摘要:

Analyst, February 1996, Vol. 121 (239-242) 239 Determination of Alkylphenol Ethoxylate Non-ionic Surfactants in Trade Effluents by Sublation and High-performance Liquid Chromatography Naaim M. A. Ibrahim and Brian B. Wheals* Chemistry Department, Brunel University, Uxbridge, Middlesex, UK UB8 3PH Alkylphenol ethoxylate (APE) non-ionic surfactants can be extracted from trade effluents into ethyl acetate by salting out and sublation (gas stripping). The dried extracts are rapidly evaporated in a Kuderna-Danish evaporator and the residue is dissolved in acetonitrile. Finally, co-extractives are precipitated out by chilling to -5 "C and the surfactant is determined in the supernatant by HPLC with fluorescence detection. An internal standard consisting of a single APE oligomer (tert-octylphenol with a nine ethoxylate unit side-chain) compensates for incomplete recovery of the surfactants.The method is applicable to APEs at concentrations of 0.1-50 mg 1-l. Keywords: Alkylphenol ethoxylate surfactants; trade efJluents; sublation extraction; high-performance liquid cn hromatograp hy Introduction Environmental concerns over the use of alkylphenol ethoxylate (APE) non-ionic surfactants are leading to a gradual phasing out of these compounds. 1.2 Nevertheless, their occurrence in trade and sewage effluents is attracting considerable interest as there are suspicions that breakdown products derived from APEs may be a significant source of the estrogenic activity now known to be present in some discharges to rivers.3 In trade effluents in particular, the presence of fats, soaps and other anionic and cationic surfactants, necessitates lengthy and complex clean-up procedures before the APE surfactants can be satisfactorily determined.4 The work that we have undertaken in developing the described method was aimed at simplifying the analysis, by exploiting an unusual HPLC procedure based on columns packed with Spherisorb silica and eluents containing water and acetonitrile.This procedure retains APEs by interaction be- tween the ethoxylate moiety of the surfactant and silanol groups of the silicas, probably by a hydrogen-bonding mechanism.5 Cationic and anionic surfactants are not retained under the conditions used and do not interfere with the APE separation. A variety of methods have been used to determine non-ionic surfactants in waste water and Andrew,6 in describing an FTIR spectrometric method, cites a selection of appropriate proce- dures.Generally, HPLC has become the method of choice for characterizing APEs in water extracts as it has the capability of separating individual surfactant oligomers from the complex mixtures found in commercial products. A recent paper7 provided a review of the many published HPLC procedures used for separating APEs. In general, normal-phase chromato- graphy on amino-, cyano- or diol-bonded silica or silica itself has been used to achieve oligomer separations. Under reversed- * To whom correspondence should be addressed. phase conditions, columns packed with Clg-silica gave no oligomer separation, but Wang and Fingass obtained a good separation on a C1-silica.Our studies have shown that with Spherisorb silica, eluents containing 80-50% of pH 3 buffer in a mixture with acetonitrile provide excellent oligomer separa- tions. Such highly aqueous eluents avoid the problems asso- ciated with normal-phase chromatography such as deactivation due to water injected with the sample, and also cause ionic surfactants to elute at the solvent front. Although APEs display UV absorbance, a significant im- provement in sensitivity is obtainable by using fluorescence for their monitoring.9 The fluorescence is associated with the aromatic ring in the surfactants, and is not unique to APEs, but under the elution conditions described no co-elution of fluorescent co-extractives occurs. A major advantage of fluorescence monitoring over UV absorbance monitoring for examining trade effluent extracts is that more dilute solutions can be injected and non-specific absorbance avoided.Gas stripping (sublation)IoJ 1 has become the accepted way to remove selectively surfactants from waste waters by concentrat- ing them in an ethyl acetate layer. With trade effluents, however, variable recoveries are obtained owing to emulsion formation and the high levels of co-extractives present. In the described method, addition of a single APE oligomer with nine ethoxylate units (9EO) permits the recovery to be determined. In combination, the unusual chromatographic conditions, fluorescence detection and the use of an interactive internal standard permit the analysis of very dirty trade effluents with minimal clean-up. There is virtually complete recycling of the ethyl acetate extractant and reliable analytical data are ob- tained.Experimental Equipment A Varian Star liquid chromatograph was used in conjunction with a Jasco FP 920 fluorescence detector. A 12.5 cm X 4.5 mm id stainless-steel column was slurry packed with 5-pm Spher- isorb silica. The column was eluted under the following gradient conditions at 1 ml min-I: solvent A = pH 3 phosphate buffer- acetonitrile (80 + 20); solvent B = pH 3 phosphate buffer- acetonitrile (50 + 50); linear gradient from A to B over 12 min; linear gradient from B to A over 5 min; and equilibration with A for 3 min. The eluents were pre-mixed. Attempts to mix acetonitrile and the pH 3 buffer using the facilities on the instrument often led to crystal formation in and around the pumphead.The fluorescence monitoring conditions were excitation wavelength 230 nm and emission wavelength 302 nm. Chromatograms were recorded on a Kipp and Zonen flat- bed recorder set at 10 mV. For sublation, an all-glass system was constructed to the dimensions given in ref. 1 1. Extracts were taken to low volume240 Analyst, February 1996, Vol. 121 using a Kuderna-Danish evaporator (1 1 capacity) fitted with a 10 ml pear-shaped flask. Final blow-down was undertaken on a Pierce Model 18780 Reacti-Vap under a stream of nitrogen (ox ygen-free). Reagents and Standards Analytical-reagent grade ethyl acetate, sodium chloride, sodium hydrogencarbonate, sodium phosphate and granular sodium sulfate were used.Eluents were prepared by dissolving sodium phosphate in distilled water to give a 0.014 mol 1-l solution. This was adjusted to pH 3 with phosphoric acid and mixed with acetonitrile (HPLC grade) in proportions of 80 + 20 and 50 + 50 (v/v) to give eluents A and B, respectively. Stock standard solutions of non-ionic surfactants were prepared at concentrations of 1 g 1-1 in methanol, and sub- dilutions were made with methanol. The surfactants used were either Triton X-100 (Aldrich Milwaukee, WI, USA) or Synperonic NP8 (manufacturer's sample). The 9E0 internal standard was prepared by reacting 4-tert- octylphenol with 1,2-bis(2-chloroethoxy)ethane to give the chloro derivative. This was isolated and further reacted with the sodium salt of hexaethylene glycol to give the final product.The reaction scheme is shown in Fig. 1. Following preparative chromatography, a clear liquid con- sisting of the singe 9E0 oligomer was obtained (full details of this preparation can be provided on request). A solution containing about 2 g 1-1 was prepared in methanol. Trade effluent samples were stored in glass bottles at 4°C without preservative. Procedure A 500 ml volume of trade effluent (shaken to provide a homogeneous suspension if appropriate) was placed in a 11 conical flask. Sodium chloride (100 g), sodium hydrogencar- bonate (5 g), the 9E0 stock solution (200 plr0.4 mg) and distilled water (400 ml) were added and shaken until soluble solids had dissolved. The mixture was transferred into the sublation equipment, additional water was added to adjust the level to that of the top side-arm and ethyl acetate (100 ml) was added.Nitrogen at 0.5-1.0 1 min-1 was passed for 5 min. The ethyl acetate layer was drawn off and two further extractions with ethyl acetate were performed. The bulked extracts were held in a separating funnel and the aqueous layer was discarded. The ethyl acetate was dried by adding anhydrous sodium sulfate (approximately 10 g) and filtered through a cotton-wool plug into a Kuderna-Danish evaporator fitted with a 10 ml pear- shaped flask containing a few anti-bumping granules. The evaporator was fitted with a splash-head for connection to a side-arm condenser and the system was placed on a steam- bath. When most of the ethyl acetate had distilled off for re-use, the evaporator was removed from the steam-bath.Condensing ethyl acetate vapour washed the evaporator clean and the residue in the pear-shaped flask was blown down to constant mass on the heating block at 70 "C under a stream of nitrogen. The residue was dissolved in acetonitrile (some heating on a steam-bath was occasionally required) and the extract was transferred into a calibrated flask (10 ml) and adjusted to volume. The flask was placed in a freezer (-5 "C) overnight to precipitate fats and other insoluble materials. The contents of the flask were transferred into a centrifuge tube and spun down to give a clear supernatant solution. This was removed and stored in a clean flask until all samples were ready for HPLC analysis. A 100 p1 volume of the extract was placed in a glass vial of about 2 ml capacity with dimensions appropriate to the autoinjector turntable of the HPLC system.A 900 p1 volume of methanol was then added to each vial and they were sealed with a suitable septum. The same dilution procedure was used for the stock standard solutions of surfactant and the internal standard to give surfactant solutions for injection of 0.1 mg ml-1 and lower and 0.004 mg ml-1 of 9E0. Identical aliquots (usually 10-20 pl) of each solution were injected on to the HPLC system described above. Samples were usually analysed overnight with a batch of about ten samples interspaced with standards and the internal standard solution. The elution sequence of the APE oligomers ranges from lower ethoxylate oligomers eluting first to higher ethoxylate oligomers eluting later.Each oligomer was readily assigned by reference to the internal standard peak at 9E0. The peak height/ area of each oligomer was noted. Processing of the data involved calculation of the recovery based on the recovery of the 9E0 internal standard, following by quantification which was based on a peak height/area of some other oligomer (usually 10E0,llEO or 12EO) in the sample and the standards. Where APE surfactants were present in a sample, the observed 9E0 peak was made up of two components: the oligomer derived from the original sample and the added internal standard. To calculate the latter contribution to the combined peak, reference has to be made to the peaks of other oligomers. (Ci CH,CH,OCH 2-)* CH,-C-CH,-C - L NaOH 1 CH, CH, O J 4 0 Na * CHl-C-CH,-$QO( CHZCH20)g H CH, I CH, ( C ) Fig.1 Reaction scheme for the synthesis of Triton X-100 with nine ethoxylate units (9EO). A = chloroderivative of phenol with three ethoxylate units; B = monosodium salt of hexaethylene glycol; C = 27-(4-fert-octylpheny1)-3,6,9,12,15,18,21,24,27-nonahexacosanol.Analyst, February 1996, Vol. 121 24 1 Most APE mixtures produce chromatograms with a predictable distribution pattern, often close to a Poisson distribution. Simple curve fitting, either graphically or on a computer screen, permits a reliable estimate to be made of the 9E0 peak present in the sample without internal standard enhancement. Subtraction of this estimated peak from the actual peak gives a measure of the recovered internal standard. Results and Discussion Because of the variable composition of trade effluents, it was found that unless recovery data were available the losses incurred during sublation, evaporation and freezing-out led to unreliable results.An idea of the variability of effluents was provided by consideration of the residue mass found in the pear- shaped flask after evaporation to constant mass. On testing 500 mI aliquots from 155 effluents, the masses ranged from 5 mg to 4.5 g; the mean mass was 386 mg and the standard deviation was 898 mg. In selecting an internal standard for the determination of APE surfactants, it was necessary to use a compound that would behave similarly and be detectable in the final HPLC step. The 9E0 oligomer meets both criteria, and is a low-cost option compared with using an expensive radioactive analogue, which is the best interactive alternative.Most commercial APE surfactants (with the exception of some Tritons, Nonidets, etc.) use 4-nonylphenol as the precursor whereas the 9E0 oligomer prepared as an internal standard was based on 4-tert-octylphenol. Under the chroma- tographic conditions used, in which separation was induced by the hydrophillic portion of the surfactant rather than the hydrophobe, both compounds co-eluted. With the trade effluents studied, the recoveries ranged from 0 to 120%. In practice, complete loss of internal standard was unusual and was found to occur only when the amount of co- extracted solids exceeded 4 g. More usual were recoveries in the range 60-95% and losses were as often associated with emulsification of the ethyl acetate-water interface as with the presence of co-extractives.A reassuring aspect of using an interactive internal standard to follow the whole analysis was that if the 9EO oligomer was the only one found in the final chromatogram, one could be confident that the sample con- tained no APE surfactant. The separation of APE oligomers under the described conditions is an unusual phenomenon, and only Spherisorb silica displays a useful capability. Four other HPLC-grade silicas of different manufacture were also tested but failed to give adequate oligomer resolution. The highly aqueous eluents used in the chromatography, and the reduction in retention time produced by increasing the acetonitrile content, make the separations appear deceptively like examples of reversed-phase chromatography.That such a classification is incorrect becomes apparent when it is found that the most hydrophillic oligomers are retained longest. The experimental evidence that we have generated5 suggests that hydrogen bonding between the ethox- ylate groups of the surfactant and the silanol groups of the silica is the main retention mechanism. Increasing the acetonitrile content of the eluent disrupts the bonding, leading to more rapid elution. The gradient elution conditions described above were primarily designed to give an adequate oligomer separation in the minimum time. They work well for most commercial APE surfactants and an example of the separation is shown in Fig. 2. Longer gradients would be required where a mixture contains oligomers higher than 20EO.The columns were found to re- Table 1 Analytcal data for samples of a bulked trade effluent ‘spiked’ with Triton X-100 at 10 and 1 mg 1-1 10 mg 1-1 added Mean Mean recovery (%) Concentration recovery (9%) Concentration (n = 9) found/mg l-l* (n = 9) found/mg I-]* 110 9.50 103 0.96 103 9.44 92 1.04 100 10.00 95 1 .oo 92 9.80 102 0.84 95 9.14 92 0.86 100 9.56 93 0.90 Mean: 100 9.57 Mean: 96 0.93 1 mg 1-1 added * The concentrations reported in mg 1-1 have been corrected to allow for the recovery (n = 9) of the internal standard. *9EO I * 9go I I I I I I 1 I 5 10 15 5 10 15 Time/min Fig. 2 Separation of (1) Triton X-100 (0.1 mg ml-1) and (2) Synperonic NP-8 (0.5 mg ml-I) on Spherisorb silica using gradient elution.Fluores- cence detection at attenuation 64 with the recorder at 10 mV. 9 EO I I I 1 I 1 i I I 5 10 15 5 10 15 Time/min Fig. 3 Chromatography of trade effluent extracts ‘spiked’ with 9EO internal standard. (1) 9E0 stock standard solution; (2) = trade effluent A; (3) = trade effluent B; (4) = trade effluent C. Conditions as in Fig. 2. The recovery of surfactant was calculated by measuring the 9E0 peak height above the base level mark shown on each chromatogram. Each value was then compared with the 9E0 peak in chromatogram (1) (which represents 100% recovery).242 Analyst, February 1996, Vol. 121 equilibrate rapidly and last for long periods, e.g., in excess of 6 months, despite regular use for trade effluent analysis. The principle of measuring recoveries is shown in Fig.3. The chromatograms show a 9E0 internal standard solution, and extracts from trade effluents which had been ‘spiked’ with internal standard before extraction. These chromatograms have been marked to show where the base-level9EO was postulated from curve fitting; the remaining portion of peak was attribu- table to the recovered internal standard. Comparison of the latter peak height/area with that of the 9E0 standard solution gave a direct measure of the recovery. These chromatograms also show that fluorescent co-extractives eluted before the APE oligomers. It was found by experiment that co-extractives usually eluted in the non-retained region of the chromatogram, and hence did not interfere with the APEs. To quantify APE surfactants, a comparison was made of the peak height/area of a specific oligomer in the extract with that of the same oligomer of a reference surfactant.The reference surfactant used throughout the study was Triton X-100, but in principle any other APE surfactant could be used. The chief requirement was that the oligomer range should be similar between the reference compound and the surfactants found in the trade effluents. Calibration graphs of peak height or area ~?er~ws pg ml-1 of total Triton X-100 when measured for a specific oligomer, usually IOEO, 11EO or 12E0, were linear over the range 0-200 pg ml-l. To test the validity of the method, 7 1 of bulked trade effluent were prepared from samples that had been found to contain no APE surfactants. Half the Table 2 Range of APEs detected in 150 samples that were analysed during the course of this study Rangel Range/ mg 1-1 as No.of mg 1-1 as No. of Triton X-1 00 samples* Triton X- 100 samples* <0.l+ 37 20-25 5 0.1-1 39 25-30 2 1 -2 16 30-35 1 2-5 5 35-40 - 5-10 15 40-50 - 10-15 7 > 50 5 15-20 1 * Owing to interferences, no meaningful results could be obtained on 17 samples. + Below detection limit. effluent was fortified with Triton X-100 at 100 mg 1-1 and the other half at 1 mg 1-1. Aliquots of 500 ml were taken through the procedure and the results obtained are given in Table 1; these results suggest that the procedure has acceptable accuracy and precision. Table 2 shows the distribution of the measured concentration of APE surfactant in 150 trade effluents. Conclusions The described procedure involves minimal clean-up, yet is able to provide reliable analytical data for most trade effluents.Incorporation of an interactive internal standard compensates for the unpredictable losses of APE surfactant inherent in the sublation extraction steps and in subsequent stages of the procedure. We acknowledge the assistance of Yorkshire Environmental Labservices for the loan of instrumentation and the provision of trade effluent samples. References 1 2 3 4 5 6 7 8 9 10 11 ENDS Rep., 1993, 222,9. ENDS Rep., 1995, 241, 11. Sumpter, J. P., and Jobling, S . , Lancet, 1993, 342, 125. Linear Alkylbenzene Sulphonates (LAS) and Alkylphenol Ethoxylates (APE) in Waters, Wastewaters and Sludges by HPLC, Methods for the Examination of Waters and Associated Materials, HM Stationery Office, London, 1993. Ibrahim, N. M. A., and Wheals, B. B., J . Chromatogr., in the press. Andrew, B. E., Analyst, 1993, 118, 153. Scarlett, M. J., Fisher, J. A., Zhang, H., and Ronan, M., Water Res., 1994,28, 2109. Wang, Z., and Fingas, M., J . Chromatogr., 1993, 637, 145. Holt, M. S . , McKerrell, E. H., Perry, J., and Watkinson, R. J., J . Chromatogr., 1986, 362, 419. Wickbold, R., Tenside, 1972, 9, 173. American Public Health Association, American Water Works Association and Water Pollution Control Federation, Standard Methods for the Examination of Water and Waste Waters, American Public Health Association, Washington, DC, 16th edn., 1985, p. 578. Paper 510538OC Received August I I , 1995 Accepted October 19, I995

