorption of PGS on Seralite SRA-400. Amount of resin, 0.5 g; volume of 2% PGS solution, 20 ml; and temperature, 28 k 1 "C Table 1.Quantitative separation of metal ions achieved Sample Separation No. achieved Eluent 1 Cull Citrate buffer (pH 2.5) Fell1 2% ascorbic acid in 0.1 M HCl 2 Nil1 Citrate buffer (pH 4) Fc"1 2°/0 ascorbic acid in 0.1 M HCI 3 Mn" Citrate buffer (pH 4) Fell1 2% ascorbic acid in 4 Fell1 2°/0 ascorbic acid in 0.1 M HCI 0.1 M HCl MOV' 0. I M HCIOd 5 Mn" Very dilute HCl (pH 4) VV 2% ascorbic acid in 6 VV 2% ascorbic acid in 0.1 M HCl 0.1 M HCj MeV' 0.1 M HClO, CUT' Citrate buffer (pH 2.5) Fe"' 2% ascorbic acid in 0.1 MHCI 7 Mn" Citrate buffer (pH 4) Eluate volume/ ml 5 0 40 30 40 30 3 0 30 40 40 40 30 40 40 40 30 Amount Amount Relative loaded/ recovered/ standard 100 98.2 0.84 pg* deviation.% 100 102.7 0.86 100 99.7 0.38 100 101.4 0.75 100 99.3 0.5s 100 103.5 0.40 100 100.7 0.69 100 98.2 0.77 100 99.3 0.25 100 98.9 0.43 100 99.5 0.22 100 98.7 0.37 100 99.7 0.52 100 98.5 0.84 100 102.0 0.60 * Mean of five determinations.ANALYST, APRIL 1989, VOL.114 44 1 various strengths. The elution is negligible with 0.1 M HCI: the same result was found using 0.1 M NaCl solution. However, as the concentration of the electrolyte is increased the degree of elution increases and more than 80% of the PGS is leached out by 30 bed volumes of 1 M NaCl solution. Sorption of Metal Ions on R-PGS There are two possibilities for the sorption of metal ions on R-PGS: (1) the oxygen atoms of PGS co-ordinate with the metal ions thus keeping them on the resin and (2) the metal may form an anionic complex which in turn may exchange at the cationic sites of the resin by displacing PGS; which of the two processes predominates is determined by the relative stabilities of the two complexes.The formation of an anionic complex normally requires either a strongly complexing anion such as CN- or F- or a high concentration of a weakly complexing anion. As only solu- tions of low ionic concentrations are involved here, the latter situation doe5 not arise. Most of the studies were carried out in the pH range 2-6 in C1-, NO3-, S042-, C104-, acetate or citrate media. Whenever solutions of high ionic strength were used, experiments were conducted to determine the elution pattern of PGS with these solutions. For MoV1, FeIII, V V and Be", the metal to ligand ratio varies from 1 : 1.9s for MoVI to 1 : 0.28 for Be".These metals are known to form strong complexes with 0,O-donors. This means that the ligand, although present in excess, occupies only some of the co-ordination sites of the metals, the remainder of the sites being occupied by water. The restricted mobility of the ligand prevents it from adopting a preferred orientation around the metal ion. For Ni", Cur1 and MdI ions, the low sorption capacity is a consequence of their lack of affinity for oxygen donors. Sorption of iron(ZZZ) Iron(II1) solutions (400 pg ml-1) were studied in 0.1 M citrate buffer at various pH to prevent the precipitation of hydrolysed iron. At higher concentrations of Fell', slow elution of the reagent took place. The sorption capacity increases with pH and reaches a maximum of 0.69 mmol g-' at about pH 4 (Fig.3). 0.8 0.6 c I 07 0.2 I 0 2 4 6 PH Fig. 3. Effect of pH on the sorption of metal ions on R-PGS. Amount of resin, 0.1 g; particle size, 10C200 mesh; reagent loading on the resin, 0.8 nitnol g-1; volume of solution, 10 ml; amount of metal ion, 4.0 mg; and shaking time, 24 h. A . Vv; B, Fell1; C, MoV1; D. Cu"; E, Nill; and F, Mn" Sorption of vanadium( V ) The sorption capacity of the loaded resin for VV is low at low pH. At about pH 2 it increases suddenly and reaches a maximum of 0.75 mmol g-1 at pH 3, then remains constant (Fig. 3). In fact, VV was found to show the maximum capacity among the various metal ions studied. As V V is a fairly strong oxidant, PGS. which is a polyhydroxyphenol, may be easily oxidised. However, no leaching of the reagent was observed under these conditions. Sorption of molybdenum( VZ) The sorption capacity of the loaded resin for MoVI increases with increasing pH, reaches a maximum of about 0.4 mmol g- at pH 2-4 and then decreases slightly (Fig.3). Molybdenum is also known to oxidise polyhydroxyphenols in acidic medium.12 This may be the reason for the low sorption capacity observed at low pH. Sorption of copper(ZZ), nickel(ZZ) and manganese(ZZ) The sorption capacity for Cu" increases with increasing pH and reaches a maximum of 0.2 mmol g-1 at about pH 5 (Fig. 3). The sorption capacity for Ni" and Mn" is very low on R-PGS. This is to be expected as these metal ions have a low affinity for oxygen donors. 0 40 80 Timeimin Fig.4. Kinetics of the sorption of metal ions on R-PGS. Amount of resin, 0.1 g; particle size, 10@200 mesh; reagent loading on the resin, 0.8 mmol g-1; volume of solution, 10 ml; amount of metal ion, 4.0 mg; and temDerature, 28 5 1 "C. A , ReII; A, FeIIT; 0, MoV1; 0, Vv; and e, cull ' 0.6 0.4 s! u- 0.2 i 0 200 400 600 800 1000 Eff I u ent vo I u m eim I Fig. 5. Breakthrough curves for metal ions on R-PGS. Column, 10 x 0.5 cm i.d.; particle size, 10C200 mesh; concentration of metal ion solutions, 100 pg ml-1. pH of solutions: VV and MoVI, 3; FeIII, Nil1, Cu" and Mn", 4. H, Vv; El, MoV1; A, FelI1; A , Nilr; a, CuT1; and 0, Mn". C, = Concentration of metal ions in the applied solution: C, = concentration of metal ions in the effluent442 ANALYST, APRIL 1989, VOL.114 /&----- i a ) 1 2 Nil' -- I CU" Volume of eluateiml Fig. 6. Elution curves for ( a ) Nil1 - Fell1 and ( b ) Cull - FeIII. 1 , Citrate buffer ( H 3); 2, 2% ascorbic acid in 0.1 M HCI; and 3, citrate buffer (pH 2.57 2 P- r 'OI Fell' Volume of eluateiml Fig. 7. Elution curves for (a) Fell1 - Mo"' and (6) Mn" - Fell1. 