摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1123 Comparison of Liquid Chromatographic Methods for Analysis of Homologous Alkyl Esters of Biphenyl-4,4’-dicarboxylic Acid” Joel K. Swadesh Alpha-Beta Technolog y, Innovation Drive, Worcester, MA 01060, USA Charles W. Stewart Jr.+ and Peter C. Uden$ Department of Chemistry, University of Massachusetts, Amherst, MA 0 1003, USA Homologous series of compounds are generated in systematic studies of liquid crystals, petroleum fractions and pharmaceutical compounds. The paper describes the separation of a series of homologous liquid crystalline alkyl esters of biphenyl-4,4,’-dicarbQxylic acid by high-performance liquid chromatography. Hig h-performance gel-permeation Chromatography (HPGPC) was capable of separating compounds differing by a single carbon, and is, therefore, similar in resolving power to either normal- or reversed-phase liquid chromatography.High-performance gel-permeation chromatography has the advantage of permitting the analysis of compounds of widely differing polarity. If the resolution of GPC can be increased slightly by reducing the particle size and decreasing the pore size to maximize resolution in the range of 250-500 u, it could become a primary analytical technique in the organic synthetic laboratory. Keywords : Comparative high -performance liquid chromatograph y; high -perf0 rma nce gel-pe rmea tion chromatography; biphen yldicarboxylic acid ester; mono- and diester homologous series Separation of members of a series of homologous compounds is often required in the course of studying liquid crystals,’ pharmaceutical structure-activity relationships,2 alkanes3 and chromatographic phase~,~75 and in the conductance of other activities of industrial and theoretical importance.Resolution of a compound of interest from synthetic precursors and contaminants can require the development of separate methods for different members of the series. Demands on the analytical procedure are particularly high in the characteriz- ation of a homologous series of alkyl-substituted liquid crystals. Firstly, the physical properties of homologues are extremely similar, differing by the effects of a single meth- ylene unit. Secondly, rigorous measurement of purity is desired, as the thermal properties are highly dependent on purity. Finally, predicted contaminants of a given homologue often include related homologues or structural isomers of the desired compound.The separation of homologues differing by a single meth- ylene is not an unknown problem in chromatographic science. Homologous parabens in cosmetics have been separated by reversed-phase chromatography.6 Homologous phthalate plasticizers have been separated by size-exclusion chroma- tography ,397-9 as have alkane~.3~8JO Reversed-phase liquid chromatography (RPLC) has been used to separate alkanes, alkyl benzenes and fatty acid methyl esters.4 Homologous alcohols have been separated by size-exclusion chroma- tography in micelles.11 The syntheses and thermotropic transitions of 14 liquid crystalline alkyl monoesters of 4,4‘-biphenyldicarboxylic acid (R,BDCA-H) were first described in 1981 .I The correspond- ing diesters (R,BDCA-R,), which are synthetic precursors, were found to exhibit no liquid crystalline mesophases.The monoesters, however, display a rich range of mesomorphism, including three distinct smectic phases and at least one nematic phase. At the time at which this class of liquid crystals was first synthesized, qualitative thin-layer chromatography * Presented at the 21st Northeast Regional Meeting of the American Chemical Society, June, 1991 as HPLC Analysis of the n-Alkyl Esters of Biphenyl-4,4’- Dicarboxylic Acid. t Present Address: Department of Chemistry, Hartwick College, Oneonta, NY 13820-4020, USA. * To whom correspondence should be addressed. was used for characterization. Since then, high-performance liquid chromatography (HPLC) has become more generally available.In the present work, a number of the members of the homologous series of R,BDCA-R, and R,BDCA-H were separated by HPLC. R,BDCA-R,: R’ = R” = n-alkyl R,BDCA-H: R’ = n-alkyl, R” = H Experimental Materials The synthesis of R-BDCA-R and R-BDCA has been described. Briefly, BDCA was esterified to the diester, then hemihydrolysed to the corresponding monoester. The purity of the monoester was monitored by thin-layer chroma- tography on silica, using a 75 + 25 mixture of chloroform- methanol. Some preparations were sublimed to remove BDCA and other non-volatile impurities. Chromatographic solvents were of HPLC grade. Unstabi- lized tetrahydrofuran (THF) (J. T. Baker, Phillipsburg, NJ, USA) was used for gel-permeation chromatography (GPC).Methanol, acetic acid, hexane and chloroform were from Fisher Scientific (Fairlawn, NJ, USA). For GPC, four 30 X 7.5 mm poly(styrene4ivinylbenzene) columns packed with 5 pm diameter, 50 A porosity particles (Polymer Laboratories Ltd., Church-Stretton, Shropshire, UK) were connected in series. For normal-phase liquid chromatography (NPLC), a 150 x 4.6 mm silica column (MacMod Analytical, Chadds Ford, PA, USA) in 90 + 10 hexane-chloroform was used for the separation of the diesters. It was not possitle to elute the monoesters from this column. The monoesters were instead chromatographed by reversed-phase in ion suppression mode on a 150 X 4.6 mm octadecylsilylated silica column (MacMod Analytical), using 92 + 8 methanol-water, with the apparent pH of the mixture adjusted to 2.8 with acetic acid.1124 C I I ANALYST, SEPTEMBER 1993, VOL.118 Equipment The liquid chromatographic system used for GPC was a Hewlett-Packard (Avondale, PA, USA) Model 1090 chro- matograph equipped with a diode-array detector. Elution was performed at ambient temperature and 1 ml min-1. Ultra- violet detection at 260 nm was used. The injection volume was 5 pl, with the analyte at a concentration of about 1 mg ml-1 in eluent. A small portion of toluene was added as a flow rate marker. The monoesters and diesters of BDCA were chro- matographed on this system. The instrumentation used for NPLC and RPLC consisted of a Series 10 pump and Model LC 75 UV detector (Perkin- Elmer, Nonvalk, CT, USA), equipped with a Model 7010 injector (Rheodyne, Cotati, CA, USA).The injection volume was 10 p1, with the analyte at a concentration of about 40 pg ml-1. Detection was at 254 nm. Results The principal results are summarized in two of the figures, demonstrating the separation of selected monoesters R,- BDCA by GPC [Fig. l(a)] and RPLC [Fig. l(b)] and selected diesters R,-BDCA-R, by GPC [Fig. 2(a)] and NPLC [Fig. 2(b)]. Fig. l(a) shows the GPC separation of the C3, C4, Cg, Cll and C15 monoesters of R,-BDCA from 24 and 28 min, with the toluene flow rate marker eluting at 35 min, whereas Fig. l(b) shows the C4, c6, C8, Cll and Cl2 monoesters separated by RPLC from 2 to 13 min. Fig. 2(a) shows the C1, C,, Cf, C4, Cg, Cll and c16 diesters separated by GPC from 23-30 min, with toluene eluting at 35 min.Fig. 2(6) shows the C1, C2, C3, C1o, C12, C14 and c16 diesters separated by NPLC from 5 to 25 min; biphenyl elutes at about 2 min. The void (exclusion) volume in GPC is difficult to determine precisely 0.3 9) C 0 .f! 0.2 z 2 0.1 0 22 t 8 .f! 2 n a crr 24 26 28 30 32 34 Ti me/m i n C11 36 38 due to hydrodynamic chromatographic effects; apparatus to determine it was not accessible, however, a reasonable estimate is about 20 ml, based on the manufacturer’s literature. The retention data is summarized in Table 1. Gel-permeation chromatography was found to be capable of detecting impurities at levels of less than 0.1% peak area at 260 nm, as long as these were well-resolved from the principal peak. Of the diesters examined, the following purities were noted: methyl, 98.8; ethyl, 95.0; propyl, 99.6; butyl, 100; nonyl, 100; decyl, 100; undecyl, 99.9; dodecyl, 100; and hexadecyl, 100.Of the monoesters examined by GPC, the following purities were observed: propyl, 97.4; butyl, 99.6; pentyl, 100; hexyl, 99.8; heptyl, 99.7; octyl, 99.8; octyl, 99.8; tridecyl, 99.7; and pentadecyl, 98.0. On RPLC and NPLC, detection limits were of the order of 2%, and no impurities were detected. Detectability is, of course, dependent on resolution. The resolution between adjacent homologues was estimated as R, = (t2 - f1)/4~~, where t2 and tl are the retention times of adjacent homologues, and st is the standard deviation, estimated as the peak width at half-height.12 The GPC elution times of the membranes of each homologous series were flow corrected and fitted to a third-order polynomial to estimate 0.8 0.7 18 20 22 24 26 28 30 32 34 36 1 , Time/min ., u a [ a I 1 I I I 3 0 5 10 15 20 Ti rne/rni n Fig. 2 Separation of diesters of bi hcnyl-4,4’-dicarboxylic acids by gel-permeation chromatography 8) and normal-phase chroma- tography (6). GPC column and conditions as in Fig. l(a). NPLC 150 X 4.6 mm silica column, mobile phase 90 + 10 v/v hexane-chloroform 1 ml min-1 Table 1 Retention data for mono and diesters of biphenyl-4,4’- dicarboxylic acids, chromatographic conditions as in the text Monoester retention Diester retention and peak widthlmin and peak widthlmin Compound alkyl chain GPC RPLC GPC NPLC Methyl Ethyl Propyl Butyl Pentyl Hexyl Hept y 1 Nonyl Decyl Undecy 1 Dodecy 1 Tridecy 1 Tetradecyl Pentadecyl Hexadecy 1 Octyl 29.82,0.29 24.44,1.13 28.40,0.26 20.31,l.W 27.22,0.31 27.51,0.26 16.13,0.88 26.62,0.28 2.13,0.38 26.88,0.25 26.35,0.27 26.11,0.25 2.18,0.50 26.10,0.27 25.88,0.27 4.06,0.88 24.75,0.25 24.58,0.24 7.44,0.38 25.30,0.27 7.63,l.W 24.20,0.25 12.63,1.38 24.09,0.25 6.63,0.38 24.73,0.27 6.13,0.25 24.66,0.31 23.58,0.22 23.18,0.26 5.63,0.25ANALYST, SEPTEMBER 1993, VOL.118 1125 the resolution as a function of carbon number. Taking the peak width at half-height to be about 0.25 min, the resolution of a monoester of carbon number n from homologue n + 1 is R, (mono, GPC) = 0.4272 - 4.961 x 10-2n + 2.3629 x lO-3n2 whereas the diester resolution is R, (di, GPC) = 1.278-0.1940n + 8.61 x 10-3n* The corresponding values were estimated for NPLC separa- tions and diester separations by calculating interpolated values for retention times and peak least squares.The estimated values of resolution were R, (mono, RPLC) = -0.3248 + 0.4096n - 8.896 X 10-2n2 and R, (di, NPLC) = 1.25 - 0.1521n + 5.569 x lO-3n2 The shape of the resolution curves is shown in Fig. 3. The carbon number is plotted on the x axis, and the resolution of adjacent homologues is plotted on the y axis. Values for resolution were estimated as R, = (t2 - t1)/4s,, where t2 and tl are the interpolated retention times of adjacent homologues, and st is the standard deviation, estimated as the interpolated peak width at half-height. + 5.710 x 10-3n3 Discussion and Conclusions Gel-permeation chromatography is comparable in analysis time to both RPLC or NPLC reversed or normal phase chromatography, but is inferior in resolution. All homologues elute in less than 30 min from GPC and in less than 25 min from both RPLC and NPLC.The RPLC and NPLC separa- tions could, of course, be accelerated without significant loss of resolution by the use of a gradient. The RPLC and NPLC chromatograms exhibit tailing, complicating the use of these techniques for analysis of purity. In contrast, peak symmetry in GPC is close to unity, with no tailing observed. Therefore, although resolution between adjacent homologues is greater in RPLC or NPLC than in GPC, the detectability of contaminants may be better in GPC. Also, the resolution in GPC is not greatly inferior to that observed in RPLC and NPLC. By reducing the pore size and particle size, it might be 1.75 1.50 1.25 g 1.00 2 .- Y - 0.75 0.50 0.25 1 1 3 5 7 9 1 1 1 3 1 5 Number of carbons in pendant chain Fig.3 Resolution of adjacent homologues by gel-permeation chro- matography, reversed-phase chromatogra hy and normal-phase chromatography. 0, Diesters on NPLC; 8, diesters on GPC; A, monoesters on NPLC; and A, monoesters on RPLC possible to obtain high-performance GPC, with resolution and separation time equivalent to that of isocratic NPLC or RPLC. No difference in peak widths was observed between larger and smaller homologues. This suggests that frictional drag in the mass transfer of solutes between mobile and static phases would not be a problem with reduced pore sizes. In GPC, gels of different porosities are often mixed, leading to a very wide distribution of pore size, without a concomitant loss of resolution being observed beyond that ascribed to diffusional broadening.Reproducibility of isocratic liquid chromato- graphic exclusion and partition methods might be considered to be generally similar. Comparisons or reproducibility among methods was not a part of the present study, particularly as data were obtained with different instrumental systems. It is also clear that although gradient elution reproducibilities approach those of isocratic modes if equipment quality and operation are of a sufficiently high standard, there always exists the possibility of greater irreproducibility in gradient elution by virtue of the additional operational parameters inherent in this method.Gel-permeation chromatography is extremely attractive as an analytical method for the separation of small molecules, as the only strict chromatographic requirement is that the analytes be soluble in a common solvent. Even analytes of widely different polarity can be chromatographed on a single system. It should be noted that secondary equilibria of the monoesters appears to be unimportant as little difference is observed in peak widths andor symmetries between the monoesters and diesters. Analytes of greatly different polarity are often encountered in synthetic organic chemistry. This was the case in the present work. The principal contaminants in the R,-BDCA monoesters are predicted to be the R,- BDCA-R, diester precursors and homologues of similar carbon chain length.Each diester and its corresponding product monoester is well-separated on GPC except for those homologues near C5. The possibility of co-elution of dissimilar molecules (such as monoesters with long R chains and diesters with short R chains) cannot, however, be ignored. As some measure of the practical utility of GPC, it should be noted that the GPC data described was obtained in about 36 h of laboratory time wheres the RPLC and NPLC methods took considerably more time to develop. The generosity of Professor J. C. Poirier in making available the samples used in this work is greatly appreciated. Support from Merck, Sharp and Dohme Research Laboratories is gratefully acknowledged. 1 2 3 4 5 6 7 8 9 10 11 12 References Cohen, S. D., Risinger-Diuguid, C. A., Poirier, J. C., and Swadesh, J. K., Mof. Cryst. Liq. Cryst., 1981, 78, 135. Kaliszan, R., J. Chromatogr. Sci., 1984, 22, 362. Krishen, A., J. Chromatogr. Sci., 1984, 16, 254. Colin, H., and Guiochon, G., J. Chromatogr. Sci., 1980,18,54. Colin, H., Guiochon, G., and Diez-Masa, J. C., Anal. Chem., 1981, 53, 146. Dong, M. W., and DiCesare, J. L., J. Chromatogr. Sci., 1982, 20,49. Benson, J. R., and Woo, D. J., J. Chromatogr. Sci., 1984, 22, 386. Russell, D. J., J. Liq. Chromatogr., 1988, 11, 383. Dong, M. W., and DiCesare, J. L., J. Chromatogr. Sci., 1982, 20, 517. Kirkland, J. J., and Antle, P. E., J. Chromatogr. Sci., 1977,15, 137. Terabe, S., Tanaka, H., Otsuka, K., and Ando, T., J. Chromatogr. Sci., 1989, 27, 653. An Introduction to Separation Science, Karger, B. L., Snyder, L. R., and HorvAth, Cs, John Wiley, New York, 1973, pp. 148-150. Paper 2105028E Received September 21, 1992 Accepted May 11, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801123

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1127 Assessment of Three Azophenol Calix[4]arenes as Chromogenic Ligands for Optical Detection of Alkali Metal Ions Mary McCarrick, Stephen J. Harris and Dermot Diamond* School of Chemical Sciences, Dublin City University, Dublin 9, Ireland Three novel chromogenic cone conformational calix[4]arene tetraesters bearing nitrophenylazophenol residues (mono-, di- and tetra-substituted) in the ester portion of the molecules have been synthesized and shown to display a dramatic change in absorbance spectrum upon complexation with lithium, and to a lesser extent sodium, in the presence of a base. An instantaneous colour change from yellow to red corresponding to a A,,, shift from 380 to 520 nm, is noted upon the addition of lithium perchlorate to a solution containing any of the three ligands with the intensity of the colour being concentration dependent.The best selectivity for lithium against sodium and potassium was exhibited by the mono-substituted derivative, ligand 4, (approximately 70-fold selectivity in both cases). Although this is somewhat less than that required for clinical applications, it is anticipated that it can be further improved using a variety of approaches, which are the subject of further investigations. The interference from potassium is too small to be measured. Keywords: Chromoionophore; lithium; calixarene; optrode; alkali metal ion The development of methods of detection for clinically important species such as lithium, sodium and potassium has received much recent attention.One of the main subjects of investigation is the development of chromoionophores, i. e., ionophoric macromolecules that incorporate a chromophoric group into their structure. The aim is to produce a ligand that exhibits a dramatic, concentration-dependent change in absorbance spectrum upon complexation with the target species. The nitrophenylazophenol group has been used as the chromogenic moiety in many different cation binding systems. A crowned azophenol was shown to exhibit lithium specific colour changes in the presence of an amine.1 A slightly larger crown ether, also containing a nitrophenylazophenol moiety as the chromogenic substituent, was found to change colour from yellow to purple upon complexation with potassium. However, the selectivity against sodium was too low to allow the determination of potassium levels in blood.2 A number of amine selective azophenol crown ethers have also been described, with the nature of the amine (i.e., primary, secondary or tertiary) determining the wavelength of the absorbance maximum of the complexed molecule with some of the compounds .3 A four-membered spherand containing the p-nitrophenyl- azophenol chromogenic group has been shown to display lithium selectivity under hydrophobic conditions with no interaction with any other metals being observed.4 The six-membered analogue of this spherand was found to respond to sodium and lithium ions with very little response to any other metal ions.5 Calixarene derivatives, another major class of macrocyclic receptor molecules, have recently evoked interest as a source of novel chromogenic ligands.Two calix[4]arenes bearing azophenol moieties as the chromogenic component have been recently described in the literature,6.7 and one has been found to display lithium selectivity in a solid-liquid two-phase extraction.6 A dinitrophenylazophenol with a similar calix- arene backbone was synthesized and found to respond to the lithium ion in the presence of a number of amines.7 Another calixarene, bearing the nitrophenylazophenol chromophore (which showed excellent selectivity for potassium against sodium), has also been described.8 In all three of these compounds, the phenol group (which is deprotonated in order to give the colour change upon complexation) is housed within * To whom correspondence should be addressed.the cavity into which the metal ion fits and complexation actually involves the phenolic oxygen atoms. Recently, work was presented on cone conformational chromogenic calix[4]arenes with hydroxynitrobenzyl groups as the chromogenic moieties.9 These compounds exhibited lithium selectivity against sodium and potassium and changed from colourless to yellow upon complexation with the metal ions in the presence of a suitable base. In an attempt to create compounds whose complexed form would have a longer wavelength absorbance maximum than the compounds con- taining the nitrophenol group, several tetrameric calixarenes have been synthesized, incorporating a nitrophenylazophenol group as the chromogenic moiety in the ester portion of the molecule.Compounds with longer absorbance maxima such as these are more useful from an optical sensor point of view owing to the cheap availability of blue light emitting diodes, which emit light in the range of the absorbance maxima of the azophenol group. Hence instrumentation development for these sensors is much simpler than for shorter wavelength analogues in which more complex excitation sources are required. Experimental Materials and Equipment Tetrahydrofuran (THF) , and lithium, sodium and potassium perchlorate, along with deuteriated chloroform and methanol were all obtained from Aldrich. Tridodecylamine (TDDA) was purchased from BDH Chemicals and triethylamine (TEA) and butan-1-01 were obtained from Riedel-De-Haen. The nuclear magnetic resonance (NMR) and ultraviolet- visible (UVNIS) spectra were obtained with a Bruker AC-400 spectrometer and a Hewlett-Packard 8452A diode-array spectrophotometer, respectively.Synthesis of Ligands Ligand 1 was prepared from the known p-tert-butyl calix- [4]arene tetraacetic acid.10 Treatment with thionyl chloride furnished the tetraacid chloride, which upon treatment with 2-hydroxy-5-(4'-nitrophenylazo)benzyl alcohol (2) in THF containing. pyridine furnished ligand 1 (80% yield; m.p. 111-115 "C) (Found: C, 65.6%; H, 5.6%. Calc. for C104H96024N12: C, 65.8; H, 5.1%). In a similar fashion triester monoacid 311 was transformed via its acid chloride into ligand 4 (70% yield; m.p. 60-64 "C) (Found: C, 66.33; H, 6.51; N,1128 ANALYST, SEPTEMBER 1993, VOL. 118 3.00%. Calc. for C71H8401sN3: C, 66.29; H, 6.65; N, 3.22%).The final ligand, ligand 5 was prepared from the diacetic acid derivative. Treatment with thionyl chloride gave the diacid chloride, which upon similar treatment to the above gave the required ligand 5 (72% yield; m.p. 65-68 "C) (Found: C, 4.64; N, 7.44%). The diacetic acid derivative 6 was itself prepared from its diethyl acetate derivative by hydrolysis with KOH in ethanol and subsequent acidification (94% yield) .I2 The diethyl acetate derivative in turn was prepared from 1,3-diallyloxycalix[4]arene (7) by treatment with ethyl bro- moacetate and K2C03 in acetone13 (75% yield). The diallyl ether was prepared from the parent calix[4]arene (8),14 by treatment with two equivalents of ally1 bromide in the presence of K2C03 in acetonitrilels (72% yield).At each stage, the product was purified by column chromatography using a neutral alumina column and dichloromethane as the eluent. The presence of the azo and carbonyl groups in the final product was confirmed from infrared absorbance spectra (not shown). 