ISSN:0003-2654    

DOI:10.1039/AN9962100239

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 (243-247) 243 Direct Determination of 7-Hyd roxycoumari n and 7-Hydroxycoumarin-glucuronide in Urine by Using Capillary Electrophoresis Declan P. Bogana, R. D. Thornesa, M. Tegtmeierb, E. A. Schaferb and Richard O'Kennedya" a School of Biological Sciences, Dublin City University, Glasnevin, Dublin 9, Ireland Schaper and Briimmer, 38251 Salzgitter, Germany A method has been developed for the direct determination of 7-hydroxycoumarin (7HC) and 7-hydroxycoumarin- glucuronide (7HCG) in urine without sample clean-up. Separation was carried out in 90% 100 mmol l-1 phosphate buffer, 11 mmol l-I deoxycholic acid (sodium salt) and 10% acetonitrile, on a 47 cm uncoated silica capillary at 20 kV with detection of the analytes at 320 nm. The linear detection range for concentration versus peak area for the assay is from 0 to 100 pg ml-1 for both analytes, with a limit of quantitation of 2 pg ml-l for 7HC and 5 pg ml-1 for 7HCG in urine.Inter- and intra-assay results showed s, values in peak areas of between 0.5 and 13%. The method was applied to the direct determination of 7HC and the glucuronide conjugate in urine from two volunteers administered with 250 mg of coumarin. The samples were also analysed by another CE method and by using HPLC. There was no statistical difference between the results determined by each of the methods. Up to 83% of the coumarin administered was excreted as 7HC or 7HCG. The majority of the 7HC excreted was in the glucuronide form (98%) with 2% occurring as free 7HC. Keywords: 7-Hydroxy coumarin; 7-hydroxy coumarin-glucuronide; capillary electrophoresis; urine Introduction Coumarin is a benzopyrone that occurs naturally in many plants and essential Coumarin is used clinically in the treatment of high-protein oedemas, brucellosis, and chronic infections.2 It is currently being investigated as an anti-cancer drug.3 7-Hy- droxycoumarin (7HC) and 7-hydroxycoumarin-glucuronide (7HCG), respectively, are the predominant phase 1 and phase 2 metabolites of coumarin in man4 but are not the only metabolites produced in vivo.The exact pharmacokinetic description of drug metabolism requires the determination of not only phase 1 metabolites, but also the phase 2 conjugates of parent drugs. It was difficult to obtain some of the coumarin metabolites standards commercially, thus enzymic cleavage or chemical derivatization was typically required before their determinati~n.~-~ Undesirable reactions, incomplete reactions, and lengthy preparation steps interfere with the accurate determination of many conjugates in biological samples during deconjugation and derivatization. Coumarin is a prodrug for 7HC as it is believed that 7HC is the active form of the drug, rather than coumarin itself.7HC is currently under investigation in clinical trials as an anti-cancer agent.8 However, no metabolism details or clinical side-effects * To whom correspondence should be addressed. have been reported for 7HC. 7HC can be analysed by CE,9J0 HPLC," spectrofluorirnetric assay,l2 TLC,13 and ELISA. l4 7HCG content in biological samples is typically determined by deconjugating the compound into 7HC and glucuronic acid by the action of P-glucuronidase (30 min incubation at 37 "C in a sodium acetate buffer, 1 moll-', pH 5.0).9?11 An assay for 7HC determination is then applied after further sample preparation e.g., extraction.9.11 To obtain 7HCG it must be isolated from urine, or produced enzymically in vitro.15 S harifi et a1.16 described the determination of 7HC and 7HCG by using HPLC in urine and plasma.They determined both metabolites in urine without extraction. Deconjugation and extraction of 7HC was carried out on the plasma samples for total 7HC concentration determination. The clinical relevance of 7HCG has not yet been determined. However, the glucuronide conjugates of other drugs, e.g., morphine-6-glucuronide, are known to possess clinical activ- ities.17J8 Other glucuronides have been used as prodrugs for cancer chemotherapy in vitro.19 Wang etal.found that the glucuronide conjugate was less toxic than the parent antineo- plastic drug, N,N-di-(2-chloroethyl)-4-hydroxyaniline. Initially, treatment with a p-glucuronidase-labelled monoclonal antibody to an antigen on the tumour surface was carried out. The antibody bound itself to the antigen on the surface of the tumour cell. The glucuronide conjugate was then administered and cleavage of the parent drug from the conjugate occurred. Thus, as a result there was an increased local concentration of the drug at the tumour site. CE is a family of techniques that employ narrow-bore capillaries to perform high efficiency separations of both large and small molecules.These separations are facilitated by the use of high voltages, which may generate electro-osmotic or electrophoretic flow of buffer solutions and ionic species, respectively, within the capillary. Several researchers are using this approach for the determination of coumarin and several of its derivatives in a variety of matrices, including the analysis of warfarin,20,21 and the determination of coumarin in plant extracts,Z2 urine and serum,9 and in metabolism studies23,24 in liver microsomes. CE has also been used for the determination of glucuronides in urine and in Chinese medicines.25-27 It has also been applied to the direct determination of several compounds in urine or plasma.28-3' The method described here allows for the direct determination of the two principle phase 1 and phase 2 metabolites of coumarin in human urine by CE without sample clean-up.Experimental Chemicals and Reagents 7HC and deoxycholic acid (sodium salt) were purchased from Sigma (St. Louis, MO, USA). 7HCG was kindly donated by244 Analyst, February 1996, Vol. I21 Schaper and Briimmer (Salzgitter, Germany). Control urine was obtained from a healthy volunteer who had not been admin- istered with coumarin or 7HC. Acetonitrile was obtained from Labsan (Dublin, Ireland). KH2P04 and KzHP04 were pur- chased from Riedel-de Haen (Hanover, Germany). Sample Preparation 7HC standards were prepared from a 1 mg ml-1 stock solution prepared in (20 + 80 v/v) methanol (Labscan, Dublin, Ireland)- de-ionized water.7HCG standards were also prepared from a 1 mg ml-1 stock solution prepared in de-ionized water. Dilutions (20-800 pg ml-1) were then prepared in de-ionized water and 10 pl of each standard was spiked into 80 yl of urine. Blank urine was also used, where necessary, for the dilution of samples with unknown metabolite levels to bring their values into the linear range of the assay. The control urine and electrolyte solutions were filtered through 0.22 ym filters. Two volunteers were orally administered with a 'fast release' 250 mg coumarin tablet and their urine samples were collected at specific timed intervals (0, 2,6, 10, 14, and 24 h, respectively). Urine samples were collected and stored at -20°C until analysis. The urinary volume was recorded.The samples were centrifuged at 13 000 rpm for 5 rnin before analysis. Capillary Electrophoresis Separation Separations were carried out in a fused untreated silica capillary (47 cm X 50 pm id; Beckman Instruments, Fullerton, California, USA) with a capillary to detector distance of 39.3 cm. The samples were analysed on a Beckman CE P/ACE 5500 instrument and controlled with System Gold software. The capillary was initially conditioned by rinsing with 0.1 moll-' HC1 (for 5 min), then 0.1 moll- NaOH (for 5 min) and de-ionized water (for 5 min). Separation was optimized under the following conditions. The capillary was reconditioned between each run with a 1.5 rnin 0.1 mol 1-1 NaOH wash, followed by a 3 rnin rinse with electrolyte buffer. The electrolyte solution, for the separation of 7HC and 7HCG in urine, was a (90 + 10 v/v) [lo0 mmol l-l phosphate buffer (pH 7.0)-11 mmol l-1 deoxycholic acid (sodium salt)]-[acetonitrile (Labscan, Dublin, Ireland)].Samples were injected at 0.5 p.s.i. (3447.38 Pa) for 5 s. The separation was carried out at 20 kV (rise time 0.2 min) and the typical running current was 160 yA. Detection of the analytes was carried out at 320 nm with a fixed wavelength detector and 320 nm filter. Calibration curves were prepared for each metabolite by plotting the concentration versus peak area. Inter- and intra-assay variations were assessed to determine the accuracy and precision of the technique. HPLC Analysis of Urine Samples Urine sample or standard (0.5 ml) spiked into control urine was treated with 0.5 ml of P-glucuronidase (Sigma, St.Louis, MO, USA) in 1 mol 1-1 sodium acetate (pH 5.0) and incubated at 37 "C for 30 min. Trichloroacetic acid (100 pl) (Merck, Poole, England) was then added, to precipitate the protein solution, and the samples were mixed and centrifuged at 13 000 rpm for 5 min. To 190 y1 of supernatant, 10 pl of a 1 mg ml-1 standard of 4-hydroxycoumarin (an internal standard) was added and the samples were analysed on a Phenomenex (Cheshire, England) Bondclone 10 C18 column. Separation was achieved with gradient elution. The solvents employed were A, water- methanol-acetic acid (950 + 50 + 2 v/v) and B, 100% methanol. The eluent was monitored at 320 nm. Sample (20 PI) was injected onto the column. The 1 ml min-1 gradient was as follows: 0-14 rnin loo%, solvent A -+ solvent A (50%)-solvent B (50%); 14-22 min, solvent A (50%)-solvent B (50%); 22-23 min, solvent A (50%)-solvent B (50%); + solvent A (100%); and 23-32 min, solvent A (100%).Results and Discussion Development of the CE Separation The best electrolyte solution was found to be a phosphate buffer-deoxycholic acid (sodium salt)-acetonitrile mixture. Separation of 7HC and 7HCG prepared in water is shown in Fig. 1. It is seen that the 7HC and the glucuronide are very well resolved with migration times of approximately 5 and 7 min, respectively. Initially, excellent separation was achieved in borate buffer at pH 9.0, but the 7HCG was found to be unstable at this pH as it decomposed to 7HC during the run time. No separation was observed in buffers below pH 6.0.Previous method~9?23,*~ had used phosphate buffer to separate 7HC from a variety of components in various matrices. Separation of the analytes prepared in buffer was carried out in 25 mmol 1-1 phosphate buffer, but when assessed in the biological fluids of interest it was found not to be applicable. Separation was achieved at pH 7.0 due to the respective charge difference between the two compounds of interest. The presence of several more -OH groups on the glucuronide would cause it to be more negative than the 7HC which has only one -OH group and, thus, enables separation. Several micellar systems were tried and it was found that separation was achieved with the use of a bile salt, deoxycholic acid (sodium salt). Bile salts are natural surfactants. They form helical micelles and yield selectivities significantly different from Clz-alkyl chain surfactants.32 They are particularly useful for the separation of hydrophobic compounds like 7HC.No separation was achieved with the anionic surfactant, sodium dodecyl sulfate. Deoxycholic acid is also known to complex fatty acids, and this may function in a form of on-line sample clean-up. The relatively high concentration of phosphate buffer was used to improve peak shape and separation efficiency. The acetonitrile served, primarily, in decreasing the viscosity of the buffer. The phosphate buffer-deoxycholic acid mixture was very viscous and the addition of the acetonitrile reduced this viscosity. The optimum level of acetonitrile was found to be 10% as lower or higher amounts altered peak shape and separation efficiency.Acetonitrile is also believed to aid in sample stacking.33 It is believed to counteract the deleterious effects of high concentrations of inorganic salts, and to remove proteins present in the sample. The removal of proteins and the unwanted ions also prolongs the life of the capillary. 1.5E-2 E 1.OE-2 c\I m 0) C c e 9 5.OE-3 8 O.OE+O 0.0 1.6 3.2 4.6 6.4 8.0 Tim elm i n Fig. 1 Electropherogram of the separation of A, water (solvent front); B, 7HC; and C, 7HCG. The samples were prepared in water and analysed as outlined under Experimental.Analyst, February 1996, Vol. 121 245 In previous methods 9,23924 detection was typically carried out at 210 nm. However, endogenous species in urine were found to co-migrate with the analytes of interest and it was impossible to observe those analytes at this wavelength.7HC and the glucuronide were also found to have an absorbance maxima at 320 nm, and in order to increase selectivity (with some loss in sensitivity) this wavelength was chosen as it was found that the interfering species did not absorb at this wavelength. Therefore, 7HC and the glucuronide form could be selectively isolated, resolved, and quantified in the urine. The optimization of the electrophoretic and electro-osmotic flow, and the use of 320 nm, allowed the resolving of the coumarin metabolites from the endogenous species present in urine (Fig. 2). It is clearly shown that the compounds are well resolved from any interfering endogenous species present in the urine.It was found that serum had a deleterious effect on the performance of the capillary, and hence it was not possible to analyse for 7HC and 7HCG in this matrix. The analysis of 7HC and 7HCG in plasma samples was also studied but the protein binding of the metabolites, and coumarin, interfered with their accurate determination. Plasma samples obtained from the volunteers who had been administered the 250 mg of coumarin were analysed. Peaks were observed but at longer migration times (8.6-9.4 min) as compared with spiked samples (results not shown). It was not possible positively to assign those peaks to 7HC or the glucuronide. However, there was a significant difference observed between the 0 h and 1 h samples for these peaks, as observed with HPLC analysis.Different preparation methods of the plasma samples were assessed, with analysis by CE employing the same electrolyte buffer. However, treatment of the plasma samples including protein precipitation, addition of organic buffers, sonication and/or heat did not allow the successful determination of the metabolites. The direct determination of any compound without sample clean-up is a very beneficial technique. One does not need to be concerned with errors occurring during sample preparation. Great care must be taken while making standards or if dilutions must be prepared. Separations are achieved in less than 7.5 min 1E-2 6 8E-3 7E-3 E 3 8 e 9 3E-3 0 c 5 5 5 a v) 1E-3 -lE-3 3 0.0 1.0 9.2 4.U 0.4 8.0 Time/m in Fig. 2 An overlay of the electropherograms of the 0 pg ~ m - ~ (dashed line) and 100 pg ~ m - ~ (full line) standards prepared in urine and analysed as outlined under Experimental.The components are A, solvent front containing any neutral compounds; B, 7 HC; and C, 7 HCG. and the capillary is regenerated in 4.5 min. The previous extraction method of Egan et al.7 involves an incubation step with (3-glucuronidase as well as an extraction, evaporation step, reconstitution and, finally, analysis. This process involved much labour and was time consuming. No deterioration in capillary performance for the duration of the study was found. Several hundred urine samples were passed down the capillary while optimizing the method and during the inter- and intra-assays. It is possible to analyse urine and plasma samples directly on HPLC coIumns, but this is not recommended.The washing of the capillary with 0.1 mol 1-1 NaOH, between each run, removes any residual components from the urine, thus leaving a new homogenous surface each time. The use of CE appears from this study to be very applicable to the direct determination of 7HC and 7HCG in urine. However, the method does not allow the determination of coumarin. Coumarin acts as a neutral compound at this pH and it migrates with the solvent front and with any other neutral compounds present. Limit of Quantitation and Linearity Both 7HC and 7HCG have two absorbance maxima at 210 nm and 320 nm. The maximum at 320 has a lower molar absorption coefficient than that at 210 nm. The limit of quantitation for 7HC in urine was found to be 2 pg ml-1.For 7HCG, it was found to be 5 pg ml-1 for urine. The linear detection range for both analytes was 0-100 pg ml-1. The mean equation of the line (inter-assay) from a plot of the concentration versus the mean peak area are (i) 7HC in urine was y = 0.00447 + 0.00497~; r2 = 0.9989 and for (ii) 7HCG the mean equation in urine was y = -0.00066 + 0.00236~; r2 = 0.9995. Accuracy and Precision Tables 1 and 2 show the mean concentrations calculated from the intra- and inter-assays, for the determination of 7HC and 7HCG in urine. Percentage s, values were found to be between 0.5 and 13.0%. The values were higher at the limit of quantitation of the method. The accuracy and precision for the inter-assay was determined over 6 d (n = 6) and the intra-assay over 5 replicates (n = 5) on one specific day.Clinical and Pharmacokinetic Studies Two healthy volunteers were treated with an oral dose of a ‘fast release’ 250 mg coumarin tablet. Urine samples were taken pre- administration and at specific time intervals after administration of the drug. The samples were analysed for the presence of 7HC and 7HCG by the above-mentioned method. Coumarin content Table 1 Intra-assay precision and accuracy results for the determination of 7HC and 7HCG in urine Concen- tration added/ pg cm-3 0 2 5 10 20 25 50 80 100 7HC mean conc.1 pg cm-3 s (%) 0 0 2.86 0.19 5.56 0.21 9.74 0.19 20.16 0.45 52.40 4.82 78.40 3.40 105.25 3.30 - - Sr (%) 0 6.8 3.7 1.9 2.2 8.2 4.3 3.2 - 7HCG mean conc . / pgcm-3 s(%) sr(%) 0 0 0 4.53 0.21 4.6 8.80 0.64 7.3 20.60 0.56 2.7 27.70 0.60 2.1 49.23 3.16 3.2 77.00 3.23 4.2 102.25 2.57 2.5 - - -246 Analyst, February 1996, Vol.121 was not determined, but from previous studies there is minimal coumarin excreted unmetabo1ized.l' Fig. 3 shows an electro- pherogram of a volunteer's urine sample 10 h after coumarin administration. Slight differences in peak migration times were due to individual differences between samples and standards (i.e., pH of urine, salts content). To confirm that the peak observed was 7HC or 7HCG, a sample of each was spiked into the respective sample and the respective peaks were seen to increase proportionately. No other coumarin metabolites were determined. The assay developed concentrated on the two major metabolites of coumarin in man. Fig.4 shows a plot of the levels of 7HC and 7HCG excreted by each of the volunteers. There is up to a 100-fold difference between the amounts of 7HC and 7HCG excreted in the urine. The majority of 7HC excreted is in the glucuronide form (98%) with 2% occurring as free 7HC in the urine. The urine samples were also analysed by the method of Bogan et aZ.9 and by HPLC (see Experimental above). Fig. 5 shows an overlay of the results obtained from each of the three methods. The results were compared for the total amount of Table 2 Inter-assay precision and accuracy results for the determination of 7HC and 7HCG in urine Concen- 7HC 7HCG tration mean mean added1 conc.1 conc.1 ygcm-3 ygcm-3 s(%) s,(%) ygcm-3 s(%) .s,(%) 0 2 5 10 20 25 50 80 100 0 0 0 0 0 0 1.95 0.25 13.0 - - - 4.97 0.39 7.9 4.98 0.62 12.6 10.00 0.53 5.3 9.80 1.00 10.2 20.50 0.70 3.2 20.80 1.75 8.4 - - 25.30 0.96 3.8 50.49 2.70 5.3 52.90 2.50 4.8 79.75 0.36 0.5 80.40 4.20 5.2 97.63 4.48 4.5 100.90 1.70 1.7 - 5E-3 4E-9 E 0 $j 3E-9 * a, C -e % 2E-3 2 BE-4 -2E-4 C 0.0 1.6 3.2 4.8 0.4 8.0 Time/min Fig.3 An electropherogram of volunteer A's urine sample 10 h after administration of coumarin. The peaks are A, any neutral compound and the solvent front and C, 7HCG. 7HC excreted at the specific timed intervals. Total 7HC content was directly determined or calculated from the addition of free and glucuronidated 7HC present in the urine samples. The total conjugated 7HC was determined from the 7HCG content multiplied by the molecular mass ratio between 7HC and 7HCG 3500.0 3000.0 2500.0 s 2000.0 .- c 2 c a, 1500.0 s 1000.0 500.0 0.0 0.0 5.0 10.0 15.0 20.0 25.0 Time/h Fig.4 Plot of time versus 7HCG (1) and 7HC(2) concentrations for each of the volunteers A and B, determined by capillary electrophoresis after the administration of 250 mg of coumarin. 160.0 140.0 1 F g .$ 80.0 5 E 120.0 c Q, 100.0 .- 3 c -0 .- a, 0 $ 60.0 2 f k 40.0 5 I- a 20.0 0.0 0.0 4.0 6.0 12.0 16.0 20.0 24.0 Time/h Fig. 5 Plot of time versus total 7HC excreted for each of the volunteers A and B, determined by capillary electrophoresis (1) by the method of Bogan et al.9 (2) and by HPLC (3).Analyst, February 1996, Vol. 121 247 (i.e. 162.1/338.1 X 100/1 = 47.9%). This result was then related to the urinary volume to determine the total amount of 7HC excreted.Fifty-four per cent. of the coumarin administered was excreted in the first 2 h as 7HC or 7HCG with up to 83% being excreted within 24 h. In the comparison of the results from the three methods there were no significant statistical differences (3.0%). The CE method above allowed the direct determination of each of the compounds without the addition of 6-glucuronidase and with minimal sample preparation. Results were obtained in shorter time than either the method of Bogan et al.9 or by HPLC. Conclusion CE was applied to the direct determination of 7HC and 7HCG in urine. The method developed was found to be both accurate and precise. The method was applied to the determination of the metabolites in urine samples from two volunteers who had been treated with coumarin.There was minimal sample preparation, the results from the analysis were obtained within 7.5 min for each sample. The linear range for the assay was between 0-100 pg ml-1 and the inter- and intra-assay mean concentrations and values were calculated. Improvements in limits of detection may be possible with other detection systems, e.g., electro- chemical, laser-induced fluorescence. Thus, the method was found to be very simple, practical and applicable to multiple sample analysis whilst also reducing the time involved for sample preparation. However, the method did not allow the determination of very low levels of metabolites. The advantages of CE include minimal sample and buffer waste, a large reduction in organic solvent use, straightforward method development, availability of autosamplers, computer controlled sampling, data acquisition, and data handling, and make the method much more user-friendly and efficient than other methods for 7HC analysis. We would like to thank Schaper and Briimmer (Germany), the Coumarin Research Fund, Forbairt and The Research and Postgraduate Studies Committee (Dublin City University) for their financial support.We would also like to thank Tony Killard for his invaluable help. References 1 2 3 4 Cohen, A. J., Fd. Cosmet. Toxicol., 1979, 17, 277. Egan, D., O’Kennedy, R., Moran, E., Cox, D., Prosser, E., and Thornes, R. D., Drug Metab. Reviews, 1990, 22 ( 3 , 503. Marshall, M. E., and Mohler, J. L., J. Zr. Coll. Phys. Surg., 1993, 22 (2) suppl., 6. Pelkonen, O., Raunio, H., Rautio, A., Maenpaa, J., and Lang, M., J .Ir. Coll. Phys. Surg., 1993, 22 ( 2 ) suppl., 24. 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 Ritschel, W. A., Brady, M. E., Tan, H. S. I., Hoffmann, K. A., Yiu, I. M., and Grummich, K. W., Eur. J. Clin. Pharmacol., 1977, 12, 457. Ritschel, W. A., Grummich, K. W., Kaul, S., and Hardt, T. J., Pharm. Znd., 1981, 43 (3), 271. Kaipainen, P., Koivusaari, U., and Lang, M., Comp. Biochem. Physiol., A. Comp. Physiol., 1985, 81C (2), 293. Marshall, M. E., Mohler, J. L., Edmonds, K., Williams, B., Butler, K., Ryles, M., Weiss, L., Urban, D., Bueschen, A., Markiewicz, M., and Cloud, G., J . Cancer Res. Clin. Oncol., 1994, 120 (suppl), S 39. Bogan, D. P., Deasy, B., O’Kennedy, R., Smyth, M. R., and Fuhr, U., J. Chromatogr.B., 1995, 663, 371. Schutzner, W., and Kenndler, E., Anal. Chem., 1992, 64, 1991. Egan, D., and O’Kennedy, R., J. Chromatogr., 1992,582, 137. Egan, D., and O’Kennedy, R., Analyst, 1993, 118,201. Cholerton, S., Idle, M. E., Vas, A., Gonzales, F. J., and Idle, J. R., J . Chromatogr., 1992,575, 325. Egan, D., and O’Kennedy, R., J. Ir. Coll. Phys. Surg., 1993, 22 (2) suppl., 72. Bogan, D. P., Killard, A. J., and O’Kennedy, R., J . Capillary Electrophoresis., 1995, in the press. Sharifi, S., Michaelis, H. C., Lotterer, E., and Bircher, J., J. Liq. Chromatogr., 1993, 16 (6), 1263. Portenoy, R. K., Foley, K. M., Stulman, J., Khan, E., Adelhardt, J., Layman, M., Cerbone, D. F., and Inturris, C. E., Pain, 1991, 47 (l), 9. Olson, J. A., J. Nutr., 1992, 122 (3), 615. Wang, S. M., Chern, J. W., Yeh, M. Y., Ng, J. C., Tung, E., and Roffler, S. R., Can. Res., 1992, 52 (16), 484. Gariel, P., Gramond, J. P., and Guyon, F., J . Chromatogr. B, 1993, 615, 317. Busch, S., Kraak, J. C . , and Poppe, H., J. Chromatogr., 1993, 635, 119. Morin, P., and Dreux, M., J . Liq. Chromatogr., 1993, 16 (17), 3735. Deasy, B., Bogan, D. P., Smyth, M. R., O’Kennedy, R., and Fuhr, U., Anal. Chim. Acta, 1995, 310, 101. Bogan, D. P., Deasy, B., O’Kennedy, R., and Smyth, M. R., Xenobiotica, submitted for publication. Liu, Y. M., and Sheu, S. J., J . High Resolut. Chromatogr., 1994, 17 (7), 559. Liu, Y. M., and Sheu, S. J., Anal. Chim. Acta, 1994, 288 (3), 221. Wernly, P., Thormann, W., Bourquin, D., and Brenneisen, R., J . Chromatogr. B., 1993, 127, 305. Deyl, Z., and Stuzinsky, J., J. Chromatogr. B., 1991, 569, 63. Zhang, C. X., Sun, Z. P., Ling, D. K., Zheng, J. S., Guo, J., and Li, X. Y., J . Chromatogr. B., 1993, 123, 287. Lloyd, D. K., Anal. Proc., 1992, 29 (5), 169. Chicharro, M., Zapardeil, A., Bermejo, E., Perez, J. A., and Hernandez, L., J . Chromatogr. B., 1993, 133, 103. Terabe, S., in Micellar Electrokinetic Chromatography, Beckman Instruments, Fullerton, CA, USA, 1992, p. 22. Shihabi, Z. K., unpublished work. Paper 510.54861 Received August 17,1995 Accepted October 16,1995