1.2% ascorbic acid in 0.1 M HCI; 2, 1 M HCIO,; and 3, citrate buffer (pH 4) - L 8 E .- 5 6 2 0-l 3. . .w CI C a, 2 4 u 2 0 Volume of eluateiml Fig. 8. Elution curves for ( a ) VV - MoV' and ( b ) Mn" - Vv. 1, 2% ascorbic acid in 0.1 M HC1; 2, 1 M HCIO,; and 3, citrate buffer (pH 4) Sorption of beryllium( I> In the pH range studied, the sorption capacity for Be" remains almost constant, having a maximum value of 1.02 mmol g-' at about pH 6.Beryllium(I1) has a high affinity for oxygen donors, particularly for vicinal dihydroxy compounds. Volume of eluateiml Fig. 9. 2, citrate buffer (pH 2.5); 3, 2% ascorbic acid in 0.1 M HCI Elution curves for MnlI - Cull - FeIII. 1, Citrate buffer (pH 4); Kinetics of the Sorption For all the metal ions studied, namely, FeIII, CuII, Nil', Mn", MoVI, Vv and BeII, the sorption was fast and the time taken to reach half-saturation was less than 10 min (Fig. 4). This demonstrates that R-PGS is suitable for the ion chromato- graphic separation of the metal ions studied. Column Breakthrough Studies Studies on the breakthrough capacity of the metal ions were carried out in the appropriate pH range. MnII, Cur1 and Ni" break through almost simultaneously without any significant retention on the column, whereas FeIII, MoVI and VV, as expected, start to appear in the effluent at different times (Fig.5 ) . The inflections of the breakthrough curves are sufficiently sharp and well separated, indicating that it should be possible to separate these metal ions from their mixtures on a column of R-PGS. Separation of Metal Ions From the apparent selectivity of the resin, it is clear that the separation of a number of binary mixtures of metal ions is possible by ion-exchange chromatography. Hence, mixtures of metal ions commonly encountered in analyses were separated and quantified using the proposed technique. These mixtures included Cu" - FeIII, Nil1 - Fell', Mn" - FeIII, Fe"1 - MoVI, MnII - VV and VV - MoVI. The separation of a ternary mixture containing Mn", Cull and Fe"1 was also performed successfully.The results of these separations are shown in Figs. 6-9. The sorption conditions and the eluents used are given in Table 1, from which it can be seen that Mn" can be eluted from a mixture using citrate buffer of pH 4, whereas Cu" is eluted using citrate buffer of pH 2.5. Fe"1 and Vv can be eluted as their reduced forms using 2% ascorbic acid in 0.1 M HC1. The results are presented in Table 1. The elution of MoVI was not possible with any of these eluents. Although 1 M HC104 was found to elute MoVI, its use resulted in a considerable amount of leaching of the loaded reagent. Hence in order to use the column further, the resin had to be re-loaded with the reagent by passing a solution of PGS through the column until saturation occurred. References 1. 2. 3. Brajter, K., Chem. Anal. (Warsaw), 1973, 18, 125. Kemula, W . , and Brajter, K., Chern. Anal. (Warsaw), 1968, 13, 305. Brajter, K., J . Chvomatogr., 1974, 102, 385.ANALYST. APRIL 1989, VOL. 114 443 4. 5 . 6. 7. 8. 9. 10. Brajter, K , , and Zlotorzynska, E. D., Talanta. 1980, 27, 19. Brajter, K., and Zlotorzynska, E. D., Talanta, 1983, 30, 355. Brajter, K., and Zlotorzynska, E. D., Talanta, 1986, 33, 149. Going, J . E., Wesenberg, G., and Andrejat, G., Anal. Chim. Acta, 1976, 81, 349. Lee, K. S., and Lee. D. W . , Anal. Chem., 1978. 50, 255. Chikuma, M., Nakayama, M., Itoh, T., and Tanaka, T . , Talanta, 1980, 27, 807. Puschel, R., and Lassner, E., in Flaschka, H. A., and Barnard, A . J . , Jr., Editors, “Chelates in Analytical Chemistry,” Volume 1, Marcel Dekker, New York, 1967, p. 281. 11. 12. Pollak, J . , and Fulnegg, G., Monatsh. Chem., 1027, 47, 537; Chem. Abstr., 1927, 21. 2676. Puschel, R., and Lassner, E., in Flaschka, H. A., and Barnard, A . J . , Jr., Editors, “Chelates in Analytical Chemistry,” Volume 1, Marcel Dekker, New York, 1967. p. 278. Paper 8102.722 K Received June 1 Oth, 1988 Accepted November 3rd, 1988

ISSN:0003-2654    

DOI:10.1039/AN9891400439

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST, APRIL 1989, VOL. 114 44s Fourier Transform Infrared Spectrometric Determination of Trace Amounts of Polydimethylsiloxane in Extracts of Plastics Additives Pierre Fux Central Analytical Department, FO 3.32, Ciba-Geig y Ltd., 4002 Bade, Switzerland Fourier transform infrared (FT-IR) spectrometry has been used to determine trace amounts of polydimethylsiloxane in a plastics additive. The efficiency of the proposed extraction method and the utility of the developed FT-IR method were used to determine siloxanes at the vg g-1 level in plastics additives. Various evaluation possibilities of the recorded spectra were compared both in the presence and absence of interfering absorption bands arising from simultaneously extracted components. The method was applied successfully to the determination of siloxanes in a number of other additives.Keywords : Silicones; pol ydime th ylsiloxan e determination; Fourier transform in fra red spectrometry; plastics additives; extraction The entrainment of trace amounts of polysiloxanes in indus- trial products cannot be eliminated completely. By modifying the surface tension, such polysiloxanes can affect adversely the properties of a product and hence its technical applica- tions. It is, therefore, important that analytical methods are available for the determination of trace amounts of siloxanes. Infrared (1R) spectrometry has always been a valuable method for the identification of polysiloxanes. Since the 1960s "classical" IR spectrometry has been used essentially to detect silicones in various substrates,'-' but reliable quantification has been possible only when relatively high polysiloxane concentrations are present in the matrix.6.7 Little work on the determination of trace amounts of silicones has been reported: their determination by inductively coupled plasma atomic emission spectrometry8.9 and atomic absorption spec- trometryl".ll has been described.In the latter instance only the total silicone content was determined and no information was given on the structure of the siloxane present. Few investigations into the determination of trace amounts of silicones have been carried out using infrared spec- trometry. 