67.74; H, 4.54; N, 7.10%. Calc. fbr CaH52014N6: C, 68.08; H, UVNIS Absorbance Spectra of Ligand Complexation Solutions (5 x 10-5 mol 1-I), of ligands 1, and 4 and a 6 x 10-5 mol 1-1 solution of ligand 5 , were made up in THF. A 2.5 ml aliquot of each ligand solution was taken and to this 100 yl of TDDA were added. Incremental concentrations of aqueous lithium perchlorate were added to give final metal perchlorate concentrations of 1 x 10-6 moll-1, to 0.1 moll-'. After gentle shaking, the clear yellow ligand solution changed colour to red immediately, with the intensity of the resultant colour being dependent on the metal perchlorate concentration.This colour change was examined using UV/VIS spectroscopy in the range 300-800 nm. Selectivity Coefficient Determination In order to determine the selectivity coefficients for lithium against sodium, a series of experiments was set up as described above, with the final lithium perchlorate concentration being varied in the range from 1 x 10-6 to 0.1 mol 1-1, in a fixed background concentration of interferent (0.05 moll-' and 0.1 mol 1-1 sodium perchlorate). Spectra were obtained from 3004300 nm and graphs of absorbance at 520 nm versus the log of the concentration of lithium perchlorate concentration drawn.At high concentrations, the sodium ion has the effect of reducing the response of the ligand to low concentrations of lithium as it dominates the complexation process with the ligand and swamps any lithium ion effects. However, at higher lithium ion concentrations, a response will be observed because of greater affinity of the ligand for lithium ions. From. these graphs, selectivity coefficients can be estimated from the ratio of the sodium and lithium ion concentrations at the intersection of the sodium and lithium dominant response regions of the curves. Two-phase Examination of Ligands In order to examine whether these compounds, in their complexed form could be retained in the organic phase of an organic-aqueous phase two-phase system, a number of solvents were examined (dichloromethane, butan-1-01, butan- one, l,l,l-trichloroethane and butan-2-01) by making up 5 x 10-5 moll-1 solutions of ligand 1 in each solvent. Then 20 pl of TEA was added, followed by 10 yl of 1 mol 1-1 aqueous lithium perchlorate and 2.5 ml of water.Any colour changes upon each addition were noted, as was the ability of any colour generated to remain in the organic layer. For comparative purposes, blank experiments were carried out for each solvent containing everything except the lithium perchlorate. Results and Discussion A slight colour change from yellow to bronze was noted upon the addition of the TDDA to the solution. This effect coincided with an increase in absorption at the uncomplexed wavelength absorbance maximum at 380 nm. No increase in intensity at 520 nm was noted.Fig. l(a) (6) and (c) illustrates the effect of varying lithium perchlorate concentration on the absorbance spectrum of ligands 1, 4 and 5, respectively. As anticipated, complexation led to a shift in A,,, from 380 to 520 nm with an isosbestic point at 425 nm, and the increase in the absorbance at 520 nm being dependent on the concentration of lithium perchlorate. No colour change was observed in the absence of base. This is indicative of the deprotonation of the azophenol group upon metal ion complexation being the cause of the colour change, as the presence of the base facilitates the removal of the phenolic proton. Confirmation of metal complexation was carried out using proton NMR spectroscopy for ligand 1. Upon addition of 1 molar equivalent of sodium thiocyanate in deuteriated methanol to a solution of ligand 1 in deuteriated chloroform, a complex sequence of peaks between 0.9 and 1.3 ppm due to the non-equivalence of the tertiary butyl protons in the free ligand [Fig.2(a)], are resolved into a singlet at 1.2 ppm, indicating that a more ordered symmetry is forced on the molecule on complexation [Fig. 2(6)]. The characteristic AB quartetl6J7 of the bridging methylene protons of the calix[4]arene at 3.2 and 4.6 ppm were found to shift to 3.4 and 4.4 ppm, respectively, which is again indicative of metal complexation conferring a more ordered structure on the moleculel6J7 [Fig. 2(6)]. Fig. 3(a), (6) and (c) shows the absorbance at 520 nm versus the log of the lithium perchlorate concentration for ligands 1,4 and 5, respectively, at each level of sodium perchlorate interferent .The monochromogenic ligand , 4, displayed the best lithium against sodium selectivity. Selectivity coefficients of 73.5, 50.0 and 31.6 against 0.05 mol 1-1 sodium and 73.3, 36.8 and 31.5 against 0.1 moll-' sodium for ligands 4,5 and 1, 3.0 2.5 2.0 1.5 1 .o 0.5 0 Ib) 1 1.2 1 .o 0.8 0.6 2 0.4 0.2 0 0 1.8 I , . 1.6 1.4 1.2 1 .o 0.8 0.6 0.4 0.2 0 300 400 500 600 700 800 Wavelengthlnm Fig. 1 One-phase investigation of changes in the absorbance spectrum of 2.5 ml of solutions of (a) 5.0 x 10-5 rnol I-' ligand 1, and (b) 5.0 X 10-5 moll-' ligand 4 and (c) 6.0 X 10-5 moll-' ligand 5, in THF with 100 yl of TDDA, upon addition of aqueous lithium perchlorate, with final lithium concentrations of: A, 0.1; B, 1 X 10-2; C, 1 x 10-3; D, 1 x 10-4; E, 1 x 10-5; F, 1 x 10-6 moll-1; and G, 0 moll-1ANALYST, SEPTEMBER 1993, VOL.118 1129 I I I I I I I I I L J I I I I I I I I I 9.0 8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0 6 Fig. 2 Partial roton NMR spectra for ligand 1. (a) Before addition of NaSCN. (bf After addition of 1 molar equivalent of sodium thiocyanate in deuteriated methanol to a solution of li and 1 in deuteriated chloroform. Symbols: A and A , tert-butyl; 8 and a, ArCHzAr (filled symbol = before complexation with lithium ions, open symbol = after complexation with lithium ions) respectively, were estimated. These values are an improve- ment on those obtained previously with the nitrophenol ligand.9 When potassium perchlorate was used as the com- plexing metal, no colour or spectral change was observed until a final concentration of potassium perchlorate of 0.1 moll-1 had been added, suggesting that all three ligands are very selective against this metal ion.In the two-phase studies, the best results were obtained with butan-1-01, which showed little or no leaching of the coloured complex into the aqueous phase. Hence, it was decided to examine further ligands 1,4 and 5 , in this solvent. It was found that by using TEA, a large increase in colour intensity, which was independent of metal perchlorate concentration was noticed upon addition of water to the system. It was therefore decided to use a weaker and more hydrophobic base, namely TDDA. Upon addition of TDDA to a solution of ligand 1 in butan-1-01 the colour changed from yellow to slightly orange.This colour was more intense than that observed when TDDA was added to solutions of the ligand in THF. This increased colour is indicative of a solvent dependency on coloration with, in this instance, butan-1-01 being a more basic solvent than THF. Fig. 4 shows the change in the UVNIS absorbance spectra of ligand 1 dissolved in butan-1-01, in the presence of TDDA, produced by varying concentrations of lithium perchlorate. Absorbance maxima at 500 nm and isosbestic points at 425 nm are obtained with all three ligands. Upon addition of water to the system, no colour transfer to the aqueous phase was noted even after a number of days. Of the three ligands, ligand 1 was the only one with which a discernible concentration dependent colour or spectral change was evident after the addition of the lithium perchlorate to the aqueous layer of the two-phase system.2.00 - 0.3 p 0.2 0.1 0.0 I I I I I 0.7 0.6 0.5 0.4 0.3 0.2 .~ -6.0 -5.0 -4.0 -3.0 -2.0 -1.0 Log([Lil/mol I-') Fig. 3 One-phase studies of the optical response of 5.0 X 10-5 moll-' solutions of: (a) 5.0 x 10-5 moll-1 ligand 1, and ( b ) 5.0 x 10-5 mol 1-1 ligand 4 and (c) 6.0 x mol 1-l ligand 5, in THF with 100 pl of TDDA, with varying concentrations of lithium perchlorate in the presence of fixed concentrations of sodium perchlorate. Response measured at 520 nm 1.6 i I 1.4 8 1.2 5 1.0 2 0.8 2 0.6 P Q: 0.4 0.2 0 t R t I 300 400 500 600 700 800 Wavelengthhm Fig. 4 One-phase investigation of changes in the absorbance spectrum of 2.5 ml of 5.0 x 10-5 mol 1-1 solution of ligand 1 in butan-1-01 with 100 p1 of TDDA, upon addition of lithium erchlor- ate, with final lithium concentrations of: A, 0.1; B, 1 X C, 1 X 10-3; D , 1 X 10-4; E, 1 x 10-5; F, 1 x 10-6 moll-1; and G, 0 rnol 1-1 Conclusion Three new chromoionophores that exhibit a dramatic colour change upon complexation with lithium ions, and to a lesser extent sodium ions, have been synthesized. The colour change occurs only in the presence of a base.The base chosen affects the level of absorbance of the complexed compound, as does the solvent in which it is dissolved. Further work on two phase systems and on the incorporation of these compounds into optical sensors is in progress. We gratefully acknowledge financial support for this research, which was jointly funded by the Irish Science and Technology Agency (Eolas) (grant no.SC/92/319) and Eolas/Amagruss Electrodes (Ireland) Ltd. (grant no. HEIC/91/327).1130 ANALYST, SEPTEMBER 1993, VOL. 118 Appendix Structures of Compounds Discussed in the Text N =N -n 1 \==/ 2 0 CHzCOEt 0 CH2COZH II 0 3 5 0 CH2CH=CH2 OCH~COH It 0 6 \ 7 1 2 3 4 5 6 7 8 9 10 11 References Kaneda, T., Sugihara, K., Kamiya, H., and Misumi, S., Tetrahedron Lett., 1981, 22, 4407. Moss, R. E., and Sutherland, I. O., Anal. Proc., 1988,25,272. Misumi, S . , and Kaneda, T., J. Inclusion Phenom. Mol. Recognit. Chem., 1989,7, 83. Cram, D. J., Angew. Chem., Int. Ed. Engl., 1986, 25, 1039. Cram, D. J., Carmack, R. A., and Helgeson, R. C., J. Am. Chem. SOC., 1988, 110,571.Shimizu, H., Iwamato, K., Fujimoto, K., and Shinkai, S., Chem. Lett., 1991, 2147. Nakamoto, Y., Nakayama, T., Yamagishi, T., and Ishida, S., Presented at the Workshop on Calixarenes and Related Compounds, Johannes Gutenberg-Universitat, Mainz, Germany, August 28-30, 1991, Poster 1. King, A. M., Moore, C. P., Sandanayake, K. R. A. S., and Sutherland, I. O., J. Chem. SOC., Chem. Commun., 1992,582. McCarrick, M., Wu, B., Harris, S. J., Diamond, D, Barrett, G., and McKervey, M. A., J. Chem. SOC., Chem. Commun., 1992, 1287. Ungaro, R., Pochini, A., and Andretti, C. D., J. Incl. Phenom., 1984,2, 199. Bohmer, V., Vogt, W., Harris, S. J., Leonard, R. G., Collins, E. M., Deasy, M., McKervey, M. A., and Owens, M. J., J. Chem. SOC., Perkin Trans. 1, 1990, 431. 12 Harris, S. J., MacManus, M., and Guthrie, J., European Pat., EP 0309291A1, assigned to Loctite (Ireland) Ltd., March 29, 1889. 13 Harris, S. J., Woods, J. G., and Rooney, J. M., US Pat. 4642362, assigned to Loctite (Ireland) Ltd., February 10,1987. 14 Arnaud-Neu, F., Collins, E. M., Laitner, B., Deasy, M., Ferguson, G., Harris, S. J., Lough, A. J., McKervey, M. A., Marques, E., Ruhl, B. L., Schwing-Weill, M. J., and Seward, E. M., J. Am. Chem. SOC.. 1989, 111,8681. Van Loon, J. D., Arduini, A., Verboom, W., Ungaro, R., Van Hummel, G. J., Harkema, S., and Reinhoudt, D. N., Tetrahed- ron Lett., 1989,30,2681. 16 Jin, T., Ichikawa, I., and Koyama, T., J. Chem. Soc., Chem. Commun., 1992,499. 17 Arnaud-Neu, F., Barrett, G., Cremin, S., Deasy, M., Ferguson, G., Harris, S. J., Lough, A. J., Guerra, L., McKervey, M. A., Schwing-Weill, M. J., and Schwinte, P., J. Chem. SOC., Perkin Trans. 2, 1992, 1119. 15 Paper 3/01 589 K Received March 19, 1993 Accepted May 4, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801127

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1131 Inverted Poly(viny1 chloride)-Liquid Membrane Ion-selective Electrodes for High-speed Batch Injection Potentiometric Analysis Jianmin Lu and Qiang Chen Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, NM 88003, USA Dermot Diamond* School of Chemical Sciences, Dublin City University, Dublin 9, lreland Joseph Wang* Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, NM 88003, USA Sodium and potassium ion-selective electrodes based on the ionophores valinomycin and tetramethyl-p-terf- butylcalix[4]arene tetraacetate immobilized in a poly(viny1 chloride) membrane were fabricated for use in an inverted configuration for batch injection (BI). The results suggest that BI can be used for routine assays involving these and similar electrodes.Peak shapes obtained under fast and zero stirring conditions are contrasted with those obtained using flow injection (FI). It is concluded that, unlike FI, BI peaks are primarily obtained under kinetically limiting conditions with negligible dispersion. This enables changes in over-all potential arising from surface and bulk effects to be clearly distinguished and will therefore provide an important tool for studying fundamental membrane exchange processes. The analytical performance is characterized by a fast, sensitive and reproducible response. Applicability to assays of mineral water is i I lustrated. Keywords: Batch injection; ion-selective electrode; membrane exchange; potassium and sodium; kinetics Ion-selective electrodes (ISEs) are widely used for the routine monitoring of various analytes.' Their success is based on the selective and fast response for the target species, low cost and simple, portable instrumentation.Cation-selective electrodes, particularly for Group I and I1 metal ions, are usually based on ionophores dissolved in a plasticizer that is dispersed as a liquid membrane in a poly(viny1 chloride) (PVC) supporting matrix.2 These sensors have been successfully used for many years in batch assays3 and more recently as detectors in flow injection (FI) systems.4.5 The advantages offered by FI677 such as reproducible sample handling, automation of sampling and sample processing and relative measurements (peak height) have enabled ISE determinations to be made with excellent precision and accuracy while achieving high sample through- puts.The fact that batch injection (BI) can offer similar advan- tages to FI without the problems associated with valves, tubing, detector flow cells and pumps required by FIS prompted us to investigate the possibility of using liquid- membrane electrodes for BI. The technique involves injecting small sample volumes (1&100 pl) onto a flat sensor surface in a dilution cell and monitoring the transient response produced by the arrival, passage and dispersion of the sample zone over the sensor surface. Very reproducible and fast responses had been demonstrated previously with amperometric8.9 and potentiometricl* (crystalline membrane chloride and fluoride and glass pH) sensors using an Eppendorf pipette to inject the sample.The adaptation of liquid membrane ISEs to BI reported below greatly extends the scope of BI towards numerous ionic analytes for which these sensors are available. The resulting analytical performance compares favourably with that of analogous FI-ISE experiments. The challenge of employing inverted liquid-membrane ISEs (as required by the BI apparatus) and the analytical and fundamental opportuni- ties accruing from such coupling are described below. * Authors to whom correspondence should be addressed. Experimental Electrode Design Preliminary experiments demonstrated that it was necessary to use the ISE in an inverted position in order to achieve reproducibly the 'wall-jet' effect needed in BI through precise control over the pipette tip to membrane distance and angle.This required modification of the usual PVC electrode design in order to prevent the internal electrolyte from reaching the solder joint between the cable and the internal Ag-AgC1 reference electrode. In addition, the internal compartment had to be filled completely with the electrolyte to prevent air bubbles from rising to the internal PVC membrane bobndary and affecting the electrode stability. The design used in this study is shown in Fig. 1. The polycarbonate body was fabricated in two parts, which could be screwed together. An Ag-AgC1 reference electrode was soldered to a BNC socket and the latter glued into place with epoxy resin. Silicone polymer was then used to fill the upper portion of the body completely and allowed to project out slightly from the screw-threaded area in order to prevent any cavity from forming on setting, which might allow air to be BNC socket Silver-silver chloride wire Electrode body Silicone polymer sealant -- PVC membrane . 