|
41. |
Selective determination of holmium in rare earth mixtures by second-derivative spectrophotometry with benzoylindan-1,3-dione and cetylpyridinium chloride |
|
Analyst,
Volume 121,
Issue 9,
1996,
Page 1317-1320
Nai-Xing Wang,
Preview
|
PDF (454KB)
|
|
摘要:
Analyst, September 1996, Vol. 121 (131 7-1320) 13 17 Selective Determination of Holmium in Rare Earth Mixtures by Second-derivative Spect rophotomet ry With Benzoyl indan- 1,3-dione and Cetylpyridinium Chloride Nai-Xing Wang“, Zhi-Kun Sia, Wei Jianga and Zhong-Cheng Qib a Department of Chemistry, Shandong University, Jinan, 250100, China Chemical Fertilizer Plant, Shandong Gaomi, 261500, China The absorption spectra of 4f electron transitions of the holmium complex with benzoylindan-1,3-dione and cetylpyridinium chloride were studied by conventional and derivative spectrophotometry. The molar absorptivity of the complex at the maximum absorption band was about 22 times greater (at 450 nm) than that in the absence of the complexing agent. The use of second-derivative spectra both eliminates the interference from other rare earths and improves the sensitivity for holmium.The calibration graph showed that the complex obeys Beer’s law up to 25 pg cm-3 of holmium. The corresponding value of Sandell’s sensitivity is 0.36 pg cm-2. The RSD evaluated from 11 independent determinations of 10 pg of holmium is 0.96%. The detection limit, obtained from the sensitivity of the calibration graph and for 3sb (sb = standard deviation of a blank without holmium, n = ll), was found to be 0.43 mg dm-3 of holmium. The quantification limit, obtained for losb, was 1.43 pg ~ m - ~ of holmium. A method for the direct determination of holmium in rare earth mixtures with good selectivity and accuracy is described. Keywords: Holmium determination; deterivative spectrophotometry ; benzoylindan-l,3-dione Introduction The determination of individual elements in mixtures of rare earths is one of the most difficult, because of the similarity of their chemical behaviour.1 There are several reagents for the spectrophotometric determination of the total rare earths, such as Xylenol Orange, Arsenazo I11 and 1 -(2-pyridylazo)-2- naphthol (PAN) but none of these is specific for rare earths and their selectivity is poor because of the considerable overlap between their broad absorption spectra.Several techniques have been developed for the determination of these elements. Trace amounts of rare earth element are usually determined by NAA,2 MS,3 AAS,4 ICP spectrometry,5 etc. Derivative spectrophoto- metry6+7 is a powerful tool for overcoming interferences due to spectral overlap.Therefore, it has been employed widely in biochemical, forensic, clinical and phannaceutical8~9 analysis. The derivative spectra method based on the 4f electron hypersensitive transitions10 of REEs is not only simple and selective, but also improves the sensitivity.’ 1-14 In recent studies, the derivative spectrophotometric determination of some REEs with 8-hydroxyquinolinel5 and its derivatives was investigated, and can be utilized for the determination of neodymium and erbium in mixed rare earths.1”lg In this work, the absorption spectra of the 4f electron transitions of the holmium complex with benzoylindan- 1,3-dione and cetylpyr- idinium chloride was studied, and the application of this system to the rapid, selective and sensitive second-derivative spec- trophotometric determination of the holmium in rare earth mixtures is described.Experiment a1 Apparatus A Shimadzu (Tokyo, Japan) UV-240 spectrophotometer with an Op 1-2 spectral processing attachment, which generates the derivative digitally, and 4.0 cm cells were used. The pH values of the solutions were measured with a pHS-2 pH meter (Shanghai 2nd Analytical Instrumental Factory, Shanghai, China). Reagents Standard solutions of lanthanides were prepared by dissolving their pure oxides (Johnson Matthey, Royston, Herts, UK) in hydrochloric acid and diluting with distilled water to a constant volume. A solution of benzoylindan-l,3-dione (BID) (0.02 0.075 A 0.050 0.025 0 -0.025 h/nm Fig. 1 Absorption spectra of holmium ion and complexes. [Ho] = 1 .5 X 10-4 mol dm-3; [BID] = 3.0 X 10-3 mol dm-3; [CPC] = 7.5 X mol dm-3; pH = 8.5; 4.0 cm cells.1, Ho-BID-CPC; 2, Ho-BID; 3, Ho (HoC13, pH = 6.0). Reference: 1 and 2, reagent blank; and 3, water. Table 1 Spectral characteristics of holmium species in different systems Parameter Ho~+ Ho~+-BID Ho3+-B ID-CPC A (nm) 450 449.7, 459.5 449.7, 460 E (dm3 moi-1 cm-1) 6.6 41.5, 12.5 146.5, 501318 Analyst, September 1996, Vol. 121 0.100 - ( b ) I 1 1 I I ;/------ 0.075 0.050 0.025 0 mol dm-3) was prepared by dissolving 0.273 g of BID (Chemical Reagent Plant, Shanghai, China) in 50 cm3 of 1% m/v Triton X-100 solution (prepared by dissolving 1.0 g of Triton X-100 in distilled water and diluting to 100 cm3 in a (c) 0.1001- 0.075 - 0.050 0.025 I I ,* I I .0 i calibrated flask). A solution of cetylpyridinium chloride (CPC) (0.05 mol dm-3) was prepared by dissolving 4,475 g of CPC (Fluka, Buchs, Switzerland) in distilled water and diluting to 250 cm3 in a calibrated flask. Buffer solution of pH 8.5 was prepared by mixing 30 cm3 of 0.5 mol dm-3 ammonia solution and 170 cm3 of 0.5 mol dm-3 ammonium chloride solution. Procedure Transfer a known volume of the lanthanide solution into a 10 cm3 calibration flask, add 1.5 cm3 of 0.02 mol dm-3 BID solution, 1.5 cm3 of 0.05 mol dm-3 CPC solution and 3.0 cm3 of buffer solution, dilute to volume with distilled water and mix. Record the second-derivative spectra in the range 44W60 nm against a reagent blank, with the use of 4.0 cm cells, with Ah = 1.0 nm, bandpass = 1.0 nm and scan rate = 20 nm min-I.Measure the amplitudes between the 444.5 (+) nm and 449.0 (-) nm peaks. 435 445 455 465 k/nm Fig. 2 Absorption spectra of lanthanide complexes. [Ln] = 1.5 X mol dm-3; [BID] = 3.0 X 10-3 mol dm-3; [CPC] = 7.5 X rnol dm-3; pH = 8.5; 4.0 cm cells. 1, Ho; 2, Nd; 3 , La; 4, Y; 5 , Er; 6, Ce'"; and 7, Ce"'. Reference: reagent blank. & d h2 51 I I 435 445 455 465 h/nm Fig. 3 Second-derivative spectra of lanthanide complexes. AL = 1 .O nm; bandpass = 1 .O nm; scan rate = 20 nm min- I ; other conditions as in Fig. 2. 1, Ce"'; 2, CelV; 3, Er; 4, Y; 5 , La; 6, Nd; 7, Ho; and 8, Ho + Er. 0.0g01-7=- 0.070 Results and Discussion Absorption Spectra The absorption spectra of holmium ion and its binary and ternary complexes formed with BID and CPC are shown in Fig.1. The difference between curves 1 and 2 in the spectral features shows an increase in sensitivity when the binary complex is converted into the ternary complex. As the characteristic sharp absorption peaks of the complexes are at almost the same wavelength as those of the free ion (holmium chloride), it can be concluded that the sensitivity is increased as a result of the ligands affecting the co-ordination field. The absorption bands observed involve transitions from the 5Z8 ground state to the 5G6 and 5F1 states. The spectral charac- teristics of systems are shown in Table 1. From Fig. 2, it is also evident that the interference of other lanthanides must be eliminated to allow the determination of Table 2 Compositions of synthetic samples (oxides, % m/m) and analytical results RSD Recovery Sample Present Found* (%) (%) 1 La 20.02, Ce'" 45.20, Pr 10.75, Nd 10.06, Sm 1.34, Eu 0.18, Er 0.73, Y 2.50, Ho 9.20 9.28 1.3 100.9 La 1.10, Ce'" 5.30, Pr 3.70, Nd 7.1 1 , Sm 1 S O , Eu 0.80, Er 8.80, Y 65.19, Ho 6.59 6.42 1.2 98.8 2 * Average of five determinations.PH BID/mmol dm-3 CPC/mmol dm-3 Fig. 4 Effect of variables on the absorbance (second-derivative) of the holmium complex (5.0 X mol dm-3 Ho): (a) pH; (b) BID concentration; and (c) CPC concentration. Except when the particular variable was studied, the solutions contained 3.0 X 10-3 mol dm-3 BID and 7.5 X 10-3 mol dm-3 CPC and the pH was 8.5. Reference: reagent blank. Instrumental and measuring conditions were as given in the Procedure.Analyst, September 1996, Vol.121 1319 Table 3 Analytical results for holmium in control sample Found (%)* Oxides present (70) Added (76) This work RSD Recovery XRF La 27.1 1, Ce’” 49.21, Pr 5.18, Nd 16.75, Sm 1.29, Eu 0.23, Gd 0.40, Tb 0.03, Dy 0.09, Er 0.027, Lu 0.003, Tm 0.0095, Yb 0.013, Ho 0.023 7.50 7.72 1.3 102.6 7.64 10.00 10.18 1.2 101.6 10.10 15.00 15.13 0.9 100.7 15.10 20.00 20.04 0.8 100.1 20.04 * Average of five determinations. Control sample obtained from Baotou Rare Earths Academy, Baotou, China. holmium in a mixed rare earth system. A chemical separation is tedious, whereas the use of the derivative spectra is not only simple and selective, but also improves the sensitivity for holmium because of the sharp absorption band. Therefore, the first- to fourth-derivative spectra of this system were investi- gated in addition to the variable measuring parameters of the instrument.The results showed that the optimum conditions for holmium determination are the use of the second-derivative spectrum with Ah = 1 .O nm and scan rate 20 nm min-1, as they afforded the best sensitivity. Fig. 3 shows the second-derivative spectra of representative individual lanthanide complexes and a mixture with holmium. It is clear that the optimum analytical signals are at 444.5 (+) nm and 449.0 (-) nm for holmium, where the other lanthanide complexes produce only constant signals, because their ordi- nary absorption spectra are smooth absorption curves and the second-derivative spectra had no peaks between 430 and 470 nm, which suggested that the determination of holmium in the presence of the other REEs should be possible.Optimization of Reaction Conditions The pH and amounts of BID and CPC were optimized for 5.0 X 10-5 mol dm-3 of holmium by varying them one at a time. Fig. 4 shows that the optimum conditions that yielded the greatest and most constant absorbance were the pH range 6.5-9.5 and CPC in the range 1.0 X 10-3-1.5 X 10-2 mol dm-3. The amplitude of holmium reached a maximum when 4.5 x 10-4 rnol dm-3 of BID was employed, and then 0.150 0.075 d 2 A o d h2 -0.075 -0.150 A 5 I I I I I L 440 445 450 455 460 465 h/nm Fig. 5 Spectral curves for calibration. [BID] = 3.0 X 10-3 mol dm-3; [CPC] = 7.5 X 10-3 mol dm-3; pH = 8.5, 4.0 cm cells. Instrumental parameters as in Fig. 3, except for change in holmium concentration: 1, 4.95; 2, 9.90; 3, 14.85; 4, 19.80; and 5 , 24.75 yg cm-3.remained constant. The optimum values thus selected in the determination are those stated in the Procedure. The complete complex formation took 5 min and the derivative absorbance was stable for at least 5 h. Beer’s Law, Sensitivity, Precision and Detection Limit The second-derivative spectra (Fig. 5 ) provide a linear calibra- tion graph up to 25 pg cm-3 of holmium (in the final solution) and the corresponding value of Sandell’s sensitivity is 0.36 yg cm-2. The deviation evaluated from 11 independent determinations of 10 yg of holmium was 0.96% for holmium alone and 1.8% in the presence of 80 pg of yttrium. The detection limit, obtained from the sensitivity of the calibration graph and for 3sb (sb = standard deviation of a blank without holmium, n = 1 l), was found to be 0.43 pg cm-3 of holmium.The quantification limit, obtained for lo&,, was 1.43 yg ~ m - ~ of holmium. Analytical Application In order to test the utility of the method, two artificially mixed samples were prepared and analysed and the results obtained are given in Table 2. As real samples were not available, the applicability of the procedure to the determination of holmium in mixed rare earth samples was tested by adding known amounts of holmium prior to control sample dissolution. The results are given in Table 3, from which it can be seen that recoveries are satisfactory. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Busev, I. A., Tipsotova, G. V., and Ivanov, M.V., Handbook of the Analytical Chemistry of Rare Earth Elements, Ann Arbor-Hum- phrey, London, 1970. Roelandts, I., Analusis, 1978, 6, 2. Morrison, H. G., and Kashuba, T. A., Anal. Chem., 1969, 41, 1842. Sen, C. I., Talantu, 1984, 31, 1045. Broekaert, A. J., Leis, F., and Laqua, K., Spectrochim Acta, Part B , 1979, 34, 73. O’Haver, T. C., Anal. Chem., 1979, 51, 91A. Griffiths, T. R., King, K., and Hubbard, H. V. St. A., Anal. Chim. Acta, 1982, 143, 163. Fell, A. F., Trends Anal. Chem., 1983, 28, 63. Fasanmade, A. A., and Fell, A. F., Analyst, 1985, 110, 1117. Choppin, R. G., Henrie, E. D., and Buijs, K., Znorg. Chem., 1966,10, 1743. Ishii, H., and Satoh, K., Bunseki Kagaku, 1986, 8, 704. Kang, J.-W., Chen, R-Y., and Bei, G.-B., Acta. Chim. Sin., 1984,42, 921. Wang, N.-X., Tulanta, 1991, 7, 71 1. Wang, N.-X., Chin. Chem. Lett., 1991, 1, 41.1320 Analyst, September 1996, Vol. 121 15 Wang, N.-X., Liang, W.-A., and Zhang, Z.-Z., Analyst, 1992, 117, 1963. 1993,110, 119. 16 Zhou, S.-F., and Wang, N.-X., Fenxi Huuxue, 1987, 11, 1041. 17 Wang, N.-X., Liang, W.-A., Zhou, S.-F., and Qi, P., Anal. Chim. Acta, 1992,262, 253. 18 Wang, N.-X., and Qi, P., Anal. Lett., 1992, 25, 11 19. 19 Wang, N.-X., Wu, Q.-C., Shi, J.-B., and Qi. P., Mikrochim. Actu, Paper 6f02306A Received April 2, I996 Accepted May 15, 1996
ISSN:0003-2654
DOI:10.1039/AN9962101317
出版商:RSC
年代:1996
数据来源: RSC
|
42. |
Simultaneous kinetic spectrophotometric determination of five phenolic compounds by reaction withp-aminophenol, using partial least squares data treatment |
|
Analyst,
Volume 121,
Issue 9,
1996,
Page 1321-1326
M. de la Guardia,
Preview
|
PDF (786KB)
|
|
摘要:
Anal-yst, Septeniher 1996, Vol. 121 (1321 -1326) 1321 Simultaneous Kinetic Spectrophotomet ric Determination of Five Phenolic Compounds by Reaction With pAminophenol, Using Partial Least Squares Data Treatment M. de la Guardiaa, K. D. Khalafa'*, B. A. Hasaw*, A. Morales-Rubioa, J. J. Ariash, J. M. Garcia-Fragab, A. I. Jimenezb and F. Jimenezh 50, 46 I00 Bui-jussot, Vulenciu, Spuin Department of Analytical Chemistry, Uniwrsit4, of Valencia, CIDoctor Moliner Department of Analytical Chemistry, Univer-sity of La Laguna, Tenerife, Spain A fast and accurate procedure has been developed for the simultaneous determination of resorcinol, m-aminophenol, o-cresol, phenol and m-cresol in waters. The method is based on the reaction between the aforementioned phenolic compounds, at concentration levels of ppm, and 300 ppm of p-aminophenol in 0.05 moll-' NaOH and in the presence of 0.004 moll-' KI04.The treatment of absorbance data, between 400 and 700 nm, obtained at interval times of 20 s between 40 and 120 s and 40 s between 120 and 600 s, by means of the use of Partial Least Squares 2 (PLS-2), using the UNSCKAMBLER program, suggested that the use of 7 factors can explain more than 99.5% of the residual variance. The method has been applied to the analysis of synthetic samples, not employed for calibration, and average errors lower than 4% were found. Keywords: Resor-cinol; m-aniinoplzenol; 0-ct-esol; phenol; m-cresnl; stopped j h 7 analysis; partial least squares; spectroi,lzotonietiy Introduction Phenol and phenolic compounds, such as resorcinol, m- aminophenol and cresols, are products widely employed in industrial manufacturing for the production o f synthetic resins, pharmaceuticals, disinfectants, fumigants, adhesives and so on.The industrial plants in which phenols are employed, are considered as potential sources of p~llution'.~ and, for that reason, these products must be controlled in environmental samples. The simultaneous determination of several phenols in the same sample can be carried out by chromatography and several method? have been proposed based on GC'-5 and HPLC.6-c) In the last few years we have developed a flow analysis strategy for the spectrophotometric determination of phenol and resorcinolI0 and also for the simultaneous determination of o- cresol and ni-cresol in water," based on the reaction between phenols andp-aminophenol (PAP) in an alkaline medium and in the presence of KI04.It is known that PAP is a phenol derivative reagent that is very sensitive for resorcinol and cresol determination, which provides different absorbance maxima for each one of the phenolic compounds under study, thus offering exciting possibilities for the simultaneous determination of a series of phenols in water. As we have previously indicated in the literature, a single, rapid method for the spectrophotometric determination of * Permanent address: University of Baghdad, Iraq. mixtures of compounds, such as carbamate pesticides, by the reaction of their hydrolysis products with PAP, is the use of multivariate calibration applied to the simultaneous treatment of both kinetic and spectrophotometric data,I2.l3 which allows us the resolution of mixtures of compounds with close similar spectra, even when their reaction speed with a derivative reagent are not excessively different.In this work, we have applied the partial least-squares (PLS) methodl7 to spectrophotometric and bilineal, kinetic-spectro- photometric, data obtained for the reaction between resorcinol, m-aminophenol. o-cresol, phenol and nz-cresol with PAP. Experimental Apparatus and Software A Hewlett-Packard (Avondale, PA, USA) Model HP8452A diode-array spectrophotometer, equipped with HPX9530A MS- DOS UV-visible software, with a response time of 0.1 s, was employed for the spectrophotometric determinations using a 50 pl internal volume and 1 cm pathlength flow cell for absorbance measurements.The three-channel manifold indicated in Fig. 1 was employed to transport and merge samples and reagents. PAP and KI04 are initially mixed, using a reaction coil of 45 cm length, and then merged with samples or standard solutions. A Gilson P2 minipuls peristaltic pump (Villiers Le Bel, France), equipped with flexible poly(viny1 chloride) tubes of 1.52 mm id, was used. Reaction coils and connections were made from PTFE tubes and Y-shaped merging zones with 0.8 mm id. The temperature was kept constant with the aid of an MS lambda thermostat. Data obtained were transferred to a PC/AT 386 personal computer DTK (Taiwan), equipped with an Intel 80387 mathematical coprocessor, for thcir treatment with the PUMP Standards or Samples WASTE Fig.1 with PAP. Manifold employed for the determination of phenols by reaction1322 Analyst, Scptemher 1996, Vol. 121 UNSCRAMBLER program Ix written according to published algorithms. 1 9,20 Reagents Stock standard solutions of 100 pg ml-1 of resorcinol, nz- aminophenol, o-cresol, phenol and rn-cresol were prepared from analytical grade reagents supplied by Aldrich (Steinheim, Germany) and dissolved in high purity distilled water. Working standard solutions were prepared by diluting the stock solutions in distilled water with the appropriate content of NaOH. The stock solution of PAP ( I 000 pg ml-I) was prepared daily by dissolving 0.5 g of 4-aminophenol (PAP) Fluka (Buchs, Switzerland) in 500 ml of boiled and cooled distilled water, Other solutions with low PAP concentrations were prepared by the same procedure.Boiling the distilled water for 10 min is very important to avoid the oxidation of PAP by dissolved oxygen. Stock solutions of potassium metaperiodate and NaOH were prepared by dissolving the corresponding solid products from Probus (Barcelona. Spain) in distilled water. General Procedure PAP and K104 solutions are mixed together and after that merged with alkaline solutions of resorcinol, m-aminophenol, o-cresol. phenol and rn-cresol, using the manifold indicated in Fig. 1 . The reaction between phenolates and the quinoneimine compound obtained by the continuous oxidation of a 300 pg ml- PAP solution with 0.004 moll-' KI04 is controlled in the stopped-flow mode, carrying out absorbance measurements in the wavelength range between 400 and 800 nm at interval times of 20 s between 40 and 120 s and at 40 s between 120 and 600 s.The data obtained were transferred to the DTK computer and treated by use of the UNSCRAMBLER program. Results and Discussion Reaction Between Phenols and PAP The reaction between phenols and PAP is based on the reactivity of the benzoquinoneimine form of PAP, obtained by a previous oxidation with KI04, with para-position free phenolates in an alkaline medium. Phenolates react with oxidized PAP through the electrophilic attack on C4 of the phenolate to produce the corresponding leucodyes, which are oxidized to different indodyes, which absorb in the visible range. The reaction is indicated in the following scheme: OH 0 OH 0 K I I I + - NH - N " G' I 0" 0" G' I 0" Different phenols give rise to different indodyes with PAP and, in addition, the reaction speed of phenols with PAP also varies.It has been observed that resorcinol reacts very rapidly with PAP, the colour being fully developed in 0.75 min; a reaction time of 2.3 min is required for maminophenol, 12 and 15 min for m-cresol and 0-cresol, respectively, and 45 min for phenol. The reaction conditions and spectrophotometric charac- teristics for the determination of the phenols under study, with PAP in the stopped-flow mode and in the optimum reaction conditions, have been described. 