12.13 The development of Fourier transform infrared (FT-IR) spectrometry and also FT-IR with attenuated total reflectance (ATR FT-IR) has increased the importance of these methods as a tool for the quantification of polysiloxanes, particularly within the last five years.The main areas of investigation have concerned quantitative surface analy- sis,14.15 the determination of silanol in silicones16 and the determination of polydimethylsiloxane on cotton fabrics" and on human skin. 18 Recently, diffuse reflectance infrared Fourier transform spectrometry (DRIFTS) has been used to study the deposition of siloxanes on keratin surfaces. 19 In this paper the quantification of trace amounts of polydimethylsiloxane (PDMS) in a plastics additive using FT-IR spectrometry is described. A liquid - liquid extraction method is used for additives that are insoluble in pentane. The additive is dissolved in N,N-dimethylformamide and the silicones are extracted with pentane using a continuous extractor. Experimental Instrumentation Infrared spectra were recorded in a 4-mm cell on a Perkin- Elmer 1710 FT-IR spectrometer, accumulating 20 scans with a resolution of 4 cm-1.From the recorded spectra the net absorbance values were evaluated using a Perkin-Elmer Model 3600 data station and the QUANT-SINGLE program. Liquid - liquid extractions were performed with a continu- ous liquid - liquid extractor according to Ludwig (Normag, D-6238, Hofheim am Taunus: Code 2071). Reagents Polydimethylsiloxane (Dimethicone 350) was supplied by Wacker. The following solvents, with very low silicone contents, were used: pentane, Burdick and Jackson quality (Fluka); carbon disulphide , infrared spectroscopic grade (Fluka) ; and N,N-dimethylformamide (DMF) , analytical-reagent grade (Fluka). Preparation of the Test Samples Extraction procedure Approximately 200 g of the additive were dissolved in 200 ml of DMF and extracted for 4 h at 60 "C with 500 ml of pentane in a continuous liquid - liquid extractor according to Ludwig. The pentane emerged from the stirred DMF phase through the perforated, centred tube as a stream of fine droplets; this tube allowed good interfacial contact. The solution was left to stand at room room temperature for 1 h to optimise the phase separation.To remove any remaining additive, the pentane phase was extracted a further 5-7 times with 50 ml of DMF using a separating funnel. A 250-ml portion of the purified pentane solution was rotary evaporated to a volume of approximately 10 ml.This solution was then evaporated to dryness for 2 h at about 50 "C under high vacuum. To the dry residue 0.6-2 ml of carbon disulphide were added depending on the intensity of the IR spectrum. (A blank run, i . ~ . , the extraction of a solvent mixture, was carried out in parallel with the analytical extraction in order to detect whether there was any contamination of the reagents or apparatus with sili- cones.) Preparation of Spiked Test Samples Known amounts of PDMS as a solution in pentane were added to the samples, e.g., 0.05, 0.2, 0.5 and 1 .O mg of PDMS were added to 200 g of the liquid sample. The extraction was performed as described above for the unspiked samples. Preparation of Calibration Solutions Standard solutions (0.02, 0.04, 0.06, 0.08, 0.10 and 0.20 mg ml-1) of PDMS in carbon disulphide were prepared; their IR spectra are shown in Fig.1.446 ANALYST, APRIL 1989, VOL. 114 FT-IR Measurements Polydimethylsiloxane shows a characteristic narrow absorp- tion band at 1260 cm-1, a double band at 1090 and 1020 cm-l and a broad band at 805 cm-1 (Fig. I). The band at 1260 cm-I is generally used for quantificationl.7,’2.lX.t~ because of its sharpness. For reliable quantitative analysis it is important to verify the absorbance ratios of these characteristic bands. In this work, the ratio of the absorbance of the band at 1260 cm-1 (Al26O) to that of the band at 805 cm-1 (Axos) was calculated (Table 1). If the silicone bands are actually single bands, then the ratio of A1260 to Axos will be 1.10 k 0.02.The IR spectra of the sample and calibration solutions were recorded from 700 to 2000 cm-1 in a 1- or 4-mm cell. The absorption of the solvent was compensated for by spectral subtraction. Calibration graphs were constructed at 80.5, 818 (shoulder of the band at 805 cm-I), 1020 and 1260 cm-1 by plotting absorbance versus milligrams of PDMS per millilitre of CS?. The base-line point was taken to be approximately 1900 cm-1. Results and Discussion Evaluation Possibilities Using the IR Measurements If the bands at 1260 and 805 cm-I were “proper” bands, i.e., A1260:AK05 was within the range l . C l . 2 , then the recorded spectrum of the sample could be evaluated from either the first or the second band, using the corresponding calibration graph.If the IR spectrum contained interfering bands (e.g., a narrow band at 805 cm-1 or an intense overlapping band at 1260 cm-1 so that this typical band was not visible) from simultaneously extracted components (impurities, residual traces of additive, etc.) that could affect quantification, then the evaluation was based on the absorbance at 818 cm-1. The PDMS content determined in this way has to be considered as the maximum concentration in the sample. s 85 t 2000 1800 1600 1400 1200 1000 800 Wavenu m bericm - 1 Fig. 1. Infrared spectra of PDMS standard solutions in CS2 in a 4-mm cell. A. 0.02: B. 0.04; C. 0.06: D , 0.08: E. 0.10: and F, 0.20 mg rnl I Table 1. Absorbance ratio, A Il_ho : AH(,5, for standard solutions of PDMS in CS2 in a 4-mm cell Standard w lu t i o ni mg ot PDMS per Abcorbance at Absorbance at ml of CS2 1260 cm 805 cni- 1 A I X ~ : AH05 0.0200 0.0490 0.0453 1 .