ISSN:0003-2654    

DOI:10.1039/AN9962100243

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 (249-254) 249 Colloidal Gold Supported Onto Glassy Carbon Substrates as an Amperometric Sensor for Carbohydrates in Flow injection and Liquid Chromatography Innocenzo G. Casellaa, Angelo Destradis" and Eli0 Desimonib a Dipartimento di Chimica, Universita' degli Studi della Basilicata, Via N . Sauro 85, 851 00 Potenza, Italy h Dipartimento di Fisiologia delle Piante Coltivate e Chimica Agraria, Universita' degli Studi di Milano, Via Celoria 2, 20133 Milano, Italy Gold microparticles dispersed on glassy carbon structures were characterized by SEM. Some aspects of the electrooxidation mechanism of carbohydrates in alkaline medium were investigated by cyclic voltammetry. The chemically modified electrode described was tested in flow injection and anion-exchange LC with pulsed amperometric detection.The electrode response was stable: over 8 h a 1.5-3% signal loss was observed for all of the investigated carbohydrates. Typical relative standard deviations ranged between 0.9 and 3.5%. The detector response was linearly dependent on the concentrations of polyhydroxyl compounds and monosaccharides over three orders of magnitude. Linear plots of 1/1, and l/c indicated a Langmuir-type adsorption of analyte on the catalytic sites. All investigated carbohydrates were detected at the picomole level. The determination of glucose in human biological fluids was achieved by LC in an alkaline medium. Keywords: Gold-modified glassy carbon electrode; alkaline media; carbohydrate; biochemical sensor; glucose; clinical sample Introduction Traditional methods used for the determination of carbohy- drates, such as polarimetry, refractometry, titrimetric analysis, etc., are in general expensive, unspecific and often time- consuming.On the other hand, specific enzymic methods are subject to rather fast, time-dependent activity losses. There is a need for a simple and specific method for carbohydrate analysis. The simultaneous determination of the components of complex mixtures can be achieved by LC. Several electrochemical (EC) detectors have previously been proposed for use in the flow injection (FI) or LC analysis of carbohydrates. 1 - 4 This electrochemical approach is of particular interest because carbohydrates, or in general hydroxyl-contain- ing compounds, do not exhibit intense absorption bands, thus UV and fluorescence detection methods are usually charac- terized by poor detection limits.Transition metals are extensively used as amperometric sensors for the determination of carbohydrates, alcohols, amino acids, etc., in LC.5-13 Chemically modified electrodes (CMEs) containing surface- confined specific chemical groups, or coated with conductive polymer films operating as size/charge exclusion barriers, have shown some advantages over conventional electrode sub- strates. 14-17 Metal microparticles dispersed in inert matrices have been recently recognized to have potential applications in electro- analysis. High surface area CMEs, prepared by electrodeposi- tion of transition metals on graphitic substrates, have provided greatly enhanced oxidation currents for aliphatic com- pounds.* Incorporation of Ni, Co and Pt species into glassy carbon surfaces by casting processes have made possible the electro- catalytic detection of carbohydrates, alcohols and sulfite.19-22 In this paper a gold-modified glassy carbon electrode (Au- CME) is described as an electrochemical sensor, in alkaline media, for the detection of carbohydrates. SEM and cyclic voltammetry have been used to elucidate the morphological and electrochemical properties of the Au-CME in alkaline media. The electrode was also tested by FI in PAD modes for the detection of carbohydrates. Examples of determination of carbohydrates by LC-EC in real matrices (such biological fluids) are also given. Experimental Reagents All solutions were prepared from anal ytical-reagent grade chemicals without further purification and by using doubly distilled, de-gassed and de-ionized water.Individual 2000 mg ml-1 standard solutions of carbohydrates (Aldrich, Milwaukee, WI, USA) were prepared freshly, daily, in distilled water. Appropriate dilutions were performed with 0.2 mol 1-1 NaOH. Unless otherwise specified, experiments were performed by using 0.2 mol 1-1 NaOH as background electrolyte. Alkaline solutions were protected from oxygen and carbon dioxide by purging with high-purity nitrogen. Real samples were centri- fuged for 5 min at 3000 rpm and diluted 1 + 99 with mobile phase (0.2 mol 1-1 NaOH). All chromatographic and FI experiments were carried out at ambient temperature. Apparatus A Model 273 EG & G Princeton Applied Research (PAR EG & G, Princeton, NJ, USA) potentiostat/galvanostat controlled by the PAR EG & G 270 Electrochemical Analysis Software was used for electrochemical measurements.Cyclic voltammetry (CV) was performed in a three-electrode cell using an Au-CME working electrode, an SCE (reference) (4 mol 1-1 KC1) and a platinum foil counter-electrode. To avoid chloride contamina- tion of the working electrode, the reference electrode was connected through a Luggin capillary. The glassy carbon electrode used in CV, having a geometric area of 0.125 cm2, was purchased from EG & G PAR. All current densities reported in this paper are quoted in terms of effective Au surface area: the areas of the Au-CME and bulk gold electrodes were estimated from the amount of charge for the reduction of the Au oxide monolayer in 0.1 moll-' H2S04 solution.23250 Analyst, February 1996, Vol.I21 Thermostated electrochemical experiments (CV) were car- ried out in a circulating water-bath (Colora Model WK 4 DS, Colora Messtechnik, Lorchwurtt, Germany). Amperometric measurements (in PAD mode) in a flowing stream were performed using an EG & G PAR Model 400 Electrochemical Detector and a flow-through thin-layer elec- trochemical cell consisting of dual Au-GC working electrodes in serial configuration (EG & G PAR), and Ag/AgC1(4 moll-' KC1) reference electrode and a stainless-steel counter electrode. The output signal was recorded by a Model 868 Amel recorder (Amel Instruments, Milan, Italy). Flow injection experiments were carried out with a Varian (Palo Alto, CA, USA) 2510 pump equipped with a Model 7125 Rheodyne (Cotati, CA, USA) injector using a 50 p1 sample loop.The mobile phase was purged from oxygen by flowing through an Online Degasser (Hewlett-Packard Series 1050; Avondale, PA, USA). Chromatographic separations were made by using a Carbo- pac PA1 (250 X 4 mm id, Dionex, Sunnyvale, CA, USA) anion- exchange column and a Carbopac MA1 (250 X 4 mm id, Dionex) anion-exchange column; the carrier electrolyte was 0.2 and 0.4 moll-' NaOH, repectively. Scanning electron micrographs were obtained using a Cambridge Stereoscan Microscope. Electrode Preparation Each Au-CME was prepared by depositing 10 pi of 30 mmol 1- A u ~ + on the surface of the glassy carbon electrode.The electrode was dried (face up) in a fan oven at approximately 60 "C for about 30 min, washed with water and electrochemic- ally conditioned in 0.2 mol 1-I NaOH by cycling the potential between -0.6 and 0.4 V versus SCE for 30 min. Prior to each electrode modification, traces of gold species were removed from the glassy carbon surface by polishing with 0.05 pm (x- alumina powder on a polishing micro-cloth. Results and Discussion SEM Fig. 1A shows the electron micrograph of colloidal particles of gold obtained by deposition of 1.2 X 10-*0 mol of Au3+ on the glassy carbon electrode surface and subsequent immersion for 2 h in 0.2 mol 1-* NaOH. After each treatment, samples were thoroughly rinsed with pure water prior to being examined by SEM. The treatment with sodium hydroxide strongly influenced the morphology of the Au-CME, which was initially charac- terized by a fairly uniform deposit. Globular particles having an average diameter of 0.8-1.0 pm were formed by gradual dissolution of the initial film and subsequent nucleation on the glassy carbon.The considerable dimensions of the particles lead to the formation of large islands of particles. After longer periods of treatment in an alkaline medium (6 h), the morphology of the Au-CME did not change any further. Fig. 1B shows an Au-CME after cycling the potential for 10 min in sodium hydroxide between -0.6 and 0.9 V. The particles appear well resolved and their distribution on the graphitic substrates is more homogeneous. Moreover, the average dimension of the particles is 100 nm.After long electrochemical treatment (30 min) the morphology of the catalyst changed. Fig. 1C shows this effect: a well resolved and homogeneous distribution of particles appeared, and in this case, the average dimensions ranged between 10 and 30 nm. After longer potential cycling, the morphology did not show further change. Moreover, after application of a pulsed waveform (i.e., E l = 0.2 s) for a long time (4 h), the general morphology of the catalyst on the graphitic structure did not show appreciable variations. It has been reported that fine Au particles (7-10 nm in diameter), deposited onto a base plane of the carbon rod by V, ti = 0.4 s , E ~ = 0.65 V, f2 = 0.2 S, E3 = -0.7 V, t 3 = 0.2 means of vacuum evaporation of gold metaP4 or by reducing the citrate in the Au3+ solutions,25 with diameter of 20 nm, offer considerable catalytic activity in the electrooxidation of HCH024 and useful features concerning the preparation of multimetallic combination of particles in order to obtain new amperometric sensors.25 Electrochemical Measurements Typical cyclic voltammograms obtained at the Au-CME in alkaline solution are reported in Fig.2A. In the absence of any analyte, the electrochemical behaviour of the Au-CME is rather similar to that of a bulk gold electrode. The anodic peak al, in the approximate region -0.15 V is believed to correspond to the anodic discharge of water, with the formation of a sub- monolayer of adsorbed hydroxyl radicals (AuOH),26,27 while the oxidation wave a2 is associated with AuO formation.The Fig. 1 SEMs of the Au-CME: A, after immersion in 0.2 moll-' NaOH for 2.0 h and rinsing with water; B, after cycling the potential between minus 0.6 and 0.9 V for 10 min; and C , after cycling the potential for 30 min.Analyst, February 1996, Vol. 121 25 1 1.2 - 1.0 - 0.8 - ' 0.6 - Q E -0 N . 0.4 - 0.2 - cathodic peak c l , in the negative sweep, is associated with the reduction of Au oxide species. In the presence of glucose, three new anodic peaks, Ia, IIa and IIIa, appeared at -0.01,0.15 and 0.405 V, respectively. Moreover the electrooxidation processes are operative at both low and high potentials (i.e., -0.6 and 0.9 V) during the anodic and cathodic sweeps. Similar cyclic voltammograms were observed for all of the examined alditols and sugars, the only differences being the sensitivity and current intensity ratios (Ia : IIa : IIIa).Cyclic voltammograms obtained at the bulk gold electrode are compared in Fig. 2B. Although the voltammograms, obtained under the same experimental condi- tions at the bulk gold electrode are qualitatively similar, (compare Fig. 2A and B), the oxidation current densities at the Au-CME, in both the positive and negative sweeps, are appreciably larger than those observed at the bulk electrode. Moreover, anodic peaks Ia and IIa are better resolved at the Au- CME than at the bulk electrode. This catalytic enhancement is a direct consequence of the micro-dispersion of the catalyst on the graphitic substrate. It is interesting to observe that the Au- CME shows significant oxidation current at high potentials (relevant to the glucose electrooxidation), while at the bulk gold electrode glucose acts as an inhibitor for the oxygen evolution process.Effect of Concentration The peak current of peaks Ia, IIa and IIIa increased linearly with increasing concentration of analytes. The peak potential Ia cathodically shifted by about 1.6 mV mmol-1 in the presence of increasing concentration of glucose. The reaction order deduced for glucose is 1.04, 0.82 and 0.56 for the peaks Ia, IIa and IIIa, respectively. Tafel slopes measured at the beginning of peak IIa (scan rate, 5 mV s-1) changed consistently with the carbohy- drates concentration. For example the Tafel slopes relevant the electrooxidation of fructose and mannitol, changed from 116 k 3 mV decade-' to 220 k 5 mV decade-' on increasing the concentration from 1.2 to 15 mmol 1-1.The change of Tafel slopes with the increase of the concentration supports the assumption that the adsorption process affects the reaction kinetk28 N 5 a E 2 2 0' L I I 1 I I : ' I - 0.00 ] 0.23 - 0.00 I I I I I I -0.800 -0.400 -0.000 0.400 0.800 1.200 EN versus SCE Fig. 2 Cyclic voltammograms at A, Au-CME; and B, bulk gold electrode in de-gassed solutions containing 0.2 mol 1-1 NaOH and 4.0 mmol 1 - 1 glucose (dashed curves); scan rate, 100 mV s- l . For further explanation see text. Effect of the Scan Rate In the presence of 5 mmol l-1 glucose, anodic peaks currents IIa and IIIa increased with increasing scan rate. Plots of the logarithm of peak current density versus the logarithm of sweep rate (not shown here) gave straight lines; the slopes were 0.64 and 0.67 for peak IIa and IIIa, respectively.This finding suggests that the electrochemical process is not an ideal diffusion-limited step, but electrocatalysis involves a relatively slow adsorption or desorption of the sugar molecular at the CME surface. Moreover peak potentials IIa and IIIa shifted towards positive potential and peak potential Ia shifted cathodically when increasing the sweep rate. For example, peak potentials Ia, IIa and IIIa were, respectively, 0.05, 0.130 and 0.394 V at 20 mV s-1, and -0.06, 0.147 and 0.416 V at 500 mV s-*. This behaviour is characteristic of an irreversible charge-transfer process. Effect of p HIHydroxide Concentration The effect of pH on the electrode signal was investigated by cyclic voltammetry of 3 mmol 1- 1 glucose solutions maintained at 1.0 mol 1-1 ionic strength with NaC104.Fig. 3 shows the current densities relevant to oxidation waves Ia, IIa and IIIa in the range pH 10-14. A remarkable increase of the oxidation current was observed in the pH range 10-13. An increase of the reaction rate with increasing pH suggests a significant role of OH- ions in the oxidation of carbohydrates on gold elec- tr0des.~9?30 Adsorbed hydroxyl radicals (AuOH) formed at potential values considerably below the onset of anodic formation of the AuO, were directly involved in the electroox- idation mechanism of the carbohydrates. In particular, carbohy- drates are preliminarily adsorbed on AuOH and this would take place through the formation of hydrogen bridges between the hydroxyls adsorbed on the catalyst and the hydroxyl group of the anal~te.