1-1- Tight-fitting ring Fig.1 Schematic diagram of electrode design developed for use in BI studies1132 ANALYST, SEPTEMBER 1993, VOL. 118 trapped. Discs (diameter 10 mm) were cut from PVC membranes (see below) with a cork borer and securely clipped into place on the electrode tip using a tight-fitting poly- carbonate ring to give a final exposed membrane disc of diameter 5 mm. A slight bevel on the inside edge of the ring prevents cutting of the PVC membrane as the ring is pressed onto the electrode body.The PVC serves as its own gasket and effectively seals the internal electrolyte from the external solution. The lower electrode compartment was then back-filled with the internal electrolyte (0.1 mol 1-1 chloride solution of the primary ion) and, when completely filled, screwed to the upper part of the body so that air was completely excluded. This electrode can then be used in an inverted position for extended time periods without adversely affecting the signal stability. Following attachment of a new membrane, elec- trodes were left to condition in a 0.1 mol 1-1 solution of the primary ion chloride for at least 1 h prior to use. New electrodes were then checked in a series of primary ion solutions by means of conventional beaker-to-beaker steady- state measurements to ensure a satisfactory slope function (S > 50 mV per decade change in primary ion activity) prior to insertion in the BI cell.Electrode bodies were graduated for length to a resolution of 0.5 mm to facilitate pipette tip to ISE membrane separation studies. With this design, PVC mem- branes can be changed in a few seconds if problems arise with the ISE response. PVC Membranes Sodium- and potassium-selective PVC membranes were prepared in the following manner. A 10 mg amount of the ionophores (tetramethyl-p-tert-butylcalix[4]arene tetraacetate for sodium and valinomycin for potassium) and 2 mg of ion exchanger [potassium tetra-p-chlorophenylborate (KTpClPB)] were dissolved in 1 g of plasticizer [o-nitrophenyl octyl ether (o-NPOE)] and 0.5 g of high relative molecular mass PVC was added to give a slurry.Tetrahydrofuran (THF) was added dropwise until a clear solution was obtained and the resulting, slightly viscous solution was poured into a glass Petri dish and left covered with a loosely fitting lid for 24 h. Evaporation of the THF left clear PVC membranes of approximately 0.2 mm diameter, from which the sensor discs were cut. All the above materials were of Selectophore grade and obtained from Fluka, except for the sodium ionophore,ll which was synthesized and purified as described previously. 12 BI Cell and Procedure The 600 ml volume cylindrical BI cell (see Fig. 2) was constructed from Plexiglas and is similar in design to those reported earlier,13 with the exception of the use of an electronic stirrer consisting of a glass rod driven by a small d.c.motor (Model 273-223; Radio Shack) instead of a magnetic stirrer in order to minimize noise. In addition, a tapered injection port was incorporated to give very reproducible pipette positioning and hence reproducible angles and dis- tances between the electrode surface and the pipette tip, a precondition if the constraints of the wall-jet hydrodynamics are to be met.14 Injections are easy to perform and, given the short timescale involved in generating a result (about 30 s), large numbers of replicate measurements can be performed on relatively small sample volumes, giving high confidence in the final assay. These assays were carried out manually as described in earlier BI studies.8-10 Injections (usually 100 p1) were made using an Eppendorf micropipette placed 2 mm from the ISE membrane.Owing to the very high selectivity of these ligands against lithium ions,15 LiCl was used in most experiments as an ionic strength adjustment buffer in samples, standards and the cell filling solution, usually at a concentra- tion of 0.1 mol 1-1. Measurements of physiological levels of n- Ermendorf PiDette Stirrer m o t o r - 0 I P E J/ Tapered inject'ion port Fi I I-em pty hole - 70 mm Filling solution - Electrode port To buffer amplifier 130 mm Fig. 2 Schematic diagram of cylindrical BI cell sodium also employed 9.8 mmoll-1 NaCl in the cell solution. As in FI, ionic strength buffering is an important consideration if real responses are to be distinguished from electronic artifacts. Reagents and Instrumentation Potassium (Fisher Scientific), sodium, ammonium, magne- sium and lithium chloride, calcium nitrate and potassium hexacyanoferrate(I1) (Baker Chemical) were of analytical- reagent grade.Mineral water samples (Naya Canadian Natural Spring Water, Gerolsteiner Sparking Mineral Water and Deming Waters 'Low-Sodium' Deep Well Mineral Water) were obtained in a local supermarket and were used as received. Cell potentials were measured using a Bioanalytical Systems (BAS) X-Y-t recorder after impedance conversion with a laboratory-made voltage follower (input impedance >lo12 Q). Two Ag-AgC1 reference electrodes, BAS Model RE-1 and Cole-Parmer No. G-05992-20, were used for the potentiometric measurements. Amperometric measurements were made using a BAS glassy carbon electrode (MF 2012) with a platinum wire auxiliary electrode and a BAS Model RE-1 Ag-AgC1 reference electrode, linked to a BAS CV27 potentiostat .All the amperometric measurements were obtained using a working electrode potential of +0.8 V versus the Ag-AgC1 reference electrode. Prior to use, the glassy carbon electrode surface was prepared by polishing with 0.05 pm alumina slurry, rinsing with de-ionized water, sonicating for 5 min and then drying in air. Results and Discussion Stability of Sodium ISEs An example of the high sample throughput coupled with excellent reproducibility obtained with the sodium electrode is illustrated in Fig. 3. Fig. 3(a) shows the peaks produced by alternately injecting 100 p1 solutions of 16 and 12 mmol 1-1 sodium chloride (pipette tip to ISE separation 2 mm) into a cell solution of 9.8 mmol l-1 sodium chloride (0.1 moll-1 LiCl was used as an ionic strength adjustment buffer for cell and injection solutions).Fast stirring was employed during these measurements, producing peaks which are reminiscent of those obtained using FI, i.e., a sharp profile with some tailing occurring over a time scale of around 30 s. This represents the detector output as sodium ions are transported towards and into the membrane (sharp rise) followed by efficient washout and dilution (dilution factor 6000) of the sample zone into the bulk filling solution (rapid fall). These characteristics are discussed in more detail under Dynamic Response Charac-ANALYST, SEPTEMBER 1993, VOL.118 1133 7 t - m Time - Fig. 3 (a) Carr over and (b) precision studies with the Naf ISE. Cell filling solution &O ml of 9.8 mmoll-l NaCI-O.l mol 1-l LiCl in both instances. Transient signals produced with a 2 mm separation distance from the pipette tip to the ISE membrane under fast stirring with 100 pl in'ections of (a) 16 mmol 1-1 NaCl (high) and 12 mmol 1-1 NaCl (1 ow and ( 6 ) 16 mmol 1-l NaCl (part of a sequence of 40 injections, s, = 1.7%) teristics. Most important is the lack of sample carryover between the two solutions. These results demonstrate that it is possible to perform many sample injections without adversely affecting the peak profile. As a result, replacement of the filling solution is required only after several hundred injec- tions. Fig. 3(b) demonstrates the good reproducibility obtained over a sequence of twenty 100 pl injections of 16 mmol 1-1 NaCl (pipette tip to ISE separation 2 mm).The relative standard deviation (s,) obtained over the entire sequence was 1.7% ( n = 40), which is excellent considering the likely kinetic-limited nature of the response (the theoretical responses for injections of 16.0 and 12.0 mmoll-1 NaCl into 9.8 mmol 1-1 NaCl are 12.6 and 5.2 mV, respectively, compared with averages of 5.18 and 1.56 mV obtained). Note also the stable baseline obtained during this prolonged experiment. These responses cover the sodium range nor- mally found in blood samples after 10-fold dilution. We are therefore confident that this method could be used to determine blood sodium with good precision and accuracy.This would require proper attention to the disposal of the blood samples accumulated in the cell solution (on its replacement). Optimization of Cell Parameters Pipette tip to ISE separation For these experiments, the cell was filled with 600 ml of a solution of 0.1 moll-' LiCl and 9.8 mmoll-1 NaCl. Using the graduated scale on the ISE body, the electrode was placed at positions from 5.0 to 1.0 mm (in 0.5 mm steps) from the pipette tip. Four injections of 16.0 mmol 1-1 NaCl (in 0.1 mol 1-1 LiCl) were made at each position. The results are shown in Fig. 4(a). It is clear that the separation should be no more than 2.0 mm if one is to obtain adequate sensitivity. However, as the separation distance decreases from 2.0 mm, there is a decrease in precision, possibly owing to minor variations in the pipette tip position during each injection, which become significant at shorter distances.Hence a separation of 2.0 mm was adopted for most measurements. Injection volume Fig. 4(b) shows the effect of varying the injection volume of 16.0 mmol 1-1 NaCl samples into 9.8 mmol 1-1 NaCl-O.l rnol 1-1 LiCl from 10 to 100 pl under conditions of constant pipette tip to ISE separation (2.0 mm) and fast stirring. There 6 3 2 0 1 2 3 4 5 10 20 30 40 50 60 70 80 90 100 Distancelmm Injection vo I u melyl Fig. 4 (a) Optimization of the distance from the pipette tip to the ISE. Injections were of 100 p1 16 mmol 1-1 NaCl into a cell filling solution of 9.8 mmol 1-1 NaCl with 0.1 mol 1-1 LiCl used as ionic strength adjustment buffer for both injection and cell filling solutions. (b) Optimization of injection volume, solutions used as for (a) 2 min * Time - Fig.5 Precision and carryover studies with the valinomycin K+ ISE. Background electrolyte rnol 1-1 LiCl in all solutions; injection volume 100 pl, pipette tip to ISE separation 2.0 mm, fast stirring, high response mol 1-l KCl, s, = 1.27%, n = 10, average response 66.15 mV; low response 10-5 moll-1 KCl, s, = 5.9%, n = 10, average response 16.85 mV is an exponential increase in the response, levelling off towards 100 pl although still increasing, suggesting that the theoretical response (12.6 mV) will eventually be reached. This volume represents a good compromise between sensitiv- ity and amount of material and was used for most of the experimental work.Performance of Valinomycin-based K+ ISE Experiments were also carried out using a valinomycin-based K+ ISE to investigate whether the electrode design would perform satisfactorily with other ionophores. Fig. 5 shows the results obtained for alternating 100 pl injections of 10-4 moll-' (high) and 10-5 moll-' KCl (low) into a 10-3 moll-1 LiCl cell filling solution (pipette tip to ISE separation 2 mm, fast stirring). These results strikingly demonstrate that even at the low concentrations used, the response is sensitive and reproducible (s, = 1.27% for 10-4 mol 1-1 and 5.9% for 10-5 mol 1-I injections, n = lo), with no observable carryover effects from high to low injections. An almost theoretical Nernstian difference (49.3 mV) is observed between the 10-5 and 10-4 mol 1-1 injections.Dynamic Response Characteristics The effect of stirring on the response obtained from injecting 100 ~1 of 10-2 moll-' NaCl into 0.1 moll-' LiCl is illustrated in Fig. 6(a). Clearly, there is a dramatic decrease in the tailing of the peak as one moves from zero stirring (A) to gentle stirring (B) through to fast stirring (C). Interestingly, there is little effect on the peak height, suggesting that the peak1134 .- CI 0, L w ANALYST, SEPTEMBER 1993, VOL. 118 Log aNa I (b) I A L Time - Fig. 6 (a) Response of Na+ ISE to injections of 100 pl of 10-2 moll-1 NaCl in 0.1 moll-' LiCl into 0.1 moll-' LiCl. A, No stirring; B, slow stirring; and C, fast stirring. (b) Response shown on expanded time scale under static (unstirred) conditions of A, K+ ISE (solid line; 100 pl of 10-3 rnol I-' KCI in mol 1-1 LiCl injected into 10-3 moll-' LiC1); and B, bare glassy carbon electrode [broken line; 100 p1 10-3 rnol I-* Fe(CN)64- injected into rnol I-' LiCl at +0.8 V versus Ag-Ag Cl] response is primarily controlled by the active transport towards and away from the membrane provided by the wall-jet effect, whereas baseline recovery is controlled by the dynam- ics of diffusion into the bulk filling solution.This is consistent with the improvement in the washout conditions existing in stirred solutions which reduces the thickness of the Nernst diffusion layer at the electrode membrane and provides efficient mixing of the sample zone with the bulk filling solution. Similarly, the stirrer position had little effect on the peak height or shape (provided that it was at least 2 cm away from the ISE).The opportunity offered by BI for fundamental studies is clearly evident from the transient profiles presented in Fig. 6(6), which compares the dynamic characteristics of the potentiometric response obtained with a valinomycin K+ liquid membrane electrode (100 p1 of 10-3 mol 1-1 KCl injected into 10-3 moll-1 LiCl) to the amperometric response of a glassy carbon electrode [lo0 pl of mol 1-1 F~(CN)G~- injected into 10-3 mol 1-1 LiCl at +0.8 V versus Ag-AgCl] with no stirring. Unlike FI, the sample zone is not in contact with the walls of the tubing, and there is almost instantaneous transfer of the sample to the detector surface and hence nearly zero dispersion. Consequently, the detector output is a true record of its response to an undistorted 'square-wave' concentration impulse.These results demonstrate the almost instantaneous response of the valinomycin ISE to the arrival of the sample, which is even faster than that of the amperometric response of the well polished glassy carbon electrode. This is surprising, considering that the generation of the potentiometric response involves movement of primary ions into the diffusion (Nernst) layer at the ISE membrane boundary, complexation of the metal ions by the ionophore and transfer of the metal complexes into the ISE membrane. The immediate conclusion one can draw from this experiment is that the kinetics of metal ion uptake and complexation by the valinomycin membrane are extremely fast, even in comparison with the electron transfer processes occurring at '1 Time - Fig.7 Calibration graph obtained with Na+ ISE, 10-3 moll-' LiCl background in all solutions and fast stirring. The electrode slope found by regression over the range from 10-4.5 to 0.1 rnol 1-1 NaCl is 35.7 mV per decade compared with 54.3 mV per decade obtained in steady-state batch measurements. The limit of detection is around 10-5 rnol 1-1 Na+ the glassy carbon electrode surface. The wall-jet effect of the sample impinging normally onto the sensor surface also aids the dynamic response, as the ions are actively transported into the diffusion layer rather than passively diffusing from the bulk solution. 16 In contrast, the decreasing signal indicating the diffusion of the sample into the bulk solution, and the consequent removal of primary ions from the diffusion layer and the ISE membrane, is much slower than the glassy carbon response to the F~(CN)G~- injection.This can be interpreted as follows. With the glassy carbon electrode, oxidation occurs only at the electrode surface. Hence, as the oxidation of Fe(CN)& proceeds, its concentration in the diffusion layer decreases. With no replenishment available owing to the rapid dilution of the sample plug into the bulk solution, the signal decays relatively quickly. In contrast, K+ ions can diffuse into the diffusion layer from the PVC liquid membrane phase in response to diffusion of K+ into the bulk solution from the diffusion layer. Hence the PVC liquid membrane can supply primary ions to the diffusion layer as the concentration profile reverses on passage of the sample zone.In addition, these ions are exchanging reversibly across the membrane boundary and are still active in signal generation, unlike the Fe(CN)63- produced at the glassy carbon surface. Overall, these data demonstrate the suitability of BI for exploring the dynamic response behaviour of ISEs. Microcomputer-controlled, high- speed data capture using I/O cards will be used in further studies to examine these features more precisely. Linear Range Fig. 7 shows the response obtained to 100 pl injections of Na+ covering the normal dynamic range of PVC membrane electrodes. Unlike the experiments described above, the cell was filled with 10-3 mol 1-1 LiCl. This represents a compro- mise between the need for a low-resistance solution to minimize noise and accurate determinations of the sensor limit of detection and linear range.The calibration graph (inset) shows the response to be linear from about 10-4.5 mol 1-1 upwards, with some decrease in slope possibly occurring at the upper end of the range. Interestingly, the slope is sub- Nernstian at 35.7 mV per decade compared with 54.3 mV per decade obtained in steady-state batch measurements. This strongly suggests that the electrode is functioning in a kinetic-limited regime under which the characteristics may be different compared with the normal Nernstian behaviour.17 From the calibration graph, the limit of detection can be estimated to be around 10-5 moll-' Na+.ANALYST, SEPTEMBER 1993, VOL.118 1135 2 min - B E I Time - Fig. 8 Analysis of mineral water samples for sodium content. Injection volume 100 pl, pipette tip to ISE separation 2.0 mm, 0.1 moll-' LiCl filling solution, samples diluted 50 + 50 with 0.2 moll-' LiCI. A, Naya (labelled 2.6 x 10-4 mol I-' Na+); B, Deming (labelled 1.2 x 10-3 moll-1 Na+); C, Gerolsteiner (labelled 4.8 x 10-3 moll-' Na+); D, Na+ standard mol 1-1) mol 1-l); and E, Na+ standard Table 1 Results of analysis of mineral water samples Sodium/mol I-' Sample Labelled Found Naya 2.6 x 10-4 1.8 x 10-4 Deming 2.7 x 10-3 1.2 x 10-3 Gerolsteiner 5.8 x 10-3 4.8 x 10-3 Fig. 8 illustrates some preliminary analytical data obtained with mineral water samples. Note again the rapid, sensitive and reproducible response obtained using the BI technique.There is relatively good agreement between the amounts specified on the bottle labels and our results (Table 1). These data suggest that BI coupled with liquid membrane ISEs can provide an inexpensive, reliable and rapid method for screening samples, even at the relatively low concentrations involved in this study. Conclusions From the results, we are confident that BI can be effectively used with any PVC liquid membrane electrode suitably modified for use in an inverted configuration. The unique almost zero dispersion conditions offered by the BI technique and the simplicity of the equipment involved will make it a valuable additional tool for both fundamental studies on liquid membrane and other selective sensors and for developing analytical methodologies for many applications in which traditional batch steady-state measurements are currently applied. The robust form of the cell coupled with the ease with which sensors and filling solutions can be changed and simple, inexpensive equipment will appeal to many industries.The performance can be expected to be particularly good when scavenging agents can be added to the cell filling solution in order to prevent the slow build-up of primary ion concentra- tion during a long sequence of injections. Even when these are not readily available (as in the ion studies in this research), larger cells can be fabricated with the required dilution factor relatively inexpensively or, alternatively, the cell design can be modified to allow continuous addition of filling solution.18 We are continuing our BI investigations into the dynamic response behaviour and applications of these and other PVC liquid membrane electrodes.Arrays of ISEs are also being explored in connection with multi-channel pipettes and a BI operation. We acknowledge financial support from Dublin City Univer- sity for D. D. to work at NMSU. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 References Arnold, M. A., and Solsky, R. L., Anal. Chem., 1986, 58, 84. Moody, G. J., and Thomas, J. D. R., in Chemical Sensors, ed. Edmonds, T. E., Chapman and Hall, New York, 1988, ch. 3, p. 76. Koryta, J., Anal. Chim. Acta, 1984, 159, 1. Meyerhoff, M., and Kovach, P., J. Chem. Educ., 1983,60,766. Telting Diaz, M., Diamond, D., and Smyth, M. R., Anal. Chim. Acta, 1991, 251, 149. RfiiiEka, J., and Hansen, E., Flow-Injection Analysis, Wiley, New York, 2nd edn., 1988. Valcircel, M., and Luque de Castro, M. D., Flow-Znjection Analysis, Ellis Horwood, Chichester, 1987. Wang, J., and Taha, Z., Anal. Chem., 1991, 63, 1053. Chen, L., Wang, J., and Angnes, L., Electroanalysis, 1991, 3, 773 1 Wang, J., and Taha, Z., Anal. Chim. Acta, 1991, 252,215. Diamond, D., Svehla, G., Seward, E., and McKervey, M. A., Anal. Chim. Acta, 1988, 204, 223. McKervey, M. A., Seward, E. R., Ferguson, G., Ruhl, B., and Harris, S. J., J. Chem. SOC., Chem. Commun., 1985, 388. Wang, J., Microchem. J., 1992, 45,219. Glauert, M. B., J. Fluid Mech., 1956, 1, 625. Forster, R., Regan, F., and Diamond, D., Anal. Chem., 1991, 63, 876. Gunasingham, H., and Fleet, B., in Electroanalytical Che- mistry, ed. Bard, A. J., Marcel Dekker, New York, 1989, vol. 16, p. 99. Diamond, D., and Forster, R. J., Anal. Chim. Acta, 1993,276, 75. Amine, A., Kaufmann, J.-M., and Palleschi, G., Anal Chim. Acta, 1993, 273, 213. Paper 3101324C Received March 8, 1993 Accepted April 21, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801131