(I82 0.0400 0.092 1 0,0849 1.085 0.0600 0.1402 0.1271 1,103 0.0800 0.1861 0.1682 1.106 0.1000 0.2339 0.2 1 3 8 1.094 0.2000 0.4521 0.4027 1.123 Recovery of PDMS Extracted from DMF Solutions Solutions of PDMS in DMF in the concentration range 0.1-1.5 pg g-1 were extracted with pentane as described under Extraction procedure; the average recovery was greater than 98%.The various concentration steps necessary to obtain a dry residue were carried out with no loss of PDMS. Reproducibility and Recovery of the Sample Extraction About 200 g of the test substance A1 were extracted alone and also after the addition of 0.05,0.52 and 1.03 mg of PDMS.The results are given in Table 2. For the unspiked sample, only the band at 1260 cm-1 was used in the calculation procedure because the intensity of the band at 805 cm-1 was substantially different from that of the former and from the intensity of the band at 1090 cm-1. For the spiked samples no interfering bands were visible in the recorded spectra, A,26o : Axo5 varying from 1.085 to 1.207. Quantification could be performed at either 805 or 1260 cm- l. The recovery of PDMS from an extracted sample could therefore be determined from these experiments. Extrapol- ation to zero PDMS added, based on a linear regression analysis of the six points (correlation coefficient = 0.9986), gave a PDMS concentration of 1.61 pg g-1 in the sample. Direct extraction of the unspiked sample gave a PDMS content of 1.57 pg g-1.This means that a recovery of about 98% for the silicones extracted from the plastics additive was achieved. To check that the extraction was quantitative, a series of unspiked and spiked samples were extracted. The results are summarised in Table 3. Impurities which changed the shape of the broad absorption band at 805 cm-1 were Table 2. Determination of PDMS content for unspiked and spiked additive sample A1 Amount of sample A 1 extracted/ & 206 222 20 1 205 Maximum PDMS con- PDMS con- PDMS con- Amount of tent found by tent found by tent found by PDMS evaluation at evaluation at evaluation at added1 1260 cm-11 805 crn-11 818 cm-’1 v& g- ’ vgg M g ’ w g-‘ - 1.57 Evaluation not possible because of interfering bands 0.23 1.97 1.82 a2.74 2.57 4.24 4.07 15.22 5.02 6.69 6.89 17.36 Table 3.Rccovery of PDMS from the extraction of plastics additive samples Concentration found for the unspiked sample .‘ i Sample pgg-1 A2 . . S1.03 A3 . . S0.45 A4 . . S0.88 No. ( a ) A5 . . a1.78 A6 . . a1.45 A7 . . a0.98 A8 . . G1.11 Amount of PDMS added1 P& & - I (c) 1.40 1.43 1.27 1.09 1.09 1.09 0.33 Concentration found for the Recovery, spiked %, 2.58 110.7 1.93 103.5 2.23 106.3 2.92 104.6 2.61 106.4 2.09 101.8 1.44 100.0 The PDMS content given is the maximum concentration in the sample.ANALYST, APRIL 1989. VOL. 114 447 100 80 60 40 20 8 a, C m 4z 4- .- 100 t I- 2 80 60 40 20 W 2000 1800 1600 1400 1200 1000 800 Wavenum berlcm-1 Fig. 2. Fourier transform infrared spectra of (uj unspiked and ( h j spiked sample A5 (containing approximately 0.2 mg of PDMS in 200 g of suhtance) recorded from 700 to 2000 cm-I in a 4-mm cell (diluted mith 1 ml of CS?) extracted simultaneously for all the samples studied.This band therefore became abnormally sharp. Hence, quantifica- tion was performed by measuring the absorbance of its shoulder at 818 cm-I. The shoulder was clearly visible in all the recorded spectra (see, for example, the spectra of spiked and unspiked sample A5 shown in Fig. 2). From this series of samples, the average recovery was found to be approximately 105 O/” . Comparison of Quantification at 805 cm-1 and “Shoulder” Evaluation at 818 cm-1 Spiked samples of the additive A1 were extracted and the IR spectra evaluated using the band at 805 cm-1.These spectra were also evaluated using the shoulder (at 818 cm-1) of this band. The results are summarised in Table 2. Extrapolation to zero PDMS based on a linear regression analysis (correlation coefficient = 0.9983), gave a PDMS content of 2.58 pg g-1. Quantification using the shoulder at 818 cm-1 resulted in an increase in the PDMS content found of approximately 60% compared with quantification at 805 cm-1. Removal of Interfering Components by Difference Spec- trometry A series of ten samples were extracted and all the recorded IR spectra showed the presence of interfering bands. Fig. 3 shows typical IR spectra of samples A l l and A14 in which strong absorbing bands mask the characteristic silicone bands. To 100 80 60 40 20 $ ai 9 L Y m i I ’ +- ._ $ 100 K F I- 80 60 40 20 2000 1800 1600 00 1200 1000 800 Fig.3. Fourier transform infrared spectra of ( u ) sample A14 and ( h ) sample A1 1 recorded from 700 to 2000 cm I in a 4-mm cell (diluted with 0.8 ml of CS?) remove these interfering bands, difference spectra were generated. A sample from this series, containing less than 0.4 ug g-1 of PDMS but showing qualitatively the same interfering bands, was purified as described under Extraction procedure. This sample was taken as a pure reference sample (“AR”), i.e., silicone-free (containing <0.2 pg g-1 which is the approximate quantification limit of the method). Difference spectra were calculated from the recorded spectra minus this reference spectrum. The absorbance value was defined by measuring the absorbance at 818 cm-1 because this was