~9--3* The decrease of the oxidation current relevant to peak la at higher pH values is most likely associated with the considerable formation of oxide species (such A u ~ O ~ ) , which act as inhibitors for the adsorption of organic species.Effect of Temperature Fig. 4 shows the effect of temperature on the electrooxidation process of glucose in 0.2 moll-1 NaOH. The rise in temperature caused a noticeable increase of peak currents Ia, and IIa, in agreement with their kinetic character. At the same time, peaks IIa and IIIa were shifted to more cathodic potential values, while peak Ia was shifted to more anodic potential values.Similar results were obtained for all sugars and alditols investigated. Assuming an Arrhenius-type temperature dependence, the peak current densities can be expressed by: (1) log I = k - E/(2.3 RT) la Ila llla 0.0 I , 1 I I I I 10 11 12 13 14 15 Fig. 3 Plots of current densities of 3.0 mmol l-1 glucose at the Au-CME as a function of the pH. Constant ionic strength NaOH + NaC104 (I = 1.0). Other conditions as in Fig. 2. For further explanation see text.252 Analyst, February 1996, Vol. 121 y E a g 5 - $ - where E is the apparent activation enthalpy. The apparent activation enthalpies of several carbohydrates calculated ac- cording to eqn. (1) are given in Table 1. It can be seen that, for a given potential range, the values of apparent activation enthalpies are almost independent of the molecular mass of the sugars. Furthermore, since experimental values for the apparent activation enthalpies of alditols, monosaccharides and polysac- charides are quite similar, it is also likely that these compounds react by a similar electrooxidation mechanism. The values of the apparent activation energies relevant to peaks IIa and IIIa are very small, thus suggesting that in these potential regions, sorption processes are the rate-determining steps.In contrast, the rather larger value of the energy of peak Ia suggests that the oxidation mechanism is under the mixed control of both sorption/desorption and chemical processes. -1.54 0.00 Flow Injection Measurements Flow injection experiments showed a continuous decrease in the amperometric signal for applied potentials ranging between 0.0 and 0.3 V versus Ag/AgCl.As expected,5-7,9.32-34 the strong adsorption of reactants, stable intermediates and reaction products on the electrode surface led to the fouling of catalytic sites. Thus, pulsed amperometric detection techniques are necessary to preserve the catalytic performance. The appro- priate waveform for triple-pulse amperometry was experimen- tally obtained using FI at 1.0 ml min-1, with 0.2 moll-' NaOH as the mobile phase. Hydrodynamic voltammograms reported in Fig. 5 are relevant at 0.444 mmol 1-I of glucose, 0.432 mmol l-1 of xylitol, and 0.351 mmol 1-1 of lactose. The maximum response of the limiting current was found in the potential range 0.2-0.3 V versus Ag/AgCl. The same behaviour was observed for all of the investigated carbohydrates.Curve B, relevant to the background current, shows a constant value (I 1-12 pA) in the potential range between 0.2 and 0.25 V, while at higher potentials increase the current drastically.8 The selected waveforms used in the flow systems were: cleaning step E2 = 0.65 V for t2 = 200 ms; activation step E3 = -0.70 V for t 3 = 200 ms; and detection step El = 0.2 V for tl = 400ms. I I I I I I I 8 "C I I I I I I -0.800 -0.400 0.00 0.400 0.800 1.200 EN versus SCE Fig. 4 Cyclic voltammograms at Au-CME in 0.2 mol 1-1 NaOH plus 2.5 mmol 1- 1 glucose at different temperatures. Experimental conditions as in Fig. 2. Table 1 Apparent activation enthalpies (kJ mol-l) of carbohydrate electrooxidized at the Au-CME* Peak Ia Peak IIa Peak IIIa Glucose 55.0 14.1 4.2 Lactose 59.1 18.9 6.4 Fructose 60.0 18.6 4.7 Arabinose 54.5 19.2 5.9 Sorbitol 56.6 18.9 5.2 Mannitol 53.5 15.6 4.8 * Apparent activation enthalpy of peak Ia was measured on the cathodic sweep. Potential was scanned between -0.6 and 0.9 V at 100 mV s-1.Supporting electrolyte, 0.2 mol 1-1 NaOH. Repetitive FI of 1.0 mmol l-1 glucose, at 1.0 ml min-1, was performed to check the stability of the amperometric signal. Over 4 h of analysis, a 34% diminution of the oxidation current was observed. The progressive erosion of the catalyst onto the glassy carbon, when used in flow-through systems in PAD mode, is the main cause for the decreased amperometric signal. A better stability was obtained by using a mobile phase containing 1.0 pmmol 1-1 Au3+: after 8 h of operating time, corresponding to the analysis of about 80 injections of 1.0 mmol l-1 glucose, an average of 1.7% decrease in electrode response was observed.The effect of flow rate on the response was determined by measuring the electrocatalytic oxidation current of 0. I and 1 .O mmol 1-1 glucose solutions. The peak current increased with flow rate up to 0.8 ml min-l and then practically did not change up to a flow rate of 1.6 ml min-1. At flow rates higher than 2.0 ml min-1, the current was found to be inversely related to flow rate. These results are in agreement with the assumption that oxidation process kinetics are independent of mass transport and can be interpreted in terms of a surface-confined rate- limiting step. In order to provide a clear overview of the electrode capability towards catalytic oxidation of carbohydrates, the molar response factors of some representative compounds were evaluated.The results indicate that the molecular size of the analytes is of primary importance. Consequently, alditols (such as mannitol, sorbitol, xylitol, etc.) and monosaccharides (glucose, galactose, etc.) show a molar response that is higher than lactose and raffinose. For example, the ratio of molar responses of lactose to glucose is about 0.6 and of raffinose to glucose is 0.4. Once again, it is suggested that steric hindrance, adsorption/orientation probably determine the over-all oxidation process. Chromatographic Separations, Calibration, Detection Limit and Reproducibility Since all carbohydrates are weakly acidic with pK, values in the range 12-1 4, anion-exchange chromatography in strongly basic solutions represents a reasonable separation approach for these compounds.Fig. 6 shows an example of a chromatogram obtained with a standard mixture of several representative alditols and carbohydrates by using a Carbopac PA1 column with 0.2 moll-' NaOH solution containing 1.0 pmol 1-1 Au3+ as mobile phase, with a flow rate of 0.6 ml min- I . Since glucose and fructose were not well separated from galactose and ribose, respectively, and xylitol, mannitol and sorbitol were partially 40 30 / 25 - - l o ! n I I I I , I , , , , , I , I I , , -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 EN versus AgJAgCI Fig. 5 Hydrodynamic voltammograms in FI at the Au-CME. Injection volume, 50 p1; electrolyte/carrier, 0.2 mol 1-1 NaOH plus 1.0 pmol 1-I Au3+; flow rate, 1.0 ml min-1. E2 and E3 potential values, 0.65 and -0.70 V, respectively; relevant time values, tl = 400 ms, 12 = 200 ms, r3 = 200 ms.A, 0.444 mmol 1-1 glucose; B, 0.432 mmol 1-1 xylitol; C, 0.351 mmol 1-I lactose; D, background current.Analyst, February 1996, Vol. 121 253 overlapped (particularly at high concentration), the separation was also accomplished with an anion-exchange Carbopac MA1 column. Fig. 7 shows the chromatogram of several carbohy- drates obtained using a 0.4 mol 1-1 NaOH solution containing 1.0 pmol l-1 Au3+ as the mobile phase, with a flow rate of 0.4 ml min-1. The analytical results obtained from the chromato- graphic analysis using a Carbopac PA1 column are summarized in Table 2.As reported in Table 2, the response of the alditols d . - 2 6 10 Timelmin Fig. 6 Liquid chromatogram for a mixture of alditols and carbohydrates (in mmol 1-I): a, 0.324 xylitol; b, 0.278 sorbitol; c, 0.319 arabinose; d, 0.333 glucose; e, 0.335 fructose; f, 0.263 lactose; and g, 0.226 sucrose. Column, Carbopac PAl; flow rate, 0.6 ml min-1; carrier electrolyte, 0.2 mol 1-1 NaOH plus 1.0 pmol 1-1 Au3+; pulsed detection as in Fig. 5. e I 10 20 30 40 Time/min Fig. 7 Liquid chromatogram for a mixture of alditols and carbohydrates (in mmol 1-I): a, 1.479 xylitol; b, 1.236 sorbitol; c, 1.236 mannitol; d, 1 S O glucose; e, 1.574 galactose; f, 1.225 fructose; g, 1.44 ribose. Column, Carbopac MAl; flow rate 0.4 ml min-1; carrier electrolyte, 0.4 mol 1-1 NaOH plus 1.0 pmol 1-1 Au3+; pulsed detection as in Fig.5. and monosaccharides was linear over about three orders of magnitude above the detection limits. For polysaccharides (such as lactose and sucrose) the linear dynamic range was over about two orders of magnitude. It is interesting to observe that the extension of the linear range depends on the molecular size of the analytes. This finding confirms that the adsorption/ desorption processes of the sugars on catalytic sites are directly involved in the electrooxidation kinetics. On considering that an adsorption process takes place prior to the electrooxidation step, and by invoking a Langmuir-isotherm type behaviour, the l/Zp versus l/c relationship should be linear. For all carbohydrates under investigation, l/Zp versus l/c plots exhibit a greater linear range with good correlation coefficient (0.999).The limit of detection (LOD) was determined using a signal- to-noise ratio of 3 from the lowest injected concentration. Although the detection limits at the Au-CME are higher than those obtained with other amperometric sensors,2,4~2*,35 the Au- CME is more convenient for the determination of carbohydrates in real matrices. Precision, estimated in terms of relative standard deviation (3,) by six repetitive chromatographic analyses with a Carbopac PA1 column (over 3 h) of a solution 0.216 mmoll-l of xylitol, 0.186 mmoll-1 of sorbitol, 0.212 mmoll-1 of arabinose, 0.222 mmol I-' of glucose, 0.222 mmol 1-l of fructose, 0.176 mmoll-1 of lactose and 0.150 mmol 1-1 of sucrose, range between 0.8 and 3.5%.The detector stability was chromatographically tested (using a Carbopac PA1 column) by measuring the decrease in the amperometric signal after 8 h of operating time, corresponding to the analysis of about 25 consecutive injections of 0.052 mmoll-l of xylitol, 0.077 mmol 1-l of glucose and 0.064 mmoll-l of lactose. After the initial stabilization of the electrode response, generally in 30 min, an average decrease of about 1.5 to 3.0% of the amperometric signal, was observed. It is interesting to observe that the bulk gold electrode shows an increase in electrode response, over 8 h of operation, of about 15%.5,33 This phenomenon was previously explained as being owing to the continuous roughening of the electrode surface.Analytical Applications As an example of the analytical applications of the Au-CME, the determination of glucose in biological fluids (such as human urine and blood serum), using a Carbopac PA1 column was performed. A chromatogram of a human blood serum obtained in anion-exchange mode by using 0.2 mol 1-l NaOH as the mobile phase is shown in Fig. 8A. Real samples were centrifuged for 5 min (3000 rpm) and diluted 100-fold with 0.2 mol 1-1 NaOH. The concentration of glucose was determined by the standard additions method. Fig. 8B shows the plot of amperometric signal versus added glucose. Recovery and precision data are summarized in Table 3. The results for the determination of glucose by LC-EC using the Au-CME were compared with those obtained by an enzymic method (glucose Table 2 Quantitative parameters of carbohydrates at an Au-CME* ST Linear range LOD/ng Linear range/moll-' r 2 (n = 6) 1/1, versus l/c X y litol 6 7.8 X 10-7-6.4 X 0.998 1.5 7.8 X 10-7-1.7 X 10-3 Sorbitol 9 9.9 X 10-7-7.4 X 0.998 1.6 9.9 x 10-7-1.8 X 10-3 Arabinose 6 8.0 x 10-7-4.2 x 10-4 0.997 2.2 8.0 X 10-7-1.7 X 10-3 Glucose 4 4.4 x 10-7-2.2 x 10-4 0.998 1.9 4.4 X 10-7-1.65 X Fructose 10 1.0 x 10-6-3.3 X 0.997 3.5 1.0 X 10-6-1.66 X 10-3 Lactose 9 5.0 X 10-7-7.5 X 0.999 0.8 5.0 X 10-7-1.37 X 10-3 Sucrose 12 7.0 x 10-7-7.5 X 10-5 0.999 2.4 7.0 X 10-7-1.13 X 10-3 * Liquid chromatographic analysis: Carbopac PA1 column; mobile phase, 0.2 mol 1-I NaOH plus 1.0 pmol 1-1 Au3+; flow rate, detection: E l = 0.2 V , t l = 0.4 s, E2 = 0.65 V , t2 = 0.2 s, E3 = -0.7 V , t3 = 0.2 s.r2 0.999 0.998 0.999 0.999 0.999 0.999 0.999 0.6 ml min-'; pulsed254 Analyst, February 1996, Vol. 121 oxidase) with a Beckman oxygen electrode.36-37 As shown in Table 3, the high level of recovery and the good agreement with the enzymic method are promising features for the use of the Au-CME in chromatographic systems for the determination of carbohydrates in real matrices. Conclusion Gold particles finely dispersed on glassy carbon substrates were investigated by SEM and CV in an alkaline medium. The performance of the Au-CME in the electrooxidation of carbohydrates was evaluated. The oxidation takes place through the preliminary adsorption of the analyte on the AuOH sites. The apparent activation enthalpies confirm that oxidation processes are surface-confined rate-limiting steps with a mixed control of both sorption/desorption and chemical processes.The suitability of the Au-CME as an amperometric detector for carbohydrates in flowing streams, including in liquid chromatography, has been explored. The Au-CME shows good sensitivity and stability over time. Determination of glucose in human biological fluids by LC- EC showed good precision and recovery, confirming the suitability of the Au-CME as an amperometric sensor for the determination of carbohydrates in real matrices (see Table 3). I Tirnehin Fig. 8 A, Liquid chromatogram relevant to human blood serum diluted 1 + 99 with 0.2 mol I-* NaOH solution; a, glucose. The sample was centrifuged for 5 min at 3000 rpm. Column, Carbopac PAI. Other experimental conditions as in Fig.6. B, Inset shows the plot of the amperometric signal versus glucose added. Table 3 Determination and recovery of glucose in human biological fluids. Comparison with an enzymic method* Found/ Recovery s, (%) Enzymic mg dl-I (%) (n = 4) test Blood serum 76.5 - 2.2 78.0 Blood serum + 94.3 97.7 2. I - Blood serum + 119.2 102.1 2.1 Urine 255.4 - 1.8 250.0 Urine + 291.2 102 1.9 20 mg dl-1 40 mg dl-1 - - 30 mg dl-1 * Liquid chromatographic analysis: Carbopac PA 1 column; the sam- ples were centrifuged at 3000 rpm. Other conditions as in Table 2. The ease of preparation of the gold microparticles dispersed onto the graphitic structure and the possibility of exploring several metal and adatom combinations confirms the potential interest of this CME system for the electrochemical detection of organic compounds in LC.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 27 28 29 30 31 32 33 34 35 36 37 Santos, L. M., and Baldwin, R. P., Anal. Chem., 1987, 59, 1766. Wang, J., and Taha, Z., Anal. Chem., 1990,62, 1413. Zadeii, J. M., Marioli, J., and Kuwana, T., Anal. Chem., 1993, 63, 649. Reim, R. E., and Van Effen, R. M., Anal. Chem., 1986, 58, 3203. Rocklin, R. D., and Pohl, C. A., J . Liq. Chromatogr., 1983, 6(9), 1577. Johnson, D. C., Dobberpuhl, D., Roberts, R., and Vandeberg, P., J. Chromatogr., 1993, 640, 79. Neuburger, G. G., and Johnson, D. C., Anal. Chim. Acta, 1987,192, 205. Roberts, R. E., and Johnson, D. C., Electroanalysis (N.Y.), 1994, 6, 269. Neuburger, G.G., and Johnson, D. C., Anal. Chem., 1987,59,203. Le Fur, E., Etievant, P. X., and Meunier, J. M., J. Agric. Food Chem., 1994, 42, 320. Luo, P., Zhang, F., and Baldwin, R. P., Anal. Chem., 1991, 63, 1702. Yuan, C. J., and Huber, C. O., Anal. Chem., 1985, 57, 180. Fleischmann, M., Korinex, K., Pletcher, J., J. Chem. SOC., Perkin Trans. 2, 1972, 1396. Wang, E., and Liu, A., J. Electroanal. Chem., 1991, 319, 217. Baldwin, R. P., and Thomsen, K. N., Talanta, 1991, 38, 1. Ikariyama, Y., Heineman, W. R., Anal. Chem., 1986,58, 1803. Ohsaka, T., Okajima, T., and Oyama, N., J. Electroanal. Chem., 1986,215, 191. O’Sullivan, E. J. M., and Calvo, E. J., in Electrode Kinetics: Reactions, ed. Compton, R. G., Elsevier, Oxford, 1987, vol. 27, ch. 3. Casella, I. G., Desimoni, E., and Salvi, A. M., Anal. Chim. Acta, 1991, 248, 117. Cataldi, T. R. I., Casella, I. G., Desimoni, E., and Rotunno, T., Anal. Chim. Acta, 1992, 270, 161. Casella, I. G., Cataldi, T. R. I., Salvi, A. M., and Desimoni, E., Anal. Chem., 1993,65, 3143. Casella, I. G., and Marchese, R., Anal. Chim. Acta, 1995, 311, 199. Rand, D. A. J., and Woods, R., J. Electroanal. Chem., 1971, 31, 29. Yahikozawa, K., Nishimura, K., Kumazawa, M., Tateishi, N., Takasu, Y., Yasuda, K., and Matsuda, Y., Electrochim. Acta, 1992, 37, 453. Mulvaney, P., Giersic, M., and Henglein, A., J. Phys. Chem., 1992, 96, 10419. Bruckenstein, S., and Shay, M., J. Electroanal. Chem., 1985, 188, 131. Angerstein-Kozlowska, H., Conway, B. E., Hamelin, A., and Stoicoviciu, L., J . Electroanal. Chem., 1987, 228, 429. Holze, R., Luczak, T., and Beltowska-Brzezinska, M., Electrochim. Acta, 1994, 39, 485. Beltowska-Brzezinska, M., Electrochim. Acta, 1985, 30, 1193. Ocon, P., Alonso, C., Celdran, R., and Gonzalez-Velasco, J., J. Electroanal. Chem., 1986, 206, 179. Vitt, J. E., Larew, L. A., and Johnson, D. C., Electroanalysis (N.Y.), 1990, 2, 2 1. Neuburger, G. G., and Johnson, D. C., Anal. Chem., 1987,59, 150. Bindra, D. S., and Wilson, G. S., Anal. Chem., 1989, 61, 2566. Johnson, D. C., and Lacourse, W. R., Anal. Chem., 1990, 62, 589A. Prabhu, S, V., and Baldwin, R. P., Anal. Chem., 1989, 61, 852. Kadish, A. H., Little, R. L., and Sternberg, J. C., Clin. Chem. (Winston-Salem, N.C.), 1968, 14, 116. Morrison, B., Clin. Chim. Acta, 1972, 42, 192. Paper Sl05613F Received August 23,1995 Accepted October 13,1995