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1137 Titrations of Non-ionic Surfactants With Sodium Tetraphenylborate Using the Orion Surfactant Electrode Robyn Dahl Gallegos Technology Support Services, Ecolab lnc., St. Paul, MN 55102, USA A method for the determination of non-ionic surfactants is described. The poly(oxyethy1ene) portion of non-ionic surfactants forms psuedo-crown compounds in the presence of barium ions. The Orion surfactant electrode is used as the end-point indicator in titrations of these compounds with sodium tetraphenyl borate. Stoichiometric constants for the reaction of barium non-ionic complexes with sodium tetraphenyl borate have been established. The results of titrations of several non-ionic surfactants indicate that an average of 5.16 oxyethylene units will form a complex with one tetraphenyl borate ion.Empirical titration factors have also been established for several classes of non-ionic surfactants and used in the analysis for the non-ionic content of commercial detergent products. An average recovery of 96.4% was obtained in the analysis of three standard detergent products. The described method has been used successfully in the routine determination of no n-i o n ic su rfacta nts . Keywords: Non-ionic surfactant; sodium tetraphenylborate; ion-pair titration; Orion surfactant electrode Much research has been carried out in the area of potentiome- tric sensors for the determination of non-ionic surfactants. Vytfas et al.1 prepared a poly(viny1 chloride) (PVC) coated- wire electrode (CWE) plasticized with 2,4-dinitrophenyl octyl ether for the titration of non-ionic surfactants with sodium tetraphenylborate (NaBPh4).With CWEs, the consequence of the ill-defined solid-state internal-reference system is often an unstable reference potential from which a high drift can result. This affects the usefulness of these electrodes. Jones et a1.2 prepared a plasticized PVC membrane electrode doped with a barium-non-ionic-tetraphenylborate complex that responded to non-ionic surfactants. The electrode had a useful lifetime of less than 3 weeks. These workers suggested that new membranes are easily prepared, but when considering a sensor that will be used for routine analysis, the lack of convenience and the possibility of adversely affecting repro- ducibility make this undesirable.Given the problems cited for these electrodes, there remains a need for a stable, repro- ducible, long-lived electrode and accompanying methodology for the routine determination of non-ionic surfactants. The Orion surfactant electrode has been commercially available for many years and has proved to be very useful for the determination of anionic and cationic surfactants. We have observed that this electrode also responds to the tetraphenylborate ion. Although the response is sub-Nern- stian (36 mV decade-'), it is stable and can be used to monitor the titrations of barium non-ionic complexes with NaBPh4. The potential break at the equivalence-point (>lo0 mV) is more than adequate for an accurate end-point determination. Several classes of non-ionic surfactants have been analysed. The experimentally determined constant for the number of oxyethylene units (OEUs) forming a complex with one mole of titrant is 5.16 f 0.62, which is in agreement (within experimental error) with that reported by Vytfas et al. 1 (5.2 -+ 0.6) and Sugawara et al.3 (5.64-5.98).Empirical titration factors have been established and used in the determination of non-ionic surfactants in commercial detergent products. A comparison of the titration factors derived using two different electrodes indicates that the method is reproducible and can be used for routine analysis. Experimental Apparatus Potentiometric titrations were carried out using a Brinkmann 670 Titroprocessor (Metrohm, Herisau, Switzerland). Auto- mated titrations were performed by attaching a Brinkmann 665 Dosimat and 673 sample handler.The titrations were monitored using the Orion Model 93-42 surfactant electrode (Orion Research, Cambridge, MA, USA). The Orion Model 90-02 Ag-AgCl double-junction reference electrode, filled with saturated AgCl solution and 10% KN03 solution in the inner and outer chambers, respectively, was used as the reference half-cell . Solutions Solutions of NaBPh4 (approximately 0.008 mol dm-3) were prepared by dissolving analytical-reagent grade powder, puriss >99.5% (Fluka, Ronkonkoma, NY, USA) in de- ionized water, adjusting the solution to pH 9-10 with NaOH and diluting to 1 dm-3. Thallium(1) nitrate (Fluka) solutions (0.01 mol dm-3) were used for the standardization of NaBPh4 solutions. A 0.2 mol dm-3 solution of BaC12 was prepared by dissolving analytical-reagent grade BaCl2-H20 in distilled water.The source and type of non-ionic surfactants that were studied are summarized in Table 1. Surfactant solutions were prepared by dissolving these raw materials in distilled water. Titration Procedure An aliquot of surfactant solution was pipetted into a titration beaker and 2 cm3 of 0.1 mol dm-3 HCl were added. Approximately 150 cm3 of distilled water were added from a graduated cylinder. The titrator was programmed to dose 5 Table 1 Non-ionic surfactants studied Surfact ant Igepal CO Igepal CA Surfonic N Tergitol NP Tergitol Triton X Neodol Plurofac Manufacturer Rhone-Poulenc Rhone-Poulenc Texaco Union Carbide Union Carbide Union Carbide Shell BASF Type NPE' OPEt NPE NPE sec.alc-EO* OPE alc-EO§ alc-EO-PO OEU 9-20 7-12 6-9.3 6-9.5 9.0 9.5 6.5 8.9 * NPE is a nonylphenoxypoly(ethy1eneoxy)ethanol. t OPE is an octylphenoxypoly(ethy1eneoxy)ethanol. Tergitol15-s-9 is an oxyethylenated alcohol with average alkyl and 9 Neodol 23-6.5 is an oxyethylenated linear primary alcohol with OEUs of C13 and 9, respectively. average alkyl and OEUs of C12 and 6.5, respectively.1138 150 > 100 E > y: E rJ 50 0 ANALYST, SEPTEMBER 1993, VOL. 118 . ' ' cm3 of 0.2 mol dm-3 BaC12 into the titration beaker at a rate of 5.0 cm3 min-1. The solution was then stirred for 240 s before the start of the titration. The titrations were performed according to a dynamic method in which the dose rate is governed by the change in potential so that the rate is decreased in the vicinity of the end-point .The maximum dose rate was set at 4 cm3 min-1 with a drift requirement of 8 mV min-1. The titrations were stopped with the identifica- tion of one equivalence-point, and the sample handler was then advanced to a conditioning solution. The conditioning solution was prepared by adding 2 cm3 of 0.1 mol dm-3 HCl to approximately 150 cm3 of distilled water. The electrodes were conditioned in this beaker, with stirring for 10 min, before the next titration. The ternary complex, barium-non-ionic-tetraphenylbor- ate, which is precipitated during the titration, is very insoluble in water and adsorbs onto the surfactant electrode materials. After several titrations have been performed, the electrode body and membrane can become coated with this complex.When the coating is dried on the membrane and not removed, both the accuracy and precision of the method are adversely affected. In order to improve the reproducibility and to ensure proper electrode care, the following analysis protocol was established. The titrations were performed in groups (all in 1 d rather than sporadically over the course of several days). All of the titrations were performed using an automatic titrator equipped with a sample handler. After a group of samples had been analysed, the electrode was allowed to air-dry. The membrane was then cleaned with gentle wiping and forced air before subsequent use. The reconditioning steps recommen- ded by the manufacturer were also followed. We have observed an electrode lifetime of greater than 7 months and this can be extended with proper electrode care.Results and Discussion Electrode Response The response of the Orion surfactant electrode to the ionic species involved in the determination of non-ionic surfactants is illustrated in Fig. 1. The slopes determined for the species that the electrode was designed to monitor, i.e., sodium lauryl sulfate (SLS) and (diisobutylphenoxyethoxyethy1)dimethyl- benzylammonium chloride (Hyamine 1622), are 46.0 and 51.1 mV decade-', respectively. The sub-Nernstian response to the tetraphenylborate ion is acceptable when considering the near-Nernstian reponse of the electrode to SLS and Hyamine 1622. The response to the tetraphenylborate ion is stable with less than 1 mV decade-1 difference between the slopes determined using two different electrodes.Given that and the fact that a Nernstian response is not required when the electrode is used as an end-point indicator, we have found the Orion surfactant electrode suitable for this application. It is known that after regular use with a strongly interfering ion the selectivity of an ion-selective electrode can change.4 Because the Orion surfactant electrode is designed for the determination of anionic surfactants and it was found to respond to the tetraphenylborate ion, this ion can be considered as a strongly interfering species. An experiment was conducted in which the response of a new electrode, electrode A, was compared with the response of an electrode that had been used to monitor titrations of non-ionic surfactants for a period of 7 months, electrode B.The data summarized in Fig. 2 indicate that the response to SLS is decreased from 46.0 to 7.7 mV decade-1 with repeated exposure to the tetraphenylborate ion. This is an indication that the membrane has been chemically altered. The change in electrode response is most likely due to two key factors related to the membrane: the alteration of the bulk composition and the loss of free volume. Sugawara et aZ.3 suggested that the partitioning of the hydrophobic ions between the membrane and sample solution phase dominates 300 , 200 > 100 E > y: E d o I I I 1 J 1 2 3 4 5 6 -Log [surfactant] Fig. 1 Plot of potential versus -log [surfactant] obtained with the Orion surfactant electrode. A, NaBPh,; B, Ba-Surfonic N95; and C, thallium nitrate -200 L 2oo r------ -50 I I I I 1 1 2 3 4 5 6 -Log [surfactant] .Fig. 2 Plot of potential versus -log [surfactant] obtained with the Orion surfactant electrode. A, Electrode A with SLS; B, electrode B with SLS; and C, electrode B after reconditioning with SLS the electrode kinetics and establishes the interfacial potential. The hydrophobic tetraphenylborate ion favours the mem- brane phase and will most likely be absorbed by the membrane and dominate the measurement of the electrode potential. The long-term change in electrode response after exposure to the tetraphenylborate ion suggests that the ion is not only altering the membrane composition at the surface, but also diffusing into the bulk of the membrane. A qualitative analysis of the internal filling solution of an electrode used to monitor the titrations of non-ionic surfactants for a period greater than 5 months, using Fourier transform infrared spectroscopy, indicated the presence of NaBPh4.As it is unlikely that this compound was a formulated species in the internal solution, this is a clear indication of the diffusion of this species into the bulk of the membrane. Cutler and Meares5 suggested that the mobilities of ions in the membrane phase can be affected by the loss of free membrane volume that occurs with repeated use. The diffusion of the tetra- phenylborate ion into the membrane has affected the response to SLS by both decreasing the free membrane volume and altering the interfacial potential. If the analysis protocol that has been described is followed, the surfactant electrode will not be exposed to excess tetraphenylborate and the rate of change of the response to SLS will be greatly decreased. Moreover, as is illustrated in Fig.2, the exposure of electrode B to SLS for 48 h increased the response to this species from 7.7 to 38.7 mV decade-'. This clearly indicates that the change in electrode response is reversible.ANALYST, SEPTEMBER 1993, VOL. 118 1139 Stoichiometry Non-ionic surfactants will form complexes with a variety of monovalent and divalent cations. The stability of the metal non-ionic surfactant complex is dependent on the ionic radius of the metal ion in relation to the radius of the helical conformation of the psuedeo-crown complex that is formed by the oxyethylene portion of the non-ionic surfactant molecules.The structure of a pseudo-crown complex compared with that of a true crown ether complex is illustrated in Fig. 3. The metal ion is held in a 'cage' of the oxygen atoms of the poly- (oxyethylene) portion of the non-ionic surfactant by an ion-dipole interaction.6 Vytfas et aZ.1 have confirmed that the stoichiometry of the complex formation of non-ionic surfac- tants with divalent cations requires that approximately 11 OEUs form a complex with one metal ion. The use of barium to form complexes with non-ionic surfactants for subsequent titration with NaBPh4 has been widely reported in the literature.lJy7.8 Sanchez and Vytfas7 have shown that complex formation with Ba is preferred to that with Ca and Cd. They indicated that the potential break in titrations of metal ions with NaBPh4 in the presence of poly(ethy1ene glycol) (PEG) decreased with decreasing ionic radius of the metal ion.In fact, because of the decrease in the stability of the metal-non-ionic surfactant complex with Ca and Cd, the shallowness of the titration curve, and the deviation from the theoretical equivalence-point, it was decided that these titrations are not suitable for practical use. It was also found that Ba is preferable to Ca in the titrations with the Orion electrode. A decrease in the potential break was observed in the titrations of calcium non-ionic complexes. Also, the nature of the precipitate and its adsorption onto the electrode materials were even more unfavourable than with the barium salt.The stoichiometry of the complex-forming reaction is dependent on the ionic radius of the cation and on the chain length of the oxyethylene portion of the non-ionic surfactant, and this defines the stoichiometry of the reaction of these complexes with NaBPh4. The results for the analysis of homologous series of nonylphenoxy and octylphenoxy poly- (oxyethylene) surfactants are summarized in Table 2. The average value of 5.56 k 0.40 OEUs per mole of titrant agrees well with that reported by Sugawara et aZ.3 With the exception of Igepal CA620, the variation in stoichiometric constants within a homologous series is small. Samples from several classes of non-ionic surfactants were also titrated. The results are summarized in Table 3. An average stoichiometric constant, calculated from the results of determinations of all of the non-ionic surfactants studied, of 5.16 was obtained. Fig.3 crown ether complex Comparison of structures. ( u ) A pseudo-crown complex, ( b ) a The factors affecting the complex formation of the OEUs include the polydispersity or distribution of oxyethylene chain length molecules within the non-ionic surfactant and the structure of the hydrophobic portion of the molecule. The average stoichiometric constant of 5.16 reported in this work most closely resembles the average value reported by Vytfas et aZ.1 They obtained the average value of 5.2 from the analysis of several different classes of non-ionic surfactants forming complexes with several different cations. However, if the average constant is calculated from the data summarized in Table 2, the value obtained (5.56) most closely resembles those reported by Sugawara et aZ.3 (5.64-5.98).They reported on the stoichiometric constants of the barium salts of poly(oxyethy1ene) mono( alky1phenyl)ether surfactants, which are similar to those summarized in Table 2. Vytfas et a2.1 also pointed out that only those non-ionic species with at least five OEUs are titratable. Igepal CA620 is the only surfactant in the series summarized in Table 2 with a stoichiometric constant of 4. It has an average OEU chain length of 7.2. In fact, all of the surfactants studied with calculated stoichiometric constants of <5 have an average OEU chain length of 4.5. With a lower average number of OEUs, it is more probable that a higher percentage of the molecules have chain lengths of <5.It is possible that the use of an average number of OEUs does not provide adequate information for interpreting these results and that information regarding the polydispersity of these surfactants could be used to understand fully the stoichiometric constants that have been reported here. Levins and Ikedag suggested that the spatial arrangement of the atoms in the crystal lattice of the Table 2 Results of titrations of homologous series of non-ionic surfact ants Mean number Stoichiometric Non-ionic of OEUs constant* surfactant (Rhone-Poulenc) (nOEUlnBPh4-) Igepal C0630 Igepal C0710 Igepal C0720 Igepal C0730 Igepal C0850 Igepal CA620 Igepal CA630 Igepal CA720 9 10.4 12 15 20 7.2 9.5 12.5 5.56 k 0.24 5.83 k 0.22 5.94 f 0.18 5.90 f 0.17 5.54 f 0.14 4.66 k 0.48 5.62 -t 0.29 5.46 f 0.02 * Stoichiometric constants are stated as the mean of three determinations plus or minus the range (Xma, - Xmin).Table 3 Results of titrations of several classes of non-ionic surfactants Non-ionic surfact ant Surfonic N95 Surfonic N60 Tergitol NP-9 Tergitol NP-9.5 Triton X-100 Tergitol15-s-9 Neodol23-6.5 Plurofac B255 Mean number M,*l of OEUs g mo1-I NMR NMR) 8.3 585 5.3 453 8.6 598 8.8 607 10.0 646 8.4 569 7.6 527 8.9 755 (13C (13C PO5 3.1 Titration factor+ surfact ant) 1.87 k 0.17* 1.28 k 0.09 1.69 k 0.05 1.74 f 0.09 1.83 _+ 0.06 1.87 f 0.09 1.85 f 0.22 2.26 k 0.10 (nBPh4-ln Stoichiometric constant (nOEUln BPh4-) 4.43 k 0.40* 4.12 k 0.29 5.09 f 0.15 5.05 f 0.26 5.47 k 0.18 4.48 f 0.22 4.09 k 0.49 5.32 k 0.24 * Calculated from the average number of OEUs determined by I3C t Empirical titration factors used in the calculation of % non-ionic * The values stated represent the mean of two determinations plus 6 PO = propylene oxide.Plurofac B255 includes approximately 3 nuclear magnetic resonance (NMR) spectroscopy. surfactant in product mixtures. or minus the range (Xmax-Xmjn). PO units in addition to the OEUs.1140 ANALYST, SEPTEMBER 1993, VOL. 118 ternary complex demands a minimum stoichiometric constant of 5.2. The structure of the hydrophobic portion of these molecules could also affect the arrangement of the atoms in the crystal lattice and the stoichiometry of the complex formation. Izatt et aZ.9 studied the ability of catiohs to form stable complexes with two isomers of cyclic polyethers.They suggested that the differences in the experimentally determined thermodynamic properties between the two isomers can be explained by the steric orientation of the hydrophobic cyclohexyl groups. It is possible that the increased variation in the stoichiometric constants reported in Table 3 could be explained by the various hydrophobic end-groups