the wavelength least affected by impurities, and the base-line point was taken to be 1900 cm-1.The results of the quantification are summarised in Table 4. Evaluation of the difference spectra for samples containing 0.2-2.0 pg g-1 of PDMS gave a PDMS content that was approximately 40% lower on average compared with evalua- tion carried out without using difference spectra. Fig. 4(a) shows the difference spectrum of sample A14, which contains approximately 2 pg g-1 of PDMS and the reference substance “AR”. The spectrum is similar to a typical PDMS spectrum with its four characteristic bands. However, for sample A 11, containing only about 0.2 pg g-1 of silicones, the difference spectrum [Fig. 4(b)] shows no evidence of PDMS bands.Comparison of PDMS Quantification by FT-IR Spectrometry and Total Silicone Determination by X-ray Fluorescence Spectrometry Some of the FT-IR results obtained in this work were compared with those obtained for total silicone determination by X-ray fluorescence spectrometry. The total silicone content was determined using the extracts of the plastics additive448 ANALYST, APRIL 1989. VOL. 114 Table 4. Evaluation comparisons for ten additive extracts Amount of sample Sample extract e di No. g A9 . . . . 183.1 A10 . . . . 174.6 A l l . . . . 178.2 A12 . . . . 188.0 A13 . . . . 175.8 A14 . . . . 184.0 A15 . . . . 167.5 A16 . . . . 157.1 A17 . . . . 157.7 A18 . . . . 174.1 Maximum concentration of PDMS found by evaluation at 818 cm-1’ S2.07 S1.20 s1.10 G0.83 60.77 12.32 S1.39 G0.97 61.07 61.32 v g s I PDMS content found by difference spectrometry with a silicone-free reference substance/ v g g- 0.86 0.54 0.24 0.41 0.47 2.06 1 .05 0.68 0.88 1 .oo Table 5.Comparison of X-ray fluorescence spectrovetry and FT-IR spectrometry for the determination of PDMS X-ray fluorescence spectrometry Maximum Amount of PDMS content sample Silicone PDMS content found by FT-TR Sample cxtractcdi found*/ calculatedl spectrometryt/ No. g Ing L% g- ’ I%% I A5 . . 200 56 0.74 d I .78 A6 . . 200 64 0.84 G1.45 A7 . . 200 108 1.43 60.98 A8 . . 165 48 0.79 51.11 A19 . . 200 340 4.49 G1.66 * Total silicone found in the extract. f Values determined by reference to a calibration graph at 818 cm l . samples. The PDMS content was determined by multiplying the total silicone concentration found by a factor of 2.64, taking into account the following structure for PDMS: The results are summarised in Table 5 .It is worth comparing the results given by the two evaluation methods for the samples studied. The values determined spectrometrically for samples A5, A6 and A8 are the maximum concentrations, the actual PDMS concentrations are approximately 40-60% lower. These results are in good agreement with those obtained by X-ray spectrometry. Samples A7 and A19 contain about 3-4 times less PDMS according to the FT-IK measurements than indicated by the X-ray spectrometry results (Table 5 ) . Therefore, the determi- nation of total silicone cannot be used as a means of quantifying PDMS in additives, nor even as a means of measuring their PDMS content. Conclusion Trace amounts of PDMS can be quantified reliably by FT-IR spectrometry, after their complete extraction from industrial products, e.g., plastics additives. Interference resulting from the remaining extracting solvents, additive residues or simul- taneously extracted impurities are eliminated by spectral subtraction. The detection limit of the proposed method is approximately 0.1-0.2 pg g-1 of PDMS in the additive. These resuits concern plastics additives that are insoluble in pentane. However, a solid - liquid extraction of silicones with butan-1-01 has been carried out for solid additives that are 100 80 60 40 $? 2o a i C rn =: .- 5 100 C 2 I- 80 60 40 * O l 1 I I, I I 2000 1800 1600 1400 1200 1000 800 Wavenum bericrn - Fig.4. Difference spectra of ( a ) sample A14 and ( h ) sample A l l with reference substance “AR” (silicone-free) generated on a Perkin-Elmer 1710 spectrometer soluble in pentane; the quantitative determination of PDMS is carried out using the same FT-IR subtraction method. In this instance the detection limit of PDMS is approximately 10 pg g-1 for the additives examined, because the less selective extraction step leads to higher spectral interferences. The author thanks P. Acker for recording and evaluating the IR spectra. 1. 2. 3. 4. 5. 6. 7 . 8. 9. 10. 11. 12. 13. References Pashenkova, L. F., and Yablochkin, V. D., Gig. Sanit., 1968, 33, 55. Danielak, R . , Ludwicki, H., and Ostrowska, E., Acta Pol. Pharrn., 1978, 35, 467.Wieczorek, H . , Seijen, Oele, Fette, Wachse, 1985, 111, 115. Wilkowa, T . , Chern. Anal. (Warsaw), 1976, 21, 399. Mayhan, K. G., Thompson, L. F., and Magdalin, C. F., J . Paint Technol., 1972, 44, 85. Hauptmann, G., Keil, G., and Eberhardt, E., Schrnierst. Schrnierungstech . , 1968, 29, 38. Kojima, S . , and Oba, T., Eisei Shikenjo Hokoku, 1970,88,26. Watanabe, N., Nagase, H., Ose, Y., and Sato, E., Eisei Kugaku, 1985, 31, 391. Watanabe, N., Yasuda, Y., Kato, K., Nakamura, T., Funa- saka, R . , Shimokawa, K., Sato, E . , and Ose, Y., Sci. Total Environ., 1984, 34, 169. Nishijima, M., Kanmuri, M., Takahashi, S . , Kamimura, H., Nakazato, M., and Kimura, Y., Shokuhin Eiseigaku Zasshi. 1975, 16, 110. McCamey, D. A . , Tannelli, D. P., Bryson, L. J . , andThorpe, T. M., Anal. Chim. Acta, 1986, 188, 119. Rotzche, H., Clauss, H., and Hahnewald, H . , Pluste Kautsch., 1979, 26, 630. Sinclair, A., and Hallam, T. R., Analyst, 1971, 96, 149.ANALYST, APRIL 1989, VOL. 114 449 14. Reikichi, I.. and Koji. O., A p p f . Specfrosc., 1984. 38, 359. 15. Fuhrmann, J . , and Glanzer, K.. Pharm. Znd., 1985, 47, 652. 16. Griffith, G . W . , Znd. Eng. Chem. Prod. Res. Dev., 1984, 23, 590. 17. Shreedhara Murthy, K. S., Leyden. D. E., and D'Alonzo, R . P.. Appl. Spectrosc.. 1985, 39, 856. 18. Klimisch. H. M., and Chandra, G., J . SOC. Cosmet. Chem., 1986. 37, 73. 19. Klimisch, H. M., Kohl, G. S . , and Sabourin, J . M., J . SOC. Cosmet. Chem., 1987, 38, 247. Paper 8104076A Received October 13th, 1988 Accepted December 5 t h , 1988