1 0 9 1 1 Analytical Features of the Spectrophotometric Determination of Phenols With PAP Under the experimental conditions indicated in Table 1 , phenols can be determined at pg ml-I trace levels with limits of detection of the order of several ng ml-1.Thus, it can be concluded that PAP is a good derivative reagent to be employed for the spectrophotometric determination of all of the phenolic compounds under study, providing a sensitivity which ranges from 0.0158 absorbance units pg-l ml for rn-aminophenol to 0.0969 absorbance units pg-1 ml for phenol (see Table 1).1()~11 The differences observed in the intercept of the calibration lines corresponding to each phenol are caused by the residual absorbance of PAP, at the measurement wavelength, and the fact that different PAP concentrations are required in each case. Experimental Conditions for the Simultaneous Determination of Phenols With PAP In order to develop an appropriate method for the simultaneous determination of all of the compounds considered, the following compromise conditions were selected: an NaOH concentration of 0.05 moll- I , a 0.004 mol 1-1 KI04 solution and 300 pg ml-l of PAP.The 0.05 moll-' NaOH concentration is that employed for m-aminophenol determination and a little bit lower than that employed for cresol determination. This NaOH concentration is approximately one order of magnitude higher than that employed for phenol and resorcinol determination. However, as has been reported,lo an increase in the NaOH concentration also increases the reaction yield; thus it can provide as fast a reaction rate as possible between PAP and phenol. Regarding the KI04 concentration, a compromise value of 0.004 mol 1-1 seems the most appropriate, taking into consideration that it is higher only than that recommended for resorcinol determination.The use of an excess of PAP concentration is required to obtain as high a reaction yield as possible in as low a reaction time as possible; thus, a high PAP concentration can be necessary, especially for the determination of phenol and cresols, which present a slow reaction rate with this reagent. However, the fact that increasing the PAP concentration also increases the intercept of the calibration curves (see Table 1) can justify the use of a small excess of PAP to reduce the residual absorbance of this reagent at the measurement wavelengths. Therefore, a 300 pg ml- I concentration was selected in this study. Under the above-mentioned conditions, resorcinol reacts completely with PAP from the beginning, the reactions with rn- aminophenol, o-cresol and m-cresol being less fast.The reaction between phenol and PAP is the least favourable one and needs more than 1 h to be completed under the selected conditions. Fig. 2 shows the spectra of the reaction products between all of the phenols assayed and PAP, obtained after a reaction time of 10 min under the compromise conditions indicated before. It can be seen that the reaction takes place over an appropriate range, the contribution of the blank being an important aspect to be considered in the resorcinol determination. (Note that the compromise conditions involve the use of a PAP concentration 6 times higher than that employed for resorcinol.) Fig. 3 shows the evolution, as a function of time, of the absorbance of each species at its maximum absorbance wavelength.In this figure it can be seen that the use of compromise conditions involves an almost immediate reactionAnalyst, September 1996, Vol. 121 1323 between PAP and resorcinol, but a very slow reaction with phenol, the reaction rates of all the five compounds being considered clearly different, thus offering tremendous possibi- lities for the simultaneous determination of all of these compounds in a same sample. PLS Treatment of the Spectrophotometric Data As is indicated in Fig. 2, the absorbance maxima of the reaction products with PAP of resorcinol, o-cresol, phenol, m-aminophe- no1 and rn-cresol are located at 540, 614, 626, 576 and 632 nm, respectively, showing the high spectral overlapping presented by these products.Because of that, the only way to solve this complex mixture seems to be the use of a multivariate calibration of the spectrophotometric data obtained after reaching an appropriate reaction yield. From a calibration matrix (see Table 2) obtained with mixtures of the five compounds under study at two concentra- tion levels, thus including 25 solutions plus a blank solution (because PAP presents a residual absorbance in the wavelength 400 SO0 600 700 800 Wavelengthhm Fig. 2 Absorbance spectra of the reaction products between PAP and resorcinol (l), m-aminophenol(2), o-cresol(3), phenol (4) and m-cresol(5) and the absorbance of a blank solution containing 300 pg ml-l of PAP in 0.05 mol 1-1 NaOH and 0.004 mol 1-1 KI04 (6).All the spectra were recorded 10 min after mixing the reagents. range considered), absorbance data were taken between 400 and 800 nm at 2 nm intervals after 10 min of merging reagents and standards using the manifold indicated in Fig. 1. The model was centred about the mean and PLS calibration carried out using PLS-2, evaluating the model by cross- validation using a number of segments equal to the number of solutions included in the calibration matrix. Table 3 shows the percentage of residual variances explained, in the cross-validation process, for the concentration matrix. As can be seen, using seven factors the variance of resorcinol, m- aminophenol and m-cresol can be determined, the variance found for phenol and o-cresol being a little bit lower than 99%.The consideration of an additional factor increases the percent- age of variance for phenol and o-cresol but involves an overfitting of the model, as can be seen in the plot of scores of absorbance data matrix versus scores of concentration data matrix (Fig. 4). The objects dispersed about the regression line when eight factors are considered (Fig. 4b), which can be ascribed to the presence of noise and hence overfitting of the model. On the contrary, when only seven factors are considered (see Fig. 4a) points are aligned through a curve. To verify the efficiency of the model for the prediction of the concentration of phenols in unknown samples, four synthetic mixtures of the five phenolic compounds under study were prepared and analysed by PLS-spectrophotometry. Table 4 shows the results obtained, also including the confidence intervals found using 7 factors.It can be observed that for 0.3 I I 0.2 20 120 220 320 420 520 620 Time/s Fig. 3 Kinetic curves of the reaction between PAP and resorcinol ( l ) , m- aminophenol (2), o-cresol (3), phenol (4) and m-cresol (5). All the curves were obtained at the maximum absorbance wavelength of each species. Table 1 Analytical characteristics of phenol determination with PAP in the optimum conditions established for each single compound. C is in pg ml-1. LOD values were established for a k = 3 value. s corresponds to the relative standard deviation of 3 absorbance measurements of a solution containing 5 pg ml-' of each compound Phenolic LOD/ Rangel compound Calibration line ngml-1 s (%) r pg ml-l Resorcinol A = 0.004+0.0397 C 7.5 0.5 0.9999 2-6 m-Aminophenol A = 0.0348 + 0.0158 C - - 0.9998 2-8 o-Cresol A = 0.0257+0.075 C 11 0.7 0.9994 2-8 Phenol A = 0.1658 + 0.0969 C 64 0.5 0.9999 2-10 m-Cresol A = 0.0383 + 0.0699 C 30 6 0.9998 2-81324 co -0,0005] 0 Analyst, September 1996, Vol. 121 I analytes present at low concentrations, the errors found were higher than those obtained for the others.The process was h L .- ii E 0,0005- 4- U al Table 2 Calibration matrix employed for the PLS-spectrophotometric determination of phenols. Resorcinol (RES), m-aminophenol (MAP), o- cresol (OCR), phenol (PHE), m-cresol (MCR). In all cases a PAP concentration of 300 pg ml-1, an NaOH concentration of 0.05 moll-' and a KI04 concentration of 0.004 mol 1-1 were employed 0 0 0 0 0 0 0 0 0 8 0 Amound addedlpg ml-1 e .2 s 0 Sample 1 2 3 4 S 6 - 7 8 9 10 11 12 13 14 1s 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Blank O 0 0 C 0 0 0 0 0 RES 2 2 2 6 6 6 6 2 2 2 6 2 2 2 2 6 2 6 6 6 2 6 6 2 6 2 6 6 6 6 2 2 0 b 2 00 -0,002- MAP 7 7 2 2 7 7 2 2 7 7 7 2 7 7 2 2 7 7 2 2 2 2 7 2 7 7 7 2 7 2 2 2 0 0 0 0 0 0 0 0 0 0 0 OCR 7 2 7 7 7 2 7 2 2 7 7 7 7 7 2 2 2 7 7 7 7 2 7 2 2 2 2 2 2 2 7 2 0 PHE 9 4 4 9 4 9 4 9 4 4 9 9 9 4 9 9 9 4 4 9 9 4 9 4 4 9 4 9 9 4 4 4 0 MCR 2 7 2 7 7 7 7 7 2 2 7 2 7 7 2 2 2 2 2 2 7 7 2 7 7 7 2 7 2 2 7 2 0 O ' O o 6 1 3 0,004 ii E .- 0 0 0 a 0 0 0 c : 0,002 -0,004 -0,002 0 0,002 0,004 Score 7 (Concentration data matrix) repeated using PLS-1 as the calibration model, but similar results were found.PLS Treatment of the Bilineal Data In order to improve the results reported above, a multivariate calibration of bilineal kinetic-spectrophotometric data was carried out. In order to select the variables which provide the highest information about the system, a calibration matrix was established with an experimental design of five variables at two concentration levels, as indicated before, also including a blank solution (see Table 2). The calibration covers a concentration range between 2 and 6 pg ml-1 for resorcinol, 4 and 9 pg ml-1 for phenol and 2 and 7 pg ml-' for m-aminophenol, o-cresol and m-cresol, and absorbance data were taken between 400 and 800 nm. The most important kinetic information for m-aminophenol corresponds to the first 120 s, as can be seen in Fig.3, where the absorbance of the indodye formed with PAP is constant after the first 2 min. Thus, to evaluate a first calibration approach, the spectra obtained between 20 and 120 s, at interval times of 20 s, and those obtained between 120 and 600 s, at intervals of 40 s, were selected. For this calibration, absorbance data found between 400 and 700 nm, at 6 nm intervals, were employed. The data tensor was modified in order to reduce the number of dimensions, Thus, the results obtained were grouped by Table 3 Residual variance explained (by cross-validation) for the concentration of the validation objects as a function of the number of factors considered (non-kinetic data) Residual variance explained (%) Factors 0 1 2 3 4 5 6 7 Resorcinol 0 7.7 41.85 63.73 77.1 95.1 97.0 99.0 m- Aminophenol 0 28.6 57.82 60.0 72.1 86.2 96.3 99.2 o-Cresol 0 33.9 3 1.94 61.8 59.4 64.6 92.7 98.2 Phenol 0 10.6 6.9 6.6 65.7 63.3 93.9 97.9 m-Cresol 0 39.1 62.6 93.6 98.7 98.9 99.8 99.8 0,0015 I I 0,001 0 0 4b -0,0010 O 1 I -0,001 5 .' ~ ' -0,001 0 0,001 Score 8 (Concentration data matrix) Fig. 4 Plot of scores of absorbance data matrix versus scores of concentration data matrix for the seventh (4a) and eighth (4b) factors considered.Analyst, September 1996, Vot. 121 1325 Table 4 Results obtained for the spectrophotometric determination of phenols using data found after 10 min of reaction with PAP. Resorcinol (RES), m- aminophenol (MAP), o-cresol (OCR), phenol (PHE) and rn-cresol (MCR) Amount added/pg ml- 1 Amount found f standard deviation/pg ml-- I Sample RES MAP OCR PHE MCR RES MAP OCR PHE MCR A 2.5 2.5 6.5 8.5 2.5 2.5kO.l 2.3fO.l 6.4f0.2 8.4f0.2 2.44k0.05 B 2.5 6.5 2.5 4.5 6.5 2.3k0.2 6.7k0.2 1.950.3 5.2k0.3 6.50f0.08 C 5.5 2.5 6.5 4.5 6.5 5.5f0.2 2.3k0.2 6.3k0.3 4.4k0.3 6.35f0.08 D 4.0 4.5 4.5 6.5 4.5 4.2k0.1 4.4f0.1 4.6k0.2 6.5k0.2 4.53k0.05 Table 5 Residual variance explained (by cross-validation) for the concentration of the validation objects as a function of the number of factors considered, using kinetic-spectrophotometric data Factor Residual variance explained (%) number RES MAP OCR PHE MCR 0 1.1 32.2 63.0 66.6 95.7 99.8 99.9 0 28.9 57.4 59.8 59.6 96.8 99.5 99.8 0 29.7 30.9 64.3 59.7 98.6 99.6 99.7 0 0 0 25.3 0 56.4 0.2 74.9 92.9 86.7 95.3 99.3 98.9 99.3 99.6 99.9 placing the spectrum recorded every 20 s next to the previous one, in such a way that the resulting time-spectra vector contains all the spectra corresponding to each solution in a sequential order.After that, the model was centred about the mean and the leverage correction method was used for validation. The graphical representations of residuals versus leverage and the anomalous data found by the program allows us to eliminate the spectra obtained for t = 20 s, which have high residual variance. This is due to the fact that data obtained at 20 s correspond to the first measurements and provide a low information about phenol and cresols, which have reacted in a small extension with PAP. On the other hand it was observed that, in general, measurements carried out at wavelength values lower than 500 or higher than 700 nm had high residual variance and low leverage.The calibration was repeated using the spectra recorded at times between 40 and I20 s at 20 s intervals, and between 120 and 600 s at 40 s intervals. The model was then centred about the mean values and the cross-validation was applied using PLS- 2. Table 5 shows the percentage of the residual variance given by the model, during the validation process, compared with the number of factors considered, it being observed that the main part of the residual variance for the concentrations of the validation objects was accounted for by using only six factors, although it was necessary to use seven factors to explain more than 99.5% of the residual variance for all the phenolic compounds assayed.Fig. 5 shows the mean standard error of prediction values, expressed in kg ml-l, for each phenol as a function of the number of factors, and it can be seen that when using seven factors errors predicted are lower than 0.2 pg ml- I . The above-mentioned model was employed to predict the concentration of phenols in four synthetic samples (not employed in the calibration process), and data are summarized in Table 6, in which has also been included the confidence intervals found and the relative errors. From these data it can be concluded that the model provides errors lower than 5% for all the compounds under study, except for phenol, for which sometimes high error values can be found. This can be 0 1 2 3 4 5 6 7 Factors Fig.5 Mean standard error of prediction values for each phenol as a function of the number of factors considered. Resorcinol (l), m- aminophenol (2), o-cresol (3), phenol (4) and m-cresol (5). explained by the fact that, in the compromise conditions selected to carry out the simultaneous determination of the five compounds under study, phenol only reacts partially with PAP and, because of that, spectra found before 160 s provide low information about the phenol concentration. Additional studies carried out, focusing the model in the interval time between 160 and 600 s and using PLS-1 approach, provided better results for phenol concentration than those obtained using PLS-2, with a mean standard error of prediction for phenol concentration of 0. I pg ml- (considering 7 factors).However, when using this approach the results obtained for the rest of phenols considered were not improved. Conclusions The reaction between PAP and phenolic compounds, not substituted in the para-position, provides a general derivatiza- tion procedure for the spectrophotometric analysis of these compounds. The use of the PLS-2 algorithm, applied to bilinear spectrophotometric absorbance data as a function of time, provides accurate results for the simultaneous determination of o-cresol, m-cresol, resorcinol, phenol and m-aminophenol in the same sample. This study is the result of a strong cooperation between the Universities of La Laguna (Tenerife) and Valencia (Burjassot) and the authors acknowledge the facilities provided by both Universities.K.D.K. acknowledges the financial support of the Institute of Cooperation with the Arabic World of the Foreign1326 Analyst, September 1996, Vol. 121 Table 6 Results obtained for the kinetic-spectrophotometric determination of phenols with PAP using PLS-2 treatment. Resorcinol (RES), m-aminophenol (MAP), o-cresol (OCR), phenol (PHE) and m-cresol (MCR) Amount added/pg ml-1 Amount found k standard deviatiodyg ml-l Sample RES MAP OCR PHE MCR RES MAP OCR PHE MCR A 2.5 2.5 6.5 8.5 2.5 2.46f0.06 2.40f0.08 6.4k0.1 8.3f0.1 2.53f0.08 B 2.5 6.5 2.5 4.5 6.5 2.60k0.09 6.2k0.1 2.6k0.2 4.5k0.2 6.4k0.1 C 5.5 2.5 6.5 4.5 6.5 5.4f0.1 2.5kO.l 6.4f0.2 4.1f0.2 6.4+0.1 D 4.0 4.5 4.5 6.5 4.5 4.15f0.05 4.42k0.06 4.45k0.08 6.54f0.09 4.57f0.06 Relative error (%I) RES MAP OCR PHE MCR -1.6 -4.0 -2.2 -2.4 1.2 4.0 -4.5 2.4 -0.2 -1.2 -2.2 -0.8 -2.0 -8.2 -0.9 3.8 -1.8 -1.1 0.6 1.6 Affairs Ministry of Spain and the University of Baghdad (Iraq) to carry out the PhD studies in Spain. References 1 2 3 4 5 6 7 8 9 10 Jacobs, M.B., Analytical Chemistry of Industrial Poisons, Hazards and Solvents, Interscience, New York, 2nd edn., 1949. American Industrial Hygiene Association (Spanish Section), Vulores Limite e Indices Bioldgicos de Exposicidn, Conselleria de Trabajo y Seguridad Social Ed., Valencia, Spain, 1991. Hoshika, Y., and Murayama, N., Analyst, 1983, 108, 984. Amankwa, L. N., and Kuhr, W. G., Anal. Chem., 1991, 63, 1733. Nanni, E. J., Lovette, M. E., Hicks, R. D., Fowler, K. W., and Borgerding, M. F., J . Chromutogr., 1990, 505, 365. MacCrehan, W. A., and Brown-Thomas, J. M., Anal. Chem., 1987, 59, 477. Di-Nuncico, C., Parisi, G., Santoro, P., and Ricci, P. A., J. Chromatogr., 1987, 392, 454. Brega, A., Prandini, P., Amaglio, C., and Pafumi, E., J . Chromatogr., 1990,535, 3 1 1. Risner, C. H., and Cash, S. L., .I. Chromatogr. Sci., 1990, 28, 239. Khalaf, K. D., Hasan, B. A., Morales-Rubio, A., and de la Guardia, M., Tulanta, 1994, 41, 547. 11 12 13 14 15 16 17 18 19 20 Khalaf, K. D., Hasan, B. A., Morales-Rubio, A., and de la Guardia, M., Mikrochim. Acta, 1993, 112, 99. Garcia, J. M., JimCnez, A. I., Arias, J. J., Khalaf, K. D., Morales- Rubio, A., and de la Guardia, M., AnaZyst, 1995, 120, 313. Khalaf, K. D., Morales-Rubio, A., de la Guardia, M., Garcia, J. M., JimCnez, F., and Arias, J. J., Microchem. J., 1996, in the press. Havel, J., JimCnez, F., Bautista, R. D., and Arias Le6n, J. J., Analyst, 1993,118, 1355. Blanco, M., Coello, J., Iturriaga, H., Maspoch, S., Riba, J., and Rovira, E., Talunta, 1993, 40, 26. Cladera, A., Gbmez, E., Estela, J. M., Cerda, V., and Cerda, J. L., Anal. Chim. Acta, 1993, 272, 339. Martens, H., and Naes, T., Multivariate Calibration, Wiley, Chiches- ter, 1991. Unscramhler-ex v. 5.0, Computer-Aided Modelling, Trondheim, Norway, 1993. Wold, H., in System Under Indirect Observation, eds. Joreskog, K. G., and Wold, H., North Holland, Amsterdam, The Netherlands, 1982, Vol. 2. Lorber, A., Wangen, L. E., and Kowalski, B. R., J . Chemometr., 1987, 1, 19. Paper 6103221 D Received May 8,1996 Accepted June 10,1996
ISSN:0003-2654
DOI:10.1039/AN9962101321
出版商:RSC
年代:1996
数据来源: RSC
|
43. |
On-line solid-phase extraction–liquid chromatography–particle beam mass spectrometry and gas chromatography–mass spectrometry of carbamate pesticides |
|
Analyst,
Volume 121,
Issue 9,
1996,
Page 1327-1334
Jaroslav Slobodník,
Preview
|
PDF (1172KB)
|
|
摘要:
Analyst, September 1996, Vol. 121 (1327-1334) 1327 On4 i ne Sol id-p hase Extract ion-Liq u id Chromatography-Particle Beam Mass Spectrometry and Gas Chromatography-Mass Spectrometry of Carbamate Pesticides Jaroslav Slobodnik",* Sacha J. F. Hoekstra-Oussoren", Maria E. Jagera, Maarten Honingb, Ben L. M. van Baarc and Udo A. Th. Brinkmanu Department of Analytical Chemistry, Free University, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands h CID-CSIC, Environmental Chemistry Department, J o d i Girona 18, 080 34 Bat-celona, Spain ( Departnwnt of Organic and Imrganic Chemistry, Free University, De Boelelhan 1083, I081 I-IV Amstei-dam, The Netherlands On-line solid-phase extraction-liquid chromatography-particle beam mass spectrometry (SPE-LC-PB-MS) was used to study 32 carbamates and 11 of their transformation products in environmental water samples. The analytes were enriched from 100 ml samples on Bondesil-C18/OH, packed in a 10 mm x 3.0 mm id precolumn, eluted on-line with a gradient of methanol-0.1 mol 1-1 ammonium acetate to a CIS analytical column and detected by PB-MS. Detection limits of 12 carbamates and five degradation products were 0 .1 4 pg 1-l in surface water, using full-scan electron impact mode detection. The RSDs of the retention times were 0.05-0.2% and those of peak areas 5-35%. Twenty-eight analytes proved to be amenable to GC with detection limits of 0.05-3 ng injected on-column. When using the same mass spectrometer, the spectra obtained by LC-PB-MS and GC-MS were identical. Keywords: On-line solid-phase extraction; column liquid chronzatography; gas chromatographj~; mass specti-ometiy; carbarnate pesticides; environmental samples Introduction Many carbamates are widely used as pesticides.I Residues of these pesticides are present in crops and in ground and surface waters. With few exceptions, carbamates are believed to be non- persistent in the environment because of photolytic and hydrolytic degradation. The separation and identification of the carbamate pesticides along with their main transformation products is important because of their proven or suspected toxicity. Unfortunately, satisfactory procedures for the deter- mination of this class of compounds are limited. Carbamates in general and their transformation products in particular are often quoted as being polar, involatile and thermally labile, properties which prevent direct analysis by GC.Although supercritical fluid chromatography (SFC), with capillary2 and packed3 columns, has been used for the separation of carbamates, column LC is obviously the preferred approach. A variety of LC detection method^,^ including diode-array (DAD-UV),S elec- trochemicalh and fluore~cence~-~ detection, have been used. * On leave from the Environmental Institute, Okrvina 785/52, 9725 I KO:, Slovakia. Despite the good sensitivity achieved by post-column deriva- tization with fluorescence detection or the robustness of DAD- UV detection, the use of MS detection seems to be a key element for future LC methods l o because of the confirmation and identification potential. LC-MS of carbamates has been performed using moving belt (MB), particle beam (PB) and thermospray (TSP) interlaces and atmospheric pressure ionization-based interfacing technol- 0 g y .~ J ' ~ ' 2 So far, PB has proved to be the only LC-MS interface that opens up the full identification power of electron impact (EI) ionization mass spectrometry. 1 Unfortunately, PB suffers from a distinct lack of sensitivity (typically 10-SO0 ng should be injected) and from non-linearity of detection. The lack of sensitivity can be compensated for, at least partly, by applying on-line trace enrichment; l6,17 solid-phase extrac- tion (SPE) of 100-250 ml samples provides enrichment factors of up to about 10 000. The problem of non-linearity has not yet been overcome, although co-elution of analytes and a so-called 'carrier compound' has been reported to provide a remedy in some cases.14,1X Promising results were obtained in a recent study on EI and chemical ionization (CI) PB-MS of carbamates, which showed that the former mode should be the first choice for carbamate determination.' 