ISSN:0003-2654    

DOI:10.1039/AN9962100249

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 (255-258) 255 Pyrite as Sensor for Potentiometric Precipitation Titrations M. M. Antonijevita, B. Vukanovika and R. Mihajlovicb a Technical Faculty Bor, Belgrade University, P.O. Box 50, I9210 Bor, Yugoslavia Faculty of Science, University of Kragujevac, 34000 Kragujevac, Yugoslavia The potentiometric titration of chloride, bromide, iodide and thiocyanate with silver nitrate, using the mineral pyrite as a sensor for the detection of the end-point, is described. The potentials of the pyrite electrode were found to be established rapidly, and accurate and reproducible results were obtained. The pyrite electrode potential showed a sub-Nernstian dependence on the concentration of the ionic species investigated. In addition, lead(1r) was determined by titration with potassium chromate and chloride by titration with standard mercury(1) solution.Accurate and reproducible results were also obtained in the reverse titrations. Keywords: Potentiometric titrimetry; pyrite sensor; chloride; bromide; iodide; thiocyanate Introduction Potentiometric titration is often used as a suitable method for the determination of chloride,*-l2 bromide, 13-21 iodide2 l-29 and thiocyanate.30-33 In most instances a standard silver nitrate solution is used as the titrating agent and silver wire and silver(1) sulfide electrodes are employed as the indicator electrodes. In situations in addition to halides, the content of acids or bases present in the solutions investigated is to be determined, the titration end-point (TEP) is detected by means of a glass electrode.Our previous investigations34 have shown that the mineral pyrite can be used as a sensor for the detection of the TEP in the titration of acids and bases. The results presented in this paper show that pyrite can also be used as a sensor in the titration of halides, silver(1) nitrate, chloride with mercury(1) solution and lead@) with potassium chromate solution. Hence, pyrite represents a suitable electrode material for the detection of the TEP. Experimental Apparatus and Reagents The potential difference during the course of the titrations was measured with an Iskra MA 5740 microprocessor pH-meter. The following electrodes were used: 1. A pyrite electrode prepared as described in a previous paper.34 Chemical and spectrographic analyses of the sample showed that the chemical composition of the pyrite used was as follows: Fe, 45.32; Mn, 0.001; Al, 0.36; Cu, 0.54; Pb, 0.002; Sn, 0.001; Ni, 0.01; Ag, 0.001; Zn, 0.1; Ti, 0.05; Mo, 0.001; and SiO2, 1.86%.2. A silver electrode in the form of a plate (area, S = 1 crnz), the contact being realized with a silver wire. 3. A saturated calomel electrode (SCE). All solutions employed were prepared from analytical- reagent grade chemicals and were standardized by known volumetric procedures. Procedure An appropriate volume of a halide or thiocyanate solution of known concentration was placed in a titration vessel; the solution was diluted to 80 ml with water followed by the addition of 20 ml of 5% Ba(N03)2. The pyrite indicator electrode, connected to the SCE by means of a salt-bridge, was immersed in the solution.After introducing the electrode into the circuit the potential difference was measured, while the solution was stirred with a magnetic stirrer. The potentiometric titration was then carried out by adding the titrant in portions and measuring the potential difference after each addition. A similar procedure was used in performing the reverse titration and in titrations where a silver electrode was used as the indicator electrode. The potentiometric titration of lead(I1) with potassium chromate was performed in 30% ethanol at 5OoC, whereas chloride was determined potentiometrically with mercury(1) in 1 mol 1-1 nitric acid. The measurement of the stationary potential of the pyrite electrode was carried out in a series of halide and thiocyanate solutions with concentrations in the range from 1 X 10- 1 to 1 X 10-6 moll-1, the potential of the pyrite electrode being followed with time.The stationary potential of the pyrite electrode was also measured in silver nitrate and Hg' ion solutions, the concentrations of which ranged from 1 X 10-1 to moll-l. The potential measurements were carried out by coupling the pyrite electrode with the SCE, at 25 k 0.1 "C. Results and Discussion Stationary Potential of the Pyrite Electrode The stationary potential of the pyrite electrode is linearly related to pH34 and this paper deals with an investigation of the dependence of the pyrite electrode potential on the logarithm of the halide and thiocyanate concentrations; this investigation was performed by following the potential of the pyrite electrode with time in solutions of various chloride, bromide (Fig.l), 330 I I I 280 t 130 I pBr = 1 80 I I I I I I 0 2 4 6 8 1 0 1 2 1 4 Umin Fig. 1. solutions of various concentrations. Dependence of the pyrite electrode potential on time in bromide256 Analyst, February 1996, Vol. 121 iodide and thiocyanate concentrations, It was found that the potential was established very rapidly at both low and high concentrations of the ionic species investigated. The depend- ences E versus pX, pSCN were obtained by plotting the electrode potential measured 20 min after equilibration against the logarithm of the corresponding concentration. The slopes of the curves thus obtained are given in Table 1.Deviations from linearity were found at lower concentrations. It is known that pyrite can contain variable amounts of silver, copper, gold, etc., which would affect the potential of the pyrite electrode. Owing to the presence of silver, the pyrite electrode may behave similarly to the silver electrode and this may explain the functional relationship between the pyrite electrode potential and the logarithm of the halide and thiocyanate concentrations. However, the starting material used for the preparation of the pyrite electrode employed in this investigation showed silver and copper contents of 0.001 and 0.54%, respectively. Such a low silver content casts doubt on the explanation given above for the behaviour of pyrite in halide solutions. We believe that the adsorption of halide ions on the pyrite surface may be involved.Similarly, the dependence of the stationary potential of the pyrite electrode on silver(1) and mercury(1) ion concentrations, respectively, was also investigated (Table 1). It was found that the pyrite electrode showed a sub-Nemstian dependence with a slope of 36 mV decade-' for Ag+ and 69 mV decade-' for Hg22+, respectively (Fig. 2). This behaviour of pyrite is important as regards its use as a sensor in precipitation titrations and may be explained by the fact that in silver(1) solution the pyrite surface is coated with elemental silver(1) sulfide (Ag2S). The formation of elemental silver on the pyrite surface probably proceeds according to the following reaction: (1) 2Ag+ + FeS2 -+ 2Ago + Fe2+ + 2S0 Table 1 Slopes and linearity ranges for the ions investigated Ion Slope/ Linearity Correlation investigated mV decade-' range/mol 1-1 coefficient c1- 45 10-1-10-5 0.998 Br- 61 10-1-10-4 1 .ooo I- 44 10-1- 10-5 0.99 1 SCN- 38 10-1-10-4 0.998 Hg22+ 69 1 0-2- 10-6 0.999 Ag+ 36 104-10-5 1 .ooo 530 > 430 330 230 n = 36.05 r = 1.000 - n = 69.13 r = 0.999 1 2 3 4 5 6 PM ' Fig.2. Dependence of the pyrite electrode potential on 10g[Hg~~+] and log"4g'I. Hiskey et aE.35 investigated the reaction between Ag+ ions and pyrite in 0.25 moll-' sulfuric acid and established that elemental silver is formed according to eqn. (1): however, after the addition of large amounts of an oxidizing agent [e.g., 1 mol 1-1 Fe2(S04),], the silver dissolves according to the following equation: Fe3f + Ago + Fe*+ + Ag+ (2) These workers also found that with increasing temperature (25-80 "C, cAg+ = 1 X 10-4 moll-1) the rate of Ago formation increased.Stirring of the electrolyte was found to increase the rate of deposition of silver on the pyrite surface.36 In order to determine the products formed in the reaction of pyrite with sulfuric acid solution containing silver ions, Buckley et ~ 1 . ~ 7 used voltammetry and photoelectron spectroscopy. It was found that Ag2S is the main reaction product throughout the reaction process and that the formation of this product proceeds according to the following equation: 4FeS2 + 14Ag+ + 4H20 --., 7Ag2S + 4Fe2+ + S042- + 8H+ (3) After a longer period of time, the presence of Ago was evident on the basis of the characteristics of the voltammogram.Silver(1) could not be detected by photoelectron spectroscopy, as it appeared in the form of crystallites, which occupy a small part of the pyrite surface, covered with silver(1) sulfide. It was also concluded that sulfur atoms arising from the reacted pyrite which are not found as Ag2S, are transferred to the solution as sulfate ions. On the basis of this information an approximately Nernstian response for Ag+ ions was to be expected; however, owing to the formation of elemental sulfur [eqn. (l)], which behaves as an insulator, the pyrite electrode exhibited a smaller sensitivity to changes in Ag+ concentration than expected. Similarly, the formation of silver crystallites decreased the sensitivity. In this instance, the potential of the pyrite electrode can be defined as (4) where n is the slope (mV decade-').electrode can be defined by a similar equation In halide and thiocyanate solutions the potential of the pyrite Similar considerations are probably valid also for the interaction of pyrite with mercury(1). In this instance the slope was greater, which suggests that pyrite is not covered with a layer of sulfur, which would diminish the slope, but is coated only with mercury(1) sulfide and/or elemental mercury. The latter proba- bly causes the pyrite to behave as a redox electrode, which gives rise to an increase in the slope. The characteristics of the pyrite electrode in solutions of halide, thiocyanate, silver(1) and mercury(I), respectively, indicated that pyrite would probably be a good sensor for the detection of the TEP in precipitation reactions.Potentiomehic Titrations The results of the potentiometric titrations showed that the highest total potential changes (240-420 mV) and the highest potential jumps at the TEP(180-265 mV per 0.1 ml) (Fig. 3) were obtained in the determination of iodide with 0.1 moll-' silver nitrate solution. In the determination of chloride and bromide the total potential changes were in the range 170-3 10 and 2 10-260 mV, respectively, whereas the potential jumps at the TEP ranged from 50 to 80 mV per 0.1 ml and from 100 to257 Analyst, February 1996, Vol. 121 135 mV per 0.1 ml, respectively. In the argentimetric deter- mination of thiocyanate with potentiometric detection of the TEP the total potential changes were in the range 190-230 mV, whereas the potential jump at the TEP was 100-130mV per 0.1 ml.The height of the potential jump at the TEP is in agreement with the thermodynamic constants for the solubility products of AgI, AgBr, AgSCN and AgC1, respectively. Large total potential changes and pronounced potential jumps at the TEP make possible the determination of halides in a mixture (Table 2). In reverse titrations, the results obtained were similar to those obtained for direct titrations. The potentiometric titrations using the pyrite electrode were found to give smaller total potential changes and smaller potential changes at the TEP in comparison with those obtained with the silver electrode. 140 - 0 2 4 6 8 10 12 14 16 Volume of Ag'/ml Fig. 3.Potentiometric titration curves for the titration of C1-, Br-, SCN- and I- with a standard silver(^) nitrate solution (CX-,SCN- = 0.0025 mol 1-1). The effect of the temperature on the height of the potential jump at the TEP was also investigated because pyrite, as a semiconductor, may show significant changes in conductivity with a change in temperature. These investigations were carried out by titrating bromide in 5% barium nitrate solution with a standard silver(1) nitrate solution at various temperatures, viz., 20, 55, 70 and 80°C. The TEP was detected with the pyrite electrode, and with the silver electrode for comparison. The results obtained are presented in Table 3. From Table 3 it can be seen that with an increase in temperature the height of the potential jump at the TEP is decreased for both the pyrite and silver electrodes, indicating that, within the range investigated, the temperature has no effect on the pyrite electrode.The temperature, however, affects the system investigated: the higher is the temperature the greater is the solubility of silver(1) bromide; an increase in solubility causes the height of the potential jump at the TEP to decrease. In titrations of chloride with mercury(1) in 1 moll-' nitric acid, the potential jumps at the TEP range from 75 to 110 mV per 0.1 ml. In reverse titrations (chloride ion was the titrating agent), smaller values for the total potential changes were obtained (Fig. 4). Chromate can be successfully determined by potentiometric titration with lead(r1) in 30% ethanol at 50 "C.In these titrations the behaviour of the pyrite electrode was similar to that of the platinum electrode except that the jumps registered at the latter were higher. As in these titrations the redox potential at the TEP is changed, it can be concluded that the pyrite electrode should be applicable to the detection of the TEP in redox reactions. These considerations and the behaviour of pyrite as sensor for these reactions will be reported in a subsequent paper. The rate of potential equilibration at the pyrite electrode during the course of these titrations were fast, and in the vicinity Table 2 Results of potentiometric precipitation titrations and of reverse titrations using the pyrite electrode for the detection of the TEP Ion to be Taken/ Titrating Found+ (%) determined mol 1-1 agent Electrode E,"JmV E,t/mV c1- 0.0026 Ag+ FeS2 160-200 40-70 100.4 k 0.2 c1- 0.0063 Ag+ FeS2 170-310 50-80 100.2 f 0.1 c1- 0.0063 Ag' Ag 280-300 50-60 100.0 * 0.1 Br- 0.0025 Ag' FeS2 200-250 90- 120 99.7 _+ 0.1 Br- 0.0061 Ag+ Ag+ 0.0061 Br- FeS2 250-280 120-130 100.2 f 0.0 SCN- 0.0062 Ag+ FeS2 210-230 90-1 30 99.7 f 0.0 SCN- 0.0062 Ag+ Ag 400-4 10 160-180 99.7 +.0.0 Ag+ 0.0065 c1- FeS2 270-280 50-60 100.6 f 0.1 FeS2 210-260 70-135 99.8 f 0.1 Br- 0.0061 &+ Ag 420-430 180-210 99.8 f 0.0 SCN- 0.0025 Ag+ FeS2 190-210 100-120 100.8 k 0.2 Ag+ 0.0063 SCN- FeS2 240-250 70-100 100.4 f 0.1 I- 0.0025 Ag+ FeS2 240-420 180-265 100.3 f 0.0 I- 0.0062 Ag+ FeS2 230-400 110-180 100.4 f 0.1 I- 0.0062 Ag+ Ag 640-650 390-420 100.3 * 0.0 Ag+ 0.0063 I- FeS2 290-340 200-220 99.7 * 0.0 I- 0.0037 Ag+ FeS2 360-460 70-100 100.3 f 0.1 c1- 0.00 18 Ag+ FeS2 360-460 60-80 101.0 +_ 0.4 c1- 0.0027 Hg22+ FeS2 270-290 95- 105 100.2 f 0.2 Hg22+ 0.0026 c1- FeS2 160-180 80-90 99.9 f 0.1 m z 2 + 0.0026 c1- Ag 16@-175 90-95 100.0 +.0.0 CI-042- 0.0012 Pb2+ FeS2 130-170 40-80 99.9 f 0.1 Pb2+ 0.0013 Chromate FeS2 160-170 100-1 10 100.0 f 0.2 Chromate Pt 240-280 170-200 100.0 f 0.0 Pb2+ 0.00 13 * E, = Total potential change. t Ep = Potential jump at the TEP. * n = 6.25 a Analyst, February 1996, Vol. 121 of the end-point it was less than 1 min. Before and after the TEP was reached, the potential was established within 30 s after the addition of a portion of the titrant. The results obtained in this work are accurate and reprodu- cible; hence, under suitable conditions, the pyrite electrode can be used as a sensor electrode for the potentiometric determina- tion of chloride, bromide, iodide, thiocyanate, silver, mercury, etc.Further investigations have shown that the pyrite electrode can be successfully used in complexometric and redox titrations in water and acid-base titrations in some non-aqueous media (propylene carbonate, tetrahydrofuran, N,N-dimethylforma- mide, pyridine, etc.). The advantage of the pyrite electrode over commonly used electrode systems is its universal character. In contrast to silver(1) and silver wire electrodes, the pyrite electrode can be used for the detection of the TEP in acid-base, precipitation, redox and complexometric titrations. On account of its various applications the pyrite electrode is suitable for mass produc- tion.~ ~~ Table 3 Dependence of the potential jump (mV per 0.1 ml) at the TEP on temperature in the titration of bromide with silver(r) nitrate solution TemperaturePC 20 55 70 80 Pyrite electrode 175 125 71 70 Silver electrode 190 105 90 85 350 300 250 , E ci 320 270 220 0 1 2 3 4 Wml Fig. 4 Potentiometric titration curves for the titration of: I , chloride with mercury(1) and 2, chromate with lead(I1). 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 27 28 29 30 31 32 33 34 35 36 37 Shiner, V. J., and Smith, M. L., Anal. Chem., 1956, 28, 1043. Prokopov, T. S., Anal. Chem., 1970, 42, 1096. Malmstadt, H V., and Winefordner, J D., Anal. Chim. Acta, 1959,20, 283. Bishop, E., Mikrochim.Acta, 1956, 619. Bishop, E., Analyst, 1958, 83, 212. Bishop, E., and Dhaneshwar, R. G., Analyst, 1962, 87, 845. Tomlinson, K., and Torrance, K., Analyst, 1977, 102, 1. Frost, J. G., Anal. Chim. Acta, 1969, 48, 321. Potman, W., and Dahmen, E. A. M. F., Mikrochim. Acta, 1972, 303. Conrad, F. J., Talanta, 1971, 18, 952. Masson, M. R., Talanta, 1975, 22, 933. RPliiEka, J., and Lamm, C. G., Anal. Chim. Acta, 1971, 54, 1. Martin, A. J., Anal. Chem., 1958, 30, 233. Chou, D. H., and Sams, L. C., Microchem. J., 1969, 14, 505. Wharton, H. W., Talanta, 1966, 13, 919. Buzukova, V., Muldan, B., and Zuka, J., Collect. Czech. Chem. Commun., 1965, 30, 28; Anal. Abstr., 1966, 13, 2341. Sramkova, B., Zuka, J., and Dolexal, J., J . Electroanal. Chem., 1971, 30, 177. Oginko, V. S., Zh. Prikl. Khim., 1960, 33, 2486. Pflaum, R. T., Frofliger, J. O., and Berge, D. G., Anal. Chem., 1962, 34, 1812. Bazzelle, W. E., Anal. Chim. Acta, 1971, 54, 29. Turner, D. L.. J . Food. Sci., 1972, 37, 791. Kotkowski, S., and Lassocinska, A., Chem. Anal. {Warsaw), 1966, 11, 789. Minczewski, J., and Glabicz, U., Acta Chim. Hung., 1962, 32, 133; Anal. Abstr., 1963, 10, 1424. Hanif, M., Dolezal, J., and Zuka, J., Microchern. J., 1971, 16, 291. Pungor, E., and Hollos-Rokosinyi, H., Acta Chim. Hung., 1961, 27, 63. Pungor, E., Hovas, J., and Toth, K., Acta Chim. Hung., 1964, 41, 239. Pungor, E., and Toth, K., Analyst, 1970, 95, 625. Pungor, E., Anal. Chem., 1967, 39, 28A. Paletta, B., Mikrochim. Acta, 1969, 1210. Kolthoff, I. M., and Lingane, J. J., J . Am. Chem. SOC., 1935, 57, 2377. Bishop, E., and Dhaneshwar, R. G., Analyst, 1962, 87, 207. Hirsch, R. F., and Portock, J. D., Anal. Lett., 1969, 2, 295. Mascini, M., Anal. Chim. Acta, 1962, 62, 29. Antonijevic, M., Mihajlovic, R., and Vukanovii, B., Talanta, 1992, 39, 809. Hiskey, J. B., Phile, P. P., and Pritzker, M. D., Metall. Trans., 1987, 18B, 641. Hiskey, J. B., and Pritzker, M. D., J . Appl. Electrochem., 1988,18, 484. Buckley, A. N., Wouterlood, H. J., and Woods, R., J . Appl. Electrochem., 1989, 19,744. Paper 6/00I 95E* * Paper originally received May 23, 1994 and accepted September 12, 1994. The delay in publication is owing to the postal embargo imposed as a result of the war in Yugoslavia.