ISSN:0003-2654    

DOI:10.1039/AN9891400445

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYST, APRIL 1989, VOL. 113 45 1 Enhanced Extraction of Lanthanides with Crown Ether Carboxylic Acids of Increasing Li po p h i licity Jian Tang and Chien M. Wai* Department of Chemistry, University of Idaho, Moscow, ID 83843, USA sym-Dibenzo-I 6-crown-5-oxyacetic acid and its modified analogues can extract lanthanides efficiently in solutions with high ionic strength and complex matrices. Increasing the length of the side-arm alkyl group increases the lipophilicity of the macrocyclic polyether and enhances the distribution ratio of the lanthanide complexes in the organic phase. This type of crown ether carboxylic acid is a potential complexing agent for seiective extraction and concentration of lanthanides in natural water systems. Keywords: Lanthanide; extraction; crown ether; lipophilicity Macrocyclic polyethers, as a new generation of specific complexing agents, have found many attractive applications in separation chemistry and in analytical chemistry.Selective complexation of cations with crown .ethers can be achieved based on the cation radius - cavity size compatibility concept. For the extraction of cations with neutral crown ethers, the nature of the counter anions involved is an important factor in determining the efficiency of the extraction. The role of the counter anions can be filled by introducing anionic functional groups to the structures of the crown ethers. -4 number of functionalised crown ethers have been synthesised and have been shown to extract alkali metal ions and alkaline earth- metal ions in aqueous solutions independent of counter anions.1.2 The combination of ion-binding cavities possessing fixed dimensions with ionisable functional groups may create novel bifunctional complexing agents with extraction effici- encies and selectivities surpassing the closely related non- ionisable crown ethers. Recent reports have indicated that some crown ethers with pendant carboxylate functional groups are surprisingly efficient and selective for extracting trivalent lanthanide ions.-'.-.' Because of their high selectivity and anion independence, these crown ether carboxylic acids may have potential analytical applications for lanthanide extraction from complex aqueous systems. Neutron activation analysis .(NAA) is one of the most sensitive methods for analysing the lanthanides.5 However, direct application of NAA to lanthanide determination in natural waters is often not possible because of the low lanthanide concentrations and the matrix interferences, espe- cially in complex systems such as sea water.Therefore. separation and concentration procedures are generally required for NAA determination of lanthanides in natural systems. Besides 'being selective for lanthanide extraction, crown ether carboxylic acids have another attractive aspect for NAA, i.e., they are inert to neutron irradiation and hence would not create spectral interferences. Recently we reported the complexation of a crown ether carboxylic acid. sym- dibenzo-16-crown-5-oxyacetic acid, with trivalent lanthanide ions in a de-ionised water system.4 This macrocyclic poly- ether, with an attached carboxylate group, is effective for lanthanide complexation but has a drawback due to its weak lipophilicity which causes slow phase separation and transport of the complex to the organic phase.The lipophilicity can be enhanced by adding alkyl groups to the attached side-arm. In this work we have studied the extraction efficiencies of several crown ether carboxylic acids with modified sidearm struc- tures in order to elucidate the factors controlling the efficiency of extraction. The effects of alkyl substitution and lipophilicity of this type of crown ether carboxylic acid on lanthanide * To whom correspondence should be addressed. extraction and their potential applications to lanthanide determination in solutions with high ionic strength and complex matrices are described in this paper.Experimental Reagents sym-Dibenzo-16-crown-5-oxyacetic acid (I), 2-(sym-dibenzo- 16-crown-5-oxy)hexanoic acid (11) and 2-(dibenzo-l6-cruwn- 5-oxy)stearic acid (111) (Fig. 1) were synthesised according to the procedures given in the literature.6 The basic structure of the macrocycle was synthesised by the reaction of catechol with bis(2-chlorethyl) ether to form bis[2-(o-hydroxy- phenoxy)ethyl] ether followed by its reaction with epichloro- hydrin to form syrn-5-hydroxydibenzo-16-crown-5. The pen- dant carboxylic acid group was introduced to the niacrocycle by its reaction with bromoacetic acid, 2-bromohexanoic acid and 2-bromostearic acid to form I, 11 and 111, respectively. Three lanthanides, La-i+, Eu'f and Lu-T+, were chosen to represent the lanthanide series.The lanthanides, in nitrate or in oxide forms, were obtained from Alfa Products. Other chemicals used were all Baker Analyzed Reagents. De- ionised water was prepared by passing distilled water through R A 0 0 0 0 Fig. 1. Structures of sym-dibenzo-16-crown-5-oxyacetic acid ( I ) , 2-(sym-dibenzo-16-crown-5-oxy)hexanoic acid (11) and 2-(sjr)z-di- benzo- 16-crown-5-oxy)stcaric acid (111)452 ANALYST, APRIL 1989. VOL. 114 an ion-exchange column (Barnstead Ultrapure Water Purifi- cation Cartridge) and a 0.2-pm filter assembly (Pall, Ultipor DFA). All containers were acid-washed, rinsed with de- ionised