of the non-ionic surfactants affecting the complex formation and stoichiometry of these reactions. For instance, Tergitol NP-9 and 15-s-9 have similar average OEU chain lengths. The difference, 5.09 k 0.15 and 4.48 -+ 0.22, respectively, in the determined constant could be due to the difference in the hydrophobic portions of these molecules, nonylphenol and secondary tridecanol, respec- tively.As the manufacturer of these two surfactants is the same, a large change in polydispersity would not be expected. However, a summary of this work and the reported literature suggest that the length of the oxyethylene chain dominates the stoichiometry of these reactions. Titrations Vyti-aslo suggested that, if either the substance determined or titrant is changed, a membrane-conditioning step is needed. He chose to precede his analysis with one or two titrations. Levins and Ikeda8 also conditioned the solid-state silver electrode that they used by making a rough titration with a PEG standard. We chose to perform three titrations of a non-ionic surfactant, which preceded all analyses each day for membrane conditioning. Standard titrations of thallium ni- trate and of Surfonic N95 with NaBPh4 are illustrated in Fig.4. It is shown that the potential break at the equivalence-point (>lo0 mV) is more than adequate for precise and accurate end-point identification. In order to establish an empirical titration factor (an experimentally determined factor indicating the volume of titrant required for a specific concentration of a surfactant), three titrations of the raw material were performed and the results were averaged. These values are reported in Table 3 for various surfactants. As the reproducibility of this determi- nation affects the precision of the method, a titration factor was established for Surfonic N95 by titrating a solution five times using each of the two electrodes that have been previously described.It can be seen from the data summarized in Table 4 that the titration factors are very similar. A pooled Volume - Fig. 4 Automated potentiometric titrations of samples with NaBPh4 (about 0.008 mol dm-3) as titrant using the Orion surfactant electrode as end- oint indicator. A , Thallium nitrate (13.32 mg). B, Surfonic N95 (28mg) in the presence of barium(1r) chloride variance t-test indicates that the difference is not statistically significant at the 95% confidence level. After a titration factor has been obtained for a specific surfactant, and a sample containing that surfactant has been titrated according to this method, the factor is then used in the determination of the non-ionic surfactant content of that product. Three laboratory-prepared samples of a laundry detergent in which the non-ionic content ranged from 13 to 39% were analysed according to this method.The empirical titration factor was determined by titrating the non-ionic surfactant that was used to make the products. The results are summarized in Table 5. An average recovery of 96.4 f 4.3% was obtained. Although the data in Table 5 indicate an average recovery <loo%, the accuracy of the method is acceptable for the determination of non-ionic surfactants in product systems. For the analysis of various detergent products, quaternary surfactants will be titrated by NaBPh4 and these will interfere in the determination of non-ionic surfactants. The concentration of quaternary surfactant can be determined by another method" and then subtracted from the result of the titration with NaBPh4 in order to determine the non-ionic surfactant content. Anionic surfactants could also interfere and should be removed before analysis.Conclusion Currently, most of the literature reports the use of laboratory- prepared CWEs or PVC membrane electrodes for the determination of non-ionic surfactants. The preparation of these electrodes is not convenient when considering their use in routine analysis. The Orion surfactant electrode has been used successfully as an end-point indicator in the automated titration of non-ionic surfactants with NaBPh4. With proper electrode care, the electrode lifetime is greater than 7 months, and the apparent change in the selectivity of the electrode that occurs as a consequence of monitoring these titrations is reversible.Empirical titration factors can be established and Table 4 Comparison of empirical titration factors determined by using electrodes of different ages Orion surfact ant Titration of electrode A Surfonic N95 (titration factor*) 1 1.833 2 1.715 3 1.632 4 1.590 5 1.620 Orion surfact ant electrode B (titration factor*) 1.727 1.532 1.515 1.495 1.487 Mean 1.678 1.551 Relative standard deviation (%) 5.8 6.4 * Titration factors are expressed as moles of NaBPh4 er mole of non-ionic surfactant. Electrode A: new electrode. Egctrode B: electrode has been used to determine non-ionic surfactants for more than 7 months. Table 5 Determination of the recovery of non-ionic surfactants in formulated products according to the titration with NaBPh4 Detergent product formula Titrated (% non-ionic (% non-ionic Recovery 13 11.9 91.5 2 1.0 26 25.9 99.5 f 1.0 39 38.4 98.4 f 1.9 surfactant) surfactant) (%) * Calculated as the mean of three determinations using the titration factor for the raw material and the formulated amount as the true value plus or minus the range (x,,,-x,,,~~).ANALYST, SEPTEMBER 1993, VOL.118 1141 used in the determination of non-ionic surfactants in corn- merical detergent products. An average recovery of greater than 96% has been achieved. The author is indebted to Barbara L. Bohn for her dedication and assistance with thk experimental work sum- marized in this paper. The author also thanks Dr. Martin P. Rigney for his thoughtful supervision and technical assistance with this work. References 1 Vytfas, K., DvofAkovA, V., and Zeman, I., Analyst, 1989,114, 1435. 2 Jones, D. L., Moody, G. J., Thomas, J. D. R., and Birch, B. J., Analyst, 1981, 106, 974. 3 Sugawara, M., Nagasawa, S., and Ohashi, N., J. Electroanal. Chem. Interfacial Electrochem., 1984, 176, 183. 4 Evans, A., in Potentiometry and Ion Selective Electrodes, ed. James, A. M., Wiley, Chichester, 1987, ch. 3. 5 Cutler, S. G., and Meares, P., J. Electroanal. Chem. Interfacial Electrochem., 1977,85, 145. 6 Thomas, J. D. R., Analyst, 1991, 116, 1211. 7 Sanchez, M. L. R., and Vytfas, K . , Analyst, 1988, 113,959. 8 Levins, R. J., and Ikeda, R. M., Anal. Chem., 1965,37,671. 9 Izatt, R. M., Nelson, D. P., Rytting, J. H., Haymore, B. L., and Christensen, J. J., J. Am. Chem. SOC., 1971, 93, 1619. 10 Vytfas, K., Mikrochim. Acta, 1984,111, 139. 11 Tsubouchi, M., Mitsushio, H., Yamasaki, N., Anal. Chem., 1981,53, 1957. Paper 210.5605 D Received October 20, 1992 Accepted January 29, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801137

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1143 Frequency Characteristics of an Electrode-separated Piezoelectric Crystal Sensor in Contact With a Liquid Shen Dazhong, Lin Song, Kang Qi, Nie Lihua and Yao Shouzhuo" New Material Research Institute, Hunan University, Changsha 4 10082, People's Republic of China The equations characterizing the frequency response of an electrode-separated piezoelectric crystal (ESPC) sensor, with one side in contact with a liquid, to the properties of the liquid were derived from an electric equivalent circuit model and supported by experiments. The frequency shift of the ESPC can provide information on the permittivity, conductivity, density and viscosity of the liquid. Keywords: frequency equation; electrode-separated piezoelectric sensor; solution properties There is increasing interest in the use of piezoelectric devices as liquid microsensors.Several papers have reviewed the applications of piezoelectric quartz crystals (PQCs) in the chemistry of solutions.1-4 In most instances, the PQC is used to measure the mass change at the crystal surface, which is the basis of the work of Sauerbrey.5 Besides the mass effect, the frequency of the PQC is also affected by the properties of liquids such as density, viscosity, specific conductivity and permittivity.@ The non-mass effect is also useful in che- mistry. For example, the PQC has been used as a viscosity sensor and applied to fermentation monitoring9 and gelation monitoring in order to determine endotoxin or blood coagula- tion factors.10.11 Some applications, based on the high-resolu- tion frequency response of a PQC, to the specific conductivity of solutions were reported from this laboratory.12-17 Usually, a PQC is configured with electrodes on both sides of a thin disc of AT-cut quartz. Recently, a type of electrode-separated piezoelectric quartz crystal (ESPC) sen- sor was reported.18 The electrodes on the surface of the crystal in the ESPC were separated by a liquid layer of a few tenths of a millimetre, and the high-frequency electric field was applied to the quartz disc through the liquid layers.As a result, it had a frequency response to the liquid properties more sensitive than that of a normal PQC, especially to the permittivity and specific conductivity. The behaviour of the ESPC in solution and its applications in analysis were also reported .19-21 However, quantitative knowledge of ESPC behaviour in solution is still required if it is to be applied successfully in chemical analysis.Mo et aZ.22 measured the electric equivalent circuit parameters of this kind of piezoelectric sensor in solution. It is worth pointing out that the resonant frequency of a piezoelectric sensor, as determined by an impedance analyser, differs from the oscillating frequency of the oscilla- tor, because the oscillating frequency depends not only on the parameters of the crystal and the solution, but also on the properties of the oscillator. In this paper, we have derived equations to describe the influence of the properties of a liquid on the oscillating frequency of an ESPC. Theory The configuration of the ESPC is illustrated in Fig.1. From the point of view of an electric equivalent circuit, an ESPC can be represented by the series combination of equivalent circuits of a PQC and a liquid. The electric equivalent circuit of the ESPC, in the frequency range near the resonance of the PQC, is shown in Fig. 2. Its total impedance 2 is given by * To whom correspondence should be addressed. where R = Gl(G2 + o 2 Q ) , X , = oCJ(G2 + d Q ) , Xo = kE + Cp; G is the solution conductivity (G = l/Rs), and C, is the solution capacitance; x is the specific conductivity, and E is the permittivity of the solution; k is the cell constant of the detection cell; Cp is the parasitic capacitance of the leading wires, with a typical value of 1.2 pF; o is the angular frequency; Co, L,, C, and R, are the static capacitance, motional inductance, motional capacitance and motional resistance of the crystal, respectively; and j = m.It is well known that there are two conditions that must be satisfied for oscillation to occur. One is that the phase shift around the loop should be zero; the other is that the loop gain should be unity. The loop is a closed path from the input of the amplifier through the amplifier to its output, and back to the input through the feedback circuit element. Usually, the loop gain condition can be easily satisfied because the gain of the amplifier is much greater than unity and will automatically adjust to unity. Hence, the oscillating frequency is mainly determined by the phase-shift condition.If the amplifier used in an oscillator has a phase shift of 8, the feedback circuit element must have a phase shift of -8 in order to satisfy the phase-shift condition. When the ESPC is connected between the input and output of the amplifier, the ratio of imaginary to real parts of its impedance should be lIoC0, N = R,oCo, M = 1 + CdCq - dL,Co, G = &, C, = Fig. 1 Configuration of electrode separated piezoelectric crystal (ESPC) sensor: A, separated-electrode; B, detection cell; C, piezo- electric quartz crystal (PQC); D and E, silver electrode of PQC and leading wire; and F, water jacket Fig. 2 Electric equivalent circuit of ESPC: R,, solution resistance; C, solution capacitance; Co, static capacitance; L,, motional inductance; C, motional capacitance; and R, motional resistance1144 ANALYST, SEPTEMBER 1993, VOL.118 equal to tan( 4) according to the phase-shift condition. Therefore, the following equation is obtained: where Y = tan@) is a parameter of the oscillator, which depends on the type of oscillator and its operating conditions. The typical Y value for the oscillator used in this paper is 1.83. Because all the parameters in eqn. ( 2 ) are related to frequency, it is difficult to obtain an exact algebraic expression of the oscillating frequency of the ESPC from this equation. Fortunately, the relative change in oscillating frequency is small, only the value of M is sensitive to the slight frequency change. Therefore, the oscillating frequency can be approxi- mately obtained from the value of M. Further, the values of Xo, X , and R are calculated by using the resonance frequency of the crystal, as the relative difference between the oscillating frequency of the ESPC and the resonant frequency of the crystal is very small.Under this assumption, eqn. (2) is rewritten as PM2 - M + PW - YN = 0 (3) where P = 1 + (X, - Y R ) WCO. From eqn. ( 3 ) , there are two roots for M . It is found that only the larger root is consistent with the experimental results, as the series oscillating frequency is monitored in our oscillator. This root is expressed as M = [ l + ( 1 + 4PYN - 4@W)*]/2P (4) According to the definition of M = 1 + CdCq - wLqCo, the oscillating frequency of the ESPC is given by ] ( 5 ) F C [ 1 + ( 1 + 4 Y P N - 4 P W ) h 2P F = F , + - i - 2CO where F, = 1 / 2 n q i s the resonant frequency of the crystal in hertz and F is the oscillating frequency of the ESPC in hertz.Equation (5) is the fundamental equation characterizing the oscillation frequency of the ESPC. If an iterative method is used, the exact solution of eqn. ( 2 ) can be obtained even if only three iterations are used. In fact, the frequency calculated by eqn. (5) is very close to the results obtained from the iterative method. If a non-electrolyte solution is used to fill the space between the separated electrode and the crystal, then P = 1 + CdC, for an ESPC under conditions of very low conductivity of the solution. Therefore, we have F C [ 1+(1+4HYN-4WhR)f 2H F = F , + ~ i - 2CO where H = 1 + CdC,. Equation (7) indicates that increases in C, or the permittiv- ity of the solution correspond to a decrease in frequency. Because F, decreases with the increment of the density and viscosity of solution,16323 the oscillating frequency of the ESPC also decreases with increasing density and viscosity of the solution.In electrolyte solution, the permittivity of the solution hardly changes with the specific conductivity; hence, the value of P is mainly determined by the specific conductivity of the solution and the cell constant. According to the definition: (7) where wo = 2nF,. It can be seen that the P values decrease with the increment of specific conductivity until a minimum is reached, then increase as the specific conductivity continues to increase. According to eqn. ( 5 ) , the variation trend in the oscillating frequency of an ESPC with the specific conductivity is similar to that of P, because the value of 1 + ( 1 + 4YPN - 4@N2)4 is less sensitive to the change in P .For the convenience of discussion, the square root in eqn. (5) is developed by Taylor' series, and eqn. (5) is approxi- mately expressed as where a = 1 + ( 1 + 2HNY)/b, and b = ( 1 + 4HNY)r. derived from eqn. (8) as follows: The sensitivity to specific conductivity, i.e., aF/ax, can be -=- aF JIG k Cq (Yw8G + 2Go& - YG2) x ax It transpires that the slope of frequency versus specific conductivity depends not only on the parameters of the crystal and circuit, but also on the solution conductivity and cell constant. In solutions of low conductivity, G< woCs, eqn. (9) can be simplified as It can be seen that the slope is constant for solutions of low conductivity.That is to say, the frequency of the ESPC decreases linearly with the increment of the specific conductiv- ity of the solution or the concentration of electrolyte. Besides, the influence of the cell constant on the sensitivity is slight, because C is proportional to the cell constant. In highly conductive solutions, G >> ooC,, eqn. (9) can be simplified as It was obvious that the values of aF/ax become positive and decrease with the increase in specific conductivity. Hence, the frequency of the ESPC increases with increasing specific conductivity of the solution or the concentration of the electrolyte in highly conductive solutions. Additionally, the sensitivity is low and becomes less with larger cell constants in this range.If the conductivity of the solution is so great that the solution impedance can be neglected, the sensitivity to specific conductivity approaches zero. In fact, a PQC with one side in contact with a liquid phase (PQC in the following discussion) can be treated as the special case of an ESPC. According to the equivalent circuit shown in Fig. 2, if R, = 0, the ESPC is transformed into a PQC. Under such conditions, P = 1 and the oscillating frequency of the PQC is expressed as It seems that the oscillating frequency of the PQC is insensitive to the change in the specific conductivity and permittivity of the liquid, and depends mainly on the resonant frequency and motional resistance of the crystal. For the situation between the two extreme examples, the change in sensitivity with specific conductivity is interesting. As the specific conductivity increases, the sensitivity increases until a maximum is reached, then it decreases through zeroANALYST, SEPTEMBER 1993, VOL. 118 1145 and increases again towards positive values.When the positive maximum is obtained, the sensitivity decreases to zero again. Generally, frequency shift is used in piezoelectric measure- ments. The frequency shift (AF) used in this paper is the frequency difference between test and reference solutions, i.e., AF = Fl - Fo, where Fl and Fo are the frequencies in test and reference solutions, respectively. With the frequency in pure solvent serving as the reference, the frequency shift in electrolyte solution can be calculated as follows: If a solution with specific conductivity of xo is used as the reference solution, and the change in specific conductivity (Ax) is much less than u, the frequency shift that reflects the change in specific conductivity is given by where Go = k u .This equation reveals that the minute change in specific conductivity can be determined by an ESPC even in the presence of a highly conductive background. Moreover, the frequency shift is linearly related to the change in specific conductivity. The sensitivity can be improved if the value of the background specific conductivity is in the range where the frequency is sensitive to the change in specific conductivity. Experimental In the ESPC (Fig. l ) , a platinum disc, with a diameter of 6 mm and a thickness of 1 mm, is attached to one end of a ground-glass tube with silicone resin; this is used as the separated electrode.An AT-cut 9 MHz piezoelectric crystal (12.5 mm diameter, with silver electrodes) is attached to the base of the detection cell with silicone resin. The distance between the separated electrode and the crystal is adjustable. For the purpose of measuring the parameters of the crystal and comparing the frequency characteristics of the ESPC and PQC, the silver electrodes are not dissolved. It is found that the response characteristics of the ESPC, with or without silver electrodes, remain almost the same. When the sepa- rated electrode and silver electrode (D) are used, the sensor is an ESPC. If both of the silver electrodes are connected, the sensor is converted into a normal PQC.The oscillating frequency of the TTL-IC oscillator (made in this laboratory) is monitored by a frequency counter (Iwatsu; Model SC-7201). An impedance analyser (Hewlett-Packard, Avondale, PA, USA; Model 4192A) is used to measure the parameters of the crystal, oscillator and liquid. The solution capacitance and conductivity are measured between the separated electrode and silver electrode (E) at a frequency of 8 MHz. Analytical-reagent grade chemicals and doubly distilled water were used. Test solutions were prepared with the reference solution (see under Results and Discussion). The reference solution was introduced into the detection cell, the stable frequency (Fo) was recorded, then the reference solution was replaced by the test solution, and the stable frequency ( F l ) was again recorded.The frequency shift, AF = Fl - Fo, was measured three times and the mean value was used. The equivalent circuit parameters of the crystal were measured by the admittance diagram method.23 Results and Discussion Frequency Response of ESPC to Specific Conductivity With pure water as the reference solution, the frequency shifts of the ESPC with different cell constants, in potassium chloride solutions of known specific conductivity, are illus- trated in Fig. 3. The sensitivity, i.e., aF/ax, is also shown in this figure. The solid lines are derived from eqns. (13) and (9), which are in good accord with the experimental points. As predicted above, the sensitivity is relevant to the cell constant with specific conductivities ranging from 0.003 to 0.3 S m-1.A better sensitivity can be expected when an ESPC with a smaller cell constant is used in this range. Beyond this range, the influence of cell constants on the sensitivity is slight. Frequency Response of ESPC to Permittivity For a normal PQC operated in liquid, the frequency shift arising from permittivity change is slight8 or negligible.' These results can be explained by eqn. (12), because the electric equivalent circuit shown in Fig. 2 is also suitable for a PQC with one or two sides in contact with a non-electrolyte solution. The frequency response of a PQC, with one side in contact with the liquid, is insensitive to the permittivity, because the permittivity of the liquid has a slight influence on the equivalent circuit parameters of the crystal. Although the Co value of a PQC immersed in non-electrolyte solution changes with the permittivity of the liquid, the frequency is insensitive to the change in CO, according to eqn.