ISSN:0003-2654    

DOI:10.1039/AN9962100255

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 (259-262) 259 Piezoelectric Crystal Sensor With a Plasticized Poly(Viny1 Chloride) Coating for Determination of Trace Amounts of Ethanol Vapour Ke-Min Wang, Zhong Cao, Hui-Gai Lin, Shi-Hua Wang, Bin-Feng Wang and Ru-Qin Yu Department of Chemistry and Chemical Engineering, Hunan University, Changsha, 41 0082, China A new type of piezoelectric sensor with a coating supported by adding the active material ETH6022 (N-acetyl-N-dodecyl-4-trifluoroacetylaniline) to a plasticized PVC polymer matrix (PCS-PVC) was successfully prepared for detecting trace amounts of ethanol vapour. Compared with the traditional piezoelectric crystal sensor (PCS) using a single active material coating, the useful lifetime of the new sensor is extended strikingly from several days to more than 2 months, which is comparable to that of an ion-selective electrode.The reproducibility of PCS-PVC is also improved. The new coating method for the sensor was also applied to extending the useful lifetime of other classical gas-sensitive PCSs. Keywords: Piezoelectric crystal sensor; plasticized poly(viny1 chloride) coating; ethanol vapour Introduction A wide variety of sensing systems for ethanol have been developed because of the importance of the on-line determi- nation of ethanol in many biotechnological processes and in biological fluids. 1,2 The sensing principles included enzymic reactions,3-6 acid-base reactions,7,* host-guest reactions9 and even the surface plasmon technique.10 Among these sensing techniques, the chemical sensing layer with optical transduction presented by Seiler et al.9 was very interesting; it was based on selective and reversible recognition of alcohols as guests in aqueous solutions by a lipophilic trifluoroacetophenone deriva- tive as a host molecule.With N-acetyl-N-dodecyl-4-trifluor- oacetylaniline (ETH6022) as the host molecule, ethanol can be determined in the range 0.5-35% v/v in aqueous solutions. Here we describe a new way of incorporating the host-guest chemistry into a sensitive piezoelectric crystal sensing tech- nique for the determination of trace amounts of ethanol vapour. Studies of the traditional piezoelectric crystal sensor (PCS) through directly coating the active material on the crystal surface for performing a chemical recognition process to determine various gas substances have developed rapidly.' 1-18 However, the coating method used in many traditional PCS techniques resulted in poor useful lifetimes.This is because the useful lifetime of PCSs is related to the adsorbability of corresponding active material coated on the crystal surface and its vapour pressure. In fact, many active materials are not ideal as absorbents, which should be viscous, adsorptive materials with a low vapour pressure. This makes the useful lifetime of many PCSs very short and their use in practice is seriously affected. In the preparation of a PCS for the determination of organophosphorus compounds, Guilbault et al. l9 described a procedure for fixing the polymer permanently. First, the crystal was coated with a highly viscous polymer used as a stationary phase, then the active material was sprayed onto the polymer on the crystal surface.However, they did not report the useful lifetime of the sensor prepared by that method. Possibly a coating method using a sandwich of active material cannot solve the lifetime problem fundamentally. In a short commu- nication, Moody et a1.20 reported the use of a high molecular mass polyalkoxylate polymer, Antarox CO-880 (nonylphenox- ypolyethoxylate with 30 ethoxylate units), as a matrix for supporting pyridoxine hydrochloride used as a sorbent for the piezoelectric quartz crystal determination of ammonia. Because of the use of a polymer matrix, the useful lifetime of the sensor was extended from about 10 days to more than 53 days. As an inert polymer matrix, poly(viny1 chloride) (PVC) has been widely used in carrier membrane ion-selective electrodes and optodes.21 Here, we used a plasticized PVC membrane with the addition of an active material, ETH6022, as the coating of a PCS, that is, the active material was supported by the PVC matrix for the determination of trace amounts of ethanol vapour.The response properties of the new type of PCS based on PVC membrane containing an active material (termed PCS-PVC) and the traditional PCS based on a single active material coating (termed PCS) were compared in detail. The experimental results showed that the useful lifetime of the new PCS-PVC reported here is extended considerably because of the high adsorbability of the PVC membrane on the surface of the piezoelectric crystal and the stability of the active material in this coating.Experimental Apparatus and Reagents An AT-cut piezoelectric quartz crystal with a fundamental frequency of about 9 MHz in which Ag electrodes had been deposited on both sides was obtained from the Beijing 707 Factory. The crystal was connected with a laboratory-made TTL-IC piezoelectric crystal oscillator circuitry. All of circuits were driven at 5 V dc, and the frequency of the oscillating quartz crystal was measured with a frequency counter with a 0.1 Hz resolving power (CN3 165, Sampo, Taiwan). A gas flow system was used to produce ethanol vapour samples with different concentrations (Fig. 1). In this system, the piezoelectric quartz crystal in the measuring chamber was purged for about 5 min with nitrogen, which was dried and cleaned up by passage through concentrated sulfuric acid until the output frequency of the crystal stabilized.Then the saturated ethanol vapour was obtained by bubbling dry nitrogen through pure absolute ethanol liquid which had been freshly distilled, held in several bottles, and mixed with dry nitrogen in well defined proportions controlled by three rotameters to obtain a total flow rate of 60 ml min-1. The mixed ethanol vapour samples were detected by the crystal when they flowed through the measuring chamber. The actual concentration of ethanol260 Analyst, February 1996, Vol. 121 vapour in this mixture was measured by gas chromatography. The detection process was controlled at room temperature (25 f 0.5 "C).ETH6022 was a gift from Professor W. Simon (Swiss Federal Institute of Technology, Zurich, Switzerland). The synthesis of this compound has been described.22 PVC (high molecular mass, Fluka Selectophore), bis(2-ethylhexyl) sebacate (C.P. grade, Shanghai Chemical Reagents) and tetrahydrofuran (THF) (analytical-reagent grade, Shanghai Chemical Reagents) were used for membrane preparation. All other reagents were of analytical-reagent grade. All solvents were freshly distilled and stored in sealed bottles with some pieces of drying CaC12 before use. Preparation of Coating PCS coating The coating components consisted of 3.8 mg of ETH6022 and 1.2 mg of methyltridodecylammonium chloride (MTDDACl). These components were dissolved in 5 ml of THF that had been freshly distilled.The resulting solution was coated on both silver electrodes of the piezoelectric crystal and spread with a glass rod. This process was repeated until an approximately 5-7 kHz frequency shift of PCS was obtained after evaporation of the solvent. PCS-PVC coating The coating components consisted of 7.8 mg of ETH6022, 2.5 mg of MTDDACl, 65 mg of bis(2-ethylhexyl) sebacate and 85 mg of PVC. These coating components were dissolved in 1.5 ml of THF that had been freshly distilled. The resulting solution was diluted 10-fold with THF prior to coating. The preparation of the coating was similar to the above. The additive MTDDACl was mainly used for the purpose of enhancing the response rate, as outlined previously.9~23 Results and Discussion Principle of Operation The highly electron-withdrawing trifluoromethyl group of ETH6022 induces a drastic increase in the electrophilicity of the carbonyl carbon atom and, as a consequence, there is a reversible addition reaction between ETH6022 and ethanol? Obviously this interaction leads to an increase in the coating mass of PCS, and the frequency of the piezoelectric quartz crystal should decrease.According to the Sauerbrey equation,Z4 the extent of the frequency shift (AF) is directly proportional to the changed mass in the coating. Therefore, vapour samples with different concentrations of ethanol can be analysed from the corresponding frequency shifts of the piezoelectric crystal. Moreover, an interaction between ethanol and ETH6022 occurs in the optode membrane after the extraction of ethanol from the sample solution into the organic membrane phase.Linear Response Range and Detection Limit In a wet system with 40% relative humidity (RH), the frequency shift responses of PCS and PCS-PVC in the range 30-400 ppm ethanol vapour were recorded. Perfect linear response relation- ships were obtained [Fig. 2(a) and (b)]. The linear response ~~ ~~~~~~~ equations for PCS and PCS-PVC are AF (Hz) = 1090.3 + 7.09C (ppm) and AF (Hz) = 259.9 + 1.61C (ppm), re- spectively. The correlation coefficients for these two kinds of sensors are both 0.998 within the measuring range. The measurement of ethanol vapour for lower or higher ranges was not studied because of the limit of sample preparation, but the detection limit of PCS-PVC is 5 ppm with a confidence level of 90% according to the definition of detection limit given by IUPAC.25 However, for PCS, the detection limit is 1.7 ppm.The zero intercepts for both sensors are exceptionally high owing to the water vapour in the wet system. It is a case of going from a clean, dry system to a wet system plus the alcohol. i _,k h)lru, L - - _ _ - _ - Fig. 1 Gas flow system used to provide ethanol vapour. 1, rotameter; 2, piezoelectric quartz crystal; 3, series of bubbling bottles; 4, measuring chamber; and 5 , dry and clean-up bottle. 4000 3500 3000 2500 2000 1500 1000 2 500 % 0 0 0 3 0 80 160 240 320 400 480 1000 C 3 U 2 u 800 600 400 200 n I I I I l o a0 160 240 320 400 480 c (PPm) Fig. 2 Response curve of (a) PCS and (b) PCS-PVC based on ETH6022 in, A, wet- and, B, dry system.Analyst, February 1996, Vol.121 26 1 In the dry system, similar linear response relationships were obtained to Fig. 2(a) and (b). The linear response equations for PCS and PCS-PVC responding to 30-400 ppm ethanol vapour are AF (Hz) = 95.1 + 7.18C (ppm) and AF (Hz) = 17.0 + 1.65C (ppm), respectively. The correlation coefficients for these two kinds of sensors are 0.998 and 0.999, respectively, within the measuring range. The detection limits of PCS and PCS- PVC are 0.2 and 1.2 ppm, respectively, which are lower than those in the wet system. The zero intercepts for both sensors become much lower than those in the wet system owing to no or very little water vapour in the gas streams. The hydrophobicity of the plasticized PVC matrix causes a much greater decline of the zero intercepts for PCS-PVC than for PCS in both the wet and dry systems. Further, the decline in the wet system is sharper than in the dry system. Therefore, the bulk of this work was carried out in dry gas streams.Reversibility, Response Time and Reproducibility The changes with time of the oscillating frequency of PCS and PVC-PCS for ethanol vapour concentration changes in steps between 0 (pure nitrogen) and 100 or 50 ppm in the wet system are shown in Fig. 3(a) and (b). These curves show that the responses of PCS and PCS-PVC are both reversible. For the 50 ppm ethanol vapour sample, the response and recovery time for PCS-PVC are 4 and 2.5 min, respectively, showing larger delays than the corresponding values for PCS for the 100 ppm ethanol vapour sample, 2 and 1.5 min, respectively.It is obvious that the delay is related to the mass transmission process when the ethanol vapour is absorbed into the PVC matrix coating. Hence the measuring time for PCS-PVC will be slightly longer than that for PCS. However, the reproducibility of PCS-PVC has been greatly improved. The reproducibilities of these two sensors for triplicate analyses of several ethanol vapour samples of different concentrations are given in Table 1. It can be seen that 6000 5000 4000 N 3000 I a a 7300 7200 71 00 7000 6900 6800 100 pprn ethanol 0 100 200 300 400 500 Time/s (b) 0 pprn ethanol I 50 pprn ethanol I 0 500 1000 1500 Timels Fig. 3 Short time reproducibilities. (a) Sensing coating based on a single active material ETH6022; and (b) sensing coating based on plasticized PVC membrane containing ETH6022.the average standard deviations for PCS and PCS-PVC are 9.1 and 5.0 Hz, respectively. Selectivity The exceptional feature of this ethanol sensing system is the fact that only nucleophilic molecules are able to interact with the carbonyl group of the ligand. The frequency shifts of PCS-PVC and PCS were obtained in the presence of different gases, such as diethyl ether, THF, chloroform and acetone (Table 2). As expected, these gases, which do not contain any nucleophilic group, led to little observable interference up to concentrations of 1500 ppm. In contrast, water, which is able to form hydration compounds with carbonyl groups, can act as a real competitor with ethanol. As can be seen from Table 2, water causes large frequency changes in comparison with the same concentration of ethanol.When dry nitrogen was used instead of moisture, nitrogen with 40% RH in the gas flow system in Fig. 1, the detection limit of PCS-PVC became 1.2 ppm. A number of other alcohols, such as methanol, propan-1-01, propan-2-01 and butan-1-01, were also studied with respect to the frequency changes of these two kinds of sensors with the same concentration of ethanol. In contrast to the above-mentioned non-nucleophilic molecules, these alcohols caused some inter- ference owing to the similarity of their structures with that of ethanol, and they possessed different lengths of the alkyl group in the molecules, resulting in different selectivities. Table 1 Reproducibilities of PCS and PCS-PVC based on ETH6022 Ethanol Frequency shift/Hz concen- tration S Sensor (ppm) 1st 2nd 3rd Mean (%) PCS 28.3 1272 1254 1279 1268 12.9 64.7 1551 1540 1550 1547 6.1 105.8 1811 1815 1808 1811 3.5 121.3 1947 1961 1972 1960 12.5 217.0 2709 2733 2705 2716 15.1 247.0 3024 3025 3032 3027 4.3 PCS-PVC 33.6 308 305 298 304 5.1 66.8 370 364 379 37 1 7.5 108.9 429 432 436 432 3.5 125.5 472 463 469 468 4.6 184.2 555 568 563 562 6.5 235.9 660 655 655 657 2.9 Table 2 Effect of several common organic polar solvents on the frequency shifts PCS-PVC PCS Concen- AFI AFI tration AFI AFethanol AF1 AFethanol Solvent (PPm> Hz (%I Hz (%) Ethanol 100 183 100 Diethyl ether 1500 34 18 Chloroform 1500 17 9.3 Acetone 1500 3 1.6 Water 100 187 102 Methanol 100 46 25 Propan- 1-01 100 31 17 Propan-2-01 100 53 29 Butan- 1-01 100 10 5.5 Tetrahydrofuran 1500 20 11 809 100 217 26.8 118 14.6 116 14.3 18 2.2 935 115 199 24.6 156 19.3 265 32.8 51 6.3262 Analyst, February 1996, Vol.121 Useful Lifetime Let Fo be the fundamental frequency of the oscillator when the piezoelectric crystal is exposed in the blank sample without coating, and Fo’ the fundamental frequency of the oscillator when the coated piezoelectric crystal is exposed in the blank sample. The difference between these two frequencies is called the coating frequency (Fo - Fo’). During usage of the sensor, the coating frequency will decrease as the active material (or coating) flows out because of evaporation. Hence the coating frequency of the sensor can be used as an indicator of its useful lifetime.The changes in the coating frequencies of PCS and PCS-PVC during the use under the same experimental conditions in the dry system are given in Table 3. The data show that the active material of PCS flows out rapidly during usage (the coating frequency of this sensor decreases by 71% in 8 days), but the coating frequency of PCS-PVC hardly changes (it decreases by only 2.7% in 66 days). This means that the coating of PCS-PVC is very stable. Continuous experiments showed that PCS-PVC can be used for more than 2 months without any decrease in sensitivity. The results show that the plasticized PVC membrane is very suitable for supporting the active material ETH6022. This useful lifetime is not only much superior to that of the traditional PCS, but also comparable to that of an ion-selective elec- trode. The new plasticized PVC coating method was also used to construct several classical gas-sensitive PCS sensors.The useful lifetimes of these sensors were also extended consid- Table 3 Useful lifetimes of PCS and PCS-PVC based on ETH6022 Coating frequency, Fo-Fo’/Hz Sensor Od 2 d 4 d 8 d 15d 30d 66d PCS 7344 6986 6112 2123 (unstable) PCS-PVC 7103 7098 7090 7078 7055 6992 6908 Table 4 Comparison of the useful lifetimes of several classical PCSs with a single active material coating and plasticized PVC coating Useful lifetime/d* Single active Plasticized Analyte Active material material coating PVC coating Cyclohexane Squalane 15 (5.3%) > 72 (6.4%) Acetone PEG 20M 19 (4.5%) >81 (3.7%) H2S Tween 80 12 (3.1%) >75 (5.1%) so2 Triethanolamine 7 (2.9%) >68 (3.4%) so2 Tridodecylamine 8 (2.9%) > 65 (4.2%) so2 Diacetylcellulose 22 (3.8%) > 84 (4.6%) frequency within the useful lifetime.* The values in parentheses represent the decrease in the coating erably (Table 4). Clearly, the coating supported by plasticized PVC polymer is one of the most effective coating methods for extending the useful lifetime of piezoelectric crystal sensors. The work described in this paper was supported by the National Science Foundation of China and the Electroanalytical Labo- ratory of Changchun Institute of Applied Chemistry (Academia Sinica) . 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 Wisemann, A., TrAC, Trends Anal. Chem. (Pers. Ed.), 1988, 7, 5.Ruz, J., Fernandez, A., Luque de Castro, M. D., and Valcarcel, M., J . Pharm. Biomed. Anal., 1986, 4, 559. Wolfbeis, 0. S . , and Posch, H. E., Fresenius’ Z. Anal. Chem., 1988, 332, 255. Walters, B. S., Nielson, T. J., and Arnold, M. A., Talanta, 1988, 35, 151. Scheper, T., and Buckmann, A. F., Biosens. Bioelectron., 1990, 5, 125. Gautier, S. M., Blum, L. J., and Coulet, P. R., J . Biolumin. Chemilumin., 1990, 5, 57, Posch, H. E., Wolfbeis, 0. S., and Pusterhofer, J., Talanta, 1988,35, 89. Dickert, F. L., Schreiner, S. K., Mages, G. R., and Kimmel, H., Anal. Chem., 1988, 60, 1377. Seiler, K., Wang, K.-M., Karatli, M., and Simon, W., Anal. Chim. Acta, 1991, 244, 151. Matsubara, K., Kawata, S., and Minami, S., Appl. Opt., 1988, 27, 1160. Alder, J. F., and McCallum, J. J., Analyst, 1983, 108, 1169. McCallum, J. J., Analyst, 1989, 114, 1173. Hahn, E. C., Suleiman, A. A., Guilbault, G. G., and Cavanaugh, J. R., Anul. Chim. Acta, 1987, 197, 195. Filho, 0. F., Andrade, J. F., Suleiman, A. A., and Guilbault, G. G., Anal. Chem., 1989, 61, 746. Nieuwenhuizen, M. S., Nederlof, A. J., Vellekoop, M. J., and Venema, A., Sens. Actuators, 1989, 19, 385. Matijiasevic, V., Garwin, E. L., and Hammond, R. H., Rev. Sci. Instrum., 1990, 61, 1747. Nieuwenhuizen, M. S., and Nederlof, A. J., Sens. Actuators, 1990, B2, 97. Von Schickfus, M., and Rapp, M., Acta Phys. Slov., 1990, 40, 26. Guilbault, G. G., Affolter, J., and Tomita, Y., Anal. Chem., 1981,53, 2057. Moody, G. J., Thomas, J. D. R., and Yarmo, M. A., Anal. Chim. Acta, 1983,155, 255. Morf, W. E., Seiler, K., Sorensen, P., and Simon, W., in Ion-Selective Electrodes, ed. Pungor, E., Pergamon Press, Oxford, 1989, vol. 5, p. 141. Behringer, Ch., Lehmann, B., Haug, J. P., Seiler, K., Mod, W. E., and Simon, W., Anul. Chim. Acta, 1990, 233, 41. Wang, K.-M., Seiler, K., Haug, J. P., Lehmann, B., West, S., Hartman, K., and Simon, W., Anal. Chem., 1991, 63,970. Sauerbrey, G., Z. Phys., 1959, 155, 206. IUPAC, Pure Appl. Chem., 1976,45,99. Paper 5/04072H Received June 23,1995 Accepted August 31,1995