water and dried in a class 100 clean hood. A synthetic sea water which was prepared according to the literature7 was used to evaluate the conditions for lanthanide extraction.Extraction Procedure The extraction solutions were prepared by dissolving weighed amounts of the crown ether carboxylic acids in chloroform in beakers with magnetic stirring. After dissolution, the organic phase was shaken with an HCI solution at pH 2 to remove potential metal impurities in the system. After purification, the organic phase was kept in contact with an LiOH solution to maintain a pH level of about 7. In some experiments, radioisotopes were used as tracers to test the extraction efficiency. In other experiments, p.p.b. to sub-p.p.b. levels of La3+, E u ~ + and Lu3+ were added to water samples to study the recovery by NAA. The water samples were adjusted to a desirable pH with LiOH and acetic acid.In general. to each water sample (100 ml in a glass-stoppered flask), 10 ml of the extraction solution were added and the mixture was shaken vigorously on a mechanical wrist-action shaker (Burrell Model 75) for a fixed time (usually 2 min) at room temperature. After shaking, the mixture was allowed to stand for a few minutes for the phase separation to be completed. For the tracer experiments, 5 ml each of the organic and aqueous phases were pipetted out of the flask and placed in 10-ml glass vials with fast-turn caps for y counting. For NAA experiments, 5-8 ml of the organic phase were removed from the flask and placed in contact with 1.5 ml of HN03 solution at pH 2 in another flask. The mixture was shaken again for 2 min to back-extract the lanthanides into the acid solution.After phase separation, 0.5 ml of the acid solution were placed in a % dram (ca. 1.48 ml) polyethylene vial which was later heat sealed for neutron irradiation. Neutron Irradiation and Activation Analysis Sample irradiations were performed in a 1-MW TRIGA nuclear reactor with a steady neutron flux of 6 X 10’2 n cm-2 s-1. Samples were generally irradiated for 2 h followed by 1 d of cooling before counting. The half-lives of the isotopes produced and the y radiations used for their identification are given as follows: I-loLa 400.2 h, 487 keV; 152Eu”’ 9.3 h, 122 keV; and 1’7Lu 6.7 d, 208 keV. A large volume Ortec Ge(Li) detector with a resolution of about 2.3 keV at the 1332 keV y line from W o was used for y counting.Signals from the detector were connected to an EG&G Ortec ADCAM (Model 950A) multi-channel analyser with software and an IBM-PC for data processing. The details of sample irradia- tion, y counting and the general procedure of NAA are given elsewhere. 8 Results and Discussion Extraction of Lanthanides From Sea Water With sym-Dibenzo- 16-crown-5-oxyacetic Acid The efficiencies of extracting the lanthanides from synthetic sea water with I as a function of pH at an organic (chloroform) to aqueous volume ratio of 1 : 1 are shown in Fig. 2. For the extraction of 2 X 10-7 M Lu3+ from the synthetic sea water with 3 x 10-3 M I in an equal volume of the organic phase, the extent of extraction was about 95-96% at pH ca. 6. However, PH Fig. 2. water by the crown ether carhoxylic acids 1-111 pH dcpendence of the extraction of L u 3 from synthetic sea if the organic to aqueous phase ratio was decreased to 1 : 10, the extraction was found to be much lower, at about 65%.The distribution ratio ( D ) is defined by the following equation: Based on the percentages of extraction measured at different organic to aqueous phase ratios, the value of D for Lu3+ was calculated to be about 20. Similarly, the D values for E d + and La’+ were calculated to be 10 and 5 , respectively. Increasing the concentration of the chelating agent should enhance the percentage of extraction of the lanthanides. However, the weak lipophilic property of I, the slow phase separation and the possible micelle formation become an experimental problem when the concentration of the chelating agent becomes high.Another approach,is to use a multiple- extraction procedure by repeatedly extracting a water sample with the organic phase containing 3 X 10-3 M of the crown ether carboxylic acid. For the extraction of 2 x lo-’ M Lu-7+ from 100 ml of the synthetic sea water with triple extraction of 10 ml each of the extractant solution, a total extraction efficiency of about 92% for Lu3+ was obtained. The total recoveries for E u ~ + and La’+ after the triple extractions under the same experimental conditions were about 88 and 60%. respectively. The solubility of I in water at room temperature has been estimated to be about 5.8 x 10-4 M.’ It has also been shown that in a chloroform - water two-phase system, the concentra- tion of I in the organic phase declines rapidly as the p€I increases from 5 to 8 and then reaches a plateau for the region pH 8-12.In this latter region, only 20-30% of the initial concentration of I remains in the organic phase.’ When a butyl group is attached to the side-arm of I , the solubility of the resulting crown ether carboxylic acid I1 in water has been reported to decrease by about an order of magnitude relative to that of I.” Extraction with 2-(sym-Dibenzo-16-crown-5-oxy)hexanoic Acid Two important observations were made when I1 was applied to extract the lanthanides from the sea water: (1) the phase separation in this instance was much faster and (2) the distribution ratios for the lanthanides were much greater compared with I. Replacing H in the side-arm of I with C4H9 significantly improved the lipophilicity of the chelating agent and greatly facilitated the phase separation process.Accord- ing to Strzelbicki and Bartsch,v in a chloroform - water system, the concentration of I1 in the organic phase remains virtually unchanged even when the aqueous phase is highly alkaline. The enhanced extraction efficiency appears to be related to the increased lipophilicity of the complexing agent. The distribution ratios for La-3+, E u ~ + and L u ~ + in the synthetic,INALYST. APRIL 1989. VOL. 113 453 sea water were calculated to be 1.5 x 102, 2.9 x 10' and 3.7 X 102, respectively (Table 1). The distribution ratio for La'+ in this instance is large enough such that over 94"/0 of La?