9 In this study, we used on-line SPE-LC-PB-EI- MS to study 32 carbarnates and I 1 of their transformation products (cf, Table 2). The fact that PB-MS yields library- searchable EI mass spectra and allows solvent-independent CI- MS potentially provides an extension to the earlier developed SPE-LC-DAD-UV method;5 it affords confirmation of the target analytes and provides provisional information on the identity of co-extracted unknown compounds.20 As LC-PB-MS and GC-MS can be combined in one set-up sharing the same mass spectrometer in a so-called multi- analysis system,17 we briefly re-investigated the use of GC-MS for the determination of carbamates. Despite reported diffi- culties (see above), several carbamates have successfully been determined directly by capillary GC, viz., by avoiding thermal degradation by the use of a cold 'on-column' injection system or programmed-temperature vaporizer21-23 (PTV) injector.Gen- erally, the more thermolabile carbamates are determined with the combined use of short (3-9 m) capillary columns and low transfer line and ion source temperatures (typically I00 OC).24 We decided to use a longer column (12 m) and more common transfer line and source temperatures (300 and 250 OC, respectively) in order to cover the wide polarity range of the compounds and to have conditions similar to those used in the PB-MS work.I328 Analyst, September- 1996, Vol.121 Experimental Instrumentation LC An HP 1090 LC system (Hewlett-Packard, Waldbronn, Ger- many) equipped with an autosampler, six-port switching valve (Rheodyne, Cotati, CA, USA) and filter photometric UV detector was used for LC and flow injection (FI) experiments. Separations were carried out on a 250 X 4.6 mm id Supelco LC- 18-DB analytical column packed with 5 pm Supelcosil C I 8 (Supelco, Bornem, Belgium). Two Model 300 LC pumps (Separations, H.1. Ambacht, The Netherlands) and a universal valve-switching module (Anachem, Luton, UK) were used to deliver methanol and the sample during the trace enrichment procedure.Trace enrichment was performed on a 10 x 3.0 mm id stainless-steel precolumn packed with 40 pm, 100 A Bondesil-CIx/OH (Varian, Harbor City, CA, USA) qr a 10 X 2.0 mm id precolumn slurry-packed with 20 pm, 100 A PLRP-S styrene-divinylbenzene copolymer (Polymer Laboratories, Church Stretton, UK) and styrene-divinylbenzene-based 80-160 pm ENVI-Chrom P (Supelco). experiments. All analyses were performed in the El mode with a scan range of 50-350 u at 1 scan s-1. Chemicals and Reagents Pesticide standards were obtained from Kiedel-de Haen (Seelze, Germany), the US Environmental Protection Agency (EPA) Repository for Toxic and Hazardous Materials (Research Triangle Park, NC, USA) and Dr. Ehrenstorfer (Augsburg, Germany). Several N-methylcarbamates were a gift from Dr.A. de Kok (Food Inspection Service, Alkmaar, The Netherlands). All compounds were at least 95% pure. HPLC-grade acetoni- trile was obtained from Westburg (Leusden, The Netherlands), HPLC-grade water, methanol, acetic acid, ethyl acetate and ammonium acetate were purchased from J. T. Baker (Deventer, The Netherlands). Helium (Hoekloos, Schiedam, The Nether- lands), for the PB nebulizer and for degassing of solvents, was of 99.99999% purity. The purities of the methane and ammonia reagent gases from the same producer were 99.99995% and 99.999696, respectively. MS LC-PB-MS was performed on an HP 5989A MS Engine equipped with an HP 59980B PB interface and a 'high-energy dynode' (HED) detector (Hewlett-Packard). The connection of LC and PB was achieved with a SO cm X 0.12 mm id stainless- steel capillary. The helium flow of the PB nebulizer was optimized by a standard procedure and operated at the optimum flow of 2-2.5 1 min-1 (30-40 psi inlet pressure indication). The PB desolvation chamber was kept at 70 "C.The mass spectrometer was operated at an optimized source temperature of 200 0C19,24 and an analyser temperature of 100 "C. It was tuned with m/z 69, 219 and 502 ions of perfluorotributylamine (PFTBA), with maximum signal in- tensity tuning for m/z 502; the HED voltage was maintained at 10 kV and the multiplier voltage was set to 100 V above the tune value during data acquisition. Spectra were acquired in the mass range 65-350 u at a scan rate of 0.36 scans s - 1 in the positive- ion mode and using 70 eV ionization energy; the El manifold pressure was maintained below 5 X 10-5 Torr (6.7 X 10-3 Pa) by regulating the PB helium flow.Details of CI operation can be found elsewhere.19724 The system was controlled and data were acquired by an HP UX 98578X data system running on an HP UNIX Series 9000/345 computer. GC-MS An HP 5989 Series I1 gas chromatograph, equipped with an on- column injector and a GC-MS interface, was used. Separations were performed on an HP- 1 cross-linked silica gum ( I2 rn X 0.2 mm id, 0.33 mm film thickness) analytical column (Hewlett- Packard, Avondale, PA, USA). The on-column injector was connected to the analytical column by a 30 cm X 0.32 mm id retention gap of deactivated fused silica; the sample (ethyl acetate-methanol solution) injection volume was 1 pl.A constant flow of helium at 0.9 ml min- was maintained during analysis at a constant column head pressure of 50 psi; a typical MS manifold pressure was 1 X 1 O-s Torr ( I .3 X Pa). The GC oven temperature programme started at 70 "C (hold, 2 min) with a subsequent increase to 300 "C at 10 "C min-1 (hold, 10 min); the temperature of the interface was kept at 310 "C. The ion source and quadrupole were kept at 250 and 100 "C, respectively. With the exception of the ion source temperature, values from the LC-PB-MS tune file were used for the GC-MS Procedures Standard solutions (200 pg ml-I) of all carbamates were prepared in methanol and diluted to appropriate concentrations in mixtures and samples prior to analysis. The standard solutions were stored in the dark, at - 18 "C, and analysed by FT-PB-MS at 2 month intervals. No degradation was observed during the 1 year period of the study, except for benomyl and thiophanate-methyl.The pH of the ammonium acetate solution was adjusted to 5 by the addition of acetic acid. The eluent was degassed every day by purging with helium for 15 min. Prior to analysis, samples of surface water were adjusted to pH 3 with acetic acid to prevent degradation of carbamates.7 The linear LC gradient of methanol-0.1 mol 1-1 ammonium acetate was from 10% to 90% methanol in 45 min and then back to 10% methanol in 5 min; the flow rate was 0.4 ml min-1. The precolumn was freshly packed and exchanged after every sixth analysis. Acetonitrile was used instead of methanol when using a PLRP-S and ENVI- Chrom precolumn; in these instances UV detection at 210 nm was utilized. Prior to the SPE experiments, 1-25 p1 of solutions of each analyte were injected into a carrier stream of organic modifier- 0.1 mol 1-1 ammonium acetate (1 + 1 v/v).The value obtained was used to determine the maximum achievable analyte response in PB-MS ( c j . Table 2). In the experimental set-up (Fig. I), the HP 1090 system was used as a central control unit. The 'remote control' port of the HP 1090 started up the MS at a pre-programmed time; LC pumps and switching valve V1 were controlled by electronic event contact signals (On and Off positions). Valve V2, built-in in the HP 1090, was controlled directly via the HP 1090 instrumental software. The experimental procedure is sum- marized in Table 1.The precolumn tubings and precolurnn sorbent were first washed with 10 ml of methanol and then 10 ml of water, each at 10 ml min-1 (Pump 1). Next, 100 ml of sample (Pump 2) were enriched on the precolumn at the same flow rate. After switching the six-port valve V2, the precolumn was eluted by the LC gradient and MS data acquisition was started. All the above actions were started by a single keystroke. With this system, sequences of analyses were performed unattended. Flow injection calculations of S/N peak-to-peak height in ion traces of m/z of base peak in spectrum) were automated by a self-compiled macro program. l 9 Results and Discussion Formally, carbamates are derivatives of carbamic acid, H*NC(O)OH, with substituents at the amino nitrogen and theAnulyst, September 1996, Vol.I21 1329 hydroxy oxygen atoms. The compounds used in this study can broadly be divided into three subgroups, N-methyl aryl carbamates (Class l), N-methyl oxime carbamates (Class 2) and N-substituted aryl ester carbamates (Class 3) (Fig. 2), and a group of miscellaneous compounds (Class 4). Because of relevant differences in the SPE-LC, PB-MS and GC-MS behaviour of the parent carbamates and their degradation products, Class 4 will be discussed as a separate group (cf. Table 2). Structures of all the compounds can be found Detection limits required for the determination of carbamates in environmental water samples may range from sub-pg 1-1 (drinking water2') viu low-pg 1-I (alert/alarm level for surface water**) to 10 and 200 pg 1-1 for aldicarb and oxamyl, elsewhere.2,12,19,25.26 EC LOAD - ELUTE HP 1090 CI P t t t WASTE Fig.1 Set-up for automated on-line SPE-LC-PB-MS analysis of carbamate pesticides. HP 1090 liquid chromatograph; V I , Vz, automatic six-port switching valves; LoadElute, positions of VI and V2; EC, electronic connections; Pump 1, Pump 2, preparative LC pumps; PR, pre- column; AC, analytical column; SDS, solvent-delivery system of HP 1090 liquid chrotnatograph; PB, particle beam interface; GC, gas chromatograph; C1, chemical ionization reagent gas inlets; MS, mass spectrometer. ~ Table 1 Timetable of automation programme in on-line SPE-LC-PB-MS Time/ min 00:00 00:01 0031 0131 01:3S 0 I :40 02:40 12:40 12:45 1250 6250 v , *.t Elute Load Elute V2*,f El Ute Load Elute Load Elute P,+,t P2+J Comment Off Off Start position, start timetable, start MS data acquisition (solvent delay, 1250 min) MeOH flow rate 10 ml min- I , clean capillaries MeOH flow rate 10 ml min- 1, condition pre- column On Off Stop MeOH flow V2 start position Sample flow rate 10 ml min-1, removal of MeOH from capillaries Sample flow rate 10 nil min- 1, enrichment of 100 ml sample on pre- column On Off Stop sample flow V I start position Elute precolumn with gradient LC eluent (end of MS solvent delay, start data acquisition) End of timetable * V, six-port switching valve.t See also Fig. 1. i: P, LC pump. respectively.29 Since PB-MS detection does not provide the required sensitivity, considerable analyte enrichment by means of on-line SPE with small precolumns or cartridges containing an appropriate sorbent will be necessary."30 Selection of Sorbent and Breakthrough Volumes This study covered a wide range of compounds, i.e., from the highly polar N-methyl oxime carbamate sulfoxides to much less polar compounds such as the N-methyl aryl carbamates.Accordingly, their breakthrough volumes may also vary significantly and especially the more polar Class 2 carbamates (e.g., aldicarb, methomyl and oxamyl) will not be sufficiently retained by conventional apolar SPE phase^.^ As a performance test, the trapping efficiency of two non-selective sorbents, polymeric PLRP-S and ENVI-Chrom P, was studied with a test mixture of aldicarb, aldicarb sulfoxide and aldicarb sulfone. Unfortunately, the breakthrough volumes determined by means of on-line SPE-LC-UV were found to be only 20-50 ml for all three analytes.The capacity of the precolumn could not be increased by using 3 mm instead of 2 mm id columns, owing to the additional peak broadening thus created. Therefore, in a later stage Bondesil-CI8/OH was used. With this sorbent, both apolar and polar interactions occur, i.e., via the alkyl (C,,) chains and hydroxyl (OH) groups, respectively. This is a distinct advantage when, e.g., very polar carbamates without an aromatic ring have to be trapped.". In this instance, break- through studies were performed with spiked surface water and using PB-MS detection. It was found that the capacity of the C1 ,/OH precolumn could be increased without creating addi- tional peak broadening by using a 3 mm instead of 2 mm id precolumn.Since the analyte recoveries and/or breakthrough volumes increased compared with the polymeric phases, Bonded C1 s/OH was selected for further studies. Typical breakthrough curves3 1 for analytes with no breakthrough up to 100 ml (carbofuran and promecarb) and with breakthrough after about 20 ml (ethiofencarb sulfoxide) or about 50 ml (methio- carb sulfoxide) are shown in Fig. 3. The results for all 43 analytes can be summarized as follows. With PB-MS detection, no signals could be recorded for ten compounds (LOD > 20 pg 1-1). Of the remaining 33 analytes, eight had breakthrough volumes of up to 20 ml, a further three gave volumes of up to 50 ml and 22 did not show appreciable breakthrough even with 100 ml sample volumes (Table 2).Except for dioxacarb and methiocarb sulfoxide, all carbamates with a breakthrough of less than 100 ml are Class 2 compounds, i.e., similar behaviour can be expected also for representatives of this class not studied in this work. In order to provide the best working conditions with regard to analyte detectability, 100 ml samples were used in subsequent experiments. LC Gradient Elution In order to assess the effect of the eluent composition on system performance, methanol and acetonitrile were used as organic Class 1 Class 2 Class 3 Fig. 2 General structures of carbamates used in the study: N- methyl aryl carbamates (Class I), N-methyl oxime carbamates (Class 2), N-substituted aryl ester carbamates (Class 3). Miscellaneous compounds are denoted as Class 4.1330 Analyst, September 1996, Vol.121 modifiers. For most analytes, the results with both organic solvents were essentially the same. However, about a two-fold increased response with methanol was found for some analytes, e.g., aldicarb sulfone, oxamyl, ethiofencarb sulfoxide, methio- carb sulfoxide, butocarboxim and isoprocarb. The effect can be partly due to a shift in retention times: with the methanolic gradient the analytes are eluted with, and thus enter the PB interface in the presence of, a higher percentage of organic modifier, which is known to enhance the PB transport efficiency. 13 Notable peak broadening was observed for the degradation products ethiofencarb sulfoxide and methiocarb sulfoxide. Such an effect will markedly lower their detectability in LC as compared with FI runs.In all experiments, 0.1 moll-' ammonium acetate was added to the aqueous phase component because of the reported beneficial effect at low analyte concentration levelslx and in order to keep the method consistent with others developed in our laboratory.20 Actually, FI-PB-MS using the LC eluent with and without ammonium acetate as the carrier stream showed slight differences only. The signal intensities improved 1 S-2-fold for 16 carbamates (representatives of all four classes) while I0 compounds showed a 1.5-2-fold lower response. Because the separation of all analytes in one run was not our ultimate goal, a 45 min linear gradient was deemed satisfactory. The retention time data in Table 2 show that it creates a rough over-all separation.This approach is justified because general experience with environmental water analyses shows that it is highly unlikely that more than a few contaminants are simultaneously present at close to law-offending levels, that is, the probability of co-elution is extremely low. Analyte Detectability In the first part of this study," S/N values in FI-PB-MS were reported for most of the present test analytes. These data are included in Table 2 (500 ng injections for all but three analytes). A detailed discussion of the widely varying analyte responses and the relatively high responses invariably found for the degradation products (in both EI and CI experiments) can be found in the paper cited. Here, a note of warning should be added: translation of detection limit data from FI to an SPE-LC set-up is dangerous because of, e .g . , low breakthrough volumes, effects of LC eluent composition and exponential response curves32 and peak distortion or spectral interferences due to the sample matrix. In unfavourable situations, it can reduce the responses up to a 100-fold (Table 2). In order to obtain a first indication of the over-all perform- ance, each compound was analysed at the 20 pg 1-1 level (100 ml sample) in the full-scan EI mode. As already briefly 0 20 40 60 80 100 Sample volume/ml Fig. 3 Plot of chromatographic peak area (arbitrary units) versus sample volume ( n = 3) for trace enrichment of river Meuse water spiked with 20 1.1.6 1-1 of A, ethiofencarb sulfoxide; B, methiocarb sulfoxide; C, carbofuran; and D, promecarb on Bondesil-C18/OH.Data were obtained with SPE-LC- PB-MS. For details, see text. indicated above, even at this high level 10 analytes could not be detected at all. In some instances, e.g., with asulam and carbendazim, this can be attributed to the low breakthrough ( c f , high S/N values in FI-PB-MS), whereas in other instances, e.g., with propham and chlorpropham, the low response in PB-MS without doubt is the major cause. Next, 12 carbamates and five degradation products (see Fig. 4) were selected for the determination of repeatability and the method detection limit. Calibration curves were constructed from five data points in the range 0.01-10 pg 1-1 ( n = 2). They were of an exponential form (see the typical result for carbaryl in Fig. 5), except for cloethocarb and 3-ketocarbofuran, which gave linear plots with linear regression coefficients of 0.998 and 0.997, respectively.It should be mentioned that the second-order regression curves give satisfactory quantification results provided that there are at least five calibration points. However, a calibration method of co-eluting isotopically labelled internal standards is suggested to be most reliable for real environmental samples.15 The method detection limits for the 17 compounds were 0.1-8 pg 1-1 (Table 2). Six additional runs with samples spiked at the 10 pg 1-I level were performed within a 3 d period to obtain repeatability data. RSDs of the chromatographic peak areas from ion chromatograms of the base peak in each spectrum typically ranged from 5% (oxamyl) to 20% (promecarb). Only methiocarb sulfoxide and ethiofencarb sulfoxide showed less satisfactory results (RSD, 33% and 35%, respectively).How- ever, this is to be expected for such high-polarity analytes for which a sample volume much above the optimum was used (cf., Fig. 3). The RSDs of retention times were 0.05-0.6% (n = 18 in a 2 month period). A typical SPE-LC-PB-MS full-scan EI trace of river Meuse water spiked with about a 5 pg 1--' mixture of the 17 test analytes is shown in Fig. 4. Detection limits of compounds not included in this mixture (marked with the symbol 11 in Table 2) were estimated from analyses of river Meuse water spiked at the 20 pg I-' level using the exponential slope of the calibration curve of carbaryl (cJ, Fig. 5). Note that the detection limits can easily be improved 5-1 0-fold when using time-scheduled selected-ion monitoring (TS-SIM). For example, with this option 0.01 pg 1- of carbaryl, ethiofencarb, desmedipham and phenmedipham could be detected in surface water.However, the use of SIM limits PB-MS to target compound analysis (whilst providing little structural information); therefore, this mode of operition was not studied in any detail. Comparison of LC-PB-MS and GC-MS It was shown recently that analyte detectability in GC-MS is 2-3 orders of magnitude better than in LC-PB-MS when the same ion source (MS Engine) and similar MS conditions are used.I7J3 As a typical example, 100 ng of atrazine produced a 170-fold higher signal in GC-MS than in PB-MS.17 Obviously, the use of GC-MS for the separation and detection even of the rather polar and often thermolabile carbamates merits attention, especially since new sample introduction technologies have become available,22 which reduce thermal degradation in the injection port and allow injections of up to several hundred microlitres into a GC instrument.Using a PTV injector, a short analytical column and low gradient temperatures compounds such as aldicarb, its degradation products, oxamyl and metho- myl, considered to be typical LC-type analytes can now be readily determined by GC-MS.2' As an extreme example, aldicarb, its derived oxime and nitrile were determined by full- scan GC-EI-MS with limits of detection of 0.3, 1.2 and 0.15 ng, respectively.23 In this study, a 12 m analytical column and an on-column injector were used for the determination of 2 ng of each analyte in our test set.Surprisingly, 28 analytes (Table 2) eluted with good peak shapes and could be detected despite the high final temperature of the gradient (300 "C) and GC-MSAnalyst, September 1996, Vol. 121 1331 interface (3 10 "C). Only two degradation products (3-keto- carbofuran and 1 -naphthol) could be detected, which is not surprising in view of their relatively high polarity. The same is probably the reason why the carbamates aldicarb, asulam, oxamyl and carbendazim could not be detected, whereas phenmedipham and desmedipham are presumably thermally degraded. None of the EI mass spectra of all 28 'positive' compounds showed any ions that could be assigned to thermal degradation.They were nearly identical with our PB-MS traces and the library ~pectra,3~ as is shown for carbaryl in Fig. 6. The detection limits included in Table 2 were calculated from 2 ng injections. Since this value was close to the detection limits in most instances, it can be considered a good estimate. In contrast to FI-PB-MS, the analyte responses varied over a much smaller range, viz., from 0.05 (propoxur) to 2 ng (dioxacarb and, mexacarbate). Admittedly, these gratifying results relate to base peak ion chromatograms rather than to full-scan acquisition, and to standard injections rather than to the analysis of real samples. Still, when applying already known procedures for SPE coupled on-line to GC-MS for 10 ml samples,35 analyte detection limits (in concentration units in the actual sample) can be expected to be of the order of 0.02-0.5 pg 1-1.However, the pursuit of this approach was outside of the scope of this study. Finally, the mixture of 17 compounds selected for SPE-LC- PB-MS studies was also analysed by means of GC-MS. As was to be expected on the basis of the results presented above, only 11 compounds showed peaks (Fig. 7). A comparison of Figs. 4 (5 ng ml-I, 100 ml) and 7 (2 ng injection) again highlights the much better analyte detectability that can be achieved for the Table 2 Retention times, detection limits (S/N = 3) and relative responses of 32 carbamates and their 11 degradation products obtained by SPE-LC-PB-MS, GC-MS and FI-PB-MS* No. Curbamutes- 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 32 1 2 3 4 5 6 7 8 9 10 11 Degradation products- Compound Aldicarb Aminocarb Asulam Barban Bendiocarb Bromocarbamate B ufencarb Butocarboxim Carbanolate Carbaryl Carbendazim Carbetamide Carbofuran Chlorpropham Cloethocarb Desmedipham Dioxacarb Ethiofencarb Fenobucarb Isoprocarb Methiocarb Methomyl Mexacarbate Oxamyl Phenmedipham Pirimicarb Promecarb Propham Propoxur Thiofanox Tri allate Trimethacarb Adicarb sulfone Aldicarb sulfoxide Butocarboxim sulfoxide Ethiofencarb sulfone Ethiofencarb sulfoxide 3 -Hydroxycarbofuran 3-Ketocarbofuran Methiocarb sulfone Methiocarb sulfoxide Thiofanox sulfone Thiofanox sulfoxide Class 2 1 3 3 1 1 1 2 1 1 3 3 1 3 1 3 I 1 1 1 1 2 1 2 3 4 1 3 1 2 4 1 2 2 2 1 1 1 1 1 1 2 2 SPE-LC-PB -MS GC-MS tr/ LODi/ tr/ LOD*/ min pgl-1 min ng -n >20 - - - >20 12.7 1.2 38.5 0.7 16.5 1.2 30.5 1011 11.6 0.6 38.6 1011 13.6 2 41.7 20 12.8 0.6 27.2 1511 10.7 0.9 34.8 0.7 12.5 0.2 32.1 0.1 13.9 0.1 29.0 0.7 15.0 0.6 30.6 0.6 12.3 0.2 - >20 11.6 0.4 29.2 0.5 13.7 0.9 35.1 0.1 - 22.8 1511 13.6 3 33.1 0.1 13.5 0.2 37.0 1011 13.2 2 34.3 1511 10.2 0.1 37.8 0.6 14.5 0.2 - >20 10.4 3 - >20 13.3 0.8 14.4 1011 - - 35.7 0.1 - 33.5 1511 13.6 0.2 38.3 1 12.0 0.1 - > 20 9.3 0.6 30.4 0.6 11.0 0.05 33.6 20 11.7 2 - >20 13.5 0.1 - >20 11.6 0.2 >20 - - - - >20 - - - - 14.2 12.9 12.2 19.7 20.3 22.5 26.6 23.5 21.3 39.4 37.1 - 1511 - 1511 - 2011 - 1511 - 1011 - - - - 5 - - 0.8 13.2 0.6 5 - - 5 - - 8 - - - - 1511 - FI-PB - MS (S/N,§ 500 ng) 5 75 375 500 50 100 15 3 35 750 350 1025 200 450 675 700 200 15 2 200 10 35 250 1075 225 3 35 2 50 725 300 1075 2675 2575 1875 450 1825 825 3450 325 - - - * Data obtained in full-scan El mode.t LOD, detection limit in river Meuse water (six-point calibration curve in the range 0.01-10 pg I-I), enriched sample volume 100 ml. Calculated from 2 ng injection. W/N, signal-to-noise ratio of ion chromatogram of base peak in analyte spectrum, n = 3. 1 Dashes indicate not detected at maximum analysed concentration/amount (SPE-LC-PB-MS, 20 pg 1-1; GC-MS, 2 ng; FI-PB-MS, 500 ng). 