ISSN:0003-2654    

DOI:10.1039/AN9962100259

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 (263-267) 263 Determination of Cinnamic Acid in Human Urine by Differential-pulse Polarography Valdir S. Ferreiraa, Cristo B. Melios", Maria VaInice B. Zanonia" and Nelson R. Stradiottob a Departamento de Quimica Analitica, Instituto de Quimicdniversidade Estadual Paulista, Caixa Postal 355, 14800-900, Araraquara-SP, Brazil Latras-Universidade de ScZo Paulo, 14049, RibeirGo Preto-SP, Brazil Departamento de Quimica, Faculdade de Filosofia, Ci2ncias e The differential-pulse polarographic behaviour of cinnamic acid was studied in acetate and phosphate buffer solutions (pH 3.5-7.5). The reduction mechanism is discussed. The drug can be determined at pH 5.0 over the concentration range 5 x 10-5-1 X effect of tetraalkylammonium salts on the electroanalytical determination of cinnamic acid was investigated.The direct determination of the drug (0.7-5.5 pg ml-l) in urine samples diluted with acetate buffer (pH 5.0) can be effected in the presence of 1 X 10-3 rnol 1-1 cetyldimethylethylammonium bromide solution. The detection limit was found to be 0.1 pg ml-1. The relative standard deviation from six determinations at the 5.5 pg ml-1 level was 1 %. Keywords: Cinnamic acid; polarographic determination; differential-pulse polarography; urine moll-'. The Introduction Cinnamic acid (3 -phenylpropenoic acid) occurs naturally in plants and is also a synthetic substance of great significance in the perfume, pharmaceutical, food, photographic and polymer industries.1.2 Cinnamic acid and its derivatives are also employed in the agricultural and medical fields.' As medical products they are used in the treatment of patients with certain forms of deafness, leprosy and tuberculosis;' in herbal medicine preparation^;^ as metabolites of other drugs;4 and they can occur in human urine due to inborn errors of metabolism.5 The versatility of the use of cinnamic acid has highlighted the need to develop rapid and reliable analytical methods for monitoring the drug in diverse matrices.Common analytical techniques for measuring cinnamic acid in different samples include gas and liquid chr~matography,~-~ electrophoresis8 and UV spectroscopy.9 Its electrochemical behaviour has been the subject of some investigations.lOJ1 Brand and Fleetlo showed that cinnamic acid is polarographically reduced in 50% ethanol in a single two-electron wave within the pH range 4-8.The application of this method to the quantitative determination of cinnamic acid by direct current (dc) polarography in herbal preparations is limited to levels of 10-3 mol 1-l." An amperometric biosensor for cinnamic acid has been proposed by Wang and Naser.2 However, the method is indirect, based on a response in the anodic monitoring of coumaric acid generated in the system. No previously reported methods have been found that relate to the determination of cinnamic acid in urine by differential-pulse polarography (DPP). The present work describes the polarographic behaviour of cinnamic acid in order to obtain a differential-pulse polaro- graphic method for determining the drug. Its electrochemical * To whom correspondence should be addressed.behaviour in the presence of tetraalkylammonium salts high- lighted the feasibility of developing a direct method for the quantification of cinnamic acid in human urine. Experimental Apparatus For voltammetric measurements an EG & G Princeton Applied Research (EG & G PAR, Princeton, NJ, USA) Model 264 A polarographic analyser and an EG & G PAR Model RE 003 1 x-y recorder were used. An EG & G PAR 303A stand was used in the dropping mercury electrode (DME) mode for polarography and hanging mercury drop electrode (HMDE) mode for cyclic voltammetry. The three-electrode system was completed by means of a glassy carbon auxiliary electrode and an Ag/AgC1(3 mol 1-1 KCl) reference electrode. Unless stated otherwise the following parameters were used: pulse amplitude, 50 mV; potential sweep rate, 5 mV s-l; drop time, 1 s, temperature, 25°C.The solution pH was measured with a Micronal (Siio Paulo, Brazil) Model B 222 pH-meter. Reagents and Solutions All reagents were of analytical-reagent grade. De-mineralized water was obtained from a Milli-Q system (Millipole, Milford, MA, USA). The cinnamic acid stock standard solution (1 X 10-2 moll-]) was prepared from the dried pure substance (Merck, Darmstadt, Germany) in ethanol. Its working solutions were prepared by dilution with supporting electrolyte. Buffer solu- tions were prepared by mixing: 0.2 mol 1-1 acetic with 0.2 moll-1 sodium acetate solution to give various pH values (3.5-5.3); (b) 0.2 mol 1-1 phosphoric acid and 0.2 mol 1-1 sodium dihydrogenphosphate solution (pH 5.67.5); and (c) 0.2 mol 1-1 Britton-Robinson buffer (pH 3.0-8.0).Tetramethyl- ammonium bromide (TMAB), tetraethylammonium bromide (TEAB), tetrabutylammonium bromide (TBAB) and cetyldimethylethylammonium bromide (CEDMA) were ob- tained from Aldrich (Milwaukee, WI, USA) and used as received. Results and Discussion Cinnamic acid is polarographically reduced in a single two- electron wave between pH 4.0 and 7.5. By using the differential pulse polarographic technique, it is possible to distinguish the wave even below pH 4.0, see Fig. 1 (a). The higher peak currents and better separation from the background electrolyte264 Analyst, February 1996, Vol. 121 1400 - reduction as compared with polarography, indicate DPP to be a suitable technique for analysis of cinnamic acid (see Fig.1). -30 1 ' 1 ' 1 ' 1 ' Influence of pH The effect of the composition of the supporting electrolyte on the voltammetric measurement of cinnamic acid was examined by comparing the response in various electrolytes, such as acetate, phosphate and Britton-Robinson buffers as well as in LiCl solutions. Britton-Robinson buffer caused a drastic decrease in the differential-pulse voltammetric response. LiCl exhibited a broad and ill-defined wave. The more convenient supporting electrolytes were acetate and phosphate buffers. Fig. 2 (curve b) shows the pH influence on the peak current. Approximately three zones can be distinguished: pH 3.0-4.0, where peak current decreases as pH decreases, pH 4.5-6.0, where peak current is independent of pH, and pH > 7.0, where there is again a decrease in peak current with increase in pH.Variations of DPP peak potential as a function of pH are shown in Fig. 2 (curve a). Peak potential can be seen to shift towards more negative values as pH increases. Two linear variation zones can be observed, intersecting at approximately pH 4.5. The slopes of the linear portions are 15.3 mV pH-' (pH 3.0-4.5) and 67.8 mV pH-1 (4.5-7.5). This could indicate 1 I5PA 4 =s 0 1.1 1.3 1.5 1.2 1.4 1.6 -EN versus AglAgCI Fig. 1 Differential-pulse (I) and direct current (11) polarograms for 1 X 10-3 mol 1-1 cinnamic acid and differential-pulse (111) polarograms of blank solution at pH: 3.5 (a); 5.3 (b) and 7.5 (c). 1650 - 1600 - > 5 lssOT 1500- 1450- "7 the presence of chemical reactions with participation of protons.'* Characteristics of the Polarographic Waves Analysis of the shapes of the differential-pulse peaks indicates a degree of irreversibility, with peak half-width values varying from about 86 mV at pH < 4.0 to 100 mV at pH 4.5-7.0.Logarithmic analysis of the dc polarograms recorded with 1 X 10-3 mol 1-l cinnamic acid (drop time of 1.0 s) resulted in straight lines, and the ana values13 obtained point to an irreversible reduction process with an, = 1.44 at pH < 4.0, and an, = 1.12 at higher pH values. The estimation of the number of electrons involved in the reduction process is known to be two according to Brand and Fleet.10 Based on the an, value obtained in the logarithmic analysis of the wave and the slope for the variation in half-wave potential with varying pH, the number of protons involved in the step controlling the rate of the electrode process at pH > 4.5 is found to be 1.26.At pH < 4.5 the proton participation might not be preponderant to the process. A linear dependence between the limiting current with h1/2 was found by dc polarography at pH 4.0, 6.8 and 7.5 using natural drop times. Studies in the concentration range from 2 X to 1 X 10-3 moll-' cinnamic acid at the same pH values are always linear. These results clearly indicate that the electrode process is diffusion-controlled. Hence, with the knowledge that cinnamic acid can be reduced in aqueous solution in a two-electron process,lO involving a pre- protonation of the electroactive site at pH > 4.0, the over-all course of the olefinic bond reduction can be represented as follows: H' C,H,-CH=CH-COOH C,H,-CH=CH-COO- pH>4.5 I H' C 6 H 5 - - O - pHc4.0 z i C6H5-CH,-CH,-COOH &-y H' C6H5-CH2-CH2-COO- Over the pH range 4.5-7.5 the anion predominates in solution (PKa = 4.44).14 However, the pre-protonation reaction of the anion taking place on the ethylene bond conjugated H+[C~HS-CH=CH-COO-] is postulated as the first, rate- determining step in the reduction mechanism.10 In the pH region below 4.0 the undissociated cinnamic acid is probably reduced. The peak potentials are not dependent, which confirms this hypothesis.The break observed at around pH 4.5 probably corresponds to the PKa valuesI4 of the carboxylate function.Based on the differential-pulse polarographic behaviour of cinnamic acid a quantitative method has been developed. At pH 5.0 the calibration curve is described by the following regression line: i, = 0.08687 + 67.55 c (correlation coefficient 0.9993, n = 7, concentration range 5 X 10-5-1 X 10-3 moll-') where ip is the peak current for cinnamic acid in pA and c is the concentration (in mmol 1-1). The determination limit was 1 X 10-5 moll-'. An attempt was made to apply the above mentioned procedure to the direct determination of cinnamic acid in human urine. However, the addition of any amount of urine to a cell containing 1 X moll-' cinnamic acid in acetate buffer (pH 5.0) was enough to suppress the signal completely. Effect of Surfactants and Quaternary Ammonium Salts In order to improve the shape of the differential-pulse peaks and to extend the applicability of the proposed method for theAnalyst, February 1996, Vol.121 265 determination of cinnamic acid in urine samples, the influence of surfactants in the electrode process was investigated. The use of a non-ionic surfactant like Triton X-100 and an anionic surfactant (lauryl sulfate) did not show meaningful changes in the differential-pulse polarograms. In contrast, the addition of a 1 X 10-3 mol 1-1 tetrabutylammonium salt to a 1 X 10-3 mol 1-1 cinnamic acid solution at pH 5.0 (acetate buffer) displays an anodic shifting in the peak potential of 60 mV, but the peak current is not changed. An investigation of the effect of tetrabutylammonium salt concentrations (1 X mol 1-1- 1 X 10-3 mol 1-1) on the polarograms of cinnamic acid (1 X 10-3 moll-1) are shown in Fig.3. The height of the original peak (a) decreases gradually with increasing alkyl ammonium salt and leads to the appearance of a new wave (b) at a less negative potential. When the concentration was increased the height of the peak b increases gradually up to 6 X 10-5 mol 1-l. The peak height remained virtually constant for further butylammonium salt addition. The calibration graph was found to be linear from 2 X 10-6 to 8 X mol 1-l of cinnamic acid at pH 5.0 acetate buffer containing 1 X 10-3 mol 1-1 TBAB (see Fig. 4). The relative standard deviation for solutions of 1 X 10-3 mol 1-1 cinnamic acid containing 1 X 10-3 moll-' of TBAB was 0.88% (n = 7).The influence of chain quaternary ions on the peak potential of cinnamic acid is shown in Table 1. At pH 5.0 as the size of the tetraalkylammonium ion increases, there is fairly rapid shift of the peak potential towards more positive values. CEDMA cation induces the greatest shift in the wave. The usual behaviour of the system is shown in Fig. 5. With the concentration of TBAB and CEDMA fixed at 1 X moll-' and cinnamic acid fixed at 1 X 10-3 mol 1-1 the pH of the solution was varied over the range 3.0-7.5. For both tetraalkyl- ammonium salts the polarographic behaviour as a function of pH is identical. The plot of Ep versus pH show the same behaviour as in Fig. 2. However the peak heights were maximum and constant from pH 4.5-6.0 and decreased outside this range.No significant change was observed in the half-width peaks. Typical cyclic voltammograms of 1 X 10-3 mol 1-l cinnamic acid at pH 5.0 in the presence of 1 X 10-3 mol 1-1 CEDMA are shown in Fig. 6. The absence of anodic waves 4 a I b 6,7 a &5 1 i I 1.2 1.4 1.6 1.8 -EN versus AglAgCI Fig. 3 mol 1-1 cinnamic acid in acetate buffer (pH 5.0) in the presence of TBAB: 1, without TBAB; 2, 1, 3 x 10-5 moll-'; 3,2.5 X 10-5 moll-l; 4,3.7 X 10-5 moll-1; 5,5.0 x 10-5 mol 1-1; 6, 5.6 X 10-5 mol 1-l and 7, 6.2 X 10-5 mol 1-1. a, Reduction of cinnamic acid and b, reduction of cinnamic acid- ammonium salt ion-pair. Differential-pulse polarograms of 1 X -EN Fig. 4 Differential-pulse polarograms of cinnamic acid at concentrations of (in moll-1) 1,8 x 10-4 (a); 3.5 X 10-4 (b); 5.2 X 10-4 (c); 6.8 X (d); 8.3 x 10-4 (e); and 0 (0 in acetate buffer (pH 5.0) and in the presence of 1 x 10-3 moll-' TBAB.Table 1 Influence of tetraalkylammonium salts (1 X differential-pulse polarographic parameters of cinnamic acid 1 X moll-' in acetate buffer pH 5.0 moll-l) on some Tetraalkylammonium salt A EJmV Ip*lpA TMAB 10 69.5 TEAB 20 69.0 TBAB 60 69.5 CEDMA 200 69.1 * Ep = -1.57 V and Zp = 65.0 pA, without tetraalkylammonium salt. 4 a 1 I 114 1 :6 I a I I 1.3 1.5 -EN versus AglAgCI Fig. 5 Differential-pulse polarograms of 1 X 10-3 moll-' cinnamic acid in acetate buffer pH 4.2 (b) and pH 6.0 (a) in the presence of 1 X 10-3 mol 1-1 CEDMA (11), and without CEDMA (I). Blank solution (111).266 Analyst, February 1996, Vol. 121 shows that the compound is reduced irreversibly at the mercury electrode.The dependence of the peak current on square root of the scan rate was found to be linear, which is indicative of a diffusion-controlled process. The same behaviour was observed in the presence of TBAB. It is generally accepted that shifts of the peak potential observed for some a, (3-unsaturated carbonyl compounds in the presence of various tetraalkylammonium ions are attributed either by ion-pair formation and/or by double-layer effects. It is known that tetraalkylammonium ions are usually capillary- active and are therefore adsorbed onto the surfaces of mercury electrodes. 15-17 According to our results and the general mechanism for similar compounds, the effect of tetraalkylammonium ion on the polarographic reduction of cinnamic acid could be explained based on the basis of a model suggested by Meites.15 In the presence of a layer of adsorbed tetraalkylammonium ions on the surface mercury, the reduction of the cinnamic acid (usually 1,2-addition of electrons to the double bond) is replaced by a two-electron step involving 1,4-addition across the ion pair. + I I The unnecessary re-orientation of the molecule with respect to the electrode surface, usually required for 1,2-addition (negatively charged oxygen atom would be repelled from the electrode surface) could be, at least in part responsible for reduction of the carboxylate-ammonium salt ion-pair at less negative potentials.In order to establish the reproducibility of the response, a series of standard solutions of 1 X 10-3 moll-' cinnamic acid containing 1 X 10-3 mol 1-1 of CEDMA was tested.The polarographic response was also found to be reproducible and stable for at least 3 h. Using moll-' of CEDMA a linear dependence between the peak current and cinnamic concen- tration was observed in the range concentration of 4 X 10-6- 2 X 10-5 mol 1-1; i, = 0.209C (10-6 mol 1-1) + 0.019 (correlation coefficient, 0.999), n = 6. Hg - RN: 0-.-..C..-.C.-.C - use of TBAB is not suitable. The wave is obscured by the wave resulting from interfering substances that are electroactive at close potentials, present in the urine. However, differential- pulse polarograms of 1 x 10-3 mol 1-1 cinnamic acid containing 1 X 10-3 moll-' CEDMA (containing 0.1 to 1.8 ml of urine) in 10 ml (total volume) of pH 5.0 acetate buffer were not affected by the presence of urine.Further dilutions involving more than 2 ml of urine to 8 ml of acetate lead to a distortion of the peak by electrolyte discharge interference. Quantification of the urine content of the drug was accom- plished by multiple standard additions. Typical differential- pulse polarograms obtained from direct dilution of 1 ml of urine, spiked with 0.7-1.5 pg of cinnamic acid, by adding it to 9 ml of pH 5.0 acetate buffer solution in the voltammetric cell are shown in Fig. 7. The proposed method is also suitable for urine spiked with 1.8-5.5 pg of cinnamic acid, as shown the calibration graph in Fig. 8. The mean recoveries of cinnamic acid based on the proposed method at concentrations of 0.5 and 1.5 pg ml-1 were 91 f 0.5 and 95 +_ 0.5%, respectively.The relative standard deviation measured at a concentration of 5.5 yg ml-1 was 1% (n = 6). The detection limit based on a signal- to-noise ratio of 3 was approximately 0.1 yg ml-1 in the supporting electrolyte. Conclusion From these studies, it was concluded that differential-pulse polarography is a suitable technique for the determination of I d I e 4 a 1.2 1.4 Analysis of Urine Samples The applicability of DPP as an analytical method for the determination of cinnamic acid in urine was tested at pH 5.0 using CEDMA and TBAB. For direct measurement in urine the -EN versus AglAgCI Fig. 7 Differential-pulse polarograms of different concentrations of cinnamic acid measured directly in urine [ 1 + 9 ml in urine-acetate buffer (pH 5.0) containing 1 X lo3 moll-' CEDMA] (in pg ml-I): 0.74 (a); 0.97 (b) 1.22 (c); 1.47 (d) and 0 (e).1.2 1.4 -EN versus AglAgCI Fig. 6 Cyclic voltammograms of 1 X 10-3 mol 1-1 cinnamic acid in acetate buffer (pH 5.0) containing 1 X mol I-' CEDMA at various scan rates (in mV s-1) 1, 10; 2, 20; 3, 50; 4, 100; and 5, 200. i P I 0 1.0 2.0 3.0 4.0 5.0 C/pg mi-' Fig. 8 diluted with acetate buffer (pH 5.0; 9 ml) containing 1 X CEDMA. Calibration graph for 1 ml of urine spiked with cinnamic acid and rnol 1-IAnalyst, February 1996, Vol. 121 267 cinnamic acid in urine samples. The use of tetraalkylammonium salts, in particular salts having a longer alkyl chain, caused significant and useful changes in the peak potentials of the differential-pulse polarographic wave obtained for cinnamic acid.The strategic addition of CEDMA to the supporting electrolyte can be used as a convenient method for the selective determination of low levels of cinnamic acid in urine samples. The developed methodology provides a simple and precise determination of cinnamic acid in urine without any sample pre- treatment step, resulting in short analysis times. The authors thank the FAPESP (Brazil) for financial support. V.S.F. thanks CAPES (Brazil) for providing a scholarship and the UFMS, Camp0 Grande, Brazil, for granting leave of absence. References 1 2 3 4 Encyclopedia of Chemical Technology, ed. Grayson, M., Wiley, New York, 1981, vol. 6, p. 142. Wang, J., and Naser, N., Bioelectrochem. Bioenerg., 1992, 27,441. Wen, K. C., Huang, C. Y., and Liu, F. S., J . Chromatogr., 1992,593, 191. Yuan, J., Bucher, J. R., Goehl, T. J., Peter, M. P., and Jamelson, C. W., J . Anal. Toxicol., 1992, 16, 359. 5 6 7 8 9 10 11 12 13 14 15 16 17 Greeter, J., and Jacobson, C . E., Clin. Chem. (Winston-Salem, N.C.), 1987, 33, 473. Greenway, W., May, J., and Whatley, F., J . Chromatogr., 1989,472, 393. Janzyk, A., and Wilczynska-Wojtulewicz, I., Chem. Anal. (Warsaw), 1988, 33, 535. Fujiwara, S., and Honda, S., Anal. Chem., 1985, 58, 1811. Bounias, M., Daurade, M. H., and Lizzi, Y., Analusis, 1989, 17, 201. Brad, M. J. D., and Fleet, B., J. Electroanal. Chem., 1968, 16, 341. Orlov, Y. E., Sirenko, L. Ya., and Lichino, 0. P., Chem. Nut. Compd. (Engl. Transl.), 1992, 2715, 633. Zuman, P., in Progress in Polarography, ed. Zuman, P., and Meites, L., Wiley, New York, 1992, vol. 3, 73. Zuman, P., and Perrin, G. L., in Organic Polarography, Wiley, New York, 1969, p. 95. Handbook of Chemistry and Physics, ed. Lide, D. K., CRC Press, Boca Raton, FL, 71st. edn., 1991, pp. 8-35. Meites, L., J . Am. Chem. SOC., 1951, 73, 177. Reinmuth, W. H., Borges, L. B., and Hummeinstedt, L. E. I., J . Am. Chem. SOC., 1959,81,2947. Missau, S . R., Becher, E. I., and Meites, L., J . Am. Chem. Soc., 1961, 83, 58. Paper 51031 36B Received May 16, I995 Accepted October 10, I995