(12). As seen in eqn. (6), the frequency of the ESPC is more sensitive to the change in the permittivity owing to the term FsC,/2(Co + Cs). With 1,4-dioxane used as the reference solution, the frequency shifts of the ESPC and its crystal in a water-1,4- dioxane mixture are depicted in Fig. 4. The solid lines ( A X ) are plotted on the basis of eqn. (6) by using the values of F, (line E), R, and Co of the crystal (shown in Fig. 5 ) and of C, in this mixture. Good agreement was obtained between the results predicted theoretically and those measured experimen- tally.Frequency Response of ESPC to Density and Viscosity According to ref. 6, the frequency shift of the PQC is proportional to (pq)i, where p and 11 are the density and viscosity, respectively, of the liquid with which the crystal is in contact. The influence of density and viscosity on the frequency of the ESPC has been investigated in sucrose solutions with known density and viscosity.24 With pure water used as the reference solution, the frequency shifts of the ESPC and its crystal are depicted in Fig. 6. Again the solid lines (B and C) are derived from eqn. (6) or -lo5 5 I n I -104 I, G -- a 103 N 5 A 8 C r 0) U 2 -102 2 -10 I I io-4 10-3 10-2 10-1 1 10 Conductivity (x)lS m-1 Fig. 3 Dependence of frequency shift (AF) and sensitivity on the specific conductivity.ESPC with cell constant (m): A, 0.0086; B, 0.0159; and C, 0.02641146 ANALYST, SEPTEMBER 1993, VOL. 118 -12 N 5 -8 5 a E UJ -4 5 2 0 - r Q) U L 4 0 20 40 60 80 100 Relative permittivity Fig. 4 Dependence of frequency shift (AF) on the permittivity in a water-dioxane mixture: A, ESPC with cell constant of 0.0086 m; B, ESPC with cell constant of 0.0159 m; C, ESPC with cell constant of 0.0264 m; D, oscillating frequency of the crystal; and E, resonant frequency of the crystal 420 9 390 a!! A 1 Q m v) g 360 Y .- 330 C .- c 300 270 - 8.6 L P \ - u" - - 8.2 $ C m 0 4- .- - 7.8 $- 0 0 .- c - 7.4 ;; I I I I 1 7.0 0 20 40 60 80 100 Relative permittivity Fig. 5 Motional resistance and static capacitance of the PQC in a water-dioxane mixture: A, motional resistance; and B , static capaci- tance -10 N -8 P G E 5 g -4 -6 >.1 1.5 2.0 2.5 3.0 3.5 cm-t cph Fig. 6 Dependence of frequency shift on the density and viscosity in sucrose solutions: A, resonant frequency of the crystal; B, oscillating frequency of ESPC with cell constant of 0.0264 m; and C, oscillating frequency of the crystal (12) by using F, (line A), and R, and Cs (in Fig. 7). It appears that the frequency shifts of the PQC are in linear correlation with (pq)i, whereas the linearity of frequency shift versus (pq)t in ESPC is only approximate. Moreover, the frequency shift of the ESPC is slightly less than that of its crystal. The difference in frequency shift is due to different response characteristics 20.2 * Iooo L P om 19.8 1 I 8 m 4- '2 19.4 c1.0 C 0 0 v) .- 3 19.0 - - G 800 Q) C m c 600 'Z 2 - 0 .- 4- 400 9 18.6 2 200 1 1.5 2.0 2.5 3.0 3.5 V&gi cm-i cp6 Fig. 7 Solution capacitance and motional resistance in sucrose solutions: A, motional resistance; and B, solution capacitance between the PQC and ESPC. The difference in frequency between the ESPC and PQC, which is obtained by subtracting eqn. (12) from eqn. (6), is expressed as Equation (15) reveals that the oscillating frequency of the ESPC is slightly greater than that of its crystal under the same solution conditions. With increasing (pq)i, C, decreases, and the oscillating frequency difference between the ESPC and PQC increases. Therefore, the frequency shift in the ESPC is slightly less than that of its crystal. For the same reason, the linearity of A F versus (pq): for the ESPC is not good as that for the PQC. The frequency change of the ESPC caused by density and viscosity is about 1.5 times larger than that of the normal PQC.20 The different results could be due to the difference in the construction of the sensor and oscillator.An ESPC with two separated electrodes was used by Nomura and Yanagi- hara.20 Frequency Response of ESPC to Specific Conductivity Under Background Conductivity With NaN03 solution of various concentrations used as the reference, the frequency shifts of the ESPC in response to the concentration of electrolyte (KCl as an example) are illus- trated in Fig. 8. It is evident that the frequency shift of the ESPC can reflect a minute change in specific conductivity or in the concentration of electrolyte against a highly conducting background.In our opinion, this is the most important advantage of ESPCs over the classical conductime- tric method, with respect to determinations based on conduc- tivity. The reason why the ESPC can detect such a slight variation in specific conductivity against a highly conducting background is that it exhibits a very good signal-to-noise ratio or resolving power. The typical noise level for an ESPC is about 2 Hz and depends slightly on the conductivity, while the fundamental frequency of the crystal is 9 MHz; therefore, the small change in specific conductivity can be measured by the ESPC with ease even against a background of high conductiv- ity. As predicted from eqn. (14), the slopes of these response curves are relevant to the value of the background conductiv- ity.Because the solution conductivity depends directly on the cell constant, the sensitivity of the ESPC can be changed by adjusting the distance between the separated electrode and the crystal.ANALYST, SEPTEMBER 1993, VOL. 118 1147 4 - 1600 t /" 400 1 I I I I I I 0 2 4 6 8 1 0 1 2 'Concentration/lO -3 mol I-' Fig. 8 Response curves of ESPC in the presence of a foreign electrolyte. Concentration of NaN03 (10-3 moll-1): A, 0; B, 1; C, 2; D, 3; E, 4.5; F, 6; G, 10; and H, 15 Conclusions The characterization of an ESPC with one separated electrode is discussed theoretically. The frequency of the ESPC is sensitive to the change in permittivity and conductivity of a liquid. The frequency shift arising from the change in density and viscosity is slightly less than that of a normal PQC that operates under the same conditions.It is more important that the ESPC can detect a minute conductivity change even against a highly conducting background. This work has been supported by the National Science Foundation and the Education Commission Foundation of China. References 1 Deakin, M. R., and Buttry, D. A., Anal. Chem., 1989, 61, 1147A. 2 McCallum, J. J., Analyst, 1989, 114, 1173. 3 Buttry, D. A., in Electroanalytical Chemistry. ed. Bard, A. J., Marcel Dekker, New York, 1991, vol. 17, pp. 1-85. 4 Thompson, M., Kipling, A. L., Duncan-Hewitt, W. C., RajakoviC, L. V., and CaviC-Vlasak, B. A., Analyst, 1991,116, 881. 5 Sauerbrey, G., 2. Phys., 1959, 155,206. 6 Kanazawa, K. K., and Gordon, J. G., 11, Anal. Chim. Acta, 1985, 177,99. 7 Nomura, T., and Okuhara, M., Anal. Chim. Acta, 1982, 142, 281. 8 Yao, S. Z., and Zhou, T. A., Anal. Chim. Acta, 1988,212,61. 9 Endo, H., Sode, K., Karube, I., and Muramatsu, H., Biotechnol. Bioeng., 1990, 36, 636. 10 Muramatsu, H., Tamiya, E., Suzuki, M., and Karube, I., Anal. Chim. Acta, 1988,215, 91. 11 Muramatsu, H., Tamiya, E., Suzuki, M., and Karube, I., Anal. Chim. Acta, 1989,217, 321. 12 Yao, S. Z., and Mo, Z. H., Anal. Chim. Acta, 1987,193,97. 13 Yao, S. Z., Mo, Z. H., andNie, L. H., Anal. Chim. Acta, 1990, 229, 205. 14 Mo, Z. H., Nie, L. H., and Yao, S. Z., Sci. China (Ser. B), 1991, 1, 1. 15 Mo, Z. H., Nie, L. H., and Yao, S. Z., Anal. Chim. Acta, 1991, 246,421. 16 Wei, W. Z., Nie, L. H., and Yao, S. Z., Anal. Chim. Acta, 1992, 263, 77. 17 Wei, W. Z., Nie, L. H., and Yao, S. Z., Anal. Chim. Acta, 1992, 269, 149. 18 Nomura, T., and Tanaka, F., Bunseki Kagaku, 1990,39,773. 19 Nomura, T., and Takada, K., Bunseki Kagaku, 1991,40, 567. 20 Nomura, T., and Yanagihara, T., Anal. Chim. Acta, 1991,248, 329. 21 Nomura, T., Tanaka, F., and Yamada, T., Anal. Chim. Acta, 1991,243,273. 22 Mo, Z. H., Nie, L. H., and Yao, S. Z., J. Electroanal. Chem. Interfacial Electrochem., 1991, 316, 79. 23 Zhou, T. A., Nie, L. H., and Yao, S. Z., J. Electroanal. Chem. Interfacial Electrochem., 1991, 293, 1. 24 Handbook of Chemistry and Physics, ed. Weast, R. C., CRC Press, Boca Raton, FL, 64th edn., 1984-1985, D-270. Paper 3100353A Received January 20, 1993 Accepted March 17, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801143

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1149 Enzymic Assays of Organic Peroxides in Microemulsion Systems Joseph Wang and A. Julio Reviejo* Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, NM 88003, USA Microemulsions are shown to extend the scope of enzymic amperometric assays towards both hydrophobic and hydrophilic compounds. This possibility is illustrated for amperometric monitoring of organic peroxides in oil-in-water emulsions containing horseradish peroxidase. Enhanced sensitivity (compared with work in pure aqueous solution) is observed also for water-soluble analytes, and is attributed to changes in the local substrate concentration. The effect of the emulsion structure on the biocatalytic activity and analytical performance is explored. Both hydrophobic and hydrophilic mediators can be used.Amperometry can also be used for elucidating biocatalytic conversions in microemulsions. Future prospects are discussed. Keywords: Enzymic assay; organic peroxide; microemulsion; peroxidase The interest in organic-phase enzymic assays is growing rapidly. 192 Several useful enzyme electrodes and unique applications utilizing non-aqueous environments have already been reported .3-8 Similar analytical opportunities can accrue from the use of microemulsion systems (i.e., mixtures of water and oil separated by a surfactant-rich film). Many studies have been devoted to biocatalytic conversions in microemulsions (particularly reverse micellar systems) .942 Analytical advan- tages accrued from the coupling of enzymic<hemilumines- cence detection schemes have been documented,l3J4 but enzymic electrochemical assays utilizing these media have not been reported.The aim of this investigation was to demonstrate the analytical possibility of using amperometry for monitoring enzymic processes in microemulsions. The main analytical advantage resulting from using such environments is solubil- ity, namely, microemulsions are good solvents for both hydrophobic and hydrophilic compounds, and for compounds of intermediate character. The large (watedoil) interfacial area offers a rapid exchange of such analytes between the organic and water domains, hence facilitating their conversion by the enzyme (present in the aqueous phase). In addition, microemulsions have low mass transfer limitations and can be prepared rapidly.Finally, by avoiding direct contact with the unfavourable organic medium it is possible to stabilize the enzyme against the inactivating action of such solvents. The prospects of using microemulsions for amperometric enzymic assays are illustrated here for monitoring organic peroxides in the presence of horseradish peroxidase. Organic peroxides are of great environmental, industrial and biological interest15316 and possess a broad solubility range. In addition to its analytical significance, the work described here illus- trates the utility of amperometric electrodes for in situ probing of enzymic conversions in microemulsions Experimental Apparatus A 10 cm3 cell [Model VC-2, Bioanalytical Systems (BAS) W. Lafayette, IN, USA] was connected to a glassy carbon disc working electrode (Model MF-2012, BAS), an Ag-AgC1 reference electrode (Model RE-1, BAS) and a platinum wire auxiliary electrode through holes in its poly- (tetrafluoroethylene) cover.A magnetic stirrer and stirnng bar provided the convective transport. Amperometric measurements were performed with an EG & G Princeton Applied Research (Princeton, NJ, USA) Model 174 voltam- * Permanent address: Department of Analytical Chemistry, Com- plutense University, Madrid, Spain. metric analyser, in conjunction with a BAS X-Y-t recorder. The working electrode was polished with an alumina slurry. Reagents and Procedures All the solutions were prepared from analytical-reagent grade chemicals and distilled water. Dimethylferrocene, butan-2- one peroxide, benzoyl peroxide, tert-butyl peroxide, lauroyl peroxide, dicumyl peroxide, ethyl acetate, cetyltrimethylam- monium bromide (CTAB) and chloroform were obtained from Aldrich (Milwaukee, WI, USA), while cumene hydroperoxide, ferrocene and o-phenylenediamine were received from Sigma (St.Louis, MO, USA). Potassium hexacyanoferrate(II), Triton X-100 and sodium dodecyl sul- fate (SDS) were obtained from J. T. Baker (Phillipsburg, NJ, USA). All the peroxides were dissolved in ethyl acetate. The enzyme [horseradish peroxidase (HRP) EC 1.11.1.7, Type I, 100 U mg-1; (1 U = 16.67 nkat) Sigma] was dissolved in 0.05 mol dm-3 phosphate buffer. Amperometric measurements were performed at room temperature, by applying the desired (-0.2 V) potential, allowing the transient currents to decay, and stirring the solution.Preparation of Microemulsions Different volumes of the organic solvent (containing the mediator) and of the surfactant were placed in a 10 cm3 calibrated flask, and diluted to the mark with phosphate buffer solution (to yield final concentrations of 1 4 % and 0.05- 0.4% , respectively). Clear emulsions were obtained by shaking or sonicating these mixtures. Results and Discussion While enzymic conversions in microemulsions have tradition- ally relied on water-in-oil (w/o) systems, oil-in-water (o/w) emulsions are used here to demonstrate the analytical prospects of enzymic assays in emulsified media. Such o/w systems offer a broad solubility range, while maintaining high biocatalytic efficiencies.Ethyl acetate-in-water emulsions are, therefore, used here for peroxidase-based assays of organic peroxides. Such assays rely on the biocatalytic conversion of organic peroxides in the presence of a suitable mediator (M): ROOH + M (red) HRP ROH + H20 + M (ox) Fig. 1 shows current-time recordings, obtained at the glassy carbon electrode immersed in emulsified (A) and aqueous (B) media, on successive additions of lauroyl peroxide (a) and cumene hydroperoxide (6). The emulsion system was stabil- ized by CTAB and Triton X-100 (not shown), respectively.1150 ANALYST, SEPTEMBER 1993, VOL. 118 Determination in the pure water solution is not feasible because of the poor solubility of these highly hydrophobic peroxides. In contrast, a rapid response to these sub-milli- molar additions is observed when using the emulsified media.Steady-state currents are achieved-within 15 s (a) and 50 s (b). Such difference is attributed to different rates of substrate partition (from the oil pools towards the bulk aqueous phase containing the enzyme). The favourable signal-to-noise characteristics indicate detection limits in the micromolar range. The o/w microemulsion operation can benefit also enzymic assays of water-soluble peroxides. Fig. 2 displays calibration plots for butan-2-one peroxide (over the 0.1-1.0 mmol dm-3 range) in aqueous and microemulsion systems, and in the presence of hexacyanoferrate(i1) (a) and o-phenylenediamine (b) as mediators. Both the Triton X-100- and CTAB-based emulsion systems exhibit a sensitivity better than that of the corresponding water solutions.The sensitivity enhancement is particularly pronounced (11-fold) in the presence of CTAB. The reason for these improvements is discussed below. If enzyme catalysis in microemulsions is to be of practical analytical significance, it is essential to understand the factors influencing the biocatalytic activity in such systems. Unlike w/o emulsions, o/w systems (used in the present investigation) do not exert substantial changes in the biocatalytic activity because the enzyme is confined within the bulk aqueous phase. The activity and stability of enzymes in such systems are usually of the same order of magnitude as in an aqueous environment. However, some changes in reaction rates can result from differences in the 'local' substrate concentration in the aqueous domain.For example, the higher reaction rates T A f 2 min H/B Time - Fig. 1 Typical current-time recordings on increasing (a) the lauroyl peroxide and (b) cumene hydroperoxide concentration in steps of 0.1 and 0.5 mmol dm-3, respectively. Solutions: A, emulsion (2% ethyl acetate, 97.8% phosphate buffer, 0.2% CTAB); and B, phosphate buffer (0.05 mol dm-3, pH 7.4); both solutions contained 5 U HRP and 2 mmol dm-3 'ferrocene. Operating potential, -0.2 V; stirring rate, 300 rev min-1 8.0 1.6 0.8 0 0.5 1 .o 0 0.5 1 .o Concentrationhmol dm-3 Fig. 2 Calibration plots for butan-2-one peroxide in A, a ueous and B , microemulsion media. Mediators: (a) hexacyanoferrateqri) and (b) o-phenylenediamine. Surfactants: (a) Triton X-100; and (b) CTAB.