11 Calculated from SPE- LC runs obtained with 20 pg 1-1 of each analyte.1332 Analyst, September 1996, Vol. 121 GC-amenable compounds. This is even true for carbamates such as carbaryl and ethiofencarb, which show good responses in LC-PB-MS.Automation and Robustness Our earlier SPE-LC procedures were automated by combining them with commercially available cartridge exchange-sample preparation units, e.g., the OSP-2A (Merck, Darmstadt, Ger- many)33 or Prospekt (Spark Holland, Emmen, The Nether- lands).5.17 A less expensive means of automation was used in this study. Here, a universal valve-switching module and the sample and methanol delivery pumps were controlled by electronic event signals from the HP 1090. All actions of the automated system are governed by a timetable that is pro- grammed into the HP 1090 ($, Fig. 1 and Table 1). With this approach, the analytical part of the method is reduced to (optional) adjusting the sample pH, inserting the Pump 2 inlet tubing into the sample container and pressing the start/enter button of the HP 1090.A disadvantage is that individual parts of the set-up are not synchronized by means of a computer program and that the presence of an operator is therefore still required. A maximum of three analyses were performed in sequence without closing the PB-MS inlet valve because of an observed decrease in the MS sensitivity after 5-6 h of continuous operation. During 1 year of operation, only the PB desolvation chamber heating unit had to be exchanged because of malfunctioning. In the method development stage, the PB vacuum pumps and manifold foreline pump had to be exchanged once because of MS vacuum problems. During routine use of the automated method, the PB interface was replaced twice by a thermospray or electrospray interface for other research purposes, but the quality of the spectra could easily be reproduced after Abundance 400001 1 300004 1 20000 remounting the PB unit.At any change of the analytical conditions, e.g., gradient composition or pH of LC eluent or temperature of the ion source, a new spectrum was entered into a laboratory-made spectral library. Comparing spectra recorded after 8 months with the original entries showed good spectrum reproducibility: variations in the relative abundances of five major ions in 70 eV EI spectra typically were less than lo%, even when comparing GC-MS and PB-MS data (see Fig. 6) or the long-term performance of negative-ion CI spectra.36 Because of the well known problems with quantification in PB- MS, the system was checked daily by injecting 500 ng of diuron.As an example, in the period from October 1993 to May 1994 the diuron FI response varied about four-fold with RSDs of the mean value of over 80% (n = 32). However, acceptable results could be obtained if several injections were made each day. The RSD values then were 20-25% (n = 15-30) within 1 week and 5-10% (n = 3-6) within 1 day. Hence one may conclude that the system is rugged and the method robust. Conclusions A test set of 32 carbamates and 11 of their degradation products was studied by automated SPE-LC-PB-MS and on-column injection GC-MS. In SPE-LC-PB-MS the analytes were enriched from 100 ml surface water samples on a Bondesil-Clg/ OH precolumn, eluted on-line with methanol-ammonium acetate and separated by gradient LC on a CIS analytical column and detected by PB-MS in the full-scan EI mode.'Cost- effective' automation was achieved by means of a gradient controller, with two additional LC pumps and an automatic six- port switching valve. This set-up allowed the determination of 32 analytes in surface water at or below 20 vg 1-I. The detection limits of 17 selected compounds (12 carbamates and five degradation products) were 0.1-8 pg 1- I . As regards GC-MS, 15 12 I 'r" 10 18 10000 0 15 20 25 30 Time/min 9 16 25 40 Fig. 4 Ion chromatograms (m/z of base peak in each spectrum'9) of the mixture of 17 carbamates obtained by automated on-line SPE-LC-PB-MS analysis in full-scan EI mode. A sample of 100 ml of river Meuse water was spiked with 5 pg 1-' of each pesticide and enriched on the 10 X 3.0 mm id precolumn packed with Bondesil-C,*/OH.Analytical column, 250 X 4.6 mm id Supelco-LC-18-DB. For numbering of compounds, see Table 2. For other conditions, see text.Analyst, September 1996, Vol. 121 1333 G9 an unexpectedly high number of 28 test compounds, mainly parent carbamates, could be determined with detection limits of 0.05-2 ng (ion chromatograms of base peaks). The spectra obtained were fully comparable with those from PB-MS and standard reference MS libraries. LC-PB-MS and GC-MS were found to be complementary. As an example, the three carbarnates not detectable by means of PB-MS, propham, chlorpropham and triallate, all give good responses in GC-MS. On the other hand, desmedipham, phenmedipham and most of the degradation products do not show up in GC analysis, and have to be subjected to LC-PB-MS.All of the above demonstrates that PB-MS, with its still unique advantage of providing classical EI spectra, has a much wider application range than is often thought. On the other hand, analyte detectability as required for modern trace-level analysis is often not up to the required standard even when 100 ml SPE 201 220 270 286 333 \ ' .... ,Ill ,.,*,.,, ,A,, .I,* ,,,,, , " , I I ,,. , II ...... .,.It.. Fig. 5 Calibration curves for A, 3-ketocarbofuran; and B, carbaryl obtained by SPE-LC-PB-MS. ...... ,- ...... ..-.. . ,.......-_.,... ............................................................. "__..I ................ .........*........ Abundance 10000 57 89 ......10000 5ooo] 0 201 229 261 290 344 I ......--... ....... "... ..I# ..... . I .L ......I..... .I... ... hr ...... ..: .............. ...L .................... -.-.* .... .- <*. procedures are incorporated. Here, the main area of research should be the improvement of the design and, thus, the analyte transfer efficiency of the PB interface. 13,37 Preliminary results of on-going research38 suggest that considerable improvements can be achieved in this area. - - - - - 5000 - 0- Abundance 100000 144 115 / 89 20 1 57 / ' I I I l d . I I 1 1 I 1 1 I I 1 I I I I 1 1 I 80000 3 29 4 _I 60000- 40000- 20000- 18 1 13 i 151 21 0 8 10 12 14 16 18 Ti me/m i n Fig. 7 GC-MS ion traces (mlz of base peak in each spectrum19) of the mixture of 17 carbamates (2 ng each). MS operated in full-scan El mode.Analytical column, HP-1, 12 m X 0.2 mm X 0.33 pm cross-linked silica gum. GC-MS interface temperature, 3 10 "C; and ion source temperature, 250 "C. For numbering ofcompounds, see Table 2. For other conditions, see text. lll5 1144 ' ' ' 1 / ....... -.. .... .....-_ ..... I ................................... ......_ .-.. ............. .."__ .. I.. .............. ..- ......... ......... 115 (144 #123816: 1 -Naphthalend, methylcarbamate Mass/C harge Fig. 6 Positive-ion EI mass spectra of carbaryl obtained by (a) SPE-LC-PB-MS, (b) GC-MS and ( c ) from Wiley EI spectral library. (a) 1 pg 1-I of carbaryl in river Meuse water; enriched volume 100 ml; scan range 65-350 u; ion source temperature, 200 "C. (h) 2 ng of carbaryl injected on-column; scan range 45-350 u; ion source ternperaturc, 250 "C.1334 Analyst, September 1996, Vol.121 Finally, the present results and those in another recent paper from our group17 demonstrate the practicality and usefulness of the multi-analysis system which combines SPE-LC-DAD-UV, SPE-LC-PB-MS and SPE-GC-MS in a fully integrated unit. In such a system the fully automated so-called SAMOS SPE-LC- DAD-UV system, which is already in position at a number of monitoring stations in Europe, provides early warning informa- tion (UV spectra and retention times) about compounds present in surface water at (or close to) offensive levels. Suspect samples can then be subjected to LC-MS and/or GC-MS analysis for confirmation or provisional identification of suspected pollutants and/or unknowns. J.Slobodnik thanks the Environmental Commission of the European Union for awarding a grant (Supplementary agree- ment N '1: CIPDCT925102 to the contract N ': EV5VCT920 105). The Rhine Basin Program (Amsterdam/ Waldbronn) is acknowledged for the use of instrumentation. Dipl. Ing. Th. Hankemeier is thanked for help with the GC experiments and J. Brands (Supelco, Bornem, Belgium) for providing LC analytical columns. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 McGarvey, B. D., J . Chromatogr., 1993, 642, 89. Kalinoski, H. T., Wright, B. W., and Smith, R. D., Biomed. Environ. Mass Spectrom., 1986, 13, 33. Berry, A. J., Games, D. E., Mylchreest, I. C., Perkins, J. R., and Pleasance, S., Biomed. Environ. Mass Spectrom., 1988, 15, 105. Honing, M., Barcelo, D., van Baar, B.L. M., Ghijsen, R. T., and Brinkman, U. A. Th., in Quality Assurance for Environmental Analysis within the BCR Programme, ed. Quevauvillier, Ph., Meier, E. A., and Griepink, B., Elsevier, Amsterdam, The Netherlands, 1995, Slobodnik, J., Groenewegen, M. G. M., Brouwer, E. R., Lingeman, H., and Brinkman, U. A., Th., J . Chvomarogr., 1993, 642, 259. Krause, R. T., J . Chromatogr., 1988, 442, 333. de Kok, A., Hiemstra, M., and Brinkman, U. A. Th., J . Chromutogr., 1992,623, 265. Graves, R. L., Measurement of N-Methylcarbamoyloximes and N- Methylcarhamates in Water by Direct Aqueous Injection HPLC with Post-column Derivatizatioiz, Revision 3 .O, Environmental Monitoring Systems US EPA, Method 53 1.1, Laboratory Office of Research and Development, Cincinnati, OH, 1989.Chiron, S., and Barcel6, D., J . Chromatogr., 1993, 645, 125. Niessen, W. M. A., and van der Greef, J., Liquid Chromatography- Mass Spectrometry, Principles and Applications, Marcel Dekker, New York, 1992. Slobodnik, J.. van Baar, B. L. M., and Brinkman, U. A. Th., J . Chromatogr. A, 1995, 703, 81. Pleasance, S., Anacleto, J . F., Bailey, M. R. and North, D. H., J . Am. Soc. Mass Spectrom., 1992, 3, 378. Creaser, C. S., and Stygall, J. W., Analyst, 1993, 118, 1467. Apffel, A., and Perry, M. L., J . Chromatogr., 1991, 554, 102. pp. 535-562. 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 Ho, J. S., Behymer, T. D., Budde, W. L., and Bellar, T. A., J . Am. Soc. Mass Spectrom., 1992, 3, 662. Bagheri, H., Slobodnik, J., Marce Recasens, R.M., Ghijsen, R. T., and Brinkman, U. A. Th., Chromatographia, 1993, 37, 159. Slobodnik, J., Hogenboom, A. C., Louter, A. J. H., and Brinkman, U. A. Th., J . Chromatogr. A, in the press. Bellar, T. A., Behymer, T. D., and Budde, W. L., J . Am. Soc. Mass Spectrom., 1990, 1, 92. Slobodnik, J., Jager, M. E., Hoekstra-Oussoren, S. J. F., Honing, M., van Baar, B. L. M., and Brinkman, U. A. Th., J . Mass Spectrom., submitted for publication. Brinkman, U. A. Th., Slobodnik, J., and Vreuls, J. J., Trends Anal. Chem., 1994,13, 373. Wigfield, Y. Y., Grant, R., and Snider, N., J . Chromatogr., 1993,657, 219. Mol, H. G. J., Janssen, H.-G. M., Cramers, C. A., Vreuls, J. J., and Brinkman, U. A. Th., .I. Chromatogr. A, 1995, 703, 277. Trehy, M. L., Yost, R. A., and McCreary, J. J., Anal. Chem., 1984,56, 1281. Honing, M., Barcel6, D., Jager, M. E., Slobodnik, J., van Baar, B. L. M., and Brinkman, U. A. Th., J . Chromatogr. A, 1995, 712, 21. Voyksner, R. D., Bursey, J. T., and Pellizari, E. D., J . Chromatogr., 1984,312, 221. Stamp, J. J., Siegmund, E. G., Cairns, T., and Chan, K. K., Anal. Chem., 1986, 58, 873. EEC Drinking Water Guideline, 80/779/EEC, EEC No. L229/ 1 1-29, EEC, Brussels, 1980. Water in The Netherlands: a Time for Action, National Policy Document on Water Management, House of Commons, Meeting Period 1988-89, The Hague, No. 21250 (1-2). US EPA, National Survey of Pesticides in Drinking Water Wells, Phase I1 Report, EPA 570/9-9 1-02 I, National Technical Information Service, Springfield, VA, 1992. Hiemstra, M., and de Kok, A., J . Chromatogr. A , 1994. 667, 155. Hennion, M.-C., and Coquart, V., J . Chromatogr., 1993, 642, 21 1. Miles, C. J., Doerge, D. R., and Bajic, S., Arch. Environ. Contam. Toxicol., 1992, 22, 247. Slobodnik, J., Hoekstra-Oussoren, S. J. F., and Brinkman, U. A. Th., Anal. Methods Instrum., in the press. Wiley Standard Mass Spectral Library, Copyright 1996 by John Wiley & Sons, licensed to Hewlett-Packard, HP 59943B, Rev. A 00.00, 130 000 entries. Louter, A. J. H., Ghijsen, R. T., and Brinkman, U. A. Th., J . Microcol. Sep., 1993, 5 , 303. Ong, V. S., and Hites, R. A., J. Am. SOC. Mass Spectrom., 1993, 4, 270. Cappiello, A., Famiglini, G., and Bruner, F., Anal. Chem., 1994, 66, 1416. Kientz, C. E., Hulst, A. G., de Jong, A. L., and Wils, E. R. J., Anal. Chem., 1996, 68, 675. Paper 61021 97B Received March 28,1996 Accepted May 13, 1996
ISSN:0003-2654
DOI:10.1039/AN9962101327
出版商:RSC
年代:1996
数据来源: RSC
|
44. |
Development of a gas chromatographic method for the simultaneous determination of trace amounts of ethylenediaminetetraacetic acid and diethylenetriaminepentaacetic acid in natural waters |
|
Analyst,
Volume 121,
Issue 9,
1996,
Page 1335-1339
Jaana Sorvari,
Preview
|
PDF (808KB)
|
|
摘要:
Analyst, September I996, Vol. I21 (1335-1339) 1335 Development of a Gas Chromatographic Method for the Simultaneous Determination of Trace Amounts of Ethylenediaminetetraacetic Acid and Diethylenetriaminepentaacetic Acid in Natural Waters Jaana Sorvari, Mika Sillanpaa and Marja-Liisa Sihvonen Laboratory of Inorganic and Analytical Chemistry, Helsinki University of Technology, FIN-02150 Espoo, Finland A GC method for the simultaneous determination of ethylenediaminetetraacetic acid (EDTA) and diethylenetriaminepentaacetic acid (DTPA) in natural waters is described. The method was studied with spiked samples of distilled water, sea-water (2 matrices) and humic-containing lake water. An acidified 5-10 ml sample was evaporated and esterified with C2H50H, containing H2SO4 and CH3COOH catalysts and CH3(CH2)&N as an internal standard.The esterified sample was extracted with CsHSCH3, neutralized with KHC03, dried with anhydrous Na2S04 and analysed by GC with nitrogen-phosphorus selective detection. The best recoveries were obtained by the esterification in 100 "C for 3 h. The tetraethyl ester of EDTA and pentaethyl ester of DTPA were further identified by mass spectrometry. In the sea- and lake water matrices, the recoveries of EDTA and DTPA were SO-110%. High concentrations (above 50 rng 1-1) of iron (Fe3+) interfered considerably in the determination of DTPA. For EDTA the interference was insignificant with iron concentrations below 500 mg 1-1. Keywords: Ethylenediuminetetraacetic acid; diethylenetriaminepentaacetic acid; gas chromatography; natural waters Introduction Conventional pulp bleaching with elemental chlorine or chlorine compounds produces poorly biodegradable chlorinated compounds, some of which have toxic and also carcinogenic and mutagenic effects in living organisms.132 Therefore, processes are widely being replaced by methods using non- chlorine-containing oxidants such as hydrogen peroxide, ozone and peroxy acids. In these methods, complexing agents are used to deactivate the metals involved in the process. These metals, when in cationic form, catalyse the decomposition of the bleach, causing a high consumption of the chemical and poor fibre quality. Ethylenediaminetetraacetic acid (EDTA) and diethyle- netriaminepentaacetic acid (DTPA) are the complexing agents most commonly used for deactivation.In addition to use in wood processing, these chemicals are widely used in household washing powders, soaps and cosmetics and also in herbicides and in the preparation of photographs. Industrial applications include metal finishing and plating and textile, leather, rubber and polymer production. As free EDTA and DTPA complexes and their metal chelates are very stable and have been found to be poorly biodegradable in activated sludge treatment,3-5 increasing use will obviously result in increasing concentrations in receiving waters. The available toxicity data, at least regarding the long-term toxicity of EDTA and DTPA, are insufficient. It is most likely that EDTA and DTPA have secondary impacts in receiving waters. For trace analyses for free and chelated EDTA and DTPA, a GC method with prior derivatization is used.In the literature several derivatization methods have been described, including ~ i l y l a t i o n ~ , ~ and esterification methods. Reported methods for the esterification of EDTA are based on the formation of methyl esters,K-l7 propyl esters4,14,1* and butyl esters.19-21 A method for analysis as an ethyl ester has also been presented.22 The simultaneous detection of EDTA and DTPA has been described in only a few papers.4.13,ls Some of the methods have been reported to be inappropriate for the detection of DTPA. In this work, different esterification methods were examined for the simultaneous determination of EDTA and DTPA and a simplified method for their determination as ethyl ester derivatives was developed. Experimental Gas Chromatography A Hewlett-Packard (Avondale, PA, USA) Model 5890 Series IT Plus gas chromatograph equipped with a nitrogen-phosphorus detector (NPD) and an HP 7673 automatic liquid sampler with a 10 pl syringe was used for the analysis.The column was an HP-5 capillary column (30 m X 0.25 mm i.d.) with a 0.25 pm film thickness. Nitrogen at a constant flow rate 3.5 ml min-1 was used as the carrier gas. The column oven temperature programme was as follows: initial temperature 100 "C (held for 1 min), increased at 60 "C min-I to 200 "C (held for 5.0 min), increased at 50 "C min-1 to 250 "C (held for 8.5 min). The injector and detector temperatures were 300 "C. Injection of a 1 pl sample was performed in the splitless mode with the inlet on-time of 0.50 min.The detector gas pressures were air 300 kPa and hydrogen 130 kPa. The chromatographic data were processed with an HP 336SA ChemStation. A Hewlett-Packard Model 5860A gas chromatograph com- bined with an HP 5971A mass-selective detector (MS) was used to identify the esters of EDTA and DTPA. The GC column, temperature programme and column flow rate were the same as above. The system was equipped with a capillary inlet system and a Model 7673A high-speed automatic liquid sampler with a 10 pl syringe. The conditions were as follows: carrier gas, helium at 3.5 ml min-l; detector gases, hydrogen at 40 ml min-l, air at 450 ml min-1 and make-up gas helium at 30 ml min-l; septum purge flow rate 1-2 ml min-1; splitting ratio 1 : 20; and column inlet pressure 150 kPa.The mass spectra were recorded at an electron energy of 70 eV and a trap current of 300 pA. The ion source temperature was 180 "C and the molecular separator temperature 135 "C. The GC-MS data were processed with an HP G1030A MS ChemStation.1336 Analyst, September 1996, Vol. 121 Reagents and Solutions about 25 ml of 1 mol 1-1 KHCO? by flushing the sample vial Dried CH30H, C4H90H, KHC03, anhydrous Na2S04, concen- trated HCl, (CzH&N and concentrated H2S04 were obtained from Merck (Darmstadt, Germany), CH3COC1 and C3H70H from BDH (Poole, Dorset, UK), CH3COCH3, concentrated CH3COOH and HCOOH and FeC13.6H20 Fixanal solution from Riedel-de Haen (Hannover, Germany), CaNa3-DTPA and the internal standard ClhH&N from Fluka (Buchs, Switzer- land), Na4EDTA from Sigma (St. Louis, MO, USA), CZHSOH from Primalco (Rajamaki, Finland) and C~HSCH~, CsH12 and C6H14 from LabScan (Dublin, Ireland).All chemicals except C3H70H (HPLC grade), C16H33CN (GC grade) and CZHSOH (AaS grade) were of analytical-reagent grade. Standard solutions A stock standard solution that contained 1000 mg 1-1 each of EDTA and DTPA was prepared by dissolving 0.2737 g of EDTA and 0.2591 g of DTPA in 250 ml of distilled water. The solution was stored in the dark at room temperature for not more than 2 months. An internal stock standard solution of 1000 mg 1-1 was prepared by dissolving 100 g of C16H&N in 100 ml of C2HSOH. This solution was stored in a refrigerator not more than 1 month. FeC13.6H20 Fixanal solution was used to prepare a stock standard solution containing 5.0 g 1-I of Fe.Ester-ification reagent The esterification reagent was prepared in a 50 ml calibrated flask by mixing 2.5 ml of H2S04, 25 pl of CH3COOH and about 30 ml of C2HsOH, adding 2.5 ml of 10 mg 1- internal standard solution and diluting to volume with C2H50H. In some of the esterification studies, different amounts of the internal standard were used. Samples Samples were prepared in four different matrices: distilled water, lake water and sea-water.* The lake water was from Lake Lammaslampi (Vantaa, Finland), situated in the middle of a densely populated area. This matrix had an extremely high humic content (KMn04 number -80 mg 1-I), which was assumed to give the maximum interference in EDTA and DTPA determinations.The iron content of this lake water is on average about 0.8 mg 1-1, with an annual maximum value of 2.7 mg 1-l. The sea-water samples were from Tvarminne (Gulf of Finland) and the Gulf of Riga. Tvarminne is situated in a pristine area where the salinity of the sea-water is about 5%0. In contrast, the Gulf of Riga is influenced by heavy industry and colonization and the salinity of this sea-water sample was 5.7%~~. All the matrices were spiked with different amounts of EDTA and DTPA. Experiments with the blanks showed no detectable EDTA or DTPA. Sample Preparation The matrix solution was acidified with HCOOH to give a pH value of approximately 2.5. This pH value was found to be optimum for the analysis (see Preliminary Tests). Then 5 ml of the acidified sample in a 20 ml glass sample vial were spiked with EDTA and DTPA to give the desired concentration.After spiking, the vial was shaken vigorously. The spiked sample was dried in an oven at 115 f 5 "C. After cooling to room temperature, 1.5 ml of esterification reagent was added and esterification was executed in the vial sealed with a screw-cap. After esterification, the sample was cooled again, 1.5 ml of CGHSCH3 added and extract transferred into a 50 ml separating funnel. The extract was further mixed with with approximately 5 ml portions of ihis solution. When all the KHC03 solution had been added, the separating funnel was shaken for about 1 min, or long enough to obtain a clear solvent layer. The solvent layer was separated and dried over Na2S04. The dried sample was transferred into a GC sample vial and analysed by GC-NPD. To test the effect of iron, the desired amount of Fe"' was added to EDTA- and DTPA-spiked samples.The samples were shaken and allowed to reach equilibrium for about I h before the addition of the esterification reagent. According to the study of Bergers and Groot,*3 no significant difference in the amount of Ferl' complex was observed if the stabilization time was changed from 1 to 24 h. Treatment of Results To compare the different esterification methods and the influence of the esterification conditions, the area ratio of the individual chromatographic peak of EDTA (or DTPA) and that of the internal standard was counted. In all the tables, this 'relative peak area' is used to compare the recoveries.For a certain concentration, the variation in the amount of the internal standard in separate studies caused unequal values of the calculated area ratio. Recoveries (%) from the natural water matrices were calculated by comparing the relative peak areas of spiked sea- and lake-water samples with the respective relative peak areas of spiked distilled water samples. The detection limits were determined with the criterion of S/N = 3. Results and Discussion Preliminary Tests Preconcentration of the sample In the sample preparation described above, in addition to the esterification, evaporation is a time-consuming stage. This stage is necessary to achieve low detection limits in the analysis based on esterification. When larger amounts of sample need to be treated, a rotary evaporator with a vacuum system can be used, as reported in many earlier methods for EDTA determination.Filtration and ion exchange with a modified anion-exchange resin has been used to remove the interfering components. In this study, small sample volumes were used and drying simplified by using an oven, instead of a rotary evaporator and a nitrogen flow. Neither filtration nor ion exchange was adopted. Esterifi cation methods Different esterification methods were examined. Silylation was not adopted, owing to the risk of contamination of the NPD by the remaining reagent. In addition, some difficulties have been reported in determination of EDTA when silylation