+ in the sea water can be extracted into the organic phase at an aqueous to organic ratio of 10 : 1.The dependence of the extraction of Lu;+ by I1 on pH is shown in Fig. 2. I n comparison with I , the extraction plot shifts slightly to lower PIT. The lanthanide complexes extracted into the organic phase by I1 can be back-extracted easily into an aqueous phase at a pH level of less than 3. The rates of both the extraction and back-extraction processes are fast, requir- ing only cu. 1 min of vigorous shaking virtually to complete the extraction. In a real operation, the mixture is normally placed in the mechanical shaker for about 2 min for either of the extraction processes. According to our experiments, the volume of the organic phase to the acid solution during the b ac k - e x t r a c t i on process s h o u 1 d n c) t ex ce e d 1 0.o t h e rw i se incomplete recovery of the lanthanides can result. Therefore, the back-extraction process is capable of providing a 10-fold increase in the lanthanide concentration. A similar concentra- tion factor of 10 can also be obtained from the first step of the extraction. Using the two-step extraction procedure with I1 as the chelating agent, a total pre-concentration factor of 100 can be achieved for the lanthanides. It is notekvorthy that the sodium present in the sea water was only extracted in very small amounts (p.p.m. levels) and bromine was not detectable in the back-extracted acid solution. Based on the activity of 'JNa observed in the sea water experiments, we estimated that the relative extraction efficiencies of Na+iLu'+ under our experimental conditions were < 1 0 - 3 .This is important for analytical applications, because sodium and bromine are two of the major interfering matrix species for NAA of natural waters. Extraction with 2-(sym-Dibenzo-16-crown-5-oxy)stearic Acid The efficiency of extraction of La-'+ from the synthetic sea water with 111 in chloroform was better than 97% with an aqueous to organic phase ratio of 10 : 1. The distribution ratios for La'+, Eui+ and Lu-i+ were calculated to be 2.9 x 102. 3.8 X 102 and 1 .0 x 103, respectively (Table 1). The preference of extracting Lu'+ over La3+ was still observed in 111. The large C1(,H3; group attached to the side-arm should make the lanthanide complexes more lipophilic which is the likely cause of the enhancement of the distribution ratios.The lanthanide complexes in the organic phase can also be back-extracted into HNO; solution at pH < 3, suggesting that the exchange reaction between the complexed form and proton is not significantly affected by the large alkyl group in the side-arm. The dependence of the extraction of Lu-'+ by 111 on pH is shown in Fig. 2. The plot shifts further toward the lower pH levels relative to those of I and 11. Conclusions Increasing the length of the side-arm alkyl group in sym- dibenzo- 16-crown-S-oxyacetic acid increases the lipophilicity of the crown ether carboxylic acid and enhances the distribu- Table 1. Distribution ratios ( U ) of lanthanides with different crown ether carboxylic acids U Macrocyclic pol ycther La Eu I, u I .. . . . . . . 5.0 1 0 x 10' 2.0 x 10' I1 . . . . . . . . 1.5 x 10' 2.9 x 10' 3.7x 10' I11 . . . . . . . 2.9 x 10' 3.8 x 102 1.0 x 10' tion ratio of the lanthanide - macrocycle complexe5 in the organic phase. Preferential complexation with heavier lan- thanides was observed in this system. The macrocycles also show extremely low affinities for alkali metals. The CJ- and C,(,-substituted crown ether carboxylic acids, with high extraction efficiency and lipophihcity, are potential extraction agents for selective concentration of lanthanides from natural water systems. Since the extraction is reversible, lanthanides extracted into the organic phase can be back-extracted into an acid solution, thus providing a large pre-concentration factor (>loo) for chemical analysis. It is calculated that the two-step extraction process described in this paper combined with NAA can detect 10-1-10-3 pg of lanthanides in 200 ml of natural waters depending on the intrinsic sensitivities of different lanthanides to neutron activation. This research was supported in part by a grant frotn the Idaho State Board of Education. Neutron irradiations were per- formed at the Washington State University Nuclear Radiation Center under a Reactor Sharing Program supported by the Department of Energy. 1. 2. 3. 4. 5 . 6 7. 8 9. References Strzelbicki, J . , and Bartsch. R. A . , Anal. Chem., 1981, 53, 1894. Strzelbicki, J . , and Bartsch, K. A , , And. Chrm., 1981, 53, 2247. Manchanda, V. K., and Chang. C. A . , Anal. Chem., 1986,58, 2269. Tang. J . , and Wai, C. M., And. Chem., 1986, 58, 3233. Haskin, L. A.. Wildeman. T. R., and Haskin, M. A , , J . Radiut. Chem., 1968, 1, 337. Bartsch, R. A.. Heo, G. S . , Kang, S . I . , Liu, Y., and Strzelbicki. J . , J . (Irg. Chem., 1982, 47, 457. "Water Analy5ts by Atomic Absorption," Varian Techtron. Palo Alto. CA, 1972. Mok, W. M., Shah. N. K.. and Wai, C. M.,Anal. Chem., 1986, 58, 110. Strzelbicki. J . , and Bartsch, R. A., Anal. ('hem., 1981, 53, 225 1, Paper 8/034040 Received August 23rd, 1988 Accepted November l l t h , 1988

ISSN:0003-2654    

DOI:10.1039/AN9891400451

出版商:RSC    

年代:1989

数据来源: RSC