[Other conditons as in Fig. 1 (A)] (than in water, e.g., Fig. 2) are attributed to bringing together the substrate, mediator and enzyme in higher 'local' concen- tration near the interface. (The hydrophobic part of the enzyme molecule can come into contact with the phase of the organic micro-domain .) The effect of the microemulsion composition (its components and their level) on the response of the enzymic assay was therefore investigated. The effect of various microemulsion stabilizers, including CTAB, Triton X-100 and SDS on the sensitivity was examined for successive (0.1 mmol dm-3) concentration increments of butan-2-one peroxide (not shown). Such stabilizers are representative of cationic, neutral and anionic surfactants, respectively. The three emulsions displayed a very rapid and sensitive response to these additions of the peroxide substrate (yielding steady-state currents within 5-10 s).While no substantial differences were observed among the three surfac- tants, the CTAB-containing emulsion offered the most favourable signal-to-noise characteristics. Unlike its enzyme denaturating activity in aqueous solutions, SDS does not display such activity in the microemulsion. As shown in Fig. 3, changes in the level of different components of the emulsion system, including the percentage of the oil (A) or of the surfactant (B), and also the mediator concentration (C) and the enzyme activity (D) have little effect on the response to the peroxide substrate. No response was observed in the absence of HRP (not shown). In contrast, oil-rich w/o microemulsions, with the enzyme entrapped in the water droplets, can lead to pronounced changes in the biocatalytic activity,12J7 and can, therefore, allow tailoring of the enzyme function (by engineer- ing the emulsion composition) to meet the requirements of a given enzymic assay. Enzymic assays in microemulsions can benefit from the use of mediators with a broad solubility range.For example, Fig. 4 displays calibration data for butan-2-one peroxide in emul- sions formed with Triton X-100 (a) and CTAB (b), and containing the water-soluble hexacyanoferrate(i1) and o-phenylenediamine, and also the poorly water-soluble fer- Peroxidase units CTAB (%) 0 0.2 0.4 2.0 6.0 10.0 w - ] 6.0 9.0 L= ~ k-1 6.0 4.0 ' 2 7.0 t - 4.0 7.0 0 1 2 3 4 Ethyl acetate (%) Ferrocenehnmol dm-3 0 1 2 3 4 5 6 Fig.3 Effect of: A, ethyl acetate; B, CTAB; and C, ferrocene contents; and D, enzyme activity, on the response to 0.5 mmol dm-3 butan-Zone peroxide. [Other conditions as in Fig. 1 (A)] 12.0 QI 9.0 s" 6.0 3.0 L 3 0 0.5 1.0 0 0.5 1 .o Concentration/mmol dm-3 Fig. 4 Effect of mediator on the amperometric response to butan-2- one peroxide (0.1 mmol dm-3 additions). Mediators: A, ferrocene; B, dimethylferrocene; and C, hexacyanoferrate(r1) or D, o-phenylene- diamine. Surfactants: (u) Triton X-100, and (b) CTAB. [Other conditions as in Fig. 3 (A)]ANALYST, SEPTEMBER 1993, VOL. 118 1151 30.0 I 1 B 0 0.5 1 .o Concentration/mmol dm-3 Fig. 5 Dependence of the steady-state current on the concentration of A, benzoyl peroxide; B, butan-Zone peroxide; C, cumene hydroperoxide; D, lauroyl peroxide; E, tert-butyl peroxide; and F, dicumyl peroxide.[Conditions as in Fig. 1 (A)] rocene and dimethylferrocene species. With use of the Triton X-100 based system there is no apparent difference in the sensitivity in the presence of the three different mediators. In contrast , the barley water-soluble ferrocene compounds exhibit better sensitivity in the presence of CTAB. Such behaviour indicates that the ferrocene mediators partition into the bulk water phase containing the enzyme. Note also the high linearity observed for these six experiments. Fig. 5 illustrates the dependence of the steady-state reduction current on the concentration of six different organic peroxides (representing a broad solubility range). Convenient measurement of sub-millimolar concentrations of these perox- ides is feasible.With the exception of benzoyl peroxide, all the other peroxides exhibit linearity over the entire range. The following trend in sensitivity is observed: benzoyl peroxide > butan-Zone peroxide > cumene hydroperoxide > lauroyl peroxide > tert-butyl peroxide = dicumyl peroxide. These concentration-dependent data and the corresponding (recip- rocal) Lineweaver-Burk-type plots were used to estimate the apparent Michaelis-Menten constant ( Km,app). For example, Km,app values of 8.56, 7.14, 6.86 and 3.11 were obtained for butan-Zone peroxide, cumene hydroperoxide , benzoyl per- oxide and dicumyl peroxide, respectively. In conclusion, the above experiments confirm the expecta- tion that microemulsions can serve as suitable media for performing enzymic assays.The major advantage of such an operation is the broad substrate solubility range, in connection with ‘ ~ l a ~ ~ - ~ e l e ~ t i ~ e ’ enzymes. While the concept is presented within the framework of organic peroxides, these observations are likely to stimulate investigations into other - classes of analytes possessing a broad solubility range. The microemul- sion peroxide assays can be readily adapted to an on-line (e.g., flow injection) operation to offer fast monitoring of organic peroxides in relevant environmental or industrial samples. Although olw emulsions were used in the present work, wlo emulsions should also offer analytical advantages, particularly for manipulating the biocatalytic activity to address specific analytical needs.Such an operation should also allow a convenient recovery and re-utilization of the enzyme. 18 This work was supported in part by the US Environmental Protection Agency (Grant No. CR-81 7936-010). Mention of trade names or commercial products does not constitute endorsement or recommendation by the Agency. A. J. R. acknowledges a fellowship from the Spanish Ministry for Science and Education. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Klibanov, A. M., Chemtech., 1986, 16, 354. Saini, S., Hall, G. F., Downs, M. E., and Turner, A. P., Anal. Chim. Acta, 1991, 249, 1. Hall, G., Best, D., and Turner, A. P., Anal. Chim. Acta, 1988, 213, 113. Wang, J., Naser, N., Kwon, H., and Cho, M., Anal. Chim. Acta, 1992, 264, 7. Hall, G., and Turner, A. P., Anal. Lett., 1991, 24, 1375. Wang, J., Reviejo, A., and Mannino, S., Anal. Lett., 1992,25, 1399. Schubert, F., Saini, S., and Turner, A. P., Anal. Chim. Acta, 1992, 245, 133. Wang, J., Lin, Y., and Chen, Q., Electroanalysis, 1993, 5,23. Martinez, K., Klyachko, N., Kabanov, A., Khmelnitsky, Y., and Levashov, A., Biochim. Biophys. Acta, 1989, 981, 161. Luisi, P. L., and Larane, C., Trends Biotechnol., 1986, 4,153. Larsson, K., Ph.D. Thesis, Lund University, 1990. Hedstrom, G., Slotte, J. P., Molander, O., and Rosenholm, J., Biotechnol. Bioeng., 1992, 39, 218. Igarashi, S., and Hinze, W., Anal. Chern., 1988, 60, 446. Abdel-Latif, M., and Guilbault, G., Anal. Chem., 1988, 60, 2671. Glaze, W., Environ. Sci. Technol., 1987,21, 224. Silbert, L., in Organic Peroxides, ed. Swern, D. Wiley-Inter- science, New York, 1971, vol. 11, p. 755. Sarcar, S., Jain, T., and Maitra, A., Biotechnol. Bionerg., 1992, 39, 1992. Larsson, K., Aldercreutz, P., and Mattiasson, B., Biotechnol. Bioeng., 1990, 36, 135. Paper 3100779K Recei-ved February 9,1993 Accepted April 6, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801149

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1153 Determination of Microgram Amounts of Uranium in Thorium by Diff erential-pulse Polarography Sulobh K. Das Fuel Reprocessing Division, Bhabha Atomic Research Centre, Bomba y-400 085, India Achyut V. Kulkarni and Rarnesh G. Dhaneshwar Analytical Chemistry Division, Bhabha Atomic Research Centre, Bomba y-400 085, India A differential-pulse polarographic (DPP) method for the determination of uranium(v1) in the presence of a large excess of thorium without any prior separation of the latter is described. Disodium ethylenediaminetetraacetate and sodium carbonate-hydrogencarbonate buffer (pH 10) was used as the supporting electrolyte. The interference of several impurities expected in the Thorex process while processing irradiated thorium was overcome by using the above supporting electrolyte.A relative standard deviation of 2.4% for the determination of uranium at the 5 ppm level was obtained. The uranium(v1) values obtained by the proposed DPP method were compared with those obtained by inductively coupled plasma atomic emission spectrometry. Keywords: Uranium determination; thorium metal analysis; differential-pulse polarography The irradiation of thorium by neutrons results in the produc- tion of fissile 233U, which can be used as a fuel in future nuclear programmes. This necessitates the development of a method for the determination of small amounts of uranium in the presence of a large excess of thorium. Radiometric1 and spectrophotometric2 methods of determining uranium in thorium exist but they involve a prior separation of uranium from the thorium matrix.In the separation procedure using trioctylphosphine oxide, thorium is also extracted along with uranium and interferes in its subsequent spectrophotometric determination.3 In the inductively coupled plasma atomic emission spectrometric (ICP-AES) method,4 the total dis- solved solids in the sample must be restricted, otherwise easy transportation of the sample to the ICP is not possible. It was therefore necessary to develop an alternative method that is capable of determining uranium in the presence of an excess of thorium without its interference. Uranyl ion (U022+) is one of the few electroactive species that is only weakly complexed by disodium ethylenedi- aminetetraacetate (Na2-EDTA) and related chelating agents.The standard potentials of Uvl-U1v and ThIV-Th0 are well separated from one another.5 This factor, coupled with the strong complexation between EDTA and other metal ions present, prompted Pribil and Blazek6 and Auerbach and Kissel7 to investigate systematically the polarography of uranium in an EDTA-containing medium and to develop a method for the determination of uranium in the presence of several elements whose half-wave potentials are significantly displaced towards more negative values by chelation with EDTA. Uranium(v1) forms a strong complex with carbonate ion.8 The polarographic behaviour of Uvl has been extensively studied in a carbonate medium,9-12 although only Perec" found a linear relationship between concentration and limiting current of Uvl in the narrow concentration range from 0.5 x 10-3 to 4 x 10-3 mol 1-1 in a supporting electrolyte of carbonate-hydrogencarbonate buffer.A differential-pulse polarographic (DPP) method for the determination of uranium in a thorium matrix is reported in this paper. Advantage was taken of the weak complexation between UvL and EDTA and of the strong complexation between thorium and EDTA in an alkaline carbonate- hydrogencarbonate buffer medium to develop a DPP proce- dure for Uvl determination. This approach helped in suppress- ing the hydrolysis of ThLV, at the same time retaining the strong complexation between Uvl and carbonate ions. Several impurities ( A P , Ca", Crvl, Cu", Mg", Mn", Zn", Fell1, etc.) might be present when irradiated thorium fuel rods are dissolved in nitric acid in the Thorex process for fuel reprocessing.These impurities, if not complexed with EDTA, interfere in the DPP determination of Uvl and the supporting electrolyte chosen removes their interference together with that of ThIV. Experimental Differential-pulse polarographic experiments were performed on a Model 174A polarographic analyser in conjunction with a Model 303A static mercury drop electrode (SMDE) and a Model RE 0089 x-y recorder, all from EG & G Princetm Applied Research (Princeton, NJ, USA). pH measuremf nts were made on a Model 330A digital pH meter manufactired by EMCO (Bombay, India). A linear-sweep ramp (scan rate 5 mV s-1) superim?osed with pulses of 50 mV amplitude was applied to the FMDE (small drop size) in the mode of a hanging mercurj drop electrode (HMDE).A saturated Ag-AgC1 electrode w as used as the reference electrode and platinum wire as the alxiliary electrode. Reagents Stock solutions of uranium and thorium were prepared by dissolving their respective nitrates of Specpure grade (John- son Matthey, London, UK) in 0.1 moll-' nitric acid. All acids were of Aristar grade (BDH, Poole, Dorset, UK). All other salts were of general-reagent grade (Merck, Darmstadt, Germany). High-purity nitrogen gas of IOLAR grade (Indian Oxygen, Bombay, India) was used for removing dissolved oxygen from the polarographic solution. Uranium solutions were standardized following a modification of the method of Davies and Gray.13 Procedure A suitable aliquot of synthetic sample solution containing thorium (10-50 mg) and uranium (10-50 pg) together with other impurities, mainly Fe"' in the range 20-100 pg and Al'II, Call, Crvl, Cu", Mg", Mn", MoV1, Ni", Pb" and ZnL1, etc., each in the range 1-5 pg, was placed in a 10 ml calibrated flask.About 0.5 g of Na2-EDTA (solid) was added to the solution and warmed to dissolve it. The amount of EDTA was1154 ANALYST, SEPTEMBER 1993, VOL. 118 EP versus pH ,-. ,-. " h v e " -0.6 -0.7 -0.8 -0.9 -1.0 -1.1 -1.2 Potential N versus Ag-Ag CI Fig. 1 Differential-pulse polarograms of Uv' in 0.1 mol 1-l sodium carbonate-hydrogencarbonate buffer (pH 10) containing Na2-EDTA. A, Blank; and B, 10 ppm of Uvi 0.01 1 y 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Concentration of UvVmmol I-' Fig. 2 Differential-pulse polarogram of peak current versus concen- tration of Uvl in the presence of 0.1 mol 1-1 thorium sufficiently in excess over the stoichiometric amount required by all the metal ions (thorium and other impurities) present.The pH of the solution was raised to the alkaline region by adding NaOH solution. A 1.0 ml volume of 1.0 mol 1-1 carbonate-1 .O moll-' hydrogencarbonate mixture was added to the above solution and the pH was adjusted to 10. The volume was made up to 10 ml with distilled water and the solution was subjected to polarography . Results and Discussion Fig. 1 shows a typical differential-pulse polarogram of Uv' in 0.1 rnol 1-1 sodium carbonate-hydrogencarbonate buffer solution (pH 10) containing Na2-EDTA. The effects of changes in pH, EDTA concentration and carbonate buffer concentration were studied to establish the optimum experimental conditions for Uvl reduction.The linear rela- tionship between Uvl concentration and DPP peak current is shown in Fig. 2. It is known that the UV1-carbonate complex is very strong, the equilibrium constant being 2 x 1018 for the complex [U02(C03)3]4-.* The choice of a suitable pH for Uvr reduction was dictated by the presence of other metal ions (including Th) that are hydrolysed in the alkaline medium. The EDTA complexes of ThIV and other cationic impurities present in the solution after dissolution of the thorium fuel rods in nitric acid are stable in alkaline solutions. Hence the addition of Na2-EDTA and the use of an alkaline pH not only prevents the hydrolysis of the bulk ThIv matrix and most of the divalent and trivalent impurities, but also shifts their reduction potentials well beyond the hydrogen discharge potential (- 1.4 V versus Ag-AgC1).This effectively eliminates the interfer- ences in the determination of Uvl. ip versus pH 0 0 r\ n I- 0 " I I I I I I 0.6 8 9 10 11 12 PH Fig. 3 Effect of pH on peak potential and peak current of 0.1 mmol 1-1 of Uvi. Suporting electrolyte: 0.1 mol 1-1 of sodium carbonate-hydrogencarbonate buffer (pH 10) containing Na2-EDTA (10 mmol 1-1) (pH adjusted with dilute NaOH solution) t +- 2 3 0 Blank c/-" U=4ppm 0.05 FA I I I I I 1 I I I I -0.6 -0.7 -0.8 -0.9 -1.0 -0.6 -0.7 -0.8 -0.9 - 1.0 Potent ia IN versus Ag-Ag CI Fig. 4 Differential-pulse polarograms of U"' in the absrnce and presence of thorium in 0.1 mol 1-1 sodium carbonate-hydr 3gencar- bonate buffer (pH 10) containing 0.5 g of Na2-EDTA (solid) n 10 ml.(a) Thorium absent. (b) Thorium present (5 g 1-1) Plots of peak potential (Ep) and peak current (ip) for Uvl reduction versus pH reveal that both the potential and current are nearly constant over the pH range 8-12 (Fig. 3). This is in agreement with the observations of Jung et aZ.14 As E, and i, were both constant in the pH range 8-12, a pH of 10 was chosen for further studies. Fig. 4 shows differential-pulse polarograms of Uvl in the absence and presence of thorium. It is clear that the polarograms are not affected to any significant extent by the presence of large amounts of thorium, except for an alteration in the shapes of the differential-pulse peaks and the nature of the blank polarograms.Analysis of Thorium Samples for Uranium Content Uranium(v1) in thorium was determined by the proposed method and the values obtained were compared with those obtained by the ICP-AES method as developed in our laboratory. 15 The sample solutions provided for the ICP-AES analysis were diluted so as to give about 1 g 1-1 of thorium, whereas the polarographic solution contained as much as 5 g 1-1 of thorium. The DPP procedure yielded uranium values with a relative standard deviation of 2.4% at the 5 ppm level. Synthetic samples containing thorium and uranium in ratios from 50 to 1000 along with other impurities were analysed by the DPP and ICP-AES methods. The results are given in Table 1.ANALYST, SEPTEMBER 1993, VOL. 118 1155 Table 1 Analysis of uranium samples by ICP-AES and DPP.Samples containing thorium and other impurities. Supporting electrolyte: Na2C03 4- NaHC03 buffer containing 0.5 g of Na2-EDTA (solid) at pH 10. Thorium content: 1 g 1-1 U concentration/pg ml-1 Sample ICP- AES DPP A B C D 1.3 5.8 10.3 19.0 1.5 5.6 10.5 18.0 It should be noted that the spectrophotometric procedure involves a prior extraction step'-3 whereas the present DPP method is capable of determining uranium in solutions after dilution but without involving any prior extraction step. The thorium content in the present method does not significantly influence the differential-pulse polarograms of Uv*, as can be seen from Fig. 4, because the ThLV-EDTA complex itself serves as the inert supporting electrolyte. It is concluded that DPP can be applied to the rapid and precise determination of low levels of uranium in thorium without the need for a prior extraction or separation step.The authors thank Sri A. N. Prasad, Director, Fuel Repro- cessing and Nuclear Waste Management Group, BARC, Sri M. K. Rao, Associate Director, FR and NWM Group, and Dr. R. K. Dhumwad, Head, Laboratory Section, Fuel Reprocessing Division , BARC, for their encouragement in this work. They also thank Dr. S. Gangadharan, Head, Analytical Chemistry Division, BARC, for his interest in this work. References 1 Clayton, R. F., Hardwick, W. H., Moreton-Smith, M. and Todd, R., Analyst, 1958,83, 13. 2 Harton, C. A., and White, H. C., Anal. Chem., 1958,30,1779. 3 Mukherjee, A., and Rege, S.G., in Proceedings of Radio- chemistry and Radiation Chemistry Symposium, University of Poona, Board of Research in Nuclear Sciences, Department of Atomic Energy, Bombay, 1982, paper RA-7. 4 Thompson, M., and Walsh, J. N., A Handbook of Inductively Coupled Plasma Spectrometry, Blackie, Glasgow, 1983, p. 20. 5 Latimer, W. M., The Oxidation States of the Elements and Their Potentials in Aqueous Solutions, Prentice Hall, Englewood Cliffs, NJ, 2nd edn., 1956, p. 299-301. 6 Pribil, R., and Blazek, A., Collect. Czech. Chem. Commun., 1953, 16, 567. 7 Auerbach, C., and Kissel, G., Talanta, 1964, 11, 85. 8 Katz, J. J., and Seaborg, G. T., The Chemistry of Actinide Elements, Methuen, London, 1957, p. 190. 9 Harris, W. E., and Kolthoff, I. M., J. Am. Chem. SOC., 1947, 69,446. 10 Stabrovskii, A. I., Zh. Neorg. Khim., 1960, 5, 811. 11 Perec, M., Nukleonika, 1961, 6, 357. 12 Branica, M., and Pravdic, V. , in Polarography (1964) (Proceed- ings of the 3rd International Congress of Polarography, South- ampton, 1964), ed. Hills, G. J . , Macmillan, London, 1966, vol. 13 Chitnis, R. T., Kulkarni, R. T., Rege, S. G., and Mukherjee, A., J. Radioanal. Chem., 1978, 45, 331. 14 Jung, K.-S., Sohn, S. C., Ha, Y. K., Eom, T. Y. and Yun, S. S., J . Electroanal. Chem., 1991, 315, 113. 15 Dhumwad, R. K., Patwardhan, A. B., Joshi, M. V., Kulkarni, V. T., and Radhakrishnan, K., in Proceedings of the Radio- chemistry and Radiation Chemistry Symposium, Kalpakkam, India, 1989, paper RA-17. I, p. 435. Paper 3/01 051 A Received February 22, 1993 - Accepted April 21, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801153