is used.7 Distilled water was used as a sample matrix with a sample size of 5 ml.Before drying, the pH was adjusted with HCOOH. The methods examined were esterification with CH3OH, CzHSOH, C3H70H and C4H90H using different catalysts (CH3COC1, H2S04 and CH3COOH). Also, different esterifica- tion times (from 0.5 to 3 h), temperatures (from 75 to 100 "C), extraction solvents [C6H5CH3, CsH12 and C6H14] and solutions for neutralization (KHC03 and distilled water) were examined. Halogenated solutes were not used owing to the possible difficulties with baseline fluctuations in the NPD; this would not have been a problem if a GC-MS system were used. Also, the effect of sample pH on the recoveries of different EDTA and DTPA esters was investigated at four pH values (2.5, 3, 4 and 6). For all of the methods, the best recoveries were obtained with the lowest pH values of 2.5 and 3, where EDTA and DTPAAnalyst, September 1996, Vol.121 1337 should be present predominantly in their acidic forms, H3A- and HdA-. The preparation of the esterification reagent containing CHTOH and CH3COCl proved to be difficult as the reaction between these chemicals is very exothermic. In addition, this method did not give repeatable results and the recoveries were obviously much lower than with other methods examined. This is in accordance with the results of Cassidy et al.,' who also reported low recoveries of EDTA methyl esters. When esterified with CH3(CH2)20H, the response of DTPA varied considerably in separate runs. Some unidentified peaks were also present in the chromatograms. These peaks were probably caused by deterioration of the internal standard and incomplete esterification of EDTA and DTPA.The esterification with CH3(CH2)30H containing CH3COCl gave reasonable results, but the recoveries varied, especially with DTPA. Butyl as well as propyl esters have high molecular masses, which prolong the analysis time. Finally, CH3CH20H was selected for esterification. This method worked with simple sample preparation and produced lighter derivatives, permitting faster analysis. Also, the recov- eries were good and no interferences were observed. Studies of the Ethyl Esterification Method Stability of retention times The retention times were about 9.2 min for EDTA and 16.2 min for DTPA (Fig. 1). In natural waters, complexing agents interact with different metal ions and humic and fulvic acids, and also with sediments and organisms.Thus, at low concentration levels a stable retention time is essential for the correct identification of EDTA and DTPA. The RSD values of the retention times (M = 15) in different sample matrices are presented in Table 1 and show the retention to be stable. All the RSD values are below 0.15%. There are no significant differences between the different matrices. The RSD values for DTPA are, however, higher than those obtained for EDTA. This is mainly due to the broader peak of DTPA, which can be observed in Fig. 1. Idenrijication of the esters The esters were further identified by GC-MS. In the mass spectrum of EDTA [Fig. 2(a)], the molecular ion signal at m/z 404 corresponding the tetraethyl ester of EDTA can be seen.The most intense signal is for rnlz 202, which corresponds to fragments originating from the splitting in half of the symmet- 'p' rical molecule. The signal for mlz 130 corresponds to the fragments after cleavage of an OOCC2HS group. The weaker signal for mlz 33 1 corresponds to the cleavage of one OOCC2Hs group (mlz 73) from the original molecule. In the mass-spectrum of DTPA [Fig. 2(b)], the molecular ion signal (mlz 533) is missing, as is often found with compounds having a high molecular mass. Here again, the fragment at mlz 460 corresponds to the cleavage of one OOCC2HS group. The signal at rnlz 202 corresponds to the HCN(H2CCOOC2H& fragment. The signal at rnlz 331 corresponds to the other half 6 16 CQ r. 7- ~ r 6 16 Time/mi n Fig. 1 Chromatograms of 5 mg I-' each of EDTA and DTPA in ( a ) sea- water (Gulf of Riga) and (h) humic-containing lake water.Ethyl ester derivatives determined by GC-NPD. Table 1 RSD values of the retention times of EDTA and DTPA ethyl esters determined in different matrices (n = 15) RSD (%) Analyte Distilled water Lake water Sea-water EDTA 0.038 0.036 0.057 DTPA 0.094 0.042 0.14 I 130 I 331 I mlz Fig. 2 Mass spectrum of ( a ) EDTA ethyl ester and (h) DTPA ethyl ester.I338 Analyst, Septeniher 1996, Vol. 121 which is left after this rupture. These results indicate the presence of the pentaethyl ester of DTPA. Influence of estei-ificution time and temperature To examine the influence of esterification time and temperature, reactions were carried out at 80, 100 and 120 "C with reaction times between 0.5 and 14 h.Samples were heated in an oven or in a water-bath (approximately 100 "C). In Table 2, recoveries from esterifications using different temperatures and reaction times are presented. On average, the best recoveries of EDTA and DTPA were obtained when esterification was performed at 100 "C for 3 h. For EDTA and DTPA concentrations of 0.05 mg ] - I , there is some inconsis- tency between the results from the 3 h esterifications at 80 and 100 "C. Therefore, a check was made under these conditions with samples containing 0.1 and 0.2 mg 1-1 of EDTA and DTPA. The results verified slightly higher recoveries at 100 "C. Prolonging the esterification time resulted in lower recoveries in most instances. Since no additional peaks were identified, the lower recoveries must be due to volatilization of the analytes.Some difficulties arose because of the loss of the internal standard, added with the esterification reagent. This problem was observed especially with the lake water sample, at esterification temperatures of 3 100 "C. The added reagents might have catalysed the decomposition of the internal standard. Therefore, it is recommended to use a lower esterification temperature if C16H&N is used as an internal standard. Optiniiiution of the e..utrac-tion and neutralization stage The extraction solvents examined were C6HsCH3, C5H12 and C6Hi4. C6HsCH3 was found to be the most appropriate solvent for ethyl esters. Compared with C6H14 and CsHl2, it gave 4-10 times and 4-5 times higher recoveries of EDTA, respectively.In contrast to C6HsCH3 extraction, no DTPA could be identified at low concentrations (G0.2 mg 1-1) when other solvents were used. In most of the earlier studies, phosphate buffer or distilled water was used in the neutralization ~tage.6-13.lS-17.~~ In this study however, KHCO? was applied. Extraction with distilled water produced a turbid solvent phase that could not be injected into the GC column. As phosphate may cause serious contamination problems with the NPD, KHC03 was preferred. The influence of KHC03 concentration was further studied and the results are presented in Table 3. At low concentrations, the best recoveries for both EDTA and DTPA were obtained by the use of 1 mol 1-1 KHC03. At EDTA and DTPA concentrations >5 mg 1-1, the influence of KHC03 concentration was negligible.A volume of 25 ml was found to be adequate for neutralization. Influence of iron(zz1) It was considered that the high stability of the iron(m) complexes of EDTA and DTPA might cause interference in the analysis.12 According to Cassidy et a1,g 10-300 mg 1-1 concentrations of Fe" or Ferrl gave irreproducible and low results for EDTA determined as a methyl ester. Means et al.17 also obtained lower recoveries of EDTA methyl ester when iron was present. The effect of iron on DTPA determination has not been reported. The influence of iron on the recoveries of EDTA and DTPA was investigated by analysing spiked 5 ml samples containing 0.01, 0.05, 0.1, 0.5 and 1 .O mg 1-1 of EDTA and DTPA. The iron(1rr) concentrations applied were 5 , 10, 50, 100 and 500 mg 1-1.In Table 4, results from the experiments with lake water are presented. The results for a concentration 0.01 mg 1-1 are missing since this concentration was below the detection limit of DTPA. Also, iron concentrations of 5 and I0 mg 1--1 were too low to result in any interferences so they are not reported in the table. It can be seen that the recoveries of EDTA were lowered significantly only when the iron concentration was 500 mg 1-1. Even then, the decreases in the relative peak areas (expressing the recovery) were about 20% at the maximum. For DTPA, significant interference was observed at iron concentrations of > 50 mg 1- I . At this concentration, the decreases in recovery varied from 24 to 51%. Studies with distilled water gave the same results except that the decreases in recovery were slightly higher. This is probably due to the ability of humic material to adsorb iron and thereby decrease its free ionic form.Table 3 Effect of the concentration of the KHC03 neutralization solution on the recovery of EDTA and DTPA ethyl esters. The values presented are the counted relative peak areas of the analyte KHCO? concentration/ mol I-' Concentration/ Analyte mg 1-1 0.5 1 .0 2.5 EDTA 0.05 0.0418 0.109 0.069 DTPA 0.05 n.d.* 0.00 1 9 1 0.00 1 65 EDTA 0.5 0.422 0.461 0.306 DTPA 0.5 0.0469 0.0883 0.0269 EDTA 5 4.82 4.83 4.73 DTPA 5 2.25 2.48 2.32 n.d. = Not detected. Table 2 Recoveries of EDTA and DTPA ethyl esters under different esterification conditions. The values presented are the counted relative peak areas of the analyte.The study at a concentration of 0.01 mg I-' was executed separately and DTPA was not detected at this concentration 0.05 mg 1-I 0.5 mg 1-1 5 mg I-' 0.01 rngl-1 TemperaturePC Tim& EDTA EDTA DTPA EDTA DTPA EDTA DTPA 80 1 0.000 888 3 0.001 20 5 0.001 38 I00 1 0.004 68 3 0.008 94 5 0.005 44 120 1 3 Water-bath 3 13 * n.d. = Not detected. - - - - 0.0 107 0.0659 0.0101 0.008 37 0.0 137 0.0 177 0.004 68 0.004 35 0.009 46 0.027 1 n.d.* 0.0207 n.d. n.d. 0.003 9 1 0.009 96 n.d. n.d. n.d. 0.01 19 0.079 I 0.149 0.139 0.18 0.229 0.224 0.206 0.156 0.184 0.184 n.d. 0.0108 0.004 02 n.d. 0.0283 0.020 I 0.202 0.008 65 0.0144 0.0137 4.7s 5.2 4.77 5.96 6.21 5.48 5.28 1.92 5.78 5.37 0.685 I .22 I .64 I .79 2.79 2.58 1.65 0.319 2.08 2.13Analyst, September 1996, Vol.121 1339 Table 4 Effect of Fe"' concentration on the recoveries of EDTA and DTPA ethyl esters. The values presented are the counted relative peak areas of the analyte. The study was executed in lake water matrix EDTA/mg I- 1 DTPA/mg 1-1 Fe/mg I-' 0.05 0.1 0.5 1 0.05 0.1 0.5 1 0 0.248 0.547 3.27 7.72 0.0634 0.31 1 1.4.5 4.8 1 50 0.241 0.541 3.17 7.57 0.0479 0.152 0.874 2.59 100 0.214 0.499 4.20* 7.50 n.d. 0.0818 0.801 1.94 500 0.190 0.461 2.85 6.20 n.d.1 n.d. 0.106 0.297 * Peak area of the internal standard is about 30% smaller compared with the other runs. + n.d. = Not detected. Linearity of the Method and Quantitative Aspects EDTA and DTPA have been reported to occur in natural waters at concentrations between 0 and 200 pg 1- I .1 ~ ~ 4 , 2 5 The method is linear over this concentration range. In the studies with spiked humic containing lake water and sea-water matrices containing 10-5000 pg 1- I of EDTA and DTPA, the correlation coefficients of the linear calibration graphs were 0.997-1 .OOO for EDTA ( n = 16) and 0.988-1.000 for DTPA (n = 8). For the lake water matrix, the equations of the calibration lines were y = (0.0075 f 0.0004)~ - (0.1632 f 0.0277) for EDTA ( n = 7; x in pg 1- l ) and y = (0.0032 f 0.0007)~ - (0.1703 f 0.0414) for DTPA (n = 3). In calculations of the confidence limits for the slope and the independent term, a level of 95% was used. The repeatabilities varied between 2 and 4% for EDTA and between 4 and 1 1 % for DTPA, at the concentrations found in natural waters (10-500 pg 1-1).The reproducibilities ( n = 5 ) were approximately 10%. RSDs were used in these calculations. At the studied concentrations between 10 and 5000 pg 1-l, the recoveries of EDTA and DTPA were 80-1 10% in the sea- and lake-water matrices. At a concentration of 10 pg l-l, EDTA gave a low recovery in lake water matrix, which is most likely due to the partial adsorption on humic material. Using samples of 10 ml, the detection limits were 4 pg 1- 1 for EDTA and 13 pg 1-1 for DTPA in sea-water matrix (Tvarminne). The detailed raw data concerning the method validation will be presented in a separate paper.26 Conclusions The proposed method is simple and has been demonstrated to be suitable for the simultaneous determination of trace levels of EDTA and DTPA in lake and sea-waters.The detection limit for DTPA was clearly higher than that for EDTA. Iron(rrr) interfered in the determination of EDTA and DTPA only at concentrations far beyond those found in natural waters. In Finnish natural waters, the iron content is typically a few hundred micrograms per litre. In those lakes containing clay or during the overflow, the concentration can exceed 1 mg 1 - l . Hence the interference of iron is expected to be significant only in the analysis of heavily contaminated waters or waste waters. The analysis of natural water matrices was accomplished by using spiked samples. However, this may not simulate the situation in real environmental samples where EDTA and DTPA are presumably held in stronger adsorption sites. This is often disregarded in method development and validation. Therefore, investigations with environmental samples con- taining verified amounts of EDTA and DTPA are advisable.Jaana Sorvari and Mika Sillanpaa express their gratitude to Dr. Merja Suutari for her valuable help in performing the MS analyses and to Mrs. Kirsi Hiillos and Mr. Ari Jiirvinen for providing the sample matrices. The valuable comments of Professor Lauri Niinisto are greatly appreciated. The financial support of the Foundation of Technology (Finland) and the Maj and Tor Nessling Foundation (Finland) is also gratefully acknowledged. References 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 Crooks, R., and Sikes, J.. Appita J., 1990, 43, 67. Suntio, L., Shiu, W. Y., and Mackay, D., Chcwzosphere, 1988, 17, Alder, A.C., Siegrist. H., Gujer, W., and Giger, W., Water R a . , 1990, Randt, C., Wittlinger, R., and Merz, W., Fresenius' J . And. Chem., Saunamaki, R., Tappi J., 1995, 78, 185. Subach, D. J., and James, J. E., J . High Resolut. Chromatogr. Chromatogr. Commun., 1980, 3, 309. Snigoski, P. J., and Venezky, D. L., J . Chromatogr. Sti., 1974, 12, 359. Blank, M. L., and Snyder, F., .I. Chromatogr., 1979, 170, 379. Cassidy, R. M., Harpur, R., and Elchuk, S., .I. Chromatogr., 1980, Ribick, M. A.. Jemal, M., and Cohen, A. I., J . Pharm. Biomed. Anal , Nishikawa, Y., and Okumura T., J . Chrornatogr. A , 1995, 690, RCtho, C., and Diep, L., Z. Lebensm.-Unters. Forsch., 1989, 188, Rudling, L., Water Res., 1972, 6, 871. Schurch, St., and Diibendorfer, G., Mitt. Geb. Lehensmittelunters Hyg., 1989, 80, 324. Williams, D. T., .I. Assoc. OR. Anal. Chern., 1974, 57, 1383. Toste, A. P., and Lechner-Fish, T.J., Waste Manuge., 1993, 13, Means, J. L., Kucak, T., and Crerar, D. A., Environ. Pollut., Ser. B, Von Wanke, T., and Eberle, S. H., Actu Hydrochim. Hydrohiol., Pietsch, J., Schmidt, W., Sacher, F., Fichtner, S., and Brauch, H.-J., Brauch, H.-J., Fleig, M., and Schullerer, S., Ber. Arheitsgem. Rhein- Nguyen, D.-K., Bruchet, A., and Arpino, P., J . High Resolut. Gardiner, J., Analyst, 1977, 102, 120. Bergers, P. J., and Groot, A. C., Water Res., 1994, 28, 639. Van Dijk-Looyaard, A. M., De Groot, A. C., Janssen, P. J. C. M., and Xue, H., Sigg, L., and Kari, F. G., Environ. Sci. Technol., 1995, 29, Sillanpaa, M., Sorvari, J., and Sihvonen, M.-L., Chrornatogruphiu, 1249. 24, 733. 1993,346,728. 190, 188. 1987, 5 , 687. 109. 223. 237. 1980, 1, 45. 1992, 20, 192. Fresenius' J . Anal. Chern., 1995, 353, 75. Wassenuerke, 1991, EV 48, 31. Chromutogr., 1994, 17, 153. Wondergem, E. A., H 2 0 , 1990,23, 682. 59. 1996, 42, 578. Paper 6102724E Received April I8, I996 Accepted June 3, I996
ISSN:0003-2654
DOI:10.1039/AN9962101335
出版商:RSC
年代:1996
数据来源: RSC
|
45. |
Effect of redox state on the response of poly-N-(2-cyanoethyl)pyrrole coated thickness-shear mode acoustic wave sensors to organic vapours |
|
Analyst,
Volume 121,
Issue 9,
1996,
Page 1341-1348
Zhiping Deng,
Preview
|
PDF (2260KB)
|
|
摘要:
Analyst, September 1996, Vol. 121 (1341-1348) 1341 Effect of Redox State on the Response of Poly-M( 2-cyanoet h y I) py r role Coated Thickness-Shear Mode Acoustic Wave Sensors to Organic Vapours Zhiping Deng, David C. Stone and Michael Thompson Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario, Canada, M5S 3H6 Different redox states of poly-N-(2-~yanoethyl)pyrrole (PCPY) films were prepared by electropolymerization and subsequent oxidation or reduction on the surface of thickness-shear mode (TSM) acoustic wave sensors. The resulting films were characterized by cyclic voltammetry, impedance spectroscopy, X-ray photoelectron spectroscopy and scanning electron microscopy. The response of the coated sensors to organic vapours consisting of 9-carbon chains with different functionalities was also studied.The observed responses were dependent on both film redox state and organic vapour properties. All films showed high sensitivity to nona-3,6-dien-l-ol, an aroma component characteristic of early degradation of fresh fish. Principal components analysis allowed all five analytes to be distinguished in the discriminant plane using only different redox state PCPY films. Keywords: Thickness-shear mode acoustic wave sensor; conducting polymer; fish freshness; poly-N-(2-cyanoethyl)pyrrole; cyclic voltammetry; impedance spectroscopy; nonane; nonane derivatives Introduction Films of conducting polymers such as polypyrrole have attracted increasing interest as sensitive and selective coatings for gas sensors.1-5 Such films are readily prepared by electrochemical polymerization of the corresponding monomer onto an electrode, and can be used to produce both conductance and acoustic wave chemical vapour sensors.It has been widely demonstrated that the electronic and vapour sorption properties of these conducting polymers are strongly influenced by the particular polymerization conditions employed including ap- plied potential, monomer concentration, counter-ion and sol- ~ e n t . ~ , ~ Unlike conventional organic polymers, however, the properties of conducting polymers can be dynamically manipu- lated, after synthesis, through the application of oxidizing or reducing potentials.x Such changes in redox state of the polymer film also lead to changes in both electronic and vapour sorption properties.9.10 Control of the redox state has been exploited for chromatographic separations,' 1,12 vapour permeability stud- ies l 3 , I 4 and the fdbrication of chemical vapour sensors.l S , I 6 Our own interest in conducting polymer films is in their use as selective coatings for thickness-shear mode (TSM) acoustic wave sensors. In particular, we have recently reported the use of a variety of electropolymerized N-substituted pyrrole deriva- tives for the selective detection of aroma components that are indicative of early fish spoilage.17 As part of this work we have also described the preparation and use of poly N-(2-cyanoe- thy1)pyrrole (PCPY) films as selective coatings for chemical vapour sensors. 18 The cyano functionality gives enhanced film- vapour interactions with polar species such as nitriles and alcohols, making it a particularly effective detector for nona- 3,6-dien-l-o1, which is one of the most sensitive indicators of early fish degradati~n.'~ In the present paper we describe the characterization of different PCPY redox states by cyclic voltammetry (CV), impedance spectroscopy (IS), X-ray photo- electron spectroscopy (XPS), scanning electron microscopy (SEM) and vapour sorption studies employing coated TSM sensors.In particular, we consider the effect of redox state on the sorption of a series of nine-carbon chain compounds containing different functional groups in order to study further the selectivity of this polymer. Experimental Reagents, Materials and Methods The solvents nonane, non-3-ene, nonan- 1-01, non-3-en- 1-01, acetonitrile (AR grade, Aldrich, Milwaukee, WI, USA) and nona-3,6-dien- 1-01 (99%, Interchim, Montlucon, France) were used as supplied.1 -(2-Cyanoethyl)pyrrole (Aldrich) was vac- uum-distilled before use. Tetrabutylammonium perchlorate (TBAP) (electrochemical grade, Fluka, Buchs, Switzerland) was used as received. The TSM devices were unpolished 9 MHz AT-cut quartz crystals plated with gold electrodes and were obtained from International Crystal Manufacturing Co. (Okla- homa City, OK, USA). A combined Pt-Ag/AgCl (3 mol 1-l KCI) electrode (Metrohm, Herisau, Switzerland) was used as the counter and reference electrode, respectively. The cell used for the electrochemical deposition of PCPY onto the gold electrodes of the TSM devices has been described previously. 