ISSN:0003-2654    

DOI:10.1039/AN9962100263

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 ERRATA 269 Analysis of Organic Polymorphs A Review Terence L. Threlfall Analyst, 1995, 120, 2435 On page 2436, column 2, line 7, ‘ranitidine’ should read ‘ranitidine hydrochloride’. On page 2437, column 2, line 20, ‘ 1,3-~yclohexadiene’ should read ‘ 1,3-~yclohexadienone’. On page 2452, the second equation in the footnote should read lim Ix-y’( + 03 < p(x) p (x’) > = p* On page 2453, Fig 8(a), Polymorph I and Polymorph I1 have been transposed in the labelling of the lines. Determination of Carbosulfan in Oranges by High-performance Liquid Chromatography With Post-column FI uorescence M. W. Brooks and A. Barros Analyst, 1995, 120, 2479 On page 2479, column 2, line 16, ‘A 500 ml volume of pyridine . . .’ should read ‘A 500 pl volume of pyridine . . .’

ISSN:0003-2654    

DOI:10.1039/AN9962100269

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 27 1 Correspondence Fast Determination of Total Fluoride by Direct Potentiometry in Samples of Aluminium Fluoride and Cryolite Maria Soledad Corbillon, Mari Paz Carril, Juan Manuel Madariaga and Ignacio Uriarte Analyst, 1995, 120, 2227 The following correspondence was received regarding the above paper: Dear Sir: The paper, ‘Fast Determination of Total Fluoride by Direct Potentiometry in Samples of Aluminium Fluoride and Cryolite’ by Corbill6n, et al., which appeared in the August issue of this journal (p. 2227) is seriously flawed. It compares the behaviour of ‘Radiometer and Orion fluoride electrodes’ and reports that ‘. . . the addition of aluminium ions to the calibration solution was necessary when the Orion electrode was used in order to avoid high error levels’.The fluoride electrode has been commercially available for almost 30 years, and its mechanism has been thoroughly studied and is well understood. To those familiar with the field, this report is an inexplicable one. However, a careful reading of the paper shows that: 1. Only one electrode from each manufacturer was used. This is not a statistically significant sample. 2. There is no indication in Table 2 of the number of samples that were run to compare the two electrodes. However, the ‘high error’ reported is on the order of 2%, which corresponds to about 0.5 mV, and is about the reproducibility expected in a direct reading potentiometric method. 3. The Radiometer electrode was used with a Radiometer reference electrode; the Orion electrode was used with a Metrohm reference electrode.For the Orion electrode, this is not using the product in accordance with the manufacturer’s instructions. 4. There was never an attempt made to determine if the problem was due to the reference electrode before reporting it as due to the sensing electrode. This could easily have been checked by switching reference electrodes, and is one of the first steps in troubleshooting mentioned in the Orion electrode instruction manual. 5. The statement (p. 2228) ‘The possible existence of a liquid junction potential (Ej) in these systems, owing to the migration of fluoride ions through the solid membrance of the reference electrode, was studied,’ makes no sense. Reference electrodes do not have solid membranes, or they would not function as reference electrodes.Fluoride ions do not cause a liquid junction potential at the solid membrane of the sensing electrode, either. How much influence this reasoning has on the kinds of experiments that were done is not clear. It is difficult, from a distance, to tell exactly where the problem lay. However, an ‘error’ of about 0.5 mV sounds suspiciously like a junction potential error due to the reference electrode. The aluminium ion, as a highly charged species or as a large complex, is a likely culprit. If aluminium in the samples caused this bias with the particular Metrohm reference electrode used in this paper (and its condition and history at the time of the tests), then adding aluminium to the standards would account for the observed results.A number of simple tests could have been used by the authors to determine the source of the bias. Interchanging the references was mentioned. Using the two reference electrodes as a pair would have shown whether the millivolt difference between the two remained constant in samples and standards (as must be the case if they are working properly), or the two sensing electrodes could have been compared in the same way, creating a cell with no liquid junction where the difference between them would not be constant if the bias truly existed. Or, as is often done, a common reference electrode could have been used for the two electrodes. Martin S. Frant Orion Research Inc. October 6.1995 Dear Sir: The publication ‘Fast Determination of Total Fluoride by Direct Potentiometry in Samples of Aluminium Fluoride and Cryolite’ which appeared in the August issue of Analyst had the purpose of describing a new analytical method to determine fluoride in solid samples of A1F3 and cryolite.As far as we know, nothing has been published about the direct determination of total fluorine in such samples by using ion selective electrodes and TISAB IV to demask the fluoride complexed by high levels of aluminium. The work was not intended to be a comparison between the behaviour to two different ion selective electrodes and nothing is written in the text suggesting a comparison. We selected the electrodes from Radiometer and Orion only because they are widely used, as least in South Europe, and we presented a careful investigation on the development of a new analytical method.The great influence of the amount of NaOH added during the sample treatment, and the need for adding the same equivalent amount of NaOH to the standards, in order to have effectively the same ionic strength in both standards and test samples, are the most important aspects of the new analytical method. Nothing similar appears in the literature for fluoride determina- tions by means of ion selective electrodes. Thus, the particular characteristics of the solid A1F3 and cryolite samples, i.e., the high A1 : F ratio, are probably responsible for such effects. These two effects, which promote high systematic errors (see Table l), were observed with both electrodes. However, we272 Analyst, February 1996, Vol. 121 observed another effect when the Orion electrode was used.If aluminium was not present in the standards, systematic errors were always obtained. This is a problem detected for these particular samples and cannot be generalized to other samples and procedures. If we had said that ‘the Orion electrode shows such effect and cannot be used’ or something similar, real damage may have been done to the company. However, we show the effect, we analyse the origin of the systematic error and present the solution to avoid it. In this sense, when write the ‘Proposed Procedure’ (p. 2230) we talk about the two electrodes, Orion and Radiometer, showing only this distinction ‘when the Orion fluoride electrode is used, it is necessary to add aluminium ions to the standard solutions.If a Radiometer fluoride electrode is employed, the addition of aluminium is unnecessary’. More- over, in the ‘Discussion’ (p. 223 l ) we talk about the advantages and prevention of the new analytical method saying again ‘. . . the use of an Orion fluoride electrode requires the addition of aluminium ions to the standard solutions in order to obtain calibration graphs without systematic errors . . .’. Thus we report the same analytical procedure for the two electrodes with only one difference when the Orion fluoride electrode is used. So that, we are not making a comparison and point 1 of Dr. Frant’s letter does not apply. The number of samples that were run in Table 2 was 16, and systematic errors were obtained when Orion electrodes were employed without an addition of aluminium ions in the standards for calibration. As we stated in the manuscript (see ref.17), other authors had also advised the addition of interferences to the standard solution in order to reduce the error in the determination of fluoride. With regard to point 2, when we talk about ‘high systematic errors’ we are not looking at the value 2% in an absolute sense but in comparison with the observed random errors we had detected. The maximum error we have detected was f0.4% but the data shown in Table 3 when we compare our results with those obtained by the classical distillation method, have only a random error of +0.16%. Thus systematic errors like those reported in Table 2 of -2.5% are high errors while systematic errors of -0.4% or +0.3%, also reported in the table, cannot be considered as such because they are of the same magnitude like the overall random error.Besides that, a systematic error of -2.5% in the data of Table 3 will represent a decrease of 1.5% units in the fluoride content (% m/m) of the AlF3 samples. In point 3, Dr. Frant says that the reference electrode used for working with the Orion electrode is not in accordance with the manufacturer’s instructions. We consider this suggestion as a commercial recommendation because the reference electrode is something universal if it functions appropriately. According to the existence of a liquid junction, reference electrodes do not have solid membranes similar to those of ion selective electrode membranes, but they have a porous pin which is also solid (that is why we have employed the expression solid membrane in this case). Thus, the migration of fluoride ions, through this porous pin could be possible. However, as we stated in the manuscript, this liquid junction potential is about 1 mV at 0.1 mol 1-l of fluoride but is negligible, below 0.05 moll-’ fluoride and we always worked down to such a value, so that the appearance of systematic errors cannot be attributed to the liquid junction potential because this variable is under control. Finally, we want to stress that we are sorry about the negative situation that this publication has generated. We would like to state that the work was never intended to be unfavourable to a commercial firm. M . S . Corbilldn University of the Basque Country October 18,1995

ISSN:0003-2654    

DOI:10.1039/AN9962100271

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 273 CUMULATIVE AUTHOR INDEX Angeletti, R., 229 AntonijeviC, M. M., 255 Arias, Juan JosC, 169 Baxter, Douglas C., 19 Biancotto, G., 229 Bilitewski, Ursula, 119 Bjorklund, Erland, 19 Bogan, Declan R., 243 Boyd, Damien, 1R Brinkman, Udo A.Th, 61 Burgot, Jean-Louis, 43 Bye, Ragnar, 201 Cao, Zhong, 259 Carmona, Pedro, 105 Casella, Innocenzo G., 249 Cepas, Juana, 49 Ceramelli, Giuseppe, 219 Cerd6, A., 13 Cerdfi, V., 13 Chen, Guo Nan, 37 Chou, Shu-Fen, 71 Copeland, D. D., 173 Corti, Piero, 219 Cosano, J., 83 Craston, Derek H., 177 Cullen, Michael, 75 Cullen, William R., 223 de Jong, Dirk, 61 de Jong, Gerhardus J., 61 Desimoni, Elio, 249 Destradis, Angelo, 249 Dodd, Matthew, 223 Dreassi, Elena, 219 Economou, Anastasios, 97 Eigendorf, Guenter K., 223 Eikenberg, Oliver, 119 El-shahat, Mohamed F., 89 JANUARY-FEBRUARY 1996 El-Shorbagi, Abdel-Nasser, 183 Emara, Samy, 183 Emteborg, H&an, 19 Escobar, Rosario, 105 Facer, M., 173 Fallon, Michael G., 127 Fell, Gordon S., 189 Ferreira, Valdir S., 263 Fielden, Peter R., 97 Fleet, Ian A., 55 Forteza, R., 13 Francis, John M., 177 Frech, Wolfgang, 19 Fukasawa, Tsutomu, 89 Glennon, Jeremy D., 127 Goosens, Elise C., 61 Gordon, Derek B., 55 Grol, Michael, 1 19 Hansen, Elo H., 31 Harrison, Iain, 189 Hayashibe, Yutaka, 7 Hernhndez, Oscar, 169 Hulanicki, Adam, 133 Ibrahim, Naaim M.A., 239 Iwatsuki, Masaaki, 89 Jackson, Laurence S., 67 Jimtnez, Ana Isabel, 169 JimCnez, Francisco, 169 Karlsson, Lars, 19 Kindness, Andrew, 205 Kratochvil, Byron, 163 Lan, Zhang-Hua, 21 1 Lancashire, Susan, 75 Legouin, BCatrice, 43 Lewenstam, Andrzej, 133 Li, Hao, 223 Lin, Hui-Gai, 259 Littlejohn, David, 189 Lonardi, S., 219 Lord, Gwyn A., 55 Lu, Zheng, 163 Lund, Walter, 201 Luque de Castro, M.D., 83 Maj-Zurawska, Magdalena, 133 Marr, Iain L., 205 Masujima, Tsutomu, 183 Mathiasson, Lennart, 19 Melios, Cristo B., 263 Mieczkowski, Jbzef, 133 Mihajlovic, R., 255 Mohamed, Ashraf A., 89 Molina, Marina, 105 Monaghan, John J., 55 Moollan, Roland W., 233 Moore, Andrew, 67 Mosello, R., 83 Mottola, Horacio A., 21 1 Mulcahy, David, 127 Murphy, William S., 127 Newton, R., 173 Nielsen, Steffen, 3 1 Odman, Fredrik, 19 O’Keeffe, Michael, 1R O’Kennedy, Richard, 243 Oms, M. T., 13 Ostaszewska, Joanna, 133 Packham, Andrew J., 97 Paradowski, Dariusz, 133 Parsons, Patrick J., 195 Pkrez-Bendito, Dolores, 49 Pergantis, Spiros A., 223 Perruccio, Piero Luigi, 219 Piperaki, Efrosini A., 11 1 Piro, R.D. M., 229 Pitre, K. S., 79 Qu, Yi Bin, 139 Quevauviller, Ph., 83 Rae, Bruce, 233 Razee, Saeid, 183 Reimer, Kenneth J., 223 Rios, Angel, 1 Sayama, Yasumasa, 7 Schafer, E. A., 243 Shukla, Jyotsna, 79 Silva, Manuel, 49 Slavin, Walter, 195 Sloth, Jens J., 3 1 Smith, Robert F., 67 Smyth, Malcolm R., 1R Sokalski, Tomasz, 133 Stegman, Karel H., 61 Stradiotto, Nelson R., 263 Tang, Shida, 195 Tegtmeier, M., 243 Thomaidis, Nikolaos S., 1 1 1 Thornes, R. D. ,243 Valckcel, Miguel, 1, 83 Verbeek, Alistair, 233 VukanoviC, B., 255 Walker, P. J., 173 Wang, Bin-Feng, 259 Wang, Ke-Min, 259 Wang, Shi-Hua, 259 Wheals, Brian B., 239 WickstrBm, Torild, 201 Xu, Xue Qin, 37 Yu, Ru-Qin, 259 Zanoni, Maria Valnice B., 263 Zhang, Fan, 37 Zhi, Zheng-liang, 1 Ziegler, Torsten, 119A brand new initiative by The RSC means that for the first time we are now able to offer you six of our highly respected journals at vastly discounted personal rates if you belong to an organisation that already subscribes.Just take a look at the substantial savings we are now offering: 5 E 7 2 2 % 0, f U r wl * Offer available only to individuals working for organisations which already have a full non-member subscription at the same site. On these terms, can you afford not to subscribe today? To order please contact: The Royal Society of Chemistry, Turpin Distribution Services Ltd, Blackhorse Road, Letchworth, Herts SG6 lHN, UK Tel+44(0) 1462-672555 / Fax +44(0) 1462-480947 RSC members ordering for their own personal use are entitled to a discount on most RSC publications, and should contact: Membership Administration Department, The Royal Society of Chemistry, Thomas Graham House, Science Park, Milton Road, Cambridge CB4 4WF, UK Tel+44(0) 1223-420066 / Fax +44(0) 1223-423247 E-mail (Internet): rsc 1 @rsc.org wwweb: http://chemistry.rsc.org/rsc For further information on RSC products please contact the Sales and Promotion Department at our Cambridge address I Send for further information, or for a free sample issue of your preferred journal today!

ISSN:0003-2654    

DOI:10.1039/AN9962100273

出版商:RSC    

年代:1996

数据来源: RSC