16 Electropolymerization was achieved using a solution containing 1-(2-~yanoethyl)pyrrole (0.1 rnol 1- l) and TBAP (0.1 mol 1-1) in acetonitrile which had been deoxygen- ated by sparging with nitrogen gas for 20 min.A constant potential of + 1.20V was applied for 1 min to each side of a TSM device using a potentiostat/galvanostat (Model 273, EG & G Princeton Applied Research, Princeton, NJ, USA). The coated device was then rinsed with acetonitrile, further washed with acetone, dried with a stream of nitrogen gas and then dried at 100 "C for 1 h. Electrochemical oxidation and reduction of the as-prepared films was achieved in the same cell using a fresh, deoxygenated solution containing only TBAP (0.1 rnol 1-l) in acetonitrile.A fresh coated device was used for each reduction or oxidation potential, which was applied for 8 min. Following removal of the applied potential, the coated TSM device was left in the cell with the TBAP solution for 10 min, after which the electrode potential of the coated device (@) was measured with respect to an Ag/AgCl reference electrode using a high impedance multimeter. All devices were then washed and dried as before and stored in a desiccator prior to use. The in-air resonant frequency of the TSM devices was measured at each stage in the process, allowing the mass changes produced by film deposition and subsequent oxidation or reduction to be1342 Analyst, September 1996, Vol. 121 recorded. The preparation conditions and measured potentials are summarized in Table 1.Cyclic Voltammetry Cyclic voltammograms were recorded using the potentiostat and 0.1 mol 1-1 TBAP in acetonitrile as the supporting electrolyte. These were all obtained for a sweep rate of 100 mV s-1 using an Ag/AgCl reference electrode. Separate coated TSM devices were used for each redox state film, only one side of the device being used to obtain the voltammogram. Impedance Spectroscopy Impedance measurements were performed using the poten- tiostat combined with a lock-in amplifier from the same manufacturer (Model 530 1). The impedance-frequency data were collected on a 386 computer using software provided by the- manufacturer. The amplitude of the sinusoidal potential applied across the measurement cell was 5 mV for all measurements.The impedance characteristics obtained for each redox state were subsequently transferred to a second program in order to fit different equivalent electrical circuits to the data (EQIVCRT.PAS, version 3.96, Department of Chemical Tech- nology, University of Twente, Enschede, The Netherlands). X-ray Photoelectron Spectroscopy X-ray photoelectron spectra were obtained using a MAX 200 (Leybold, Cologne, Germany) spectrometer using an unmonochromatized Mg K a source and an analysis area of 2 X 4 mm2. Survey and low resolution spectra were obtained with a pass energy of 192 eV: high resolution spectra of the C(ls), N( I s) and Cl(2p) regions were obtained using a pass energy of 48 eV. All spectra were satellite subtracted and normalized with respect to the spectrometer transmission function and elemental sensitivity factors using software and parameters provided by the manufacturer.Charging effects were compensated for by calibrating the spectra to the main C(1s) peak position at 284.6 eV. Peak deconvolution and elemental analyses were also performed using software provided by the manufacturer. Scanning Electron Microscopy Scanning electron micrographs were obtained on a Hitachi, Model S-570, microscope fitted with a Quartz PCI image capture system for recording, storing and printing the images. In order to reduce distortion due to charging of the semiconducting polymer samples, the coated TSM devices were first sputter- coated with a thin gold film using a Polaron vapour deposition system (Model E5 100, Polaron, Watford, Hertfordshire, UK).The gold film was deposited using an argon plasma for 90 s at a current of 20 mA. All images were obtained using an accelerating potential of 18 kV. Vapour Sorption Studies Vapour generation was achieved by bubbling helium gas through the corresponding liquid at room temperature and pressure using the flow system described previously.18 The flow rates of the sample and purge streams were set to 30 ml min-1 for all experiments, using a combination of needle valves and mass flow meters. The resonant frequency of the coated TSM sensors was monitored continuously using an oscillator circuit and a universal counter (Model HP5334B, Hewlett-Packard, Avondale, PA, USA). Data was collected, displayed and stored using a Macintosh I1 computer connected via an IEEE 488 interface bus using software written in-house.Principal compo- nents analysis (PCA) was performed on the sensor response data using the SYSTAT statistical analysis program (SYSTAT, Evanston, IL, USA). The response of the coated TSM sensor was recorded by first allowing the device to stabilize in a pure helium stream. This was then switched to the vapour stream, and the frequency monitored until a steady-state plateau had been reached. The stream was then switched back to pure helium to purge the device. Three such measurements were recorded for each coated TSM sensor, and three such devices were prepared for each redox state in order to monitor both between-run and between-device reproducibilities. Sensor calibration graphs were obtained by diluting the vapour stream with different ratios of pure helium while keeping the net flow rate constant.Vapour concentrations were established by condensing the sample stream in a liquid nitrogen cold trap and weighing. The measured concentrations of the vapours are listed in Table 2 together with relevant physical data. Results and Discussion Cyclic Voltammetry The cyclic voltammograms for as-prepared, reduced (-1.20 V) and oxidized (+1.40 V) thin films of PCPY on the gold TSM Table 1 Preparation conditions and electrode potentials of the different redox state PCPY films Film + 1.20 +1.20 +1.20 + I .20 +1.20 +1.20 1 .0 1 .0 1 .0 1 .0 1 .0 1 .0 -1.20 - 1 .OO 0.00 - +1.20 + I .40 8.0 8.0 8.0 - 8.0 8.0 Reduced Reduced Reduced As-prep Oxidized Oxidized -0.13 -0.07 +0.3 1 +0.67 +0.90 +0.92 * All potentials V P ~ S U S Ag/AgCl reference electrode.Table 2 Concentrations and physical data for the test vapours and corresponding liquids Vapour Nonane Non-3-ene Nonan- 1-01 Non-3-en-1-01 Nona-3,6-dien- 1-01 Boiling-pointPC 1s 1 146 21s 2 12-21 s 195 Density/g ml-' 0.718 0.730 0.827 0.843 Unknown Molecular mass/g mol-' 128.3 126.2 144.3 142.2 139.4 C&g 1-1 at 21 "C and 30 ml min-I 4550 4010 374 382 410Analyst, September 1996, Vol. 121 1343 N -100- EIV versus AgfAgCI sensor electrode are shown in Fig. 1. The anodic peak potential for as-prepared PCPY (curve 1) is about +0.84 V while the cathodic peak potential is +0.60 V, in agreement with our earlier results.I8 These values are shifted to more positive potentials than for the parent polypyrrole (PPY).For example, values of -0.09 V and -0.29 V ( i ~ r s u s NaCl calomel electrode) for PPY on a platinum electrode have been reported for the anodic and cathodic peaks, respectively.20 While our results were obtained on gold versus an Ag/AgCl electrode (other conditions were identical), the positive shift is larger than can be explained on the basis of this difference alone. This effect has also been observed for other N-substituted polypyrroles, and is attributed to steric effects associated with the pendant side chain. Anodic and cathodic currents are highest for the as-prepared polymer, indicating that rates of oxidation and reduction are faster than for the oxidized or reduced forms. Since the polymer carries an increased positive charge in its oxidized form, C104- counter ions will become incorporated into the film from the supporting electrolyte to maintain electrical neutrality.This is supported by the frequency shifts in air of the dried polymer- coated TSM sensors: a frequency decrease is observed follow- ing oxidation, corresponding to mass uptake by the film. Upon reduction, the polymer attains a neutral state, requiring diffusion of counter ions out of the film. In this case a frequency increase is observed, corresponding to a mass loss by the film. The origin of the small anodic peak at -+0.4 V is unclear, but may be a capacitive current arising from the highly porous nature of the film surface. The voltammogram for the reduced form (curve 2) shows the anodic peak shifted to lower potential and the cathodic peak shifted to higher potential.The anodic peak shows a shoulder at = +0.55 V, which is tentatively assigned as a capacitive current. Both anodic and cathodic peaks for the oxidized form (curve 3) are shifted towards more positive potentials by about 0.20 V relative to the reduced form, as expected from charge considerations. The results for the three forms of PCPY indicate quite different electrochemical behaviour, reflective of the different charge states of the polymer backbone and the extent of counter- ion incorporation. These differences are expected to influence both the impedance and sorption properties of the films. It is also clear from Fig. 1 that applied potentials more positive than + 1 .O V will produce a highly oxidized state, while potentials more negative than +0.30 V produce a highly reduced state.In between these extremes an intermediate state will be obtained, in which the fraction of oxidized or reduced forms strongly depends on the applied potential.12 In the current study, w ; 5 : E j j kinetic i diffusion j control j control 4-----u w potentials of +I .20 or + I .40 V were applied to obtain extensive oxidation and potentials of 0.00, -1.00 and -1.20 V for extensive reduction. Impedance Spectroscopy Impedance spectroscopy is used to study the electrochemical behaviour of a diverse number of systems.21 A small alternating potential is applied to the system under test, and the impedance (defined as the ratio of voltage to current) measured as a function of frequency.Unlike cyclic voltammetry, which involves large potential perturbations, the system is only infinitesimally perturbed with respect to the steady state. This may be particularly important when studying conducting polymers, since large perturbations may induce inhomogeneous states in the films. Because impedance is a complex quantity, results are typically presented in the form of a Nyquist diagram in which the imaginary (reactive) component is plotted against the real (resistive) component with frequency as a parameter. For redox couples in aqueous electrolytes, for example, a semicircle is indicative of kinetic control by an electrochemical charge transfer step; linear behaviour at 45" to the real axis, on the other hand, is characteristic of diffusion control.Rate constants and diffusion coefficients can be calculated by relating the physical system to an equivalent electrical circuit displaying the same impedance-frequenc y characteristics. A suitable equivalent circuit for porous semiconducting polymers has been described by Albery ct cr1.,22 and is shown in Fig. 2 together with the corresponding Nyquist plot. Here RE represents the bulk electrolyte resistance, CE the electrolyte- polymer-electrode capacitance and ZD a transmission line consisting of the charge transfer resistances for the polymer and electrolyte within the film pores (RP and R,,, respectively) coupled by a distributed capacitance Cp. This latter term is associated with interfacial redox reactions involving polaron and bipolaron states at the pore wall.The analysis of Albery et (11. shows that: if RP << R,,, current flows through the polymer; if RP >> R,, current flows through the pores: and if R P = R,,, the system is symmetrical and the diffusion-controlled region vanishes. This latter condition may be explained by local coupling of charge carrier motion in both polymer and pore electrolyte. An alternative explanation is that mobility of the counter-ions depends on polymer chain motion, which, in turn, has a direct effect on polymer conductance. Results for as-prepared and oxidized (+1.20 and +1.40 V) PCPY films in contact with 0.1 moll-1 TBAP in acetonitrile are shown in Fig. 3(a). The impedance spectrum for a reduced1344 Analyst, September 1996, Vol. 121 16 14 12 10 kQ, respectively. These values are of the same order of magnitude as those obtained by Albery for lightly oxidized and reduced PPY.An estimate of CE can be obtained from the frequency corresponding to the maximum of the semicircle since this gives the characteristic relaxation time 'tR = RECE = 1/(2nfi. Note that in this case, however, the relaxation time is actually the mean of a distribution of values, as indicated by the fact that the centre of the semicircle lies below the real axis (i.e., there are distributed elements in the film-electrode system). For the oxidized PCPY films, CE =: 13 nF, while for the as-prepared and reduced films the estimated values are 40 nF and 0.16 pF, respectively. This apparent decrease in CE with increasing oxidation is also consistent with the findings of Albery et al.for PPY.22 For the highly reduced polymer a larger value of R, is indicated ( = 2 kQ), reflecting the much lower density of charge carriers in this state. The much higher impedance for the reduced polymer compared with the as-prepared and oxidized forms is again consistent with the findings of Albery et al. for PPY.22 Interestingly, the shape of the reduced PCPY character- istic suggests a partial second arc on the low frequency side of - - - - 1.2 1 .o 0.8 0.6 0.4 0.2 c 0.0 5 0.0 0.2 0.4 0.6 0.8 1 .o E y 22 20 18 / ,B / d L( / 4 0 2 4 6 8 10 12 14 16 18 20 Z d k Q the kinetic control region. The characteristic for the as-prepared polymer [Fig. 3(a), curve 31 shows a shallow slope in the same region, which could also be interpreted as either a partial second arc or a narrow diffusion control region.Such an arc could be indicative of electrode polarization, for example. X-ray Photoelectron Spectroscopy The C(1s) spectra of the as-prepared, oxidized and reduced PCPY films are shown in Fig. 4. These show three peaks at binding energies of 284.6,286.3 and 288.4 eV, respectively, the spectrum for the as-prepared polymer being in agreement with our earlier results.18 The main peaks are complex and cannot be assigned to a single species. Thus. the peak at 284.6 eV is assigned to both the a and 8 carbons of the pyrrole r i n ~ 2 ~ and 1 I I I I I I I 292 288 284 280 I I I I I 1 I I 292 28 8 284 280 I I I I I I I (c) I I $cis Picis Fig. 3 Nyquist plots for PCPY in: (a), different oxidation states; 1. oxidized (+1.40 V); 2, oxidized (+I 20 V); and 3 , as-prepared; (h), a highly reduced state (-1.20 V).Inset shows an enlargement of the region Z,, = &2 kB. Binding energy/eV Fig. 4 oxidized (+1.40 V); and (c), reduced (-1.20 V). C( 1s) X-ray photoelectron spectra for PCPY: ( a ) , as-prepared; (h),Anulyst, September 1996, Vol. 121 1345 I I I I I '4 N is' 5. N I s (c) - - - - - - - - - - - - - - I I I I I I I 406 404 402 400 398 396 394 are not resolvable under the instrumental conditions used in our experiments. The peak at 286.3 eV is assigned to the carbon of the cyano group, as discussed previously.18 The high binding energy peak at 288.4 eV can be attributed either to the carbonyl group introduced by termination of the polymerization,7 or to an interaction with the perchlorate counter ion.The two main peaks also include contributions from the methylene carbons in the side chain. The relative C( 1 s) peak intensities obtained for various oxidized and reduced PCPY films are listed in Table 3. The results show that further oxidation of the polymer in 0.1 moll-' TBAP leads to an increase in the relative intensity of the peak at 288.4 eV, while reduction leads to a corresponding decrease. The change in the higher binding energy C(1s) peak is mirrored by the chlorine content of the film. This was calculated from the XPS data to be 0, 2 and 6 atom % for the reduced, as- prepared and oxidized (+I .2 V) films, respectively. These results confirm that increased oxidation of the film results in increased incorporation of perchlorate ion.The spectra do not, however, support a simple model involving only diffusion of the perchlorate ion in and out of the film. The as-prepared polymer shows peaks at 200.8 and 207.5 eV for the 2P3/2 component in an approximately 1 : 1 ratio. The higher binding energy peak can be assigned to perchlorate ion trapped within the polymer matrix during electrochemical deposition. l 8 The film oxidized at +1.20 V shows only a trace of the perchlorate ion, while the film oxidized at +1.40 V shows no detectable perchlorate signal and a lower chlorine content (3.5 atom %). The lower binding energy peak was previously interpreted as chloride ion leaking from the Ag/AgCl reference electrode; however, the value of 200.8 eV is high for a chloride and closer to that for an organochlorine species. Alkali chlorides typically appear at 198-199 eV, for example, while dichlorides such as CuC12 and NiC12 range from 199 to 200 eV.PVC and chlorobenzene, on the other hand, show Cl(2p) peaks at 200 and 201 eV, respectively. A related observation is that the intensity of the C( Is) peak at 286.2 eV drops for the film oxidized at +1.20 V compared with the one oxidized at +1.40 V, while the chlorine content decreases from 6 to 3.5 atom %. This correlation would be expected if the observed chlorine was covalently bound to the polymer. Further work is needed to clarify this issue and determine the exact nature of the chlorine species introduced into the polymer by these oxidizing potentials. It should be remembered, however, that these films are much thicker than the depth of XPS analysis.It is therefore possible that the decrease in the perchlorate signal (and subsequent decrease in the total chlorine signal) is, at least in part, a result of the counter-ions penetrating much more deeply into the film at the highest oxidation potential. The N (1s) spectra of the as-prepared, oxidized and reduced PCPY films are shown in Fig. 5. Both the ring and nitrile nitrogen atoms are expected to give peaks between 399 and 400 eV for the neutral polymer, and the main peak, is, in fdct, found at 399.9 eV. This is in agreement both with our earlier results for PCPYlX and literature values for the ring nitrogen in poly- Table 3 Relative intensities (5%) of the C(ls) region for PCPY films in different redox states Binding energyIeV State 284.5 k 0.5 286.2 f 0.5 288.4 k 0.5 Oxid.(+1.40 V) 42.3 47.4 10.4 Oxid. (+1.20 V) 37.7 55.7 6.6 As-prep. 51.3 43.8 5.0 Red. (0.00 V) 49.8 45.0 5.2 Red. (- I .OO V) 54.0 41.8 4.2 Red. (- 1.20 V) 53.1 44.7 2.2 p y r r ~ l e . ~ ~ , ~ ~ A higher binding energy component at 402.5 eV is attributed to the nitrogen atoms in the polaron-bipolaron ~ n i t s , ~ ~ , ~ ~ and is observed for the as-prepared, oxidized (+1.4 V) and reduced (- 1 .O V) samples. The oxidized sample shows an additional component at 400.5 eV, while the reduced form shows a minor component at 397.4eV. This latter peak has also been observed for a reduced polypyrrole film,lh and was attributed in this case to the effect of hydrogen bonding interactions between pyrrole subunits in adjacent chains.Such an effect is not possible in PCPY, however, and the origin of this component remains unclear. Scanning Electron Microscopy Morphological differences are found among the different polymer oxidation states, as seen in Fig. 6. Here, the as-prepared I - 1 I I I I 1 404 402 400 398 396 Fig. 5 oxidized (+1.40 V); and (c), reduced (-1.20 V). N(1s) X-ray photoelectron spectra for PCPY: (u), as-prepared; (h),1346 Analyst, Septenzher 1996, Vol. 121 and reduced films show a high degree of similarity, while the oxidized form has a surface with a greatly reduced variation in film height. Feldheim and Elliott14 have also observed sig- nificant morphological differences between the reduced and as- prepared forms of pofy(3-methylthiophene) and poly(N- methylpyrrole) films.For the former, the as-prepared film is smoother, while for the latter, the reduced film is smoother. Clearly, then, film morphology is dependent on both the polymer type and redox state. It will also be dependent on film thickness and substrate roughness, since earlier work has shown that electrochemically-deposited films of polypyrrole and its derivatives initially follow the morphology of the underlying electrode surface.16.18 The thicker the film, the more any surface features of the underlying electrode become 'filled-in'. Film thicknesses were not measured by SEM in the present work, but were estimated to be between 1.2 and 1.4 pm based on the TSM frequency shifts in air before and after coating and the film density. Thus, the as-prepared and reduced films show a globular PCPY structure superposed on the underlying elec- troae roughness, The oxidized film appears to have swollen, however, resulting in a more uniform film surface.This latter effect can be rationalized in terms of increased counter ion incorporation. Vapour Sorption Studies TSM sensors coated with PCPY films in different redox states were exposed to nonane, non-3-ene, nonan- l-ol, non-3-en- 1-01 and nona-3,6-dien- 1 -01 vapours. The steady-state frequency shifts observed for vapour sorption by the different films are presented in Table 4. These have been normalized to constant film thickness between the different TSM sensors as discussed previously.'7,1* This is necessary since sensor response is a function of film thickness, and it is difficult to obtain consistent film thicknesses with the apparatus used in this study.It is also convenient to convert these frequency shifts to partition coefficients using the equation26 (1) where Af, and At5 are the frequency shifts due to vapour sorption and film deposition in Hz and kHz, respectively, p is the coating density in g ml-1 and C, the vapour concentration in g 1-l. A density of 1.40 g ml-1 was assumed for all three redox states18 Yalues of log K calculated from the normalized frequency shifts are also included in Table 4. The partition coefficient offers the advantage of being independent of vapour concentration for constant temperature and pressure, allowing direct comparison of the results for different test vapours. The greatest response for the as-prepared coating was to nona-3,6-dien- 1-01, which is the most sensitive indicator of early fish degradation.Fig. 7 shows the corresponding calibra- tion curves over the concentration range 14410 pg 1-1 for various film thicknesses. These show good linearity and the expected dependence of sensitivity on film thickness. For a film causing a 30kHz frequency shift, the sensitivity of the 9 MHz TSM sensor to nona-3,6-dien-l-ol is 1.45 Hz 1 pg-1. The responses of the coatings involving different redox states from the test vapours show some clear patterns. For any given redox state, for example, log K decreases in the order nona-3,6-dien- 1-01 > non-3-en-1-01 > nonan-1-01 > non-3-ene > nonane. This trend reflects the number of functional groups per molecule capable of interacting with PCPY through hydrogen bonding and n-electron overlap interactions.The unsaturated alcohols can interact through both hydrogen bonding and n- electron overlap, for example. Nonane, on the other hand, can Af" P AfS" K =- Fig. 6 Scanning electron micrographs of PCPY: (a), as-prepared; (b), oxidized (+ 1.40 V); and (c), reduced (- 1.20 V). Table 4 Normalized frequency response (AS\ = 30 kHz) and partition coefficients for the different redox state coated TSM sensors Nona-3,6-dien- l-ol Non-3-en- 1-01 Nonan- 1-01 Non-3 -ene Eappllv A W Z log K - 1.20 654 f 49 4.87 - 1 .oo 456f 18 4.72 0.00 337 f 38 4.58 +1.20 710 f 40 4.91 + 1.40 916 f 34 5.02 As-prep 612 f 27 4.84 Af,lHz log K 456 ? 3 4.75 228f 14 4.44 294 k 20 4.56 425 f 38 4.72 235 f 12 4.46 419 k 45 4.7 1 AfvlHz log K 245 k 12 4.49 134 f 17 4.22 143 f 15 4.56 189 k 8 4.37 215 f 10 4.43 269k 15 4.53 Nonane AfvlHz log K 148 f 28 3.24 106 f 8 3.09 137f 15 3.20 104f 15 3.08 142 f 20 3.22 152 f 27 3.25 Af"lHZ log K 124 f 4 3.10 165 k 14 2.82 5 9 f 13 2.78 61 It 10 2.80 7 0 f 15 2.86 58 f 9 2.77 Afv values are the mean k standard deviation for three consecutive measurements.Analyst, September 1996, Vol.121 1347 3 2 1 g 0 - g -1 N v - 0 Q 0 a Q c - .- " -2 -3 .- L -4 1200 1000 2 800 2 f $' 600 0 0- t" 400 200 0 ! ! ! ! ! ! i nonanT1-ol ; ....................... 4 ....; ............... : ............. i nonane* i ; g ; ; ............. ..... .............. ............. ....... i ............... i .............. I 8 ..__A i i ........................... ; ..............:i ........... 2 .............. ; ............................ -- .................. a-(7.; ............ -.: ......................................... OD .. i 4" fl.....; .............. 1 .............. i ............... j ............. Q - ............ p... ....................................................... 1 .............. _i , ! ndn-3-enq inon-3-en-I -01 : i nona-3,6-di& -ol i Oio i : o ' ; q ............ ...... I 0 i _ 1 fO I 1 I I 1 I 0 50 100 150 200 250 300 350 400 450 Concentration of vapourlpg I-' Fig. 7 Calibration curves for a TSM sensor coated with PCPY (as- prepared) towards nona-3,6-dien-l-o1. Film thickness is: (l), 2.3; (2), 1.8; ( 3 ) , 1.5; (4), 0.90; and (S), 0.54 pm. only interact with PCPY though dispersion interactions.It is also considerably more volatile than the alcohols, resulting in the large difference in log K values between these compounds. Similar effects arising from both redox state and the nature of the test compounds have been observed for polypyrrole chromatographic stationary phases l , various permeable mem- branes13.14 and sensor coatings.Is-lx The greatest response to nona-3,6-dien- 1-01 was obtained for the film oxidized at +I .40 V whilst the greatest response to nonane was obtained for the film reduced at -1.20 V. Indeed, the extreme polymer redox states generally show higher log K values than the intermediate and as-prepared ones. In the case of the reduced film this can be explained by increased penetration of the test vapour into the polymer due to its electrical neutrality (allowing for enhanced hydrophobic interactions'o) and re- moval of the counter-ions.In addition, the SEM results show that these films have greater macroporosity than the oxidized ones. In the case of the highly oxidized film, the enhanced response may be attributed in part to the increased charge on the polymer network giving rise to larger induced dipolar inter- actions. Principal components analysis (PCA) was applied to the response data in order to see i l variations in redox state alone would be sufficient to achieve selectivity between the different test vapours. PCA can be used to reduce the complexity of multivariate problems in which the variables are partially correlated to two or more orthogonal axes (the principal components) which best fit systems of points in space.27 The data set consisted of the normalized frequency responses for all coating-vapour combinations which were further divided by the vapour concentration in order to remove this factor from the data.Since three coated TSM devices were prepared for each redox state and the vapour response of each device was measured in triplicate, this resulted in a total of 54 values for each vapour. It was found that the five vapours could be distinguished using the first two principal components (Fig. S), which account for 59.3% and 30.7% of the variation in the data, respectively. In principle, then, an array of sensors prepared using different redox states of PCPY could be used to detect the early onset of fish degradation, Conclusions We have characterized, for the first time, poly-N-(2-cyan- oethy1)pyrrole films in various redox states in terms of both their chemical and electronic properties by using a variety of techniques.Clear differences in these properties exist across the -20 -15 -10 -5 0 5 10 15 Principal component (1) Fig. 8 PCPY films in different redox states to the five test vapours. First two principal components plot for the normalized responses of range from neutral to highly oxidized films. These differences are, in turn, reflected by both the film's morphology and its properties with respect to sorption of organic vapours. Results obtained using TSM acoustic sensors coated with PCPY films in different redox states further show that such differences in properties are sufficient to be able to distinguish various structurally similar molecules by combining the experimental measurement of sensor response with principal components analysis. Such an approach has potential application to the monitoring of nona-3,6-dien- 1-01 for the detection of early fish spoilage in the food industry.Different redox state PCPY films may also be incorporated into both acoustic and conductance sensor arrays for a variety of vapour detection and measurement applications. We are grateful to the Natural Sciences and Engineering Research Council (Canada) for financial assistance. We also thank A. Verma of the University of Toronto for assistance with the impedance spectroscopy and Professor R. H. Morris of the University of Toronto for the use of the cyclic voltammetry system. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Teasdale, P.R., and Wallace, G. G., Analyst, 1993, 118, 329. Bartlett, P. N., and Ling-Chung, S. K., Sens. Actuators, 1989, 20, 287. Slater, J . M., Paynter, J., and Watt, E. J., Analyst, 1993, 118, 379. Topart, P., and Josowicz, M., .I. Phys. Chem., 1992, 96, 8662. Kunugi, Y., Nigorikawa, K., Harima, Y., and Yamashita, K., J . Chem. Soc. Chem. Commun., 1994, 873. Feldman, B. J., Burgmayer, P., and Murray, R. W., J . Am. Chem. Soc., 1985, 107, 872. Diaz, A. F., and Bargon, J., in Hundbook of Conducting Polymers, ed. Skotheim, T. A., Marcel Dekker, New York, 1986, vol. 1, pp. 81- 11s. Zhao, H., Price, W. E., and Wallace, G. G., Pnlym., 1993, 34(1), 16. Ikariyama, Y., and Heineman, W. R., Anal. Chem., 1986, 58, 1803. Lu, W., Zhao, H., and Wallace, G. G., Anal. Chim. Acta, 1995, 315, 27. Ge, H., and Wallace, G. G., Anal. Chem., 1989, 61, 2391. Deinhammer, R. S., Shimam, K., and Porter, M. D., Anal. Chem., 1991,63, 1889. Schmidt, V. M., Tegtmeyer, D., and Heitbaum, J., Adv. Muter., 1992, 4, 428. Feldheim, D. L., and Elliott, C. M., J . Menihrane Sci., 1992, 70, 9. Feldheim, D. L.. Krejcik, M., Hendrickson, S. M., and Elliott, C. M.. J . Phys. Chem., 1994,98, 5714.1348 Analyst, September 1996, Vol. 121 16 17 18 19 20 21 Vigmond, S. J., Kallury, K. M. R., Ghaemmaghami, V., and Thompson, M., Talanta, 1992, 39,449. Deng, Z., Stone, D. C., and Thompson, M., Analyst, 1996, 121, 671. Deng, Z.. Stone, D. C., and Thompson, M., Can. J . Chem., 1995,73, 1427. Josephson, D. B., Lindsay, R., and Olafsdottir, G., in Seafood Quality Determination, Proceedings of an International Symposium Co- ordinated by the University of Alaska Sea Grant College Program, Anchorage, Alaska, 1986, eds. Kramer, D. E., and Liston, J., Elsevier, Amsterdam, 1986, pp. 2747. Diaz, A. F., Castillo, J. I., Logan, J. A., and Lee, W.-Y., .I. EIectr-oanal. Chem., 1981, 129, 115. MacDonald, J. R., and Johnson, W. B., in Impedance Spectroscopy, ed. MacDonald, J. R., Wiley, New York, USA, 1987, ch. 1. 22 Albery, W. J., Chen, Z., Horrocks, B. R., Mount, A. R., Wilson, P. J., Bloor, D., Monkman, A. T., and Elliott, C. M., Faraday Discuss. Chem. Soc., 1989,88, 247. Pfluger, P., and Street, G. B., J. Chem. Phys., 1984, 80, 544. Kang, E. T., Neoh, K. G., Ong, Y. K., Tan, K. L., and Tan, B. T. G., Synth. Met., 1990, 39, 69. Nalwa, H. S., Polymer, 1991, 32, 802. Grate, J. W., Snow, A., Ballantine, D. S., Jr., Wohltjen, H., Abraham, M. H., McGill, R. A., and Sasson, P., Anal. Chem., 1988, 60, 869. Malinowski, E. R., and Howery, D. G., Factor Analysis in Chemistry, Wiley, New York, 1980. Puper 6102500E Received April 10, I996 Accepted May 24,1996 23 24 25 26 27
ISSN:0003-2654
DOI:10.1039/AN9962101341
出版商:RSC
年代:1996
数据来源: RSC
|
46. |
High sensitivity conducting polymer sensors |
|
Analyst,
Volume 121,
Issue 9,
1996,
Page 1349-1353
A. C. Partridge,
Preview
|
PDF (1018KB)
|
|
摘要:
Analyst, September 1996, Vol. 121 (1349-1353) 1349 High Sensitivity Conducting Polymer Sensors A. C. Partridge, P. Harris and M. K. Andrews Industrial Research Limited, P.O. Box 31 -310, Lower Hutt, New Zealand The sensitivity and response speed of the resistance changes of conducting polymer films when exposed to gases is increased if the film is thin and if the measurement of the response excludes the polymer-to-metal contact resistance at the electrodes of the sensor. A method is described whereby the resistance of electrodeposited films can be measured as they are grown, enabling films of pre-determined resistances to be produced. The principal cause of film resistance variation is differences in the time taken to bridge the gap between electrodes. The experiments reported here involve four-terminal measurements, which enable the effect of electrode contact resistance to be estimated and eliminated.The contact resistance between the polymer and metal electrodes is found to be up to 50% of the total resistance of the sensor, and less affected by exposure to a gas than is the polymer resistance. Four-terminal measurements of resistance therefore give the greatest sensitivity. Keywords: Sensor; conducting polymer Introduction The development of biological and chemical sensors based on conducting polymers (CPs) has gained considerable momentum in the past decade.',* The reasons for this relate to the numerous advantages of employing CPs as opposed to the more conventionally used materials. These advantages include the availability of a diverse range of monomer types and synthetic monomer analogues, electrochemical preparation allowing the ready mass production and miniaturization of the sensors, the ease with which biomaterials (enzymes, antibodies and whole cells) are incorporated in the polymer, the ability to change the oxidation state of the polymer after deposition and thereby tailor the sensing characteristics of the film, and the ability to obtain reversible responses at ambient temperatures. Signals from a CP-based gas sensor are typically derived from changes in the resistance of the film upon exposure to the gas of interest.Numerous papers demonstrating the responses of different polymer types to common organic vapours have been published.3-7 The sensors are typically prepared by the direct electrochemical deposition of the polymer onto a sensing substrate, which consists of two parallel platinum or gold tracks, typically defined by microlithography, onto an insulating base such as oxidized silicon.A major problem of repeatability exists in the electrochem- ical preparation of the sensor, and is a direct result of the random nature of the deposition process. A random growth pattern means that sensors grown under identical conditions of time, chemical conditions and current density may show widely different final resistances. Although growth time defines the amount of polymer deposited, it gives no indication of the thickness or amount of polymer used to bridge the gap between adjacent tracks, i.e., that portion of the polymer which elicits the response. In order to overcome this problem, and retain resistance reproducibility, one would have to grow very thick films which tend to have repeatable yet sluggish responses.In order to optimize the amount of polymer deposited onto a sensing substrate, it would be desirable to have methods of monitoring film properties during electrodeposition. Kankara and Kupilag measured inter-electrode resistances of very thick films, i.e., films of low resistance. This paper describes two methods by which gas sensitivity can be optimized. The first is a method of continuously monitoring the resistance of the electrochemically deposited CP film as it grows. The sensitivity is in the megohm range, which allows the point at which the polymer bridges adjacent tracks to be established, and sub- sequent polymer growth to be carried out in a controllable fashion.The improved response of thinner films is shown below. Second, by depositing CP films onto four- rather than two-electrode arrays, we demonstrate that a four-terminal measurement, which largely eliminates electrode contact re- sistance between the metal and the polymer film, results in both greater sensitivity and reproducibility. Since the contact resistance is less affected by gas exposure than is the bulk film resistance, two-terminal measurements underestimate the true polymer sensitivity. Experimental Pyrrole and p-toluenesulfonic acid sodium salt (PTS) were obtained from Aldrich (Milwaukee, WI, USA). Pyrrole was distilled and stored under nitrogen prior to use.Distilled water was used throughout. The sensing substrate consisted of a silicon chip, S mm square, coated with 100 nm of silicon dioxide. Four parallel gold electrodes, 10 ym wide with 10 pm spacing, each 2000 ym long, were deposited onto the chip, using a 20 nm titanium coating as an adhesion layer. The gold electrodes were 0.2 ym thick. The chip was attached to an alumina substrate and wire- bonded to make electrical connection. The bond wires were protected with epoxy (Epon 82S/Jeffamine D230, with fused silica added to control viscosity); the epoxy was also used to cover the on-chip metallization as far as the start of the microband electrodes, leaving the full 2000 ym exposed. After curing the epoxy, the chips were vapour-primed with hex- amethyldisilazane-saturated nitrogen for 2 min, a conventional procedure in micro-fabrication used to render the silicon dioxide hydrophobic. This was found to encourage lateral polymer growth across the surface of the chip, in a way similar to that reported by Nishizawa et.al.9 Poly(pyrro1e) (PPy) films doped with PTS were electro- chemically deposited onto the gold tracks of the sensing substrate. Deposition was carried out by applying either a pulsed potential (750-900 mV versus Ag-AgC1) or a current (density of 1-10 mA cm-2) and monitoring the resistance of the film as it grew. The potentiostat employed was fabricated in- house and was controlled and monitored by a MacLab (ADI, Dunedin, New Zealand) system running chart (ADI) software. All films were deposited from de-oxygenated aqueous solutions of pyrrole (0.1 mol dm-3) and the appropriate dopant ion salt (0.1 mol dm-3).After deposition, the sensors were stored in a sealed container until required. Vapours were generated using the apparatus depicted in Fig. 1, which consisted of a series of mass flow controllers,1350 Analyst, September 1996, Vol. 121 solenoid valves and bubblers. Target vapours (stream B) were produced by bubbling a nitrogen carrier gas stream through a volatile liquid, thus giving a continuous flow of saturated gas, with the vapour concentration being dependent on the vapour pressure of the liquid. A dilution stream (stream A), of constant humidity, was similarly generated by bubbling nitrogen through distilled water and diluting with dry nitrogen to the desired humidity.The two streams were combined at different ratios and exposed to the sensor, while maintaining a constant flow rate over the sensor. The resistance measurements were made by passing a 1 kHz constant current set in the range 1-100 FA through the outer pair of gold tracks of the sensing substrate, and measuring the implied voltage on the inner pair of tracks. The differential voltage signal, always held below 0.1 V to eliminate self- heating effects,'() was amplified and a programmable offset and scale factor then applied before synchronous detection and subsequent low-pass filtering with a 10 Hz bandwidth. In this way small sensor voltages could provide high resolution by eliminating dc potentials and drifts, which might arise from thermal emfs and electronic offsets, and minimizing ac potentials arising from interference and llf noise.Dc currents have been shown to generate significant amounts of llfnoise in CP films.10 The programmable offset and scale were set so that the individual sensor response best utilized the 12 bit ADC range. Two octal ADCs were used to provide the capability of monitoring up to 16 channels simultaneously. A local micro- processor was employed to perform further digital filtering before passing the data (resistance measurements and gas type) back to a PC. Input commands from the PC, sent prior to the sensing experiment, set the filter length, offset, scale and gas control sequences. The parameter extracted was 6R/R, the fractional change in resistance on exposure to gas. Results and Discussion Deposition CP films were grown on the electrode arrays in either constant- current or constant-voltage mode, using the potentiometer/ resistometer depicted in Fig.2. In order to encourage uniform growth, deposition was initiated by shorting the four tracks of the electrode array via a relay to form the working electrode. Some feedback was necessary to ensure equal current sharing between the tracks. The deposition controller then applied a current or voltage pulse to the tracks using a conventional auxiliary and reference electrode set-up. Polymer growth was typically achieved by pulsing the working electrode at 750-900 mV versus Ag-AgC1, or at a current density of 1-10 mA cm-2. Following the pulse, the relay was opened and the inter-track resistance measured in either two- or four-terminal mode to Fig.1 Mixing system to provide a gas stream at a controlled humidity to the sensor manifold. Mass flow controllers are shown by squares, solenoid valves by ovals and bubblers by large circles. S is the 3-way solenoid for which the arrow indicates normally open. detect whether or not contact had been made. This measurement was carried out synchronously using a voltage sufficiently low so as not to interfere with the electrodeposition, in a way that the measured impedance would not include the capacitive shunt through the conducting solution, and so that the measuring circuit was electronically isolated from the growing circuit. Following the detection of a film across the electrode array, deposition could proceed either in pulse or continuous mode, since the synchronous resistance measurement, which is necessary to make meaningful resistance measurements when the film is thin, could be performed simultaneously with the deposition.Conventionally, a resistance measurement is performed by passing a known current through the resistor and measuring the voltage developed. Here, a current was generated which produced a small, automatically selected voltage (< 10 mV) across the sensor tracks, a value low enough not to induce further polymer growth. The combination of variable current and selectable voltage facilitates measurement over a wide range of resistances, as is encountered during deposition, Fig. 3 depicts the growth curves for two PTS-doped PPy films deposited onto two different substrates.Growth was achieved using a constant-current pulse of density 10 mA cm-2, and the voltage measured relative to an Ag-AgCl reference electrode. Although the substrates and the growing solutions were identical in every way, the time taken for a resistance measurement to appear, and hence the time taken for the polymer to bridge the gap between tracks, was 50 s in the first sensor and 95 s in the second. Once bridging was achieved, both devices took approximately the same time (45 s) to reach the A A m Auxilary Reference electrode electrode '------ Sensor Current time I I 11 > Relay excitation > Resistance Fig. 2 Potentiostat (A) and resistometer (B) used to control the fabrication of the CP sensors. The lower panel illustrates the initial operation in which current pulses and resistance measurement alternate until bridging between the sensor tracks occurs.Following bridging, deposition and measurement can proceed simultaneously.Analyst, September 1996, Vol. 121 1351 3. end point of 2 kS2. Reasons for the variation in initiation time probably relate to a combination of the condition of the inter- track silicon surface and the random nature of the polymer growth. Fig. 3 illustrates the ability to grow films to constant resistance despite the large variation in initiation time. Fig. 4 illustrates in more detail an example of the changes in the electrical characteristics as the conductive film develops during a sequence of growth pulses. The initial impedance exceeds 105 Q, before falling to a value of around 300 Q.Fig. 4 ............................................................................................ -0 -0.5 1 0.8 0.6 0.4 3 (0 c .- ki g o - 0.2 i" ..... -8 -4 .... 30 $0 90 120" I I --I Fig. 3 Data for two PPy-PTS sensors, grown to equal resistances, but requiring greatly different deposition times because of the different growth patterns. Polymerizations were carried out under constant-current condi- tions (1 niA cm-*) from aqueous solutions of pyrrole (0.1 inol dm-?) and PTS (0.1 mol dm-3). Solid line, deposition potential; and dashed line, resistance. 5.51 13 5 ................................................................... l----h .......................... / I: also shows the film conductivity and its rate of change. The conductivity rises rapidly around the fifteenth pulsc, but is then followed by a marked transition to a slower rate of increase.This general behaviour is common and from a comparison of photographs of the sensors and conductance records it is thought that the two regions represent different stages of film growth. The initial rising conductance corresponds to the initial bridging and filling of the area between the gold tracks. Following this, the film conductance rises linearly with thickness and so is approximately linearly related to time. Films grown for extended periods show that the rate of conductance increase eventually slows; presumably these are thick films where the outer layers have a decreasing contribution to the inter-track conductance. We have found such films to exhibit poor gas sensitivity. These results illustrate the value of monitoring electrical characteristics during film deposition as an indication of the progress of inter-track growth in order to produce consistent sensors.SEM images show that all the films which exhibit good gas sensitivity are substantially less than 1 pm thick. Given that the inter-track spacing is 10 pm, the inter-track conduction will essentially be in two-dimensions, not three. Film Responses Using the apparatus described above, a number of sensors were produced with different thicknesses of deposited polymer, and exposed to a variety of common organic solvents. As expected, the results showed that very thick films produced comparatively sluggish responses and lower sensitivities, as measured by the fractional change in resistance 6RIR.However, it was found that the sensitivity tended to reach a limiting value as the film thickness decreased, and there was no advantage in producing extremely thin films. Fig. 5 illustrates the response to ethanol of two sensors, one with a thick layer (A) and one with a thin layer (B). The four-electrode array enabled an estimate of contact resistance to be made. Fig. 6 shows a sketch of the three inter- track resistances, (R 1-R3), which frequently are unequal, and 0.03 - 0.025 ~- 0.02 - - 0.015 5 0.01 0.005 0 -0.005 73 -0.01 ' 0 0 0 0 0 0 0 0 0 1 o o w o w o v ) o l - - N C V O m * Tim e/s Fig. 5 Response of thick (A) and thin (B) PPy-PTS polymer films to saturated ethanol vapours in dry nitrogen at ambient temperature (flow rate of 200 cm3 min-I). Sensor A has a resistance of 10 Q and sensor B a resistance of 50 Q.1 2 3 4 Fig. 6 Model of resistance components in the four-element array. With the assumption that the polymer/metal contacts are the same, components can be calculated from the two-terminal measurements.1352 Analyst, September 1996, Vol. 121 the four contact resistances (Rc). With the assumption that all the contact resistances are the same, the film and contact resistances found from two-point resistance measurements between tracks R2 and Rc were determined for five sensors exposed to ethanol and hexane for a period of 26 min. The results, plotted as the fractional change in the resistance component, are shown in Fig. 7. The contact resistance in most instances is affected by gas exposure, but in all the ethanol responses and most of those for hexane, the fractional change in the contact resistance is less than that in the polymer component. Hence, the over-all sensitivity will be reduced by the inclusion of the contact resistance component, the exact amount depending on the relative sizes of the contact and polymer resistors.Table 1 lists the estimated film and contact resistances for the five sensors exposed to ethanol, together with the increase in Sensor 1 Ti me/min Fig. 7 Estimates of fractional change in the film and contact resistances (based on the model of Fig. 6) for five sensors when exposed to 25% ethanol (A) and hexane (B) vapours in dry nitrogen at ambient temperature. +, Contact K; H, polymer R.Table 1 Estimate of film resistance (R2) and contact resistance (Rc) extracted from five PPy-PTS sensors. The bottom line shows the increased sensitivity obtained when the film resistance alone, rather than total resistance. is used Sensor number 1 2 3 4 5 R2/Q 328 1015 550 346 3.52 RClQ 68 63 119 105 160 Increase in sensitivity to ethanol (%) 44 33 32 55 47 sensitivity implied by the data of Fig. 7 if the film resistance change, rather than the total resistance change, is employed. The contact resistances are fairly large and must to some extent depend on film thickness (since it is possible to grow thick films with lower total resistances than the values derived from these sensors). Model calculations show that the long and narrow microelectrode gold tracks contribute about I0 C2 to the value of Rc.While not insignificant, in general it accounts for only 10% of the estimated contact resistance. The total contact resistance in these examples is a significant fraction of the total sensor resistance, and the gain in sensitivity obtained by eliminating it is fairly marked. Note that the film resistance R2 derived here from a series of two-terminal measurements will not be the same as that obtained by a four- terminal measurement, because the current paths through the polymer are different. However, both eliminate the contact resistance. Fig. 8 shows the responses of three sensors to differing concentrations of ethanol vapours as both four (A) and two (B) terminal resistance measurements. Data shown are the frac- tional resistance changes to a succession of vapour pulses, each lasting 10 s, followed by a 100 s nitrogen purge.Responses are shown displaced in time for clarity. The four-terminal measure- ments give a sensitivity increase of 20-30% here, but perhaps more significantly, give less spread in responses at a particular concentration. The responses to a 90 ppm vapour concentration by two-terminal measurement are 0.06, 0.075 and 0.04 for the three sensors. The corresponding values for four-terminal measurement are 0.078, 0.078 and 0.058. The greater con- sistency is considered to reflect the changes in the true gas/ polymer response, not masked by contact resistance effects. Conclusions Some difficulties still exist in applying CPs to odour sensing. One such area is the production of reproducible sensors with optimized sensitivities.In this paper, two methods whereby improvements can be made have been considered. First, a method of monitoring the resistance of the CP film as it is deposited has been described, which enables thin films of high sensitivity to be routinely produced. Second, it has been -0.02 -1 g 0 0 0 0 0 0 0 0 0 0 0 0 0 % 0 0 0 0 0 0 0 0 0 0 0 0 - - c v c v c v c * l w ( o m o , 2 ? Z c Q O N * -0.02 I 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 h l h ( C \ I c v c v m n I ~ ( o ~ ~ ~ ~ ~ ~ O Time/s Fig. 8 Comparison of the responses of three sensors to varying concentrations of ethanol vapour in dry nitrogen (numbers indicate concentration in mg dm-?). The measurements were made as both four- terminal (a), and two-terminal (h) measurements at ambient temperature.Analyst, September 1996, Vol. 121 1353 demonstrated that contact resistance is a significant contributor to the total sensor resistance, and this contact resistance degrades the sensitivity. Significant improvements in sensitivity and consistency can be made by the use of four-terminal resistance measurements, which are insensitive to contact resistance. References 1 Neaves, P. I., and Hatfield, J. V., Sens. Actuators, 1995, B26-27, 223. 2 Peace, T. C., Gardner, J. W., Friel, S., Bartlett, P. N., and Blair. N., Anulyst, 1993. 118, 371. 3 Bartlett, P. N., Archer, P. €3. M., and Ling-Chung, S. K., Sens. Actuators, 1989, 19, 125. 4 Bartlett, P. N., and Ling-Chung, S. K., Sens. Actuators, 1989, 19, 141. 5 6 7 8 9 10 Bartlett, P. N., Archer, P. B. M., and Ling-Chung, S. K., Sens. Actuators, 1989, 20, 287. Blanc, J. P., Derouiche, N., El Hadri, A., Germain, J. P., Maleysson, C., and Robert, H., Sens. Actuators, 1990, B1, 130. Charlesworth, J. M., Partridge, A. C., and Garrard, N., J. Phys. Chem., 1993,97, 5418. Kankara, J., and Kupila, E-L., J. Electroanal. Chem., 1992, 322, 167. Nishizawa, M., Shibuya, M., Sawaguchi, T., Matsue, T., and Uchida, I., J . Phys. Chem., 1991, 95, 9042. Harris, P. D., Arnold, W. M., Andrews, M. K., and Partridge, A. C., Sens. Actuators, submitted for publication. Paper 6102637K Received April 16, 1996 Accepted May 28,1996
ISSN:0003-2654
DOI:10.1039/AN9962101349
出版商:RSC
年代:1996
数据来源: RSC
|
47. |
Cumulative author index |
|
Analyst,
Volume 121,
Issue 9,
1996,
Page 1355-1358
Preview
|
PDF (1076KB)
|
|
摘要:
Analyst, September 1996, Vol. 121 1355 CUMULATIVE AUTHOR INDEX JANUARY-SEPTEMBER 1996 Abdel-Aziz, Mohamed Shafei, Abraham, Michael H., 5 11 AbramoviC, Biljana F., 401, 425 AbramoviC, Borislav K., 401, 425 Abroskin, Andrei G., 419 Acedo Valenzuela, M. I., 547 Adam, S., 527 Adams, Freddy, 1061 Adeloju, S. B., 699 Aherne, G. Wynne, 329 Ahonen, Ilpo, 1253 Akhtar, M. Humayoun, 803 Al-Othman, Rashed, 601 Alazard, S., 527 Aldridge, Paul K., 1003 Alegret, S., 959 Aleixo, Luiz M., 559 Alexandrov, Yu. I., 1137 Almirdll, J., 959 Analytical Methods Committee, Andrade, Francisco J., 6 13 Andrews, M. K., 1355 Angeletti, R., 229 AntonijeviC, M. M., 255 Appleton, Mark, 743 Aratake, Sachiko, 325 Araujo, Pedro W., 581 Arias, J. J., 1327 Arias, Juan JosC, 169 Artjushenko, Slava, 789 Bacci, Mauro, 553 Baggiani, Claudio, 939 Balasubramanian, N., 647 Bannon, Thomas, 715 Barlett, Philip N., 715 Bamabas, Ian J., 465 Barroso, C.G., 297 Banvick, Vicki J., 691 Baxter, Douglas C., 19 Baxter, Pamela J., 945 Baya, Maria P., 303 Benedetti, A. V., 541 Benmakroha, Yazid, 521 Bentsen, Ragne K., 1191 Biancotto, G., 229 Bilitewski, Ursula, 119, 863, 877 Birch, David J. S., 905 Birch, M. Eileen, 1183 Bjorklund, Erland, 19 Blais, Jean-Simon, 483 Blanco, Marcelo, 395 Bogan, Declan R., 243 Bond, Alan M., 357 Borah, Lakhimi, 987 Borowiak, Annette, 1247 Boswell, Stephen M., 505 Bouhsain, Zouhair, 635 Boukortt, Sheriffa, 663 Boutelle, Martyn G., 761 Boyd-Boland, Anna A., 929 Boyd, Damien, 1R Branica, Marko, 1127 Brereton, Richard G., 441, 575, 581,585,651, 993 Brinkman, Udo A.Th., 61, 1069, 1327 Brown, R. H., 1171 Brown, Richard C., 1241 Brfi, E. R., 297 Buchet, Jean-Pierre, 663 Burgot, Jean-Louis, 43 Bye, Ragnar, 201 Cai, Xiaohua, 965 Callejon Mochdn, M., 681 Cammann, K., 527 Campillo, Natalia, 1043 Cao, Zhong, 259 1079 573,889 Carbonnelle, Philippe, 663 Cardoso, A. A., 541 Cardwell, Terence J., 357 Carmona, Pedro, 105 Cary, Robert A., 1183 Casajcis, Rocio, 813 Casella, fnnocenzo G., 249 Cassidy, Richard M., 839 CaviC-Vlasak, Biljana A., 53R Cazemier, Geert, 11 11 Cela, R., 297 Cepas, Juana, 49 Ceramelli, Giuseppe, 219 Cerda, A., 13 Cerdi, V., 13 Chan, Wing Hong, 531 Chatergoon, Lutchminarine, 373 Chen, Guo Nan, 37 Chiou, Chyow-San, 1107 Chou, Shu-Fen, 71 Christian, Gary D., 601 Christie, Ian, 521 Cirovic, Dragan A., 575, 581 Coello, Jordi, 395 Cole, S.Keith, 495 Collier, Wendy, 877 Copeland, D. D., 173 Corbella Tena, R., 459 Corti, Piero, 219 Cosano, J., 83 Craston, Derek H., 177 Crosby, Neil T., 691 Croteau, Louise G., 803 Crump, Paul W., 871 Crumrine, David S., 567 Cruz Ortiz, M., 1009 CuculiC, Vlado, 1127 Cullen, Michael, 75 Cullen, William R., 223 Daae, Hanne Line, 1191 Daenens, Paul, 857 Daghbouche, Yasmina, 103 1 Dalene, Marianne, 1095, 1101 de Jong, Dirk, 61 de Jong, Gerhardus J., 61 de la Guardia, M., 1327 de la Guardia, Miguel, 635, 923, de Lacy Costello, Benjamin P. J., de Oliveira Neto, Graciliano, 559 De Saeger, Emile, 1247 Dean, John R., 465, 85R Demidova, M. G., 489 Demir, Cevdet, 65 1, 993 Deng, Jiaqi, 97 1 Deng, Qing, 1123 Deng, Zhiping, 671, 1341 Desai, Mohamed, 521 Desimoni, Elio, 249 Destradis, Angelo, 249 Devi, Surekha, 807 Dey, Nibaran C., 987 Dilleen, John W., 755 Dobrowolski, R., 897 Dodd, Matthew, 223 Dolmanova, Inga F., 43 1 Dong, Shaojun, 1123 Dreassi, Elena, 219 Dumasia, Minoo C., 651 Dumschat, C., 527 Dunemann, Lothar, 845 Dunhill, Roger H., 1089 Economou, Anastasios, 97, 1015 Eduard, Wijnand, 1191, 1197 Eigendorf, Guenter K., 223 Eikenberg, Oliver, 119 El-shahat, Mohamed F., 89 El-Shorbagi, Abdel-Nasser, 183 103 1 793 Elbergali, Abdallah K., 585 Eller, Peter M., 1163 Ellwood, Jo A., 575 Emara, Samy, 183 Emteborg, H&an, 19 Endo, Masatoshi, 391 Eng, Jimmy, 65R Escobar, Rosario, 105 Essers, Martien, 11 11 Evans, Phillip, 793 Fabriks, Jean-Franqois, 1257 Facer, M., 173 Fallon, Michael G., 127 Fang, Kai-Tai, 1025 Fawaz Katmeh, M., 329 Fearn, Tom, 275 Fell, Gordon S., 189 Fernandes, Julio Cesar B., 559 Ferreira, Valdir S., 263 Fielden, Peter R., 97, 1015 Fillenz, Marianne, 76 1 Fiore, Amy A., 1265 Fischbach, Thomas J., 1163 Fitzgerald, Catherine A., 715 Fleet, Ian A., 55 Forsberg, Bertil, 1261 Forster, Robert J., 733 Forteza, R., 13 Francis, John M., 177 Frank, Gerhard, 1301 Frech, Wolfgang, 19, 1055 Fugivara, C.S., 541 Fukasawa, Tsutomu, 89 Fung, Yingsing, 369 GaB1, Ferenc F., 401,425 Gala, BelCn, 1133 Galeano Diaz, T., 547 Gamble, Donald S., 289 Gao, Xiao Xia, 687 Garcia-Fraga, J. M., 1327 Garcia, M., 959 Garrigues, Salvador, 635, 923, Gebefugi, Istvan, 1301 Genrich, Meike, 877 Georgieff, Michael, 901 Ghosh, Anil G., 987 Giannousios, A., 413 Giersch, Thomas, 863 Giraudi, Gianfranco, 939 Glennon, Jeremy D., 127 Godinho, Oswaldo E.S., 559 Goldstein, Steven L., 901 Gomez-Hens, Agustina, 1133 Gong, Zhilong, 1119 Gooijer, Cees, 1069 Goosens, Elise C., 61 Gordon, Derek B., 55 Gomer, Peter, 1257 Goto, Nobutake, 1085 Greer, James C., 715 Grol, Michael, 119 Groves, John A., 441 Guiberteau Cabanillas, A., 547 Guiraum Pkrez, A., 681 Gurden, Stephen. P., 441 Gustavsson, C. A., 1285 Haasnoot, Willem, 1 11 1 Hadjiivanov, K., 607 Haferkamp, Heinz, 129 1 Hafkenscheid, The0 L., 1249 Hagenbjiirk-Gustafsson, Annika, Halgard, Kristin, 1191 Halliwell, David J., 1089 Hammerich, Ole, 345 Hangartner, Markus, 1269 Hansen, Elo H., 31 Hansen, Erik Beck, 1291 1031 1261 Harper, Martin, 1265 Harris, P., 1355 Harris, Roy, 913 Harrison, Iain, 189 Hart, Barry T., 1089 Hartnett, M., 749 Hasan, B. A., 1327 Hauser, Peter C., 339 Hayashi, Yuzuru, 591 Hayashibe, Yutaka, 7 Hays, Lara, 65R Heeremans, Carola E.M., 1273 Hemingway, Michael A., 1241 Hendrix, James L., 799 Hemandez-Cdrdoba, Manuel, Hernandez, Oscar, 169 Hestvik, Gete, 1261 Hietel, Bernhard, 1301 Hindmarch, Peter, 993 Hoekstra-Oussoren, Sacha J. F., Honing, Maarten, 1327 Hu, Yan, 883 Hulanicki, Adam, 133 Hyland, Mark, 705 Ibrahim, Naaim M. A., 239 Idriss, Kamal A., 1079 Jnagawa, Jun, 623 Iiiiguez, Montserrat, 1009 Ioannou, Pinelopi C., 909 Irwin, G. W., 749 Ishida, Yasuyuki, 853 Ishihara, Masahito, 391 Isomura, Shinichi, 853 Iturriaga, Hortensia, 395 Ivanova, Elena K., 419 Iwatsuki, Masaaki, 89 Jackson, Laurence S., 67 Jager, Maria E., 1327 Jaselskis, Bruno, 567 Jiang, Chongqiu, 317 Jiang, Wei, 1317 Jimenez, A.I., 1327 JimCnez, Ana Isabel, 169 Jimenez, F., 1327 JimCnez, Francisco, 169 JimCnez-Prieto, Rafael, 563 JimCnez Sanchez, J. C., 681 Johnson, Mark, 1075 Jonsson, B. A. G., 1279, 1285 Jurkiewicz, M., 959 Kalish, N. K., 489 Karayannis, Miltiades I., 435 Karlsson, Dons, 1261 Karlsson, Lars, 19 Kennedy, Eugene R., 1163 Kenny, Lee C., 1233 Kettling, Ulrich, 863 Kettrup, Antonius, 130 1 Khalaf, K. D., 1327 Kimbrough, David Eugene, 309 Kimoto, Takashi, 853 Kindness, Andrew, 205 Kirchner, Manfred, 1269 Knoll, M., 527 Kolotyrkina, Irina Ya., 1037 Konstantianos, Dimitrios G., 909 Korda, T. M., 489 Kozik, Andrzej, 333 Kratochvil, Byron, 163 Kuznetsova, Vera V., 419 Kvasnik, Frank, 1 1 15 Kwong, Daniel W. J., 531 Lan, Zhang-Hua, 21 1 Lancashire, Susan, 75 Lancia, Antonio, 789 Laurie, David, 951 Lawrence, Chris M., 755 1043 13271356 Analyst, September 1996, Vol.121 Le, Quyen T. H., 1051 Lee, Albert W. M., 531 Legouin, Beatrice, 43 Lei, Chenghong, 971 Levin, Jan-Olof, 1177, 1273 Lewenstam, Andrzej, 133 Li, Hao, 223 Liang, Yi-zeng, 1025 Lightbody, G., 749 Lin, Hui-Gai, 259 Lindahl, Roger, 1177, 1273 Lindh, C. H., 1285 Lindskog, Anne, 1295 Link, Andrew J., 65R Lipkovska, N. A., 501 Lison, Dominique, 663 Littlejohn, David, 189 Lonardi, S., 219 Lopes, Teresa I. M. S., 1047 Lopez Carreto, Maria, 33R Lopez-Cueto, Guillermo, 407 Lopez-Erroz, Carmen, 1043 Lopez, Martin, 905 Lord, Gwyn A., 55 Loakas, Yannis L., 279 Lowry, John P., 761 Lowthian, Philip, 743, 977 Lowy, Daniel A., 363 Lu, Bin, 29R Lu, Changyin, 883 Lu, Xiao-Quan, 1019 Lu, Zheng, 163 Lund, Walter, 201 Luo, Yongyi, 601 Luque de Castro, M.D., 83 Lyons, Michael E. G., 715 McAdams, Eric T., 705 McAlernon, Patricia, 743 McAteer, Karl, 773 McCormack, Ashley L., 65R MacCraith, Brian D., 785, 789 McDonagh, Colette M., 785 McEvoy, Aisling K., 785 McKelvie, Ian D., 1089 MacLachlan, John, 11R McLaughlin, James A., 705 McNaughtan, Arthur, 11R Madsen, Gary L., 567 Magdic, Sonia, 929 Maines, Andrew, 435 Maj-Zurawska, Magdalena, 133 Malahoff, Alexander, 1037 Mannaert, Erik, 857 Marr, Iain L., 205 Marshall, William D., 289, 483, Mhrtensson, Maud, 1 177 Martin, Patricia, 495 Martinez-FBbregas, E., 959 Martinez-Lozano, Carmen, 477, Mason, Andrew J., 951 Maspoch, Santiago, 395, 407 Masujima, Tsutomu, 183 Mathiasson, Lennart, 19 Matsuda, Rieko, 59 1 Matsui, Masakazu, 105 1 Meaney, Mary, 789 Melbourne, Paul, 1075 Melios, Cristo B., 263 Mieczkowski, Jozef, 133 Mierzwa, J., 897 Mihajlovic, R., 255 MilaEiE, Radmila, 627 Mills, Andrew, 535 MilosavljeviC, Emil B., 799 MitroviL, Bojan, 627 Mizgunova, Ulyana M., 43 1 Mo, Jin-Yuan, 1019 Mo, Songying, 369 Moane, Siobhan, 779 Mocak, Jan, 357 Mohamed, Ashraf A., 89 Molina, Marina, 105 Monaf, Lela, 535 Monaghan, John J., 55 Montelongo, F.Garcia, 459 817 813 Moollan, Roland W., 233 Moore, Andrew, 67 Morales-Rubio, A., 1327 Mosello, R., 83 Motomizu, Shoji, 1085 Mottola, Horacio A., 21 I , 381 Mounsey, Andrew, 955 Mowrer, Jacques, 1249, 1295 Mulcahy, David, 127 Muller, Beat, 339 Munro, C. H., 835 Murphy, William S., 127 Nakamura, Masatoshi, 469 Nakamura, Motoshi, 469 Nakanishi, Masami, 853 Newton, R., 173 Nie, Lihua, 883 Nielsen, Steffen, 31 Nolte, Joachim, 845 Noreiia-Franco, Luis E., 11 15 Norris, Timothy, 1003 Not@, Hilde, 1191 Nygren, Olle, 1291 Obendorf, Dagmar, 35 1 ObradoviC, Danilo M., 401 Odman, Fredrik, 19 Ohtani, Hajime, 853 O’Keeffe, Michael, 779, 1R O’Kennedy, Richard, 243, 767, O’Lear, Christina, 1265 Olmi, Filippo, 553 Olsen, Erik, 1155 Oms, M.T., 13 O’Neill, Robert D., 761, 773 Oniciu, Liviu, 363 Oosten, Koos van, 1273 Orlando, Andrea, 553 Oshima, Mitsuko, 1085 Osipova, Natdliya V., 419 Ostaszewska, Joanna, 133 Owen, Susan P., 465 Packham, Andrew J., 97, 1015 Papadopoulos, C., 413 Paradowski, Dariusz, 133 Pardue, Harry L., 385 Park, Chang J., 1311 Parsons, Patrick J., 195 Partridge, A.C., 1355 Patel, Sunil U., 913 Patterson, Knstine Y., 983 Paulls, David A., 831 Pawliszyn, Janusz B., 929 Pedrero, Maria, 345 Perez-Bendito, Dolores, 49, 563, Perez-Bustamante, J. A., 297 PCrez-Ponce, Amparo, 923 PCrez-Ruiz, Tomas, 477, 8 13 Pergantis, Spiros A., 223 Perruccio, Piero Luigi, 219 Pfaffli, P., 1279, 1285 Piggott, Nighel H., 951 Pihlar, Boris, 627 Pingarron, Jose, 345 Piperaki, Efrosini A., 11 1 Piro, R. D. M., 229 Pitre, K. S., 79 Poe, Russell B., 591 Poole, Colin F., 5 11 Potter, Annika, 1295 Power, J. F., 451 Prodromidis, Mamas I., 435 Proinova, I., 607 Proskurnin, Mikhail A., 419 Pui, David Y. H., 1215 Puster, Thomas, 1291 Pyrzynska, Krystyna, 77R Qi, Zhong-Cheng, 1317 Qu, Yi Bin, 139 Quevauviller, Ph., 83 Quinn, John G., 767 Rader, W. Scott, 799 Rae, Bruce, 233 Raghunath, A.V., 825 Rahmani, Ali, 585 Ramachandran, Gurumurthy, 1225 29R 1133,33R Ramanaiah, G. V., 825 Rangel, Antonio 0. S. S., 1047 Ratcliffe, Norman M., 793 Razee, Saeid, 183 Redon, Miguel, 395 Regan, Fiona, 789 Reimer, Kenneth J., 223 Reinartz, Heiko W., 767 Rigby, Geraldine P., 871 Riipinen, Hannu, 1253 Rios, Angel, 1 Rodriguez Delgado, M. A., 459 Rodriguez-Medina, JosC F., 407 Rohm, Ingrid, 877 Roos, Aappo, 1253 Rowell, Frederick J., 95 1, 955 Rowell, Vibeke, 955 Rozendom, Eduard J. E., 1069 Rubio, Soledad, 33R Ruzicka, Jaromir, 601, 945 Sadler, Peter J., 913 Sakslund, Henning, 345 Salden, Martin, 1 1 11 Saleh, Gamal A., 641 Salinas, F., 547 San Martin Fernandez-Marcote, Sanchez, Ma, J., 459 Sandstrom, Thomas, 1261 Santamaria, Fernando, 1009 Santos, Jose H., 357 Sanz, Antonio, 477 Sarabia, Luis A., 1009 Sartini, Raquel P., 1047 Sasaki, Takayuki, 105 1 Sato, Hidetoshi, 325 Satyanarayana, K., 825 Sayama, Yasumasa, 7 Schafer, E.A., 243 Schieltz, David, 65R Schmid, Rolf D., 863 Schnelle, Jurgen, 1301 Schoeps, Karl-Olof, 1203 Schoppenthau, Jorg, 845 Scobbie, Emma, 575 Scudder, Kurt, 945 Sedaira, Hassan, 1079 Seebaum, Dirk, 1291 Seibert, Donna S., 51 1 Sekino, Tatsuki, 853 Seviour, John, 95 1 Shah, Rupal, 807 Shanthi, K., 647 Shi, Renbing, 1311 Shih, Jeng-Shong, 1107 Shijo, Yoshio, 325 Shiraishi, Haruki, 965 Shpigun, Lilly K., 1037 Shukla, Jyotsna, 79 Shulman, R. S., 489 Shulman, Stanley A., 1163 Si, Zhi-Kun, 1323 Sihvonen, Marja-Liisa, 1335 Sillanpaa, Mika, 1335 Silva, Manuel, 49, 563 Siskos, Panayotis A., 303 Skarping, Gunnar, 1095, 1101 Slater, Jonathan M., 743, 755 Slavin, Walter, 195 Slobodnik, Jaroslav, 1327 Sloth, Jens J., 31 Smith, Clayton, 373 Smith, Dennis C., 53R Smith, Robert F., 67 Smith, Roy, 321 Smith, W.E., 835 Smyth, Malcolm R., 779, 1R, 29R Smythe-Wright, Denise, 505 Snell, James P., 1055 Sokalski, Tomasz, 133 sole, s., 959 SolujiC, Ljiljana, 799 Somsen, Govert W., 1069 Song, Ruiguang, 1163 Sorvari, Jaana, 1335 Spanne, MArten, 1095, 1101 Spear, Terry M., 1207 M., 681 Stathakis, Costas, 839 Stegman, Karel H., 61 Stegmann, Werner, 90 1 Stein, Kathrin, 13 1 1 Stevenson, Derek, 329 Stone, David C., 671, 1341 Stouten, Piet, 11 11 Strachan, David, 951, 955 Stradiotto, Nelson R., 263 Streppel, Lucia, 1 11 1 Stuart, Iain A., 11R Stubauer, Gottfried, 35 1 Subramaniam, K., 825 Suffet, I.H. ‘Mel’, 309 Sukhan, V. V., 501 Suliman, Fakhr Eldin O., 617 Sultan, Salah M., 617 Sumodjo, P. T. A., 541 Susanto, Joko P., 1085 Svanberg, Per-Arne, 1295 Sweedler, Jonathan V., 45R Symington, Charles, 1009 Szklar, Roman S., 321 Tam, Wing Leong, 53 1 Tan, Yanxi, 483 Tang, Bo, 317 Tang, Shida, 195 Tegtmeier, M., 243 TepavCeviC, Sanja D., 425 Thastrup, Ole, 945 Thomaidis, Nikolaos S., 11 1 Thomassen, Yngvar, 1055 Thompson, Michael, 275,285, 671,977, 1341,53R Thornes, R. D. , 243 Thorpe, Andrew, 1241 Tian, Baomin, 965 Timperman, Aaron T., 45R Tinnerberg, H&an, 1095, 1 10 1 Tomas, Virginia, 477, 813 Torgov, V. G., 489 Townshend, Alan, 83 1 Troccoli, Osvaldo E., 613 Tsuge, Shin, 853 Tsurubou, Shigekazu, 105 1 Tudino, Mabel B., 613 Tyson, John D., 95 1,955 Tzouw ara-Karay anni, Stella M., Ubide, Carlos, 407 Uehara, Nobuo, 325 Umetani, Shigeo, 1051 Vadgama, Pankaj, 435,521,871 Vaggelli, Gloria, 553 Valcarcel, Miguel, 1 , 83 van Baar, Ben L.M., 1327 Van Mol, Willy, 1061 van Wichen, Piet, 11 11 Vassileva, E., 607 Veillon, Claude, 983 Velthorst, Nel H., 1069 Verbeek, Alistair, 233 Viles, John H., 913 Villegas, Nuria, 395 ViAas, Pilar, 1043 Vincent, James H., 1207, 1225 Vos, Johannes G., 789 VukanoviC, B., 255 Wahlberg, Sonny. 1261 Wake, Derrick, 1241 Walker, P. J., 173 Wallace, G. G., 699 Walsh, James E., 789 Walsh, Peter T., 575 Wang, Bin-Feng, 259 Wang, Chen, 317 Wang, Jin, 289, 817 Wang, Joseph, 345, 965 Wang, Ke-Min, 259, 531 Wang, Nai-Xing, 1317 Wang, Shi-Hua, 259 Watanabe, Kazuo, 623 Welinder, H., 1279, 1285 Werner, Herbert, 1269 Werner, Mark A., 1207, 1225 WessCn, Bengt, 1203 Wheals, Brian B., 239 435Analyst, September 1996, Vol.121 1357 White, P. C., 835 Whiting, Robin, 373 Wickstrom, Torild, 201 Wilmot, John C.. 799 Witschger, Olivier, 1257 Wittmann, Christine, 863 Wolf, Kathrin, 130 1 Wood, Roger, 977 Woolfson, A. David, 7 11 Wu, Weh S., 321 Xin, Wen Kuan, 687 Xu, Xue Qin, 37 Xu, Yuanjin, 883 Yamada, Shinkichi, 469 Yan, Xiu-Ping, 1061 Yao, Shouzhuo, 883 Yates 111, John R., 65R Yu, Ru-Qin, 259 Zagatto, Elias A. G., 1047 Zanker, Kurt, 767 Zanoni, Maria Valnice B., 263 Zaporozhets, 8. A., 501 Zhang, Fan, 37 Zhang, Xiaogang, 3 17 Zhang, Zhanen, 971 Zhang, Zhujun, 11 19 Zhi, Zheng-liang, 1 Zhou, Dao-Min, 705 Ziegler, Torsten, 119 Zolotova, Galina A., 43 1 Chemicals = Business Opportunities in Africa International Technology and Joint Venture Forum: Dar-Es-Salaam, Tanzania, 5-8 November 1996 The Technology Exchange Ltd.in Bedford, in co-operation with the United Nations and the Tanzanian Government have compiled a detailed compendium of the partnership and technology requests from 230 Tanzanian firms. This is available free to organisations willing to co-operate with a Tanzanian firm. The requests cover the following industries: light engineering, textiles and leather, wood products, medical and pharmaceuticals, construction and building materials, chemicals, agriculture and food processing, infrastructure and transport, tourism. If you can supply technology assistance, training, machines or investment in any of these areas then the Technology Exchange will include your offer free of charge in a UNIDO Compendium of offers which will be circulated to Tanzanian industry before November - entries must be received by 10 September.UNIDO are offering FREE booths at the exhibition in Dar-Es-Salaam to all UK firms who respond in time, and some travel grants may also be available (funds permitting). For entry forms, you should write immediately to: The Technology Exchange Ltd., Wrest Park, Silsoe, Bedford MK45 4HS. Fax: 01 525-860664 E-mail: tech-ex @ dial. pipex.comAnalyst, September 1996, Vol. 121 1357 White, P. C., 835 Whiting, Robin, 373 Wickstrom, Torild, 201 Wilmot, John C.. 799 Witschger, Olivier, 1257 Wittmann, Christine, 863 Wolf, Kathrin, 130 1 Wood, Roger, 977 Woolfson, A.David, 7 11 Wu, Weh S., 321 Xin, Wen Kuan, 687 Xu, Xue Qin, 37 Xu, Yuanjin, 883 Yamada, Shinkichi, 469 Yan, Xiu-Ping, 1061 Yao, Shouzhuo, 883 Yates 111, John R., 65R Yu, Ru-Qin, 259 Zagatto, Elias A. G., 1047 Zanker, Kurt, 767 Zanoni, Maria Valnice B., 263 Zaporozhets, 8. A., 501 Zhang, Fan, 37 Zhang, Xiaogang, 3 17 Zhang, Zhanen, 971 Zhang, Zhujun, 11 19 Zhi, Zheng-liang, 1 Zhou, Dao-Min, 705 Ziegler, Torsten, 119 Zolotova, Galina A., 43 1 Chemicals = Business Opportunities in Africa International Technology and Joint Venture Forum: Dar-Es-Salaam, Tanzania, 5-8 November 1996 The Technology Exchange Ltd. in Bedford, in co-operation with the United Nations and the Tanzanian Government have compiled a detailed compendium of the partnership and technology requests from 230 Tanzanian firms. This is available free to organisations willing to co-operate with a Tanzanian firm. The requests cover the following industries: light engineering, textiles and leather, wood products, medical and pharmaceuticals, construction and building materials, chemicals, agriculture and food processing, infrastructure and transport, tourism. If you can supply technology assistance, training, machines or investment in any of these areas then the Technology Exchange will include your offer free of charge in a UNIDO Compendium of offers which will be circulated to Tanzanian industry before November - entries must be received by 10 September. UNIDO are offering FREE booths at the exhibition in Dar-Es-Salaam to all UK firms who respond in time, and some travel grants may also be available (funds permitting). For entry forms, you should write immediately to: The Technology Exchange Ltd., Wrest Park, Silsoe, Bedford MK45 4HS. Fax: 01 525-860664 E-mail: tech-ex @ dial. pipex.com
ISSN:0003-2654
DOI:10.1039/AN9962101355
出版商:RSC
年代:1996
数据来源: RSC
|
|