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1157 Electrochemical Reduction at Mercury Electrodes and Differential-pulse Polarographic Determination of Pentamidine lsethionate M. Valnice B. Zanoni and Arnold G. Fogg* Ch em is tr y Department, Lo ugh bo ro ug h University of Tech no logy, Lo ugh bo ro ug h, L eiceste rsh ire, U K LEI1 3TU Reduction processes are observed for pentamidine isethionate at a dropping mercury electrode above pH 7: the reduction potential is independent of pH below about pH 10. Below pH 10, adsorption of the reduced species is observed, whereas above pH 10 there is a large contribution owing to adsorption of pentamidine isethionate. For the polarographic determination of pentamidine isethionate, a pH of 8-9 is recommended with the addition of Triton X-100 as a maximum suppressor.Pentamidine isethionate can be determined by differential-pulse polarography down to about 5 x 10-6 mol 1-1. Keywords: Pentamidine isethionate determination; differential-pulse polarograph y Amidines are important medical and biochemical agents and can be identified in numerous cyclic derivatives of biological interest. Many of the compounds of this type have been examined as potential antibacterial and antiprotozoal drugs in humans and domestic animals.' Most of the active compounds possess two benzamidine residues, which are separated by a structural unit containing one or more atoms. A striking feature is that all of these compounds contain an unsubstituted amidine group. Pentamidine isethionate [4,4'-(pentamethy1enedioxy)di- benzamidine bis(2-hydroxyethanesulfonate)] belongs to this class of compound, and is used in the treatment of patients with pneumocystis carinii pneumonia.2 This is a parasitic infection that attacks patients with severe immunosuppres- sion.Since 1981, the occurrence of acquired immunodefi- ciency syndrome (AIDS) has been accompanied by a dramatic increase in the incidence of pneumocystis carinii pneumonia. A corresponding increase in the use of pentamidine isethio- nate has highlighted the need to develop rapid and reliable analytical methods for evaluating the quality of the drug and with sufficient sensitivity to quantify it at low levels in human patients. Several analytical methods have been employed for these purposes, viz . , thin-layer chromatography, liquid chromato- graphy, ultraviolet (UV) spectrophotometry , mass spec- trometry and infrared spectroscopy.3.4 A spectrofluorimetric method for determining pentamidine isethionate in plasma, urine and tissue,5 and a high-performance liquid chromato- graphic method for its determination in blood serum6 and whole-blood plasma and urine7-9 have also been published.Although some amidines and related compounds can be reduced or oxidized at electrodes,lJo only in recent years has increasing attention been paid to electrochemical studies of cyclic amidines, particularly the biologically active amidines. However, few electrochemical studies involving benzamidines in aqueous solution are found in the literature.llJ2 The reduction of substituted benzamidines is postulated to occur in two diffusion-controlled two-electron waves,12 in contrast to earlier results for unsubstituted benzamidine,ll which is reduced in a single four-electron step.The over-all process is usually pH-dependent and involves the saturation of double bonds and a cleavage of C-N bonds. The aim of the present work was to investigate the possibility of using polarography for the determination of pentamidine isethionate in aqueous solution. * To whom correspondence should be addressed. Experimental Apparatus For differential-pulse polarographic and cyclic voltammetric measurements a Metrohm E612 VA scanner, an E611 VA-detector and a 663 VA stand were used in conjunction with a Houston Instruments Model 2000 x-y recorder. The multimode electrode in the VA stand was used in both the dropping mercury electrode and the hanging mercury drop electrode modes.The three-electrode system was completed by means of a glassy carbon auxiliary electrode and an' Ag-Age1 reference electrode. All potentials are given relative to this Ag-Age1 (3 mol 1-1 KCl) electrode. The d.c. and a.c. polarographic studies were carried out using a Metrohm E506 Polarecord with the 663 VA stand. All pH measurements were made with a Corning combined pH reference electrode using a Radiometer PHM 64 pH meter, previously calibrated. Reagents All chemicals were of analytical-reagent grade. Water was obtained from a LiquiPure system. Pentamidine isethionate was supplied by Fisons Phar- maceuticals. Stock standard solutions of the drug (1.0 x 10-3 mol 1-1) were prepared from the dried pure substance.All stock solutions were freshly prepared weekly and more dilute solutions were prepared as required from the stock solutions. The supporting electrolyte was Britton-Robinson (B-R) buffer prepared in the usual way, i.e., by adding to a solution 0.04 moll-1 orthophosphoric acid, 0.04 moll-1 in acetic acid and 0.04 moll-' in boric acid with the appropriate amount of 0.2 moll-1 sodium hydroxide solution. Procedures The general procedure adopted for obtaining polarograms was as follows. An aliquot (10 ml) of B-R buffer or sodium hydroxide solution was placed in a clean, dry voltammetric cell and the required volume of standard pentamidine isethionate was added by means of a micropipette. The solution was purged for 15 min and the polarogram was recorded.The differential-pulse mode was used with a pulse ampli- tude of 50 mV, a drop time (id) of 1 s and a scan rate of 3 mV s-1, unless stated otherwise. The phase-selective a x . polarographic behaviour was studied using a superimposed alternating voltage (U) of 8 mV and a phase angle of 0 or 90".1158 ANALYST, SEPTEMBER 1993, VOL. 118 Results and Discussion D.c. and Differential-pulse Polarography Pentamidine isethionate contains reducible amidine groups that provide the basis for its polarographic determination. In B-R buffer solutions, pentamidine isethionate is readily reduced in the pH range 7.0-12.0: the half-wave potentials (Eh), wave heights, shapes and the number of waves, depend on the pH, the drop time and the pentamidine isethionate concentration.In acidic solution (i.e., pH <6.0), the reduc- tion wave is masked by the electrolyte discharge. Effect of pH Differential-pulse polarograms of 0.5 x 10-5 moll-1 pentami- dine isethionate solutions, recorded at a drop time of 2 s and a pulse amplitude of 50 mV, exhibit one cathodic peak in the pH range 7.0-9.0. This peak becomes broader as the pH increases and becomes a double peak in solutions of pH 12.0, as is shown in Fig. 1. Plots of peak potential versus pH show two distinct regions (see Fig. 2): over the pH range 7.0-10.0 the peak potential is independent of pH, but above pH 10 it shifts linearly towards more negative values (slope of 38 mV per pH unit). The effect of the drop time on this peak, also shown in Fig. 2, is discussed below.The height of the peak decreases continuously above pH 9, being only about one-third of its original height at pH 12 (see Fig. 3). The heights of the d.c. wave and differential-pulse peaks within the pH range studied remained constant at room temperature for a period of at least 12 h after preparation of the solution for polarography. -1.60 -1.60 -1.61 -1.62 -1.64 -1.60 -1.70 E d v Fig. 1 Differential-pulse polarograms of pentamidine isethionate in B-R buffers. Pentamidine isethionate concentration = 0.5 X 10-5 moll-'. pH: A, 7.0; B, 8.0; C, 9.0; D, 10.0; E, 11.0; and F, 12.0. Drop time = 2 s, pulse amplitude = 50 mV, scan rate = 3 mV s-l 1.70 . c m I 7 8 9 10 11 12 PH Fig. 2 Effect of pH on the peak potentials of the differential-pulse polarographic peaks obtained for a 0.5 x 10-5 mol 1-1 solution of pentamidine isethionate in B-R buffer.A, Potential of single peak obtained with a drop time of 2 s; B, potential of second (main) peak (dro time = 1 s); and C, potential of first (pre-) peak (drop time = 1 s? Analysis of the shapes of the differential-pulse peaks indicates a degree of irreversibility, with peak half-width values of about 78 mV at pH 7-9, and even higher half-width values at higher pH values (82.7 and 102.4 mV at pH 10.0 and 11.0, respectively). The irreversibility of the system was confirmed by applying the criterion of Birke et aZ.,13 when comparing forward and reverse differential-pulse scans. Dif- ferential-pulse polarograms obtained for pentamidine ise- thionate at a pulse amplitude of 50 mV (cathodic peak, forward scan) and 50 mV (anodic peak, reversescan) yielded a ratio of the heights of the cathodic (iPc) and anodic (ipa) peaks of about 0.35 (pH 7.0-9.0) and a value of Epc - Epa (Epc = cathodic peak potential; E,, = anodic peak potential) (about 39 mV) of less than the pulse amplitude.In strongly alkaline solutions, values of ipa/ipc and Epc - Epa became slightly larger. However, although two peaks are observed using a pulse amplitude of 50 mV at pH 12.0, only one peak is observed in the reverse scan (pulse amplitude, 50 mV). An interesting effect was observed when differential-pulse polarograms were recorded using a drop time of 1 s, in that it was possible to distinguish two peaks between pH 7.0 and 12.0. Whereas the second (main) peak shows the same behaviour as that observed with a drop time of 2 s (see Fig.2), the peak potential of the pre-peak (at the less negative potential) is independent of pH over the pH range 7-12, and its peak height (see Fig. 3) is slightly smaller than that observed for the second peak. However, the heights of both peaks are affected equally by variation of pH. These results could indicate either that the electrode process occurs in two cathodic steps having very close reduction potentials, or that the reduction is associated with a strong adsorption process. In order to distinguish between these two alternatives the reduction of pentamidine isethionate was studied further by d.c. and a.c. polarography and by cyclic voltammetry. Typical polarograms and voltammograms in B-R buffer of pH 9.0, which are representative of the pH range 7.0-9.0, are shown in Fig.4. The d.c. polarographic parameters (Et; il; Ei - Ei and an,) (an, = electron transfer coefficient) were obtained from polarograms recorded with 0.5 X 10-4 moll-1 pentamidine isethionate with a drop time of 0.4 s (see Table 1): these experimental conditions favoured the occurrence of a single well-defined wave over the pH range studied. As indicated above, in the pH range 7.0-9.0, the half-wave potential was almost independent of pH and the limiting current was virtually constant. Logarithmic analysis of the waves resulted in straight lines, and, by applying the treatment of Meites and Israel,14 the an, values point to an irreversible reduction process with an, = 1.38. At higher pH values the 17 \.\.', '. ' 100.0 50.0 '\ 3 150.0 \ \ t 1 0.0 7 8 9 10 11 12 PH Fig. 3 Effect of pH on the peak currents of the differential-pulse polarographic peaks obtained for a 0.5 x mol 1-I solution of pentamidine isethionate in B-R buffer. A, Peak obtained at a drop time of 2 s; B, second (main) peak obtained at a drop time of 1 s; and C, first (pre-) peak obtained at a drop time of 1 sANALYST, SEPTEMBER 1993, VOL. 118 1159 1-1-40 ~~ c------- /@ I I I I I B - 1.40 - 1.50 -1.60 -1.70 E N Fig. 4 Response of pentamidine isethionate (0.5 X 10-4 moll-l) in B-R buffer (pH 9.0); drop time = 1 s: A, d.c. polarography; B, a.c. polarography, phase angle 0" and U = 8 mV; C, a.c. polarography, phase angle 90" and U = 8 mV; and D, cyclic voltammogram of pentamidine isethionate at a concentration of 5.0 X 10-4 moll-' at a scan rate of 20 mV s-1 Table 1 Effect of pH on d.c.polarograms of 0.5 X lov4 mol 1-1 pentamidine isethionate in B-R buffer (td = 1 s), and a.c. polaro- graphic waves of 1 x 10-4 mol 1-l pentamidine isethionate obtained with phase angles of 90 and 0" and a superimposed potential of 8 mV 0" 90" -Ed ill id -Ed id -Epl pH V mA an, mA V mA V 7.0 1.610 0.86 1.38 0.27 1.64 0.32 1.52 8.0 1.608 0.73 1.38 0.20 1.63 0.37 1.51 9.0 1.615 0.73 1.38 0.28 1.64 0.33 1.51 10.0 1.620 0.77 1.17 0.16 1.64 0.75 1.51 11.0 1.644 0.39 0.81 0.07 1.68 0.81 1.52 12.0 1.690 0.37 0.65 - - 1.08 1.53 half-wave potentials are shifted linearly to more negative values (see Table l), and the slope of the relationship Et versus pH can be expressed as E4 = -1.266 - 0.039 pH.As indicated in Table 1, the change in the character of the reduction process observed above pH 10.0 is accompanied by an increase in the degree of irreversibility, and the waves diminish in height. Hence, it can be suggested that at pH values >10.0, the reduction of pentamidine isethionate probably involves preceding chemical reactions with participa- tion of protons and/or strong adsorption effects. Effect of Pentamidine Isethionate Concentration, Pulse Amplitude and Drop Time In order to obtain further information on the nature of the electrode reaction, the influence of concentration and drop time was investigated at two pH values, viz., 9.0 and 12.0, which are representative of the polarographic behaviour in the pH ranges 7.0-9.0 and 10.0-12.0, respectively. It was ob- served that the drop time directly affected the polarographic response at drop times from 0.4 to 3.0 s.Differential-pulse polarograms obtained at pH 9.0 show two cathodic peaks when shorter drop times are used [Fig. 5(a)]. In the same way, when the concentration is decreased and adsorption effects are minimized, the polarograms change from one with two peaks to one with a single peak [Fig. 6(a)]. Also, the differential-pulse polarographic response is influ- enced by pulse amplitude. For the same drop time (Id = 1 s) there is a coalition of the two peaks when the pulse amplitude is decreased and only one peak can be observed at more negative potentials for pulse amplitudes less than -40 mV. -1.30 -1.50 -1.70 -1.30 -1.50 -1.70 E N Fig.5 Influence of the drop time of the DME on the differential- pulse polarographic reduction of 1 x mol 1-1 pentamidine isethionate in (a) B-R buffer (pH 9.0) and (6) B-R buffer (pH 12.0) at drop times of: A, 0.4; B, 0.6; C, 1.0; and D, 2.0 s -1.70 -1.30 -1.50 -1.70 - 1.30 - 1.50 EN Fig. 6 (a) Differential- ulse polarograms of pentamidine isethionate in B-R buffer (pH = 9.07 at different concentrations: I, 0.6 X 2, 1.2 x 10-6; 3,2.0 x 4,2.6 x 10-6; 5,3 x 6,4 x and 7, 5 x 10-6 mol 1-1. (b) Differential-pulse polarograms obtained in B-R buffer (pH 12.0): A, 0.5 x 10-5; B, 1.0 x 10-5; C, 2.0 X lo-? and D, 3.0 x mol 1-l pentamidine isethionate A comparison was made between a differential-pulse polarogram with a negative-going potential scan and a differential-pulse polarogram recorded with a positive-going potential scan.Although two cathodic peaks are observed on the forward scan for shorter drop times (td = 0.4 s), only one anodic peak is obtained on the reverse scan. In contrast, although one cathodic peak is observed on the forward scan (td = 2 s), this is split into two peaks at higher pulse amplitudes. Hence the behaviour is not that reported in the literature for an electrode process controlled either kinetically or by diffusion.15-17 These results can be explained by the occur- rence of a strong adsorption phenomenon coupled with electron transfer. l 8 ~ 9 Double peaks can occur in differential- pulse polarograms when saturation of the electrode occurs before the application of the pulse. Hence at longer pulse amplitudes and higher concentrations, the main diffusion- controlled peak appears next to the adsorption peak.In1160 ANALYST, SEPTEMBER 1993, VOL. 118 contrast, if the pulse application and sampling times are kept constant, two peaks are observed at shorter drop times, when the surface area is such that it promotes the situation where the sampling time is shorter than the time required to achieve maximum coverage. Direct current polarograms cohfirm this diagnosis. Studies in the concentration range from 0.2 x 10-4 to 2.0 x 10-4 mol 1-1 pentamidine isethionate, at pH 9.0, lead to the appearance of a new wave at a less negative potential when the concentration is increased to 0.7 X 10-4 mol 1-1 (see Fig. 7, curve B). When the concentration was increased further, the height of the main wave (i) increased steadily and the height of the wave at the less negative potential (i') remained virtually constant.This indicates that the electrode reaction is con- trolled by adsorption of the reduced form of the compound.20 Polarograms obtained in the same concentration range but using a drop time of 0.4 s show a single wave (Fig. 7, curve A), for which the peak height increases linearly with an increase in the pentamidine isethionate concentration in the same con- centration range discussed above. At pH 9.0, d.c. polarograms obtained at lower concentra- tions (0.2 x 10-4 mol 1-1) of pentamidine isethionate exhibited a diffusion-limited current, as the limiting current varied in a direct proportion to tt. However, at 1.0 x 10-4 mol 1-1, the wave obtained at a drop time >l.O s splits into two waves (see Fig.7, curve 11), for which the total limiting current decreases linearly with t-+, indicating that the current is adsorption-controlled.20 However, at pH 12.0, two peaks are observed in the pulse polarograms at any drop time or concentration, as is shown in Figs. 5(b) and 6(b). The linearity between the peak current (ip) and t", where t, the drop time, varies between 0.4 and 3.0 s, suggests that the first electrode process is diffusion-controlled. The same conclusion was reached from the fact that the peak height increased linearly with concentration in the range from 1.0 x 10-5 to 7.0 x 10-5 moll-'. The second peak is broader than the first peak and the peak current showed deviations from linearity at high concentrations, which clearly indicates the presence of a strong adsorption phenomenon.Direct current polarograms at pH 12.0 exhibit polaro- graphic maxima (Fig. 8, curve A), which become more distorted at high concentrations. This behaviour is typical in the presence of an adsorption process, and the polarographic maximum hindered accurate current evaluation. In an attempt to finalize the conclusions, the effect of the addition of Triton X-100 on the differential-pulse polarograms was studied for conditions where two peaks are observed. Polarograms recorded at pH 8.50 and at a drop time of 1.0 s in the presence of Triton X-100 show only one well-defined peak at all concentrations [Fig. 9(a)]. It is probable that the effect of the surface-active agent is to contribute to complete coverage of the electrode surface and to promote the suppression of the ~~ B iBT0.25 mA ......................... A.-EN Fig. 7 (11) D.c. polarograms of 1 X mol 1-l pentamidine isethionate measured in B-R buffer (pH 9.0) at a drop time of: A, 0.4 and B, 2.0 s. (I) D.c. polarograms obtained at pH 9.0 (drop time 0.4 s) in A, 0.5 x and B, 2.0 X moll-' pentamidine isethionate adsorption process coupled with the reduction of the pentami- dine isethionate. In contrast, differential-pulse polarograms obtained at pH 12.00 in the presence of Triton X-100 exhibit two peaks. The first peak is slightly larger and the second peak is smaller and I I 3 / \ \--I---.# --0-.- I -____--- --- 1 I -1.40 -1.50 -1.60 -1.70 -1.80 -1.90 EN Fig.8 Response of pentamidine isethionate (1.0 x moll-1) in B-R buffer (pH 12.0) (drop time 1 s): A, d.c. polarography; B and C, first and second cyclic voltammograms at 5 x mol 1-1 pentamidine isethionate at a scan rate of 20 mV s-l; and D, a.c. polarography, phase angle 90" and U = 8 mV at a pentamidine isethionate concentration of 1,0.5 x 2,l.O x 10-4; and 3,1.5 x 10-4 moll-1 500.0 400.0 300.0 P .> 200.0 100.0 0 EN - 1.30 b) /'o / /I /d/ // dl / // d d /@I / d / I I Concentration/lO-6 rnol I-' 4.0 8.0 12.0 Fig. 9 (a) Differential-pulse polarograms of pentamidine isethionate in B-R buffer (pH 8.50) (drop time 1 s) at different concentrations: A, 0.2 x 10-6; B, 0.9 x 10-6; C, 1.9 x 10-6; D, 2.9 x 10-6; and E, 3.9 x mol 1-l. (b) Calibration graph obtained at pH 8.50 in the presence of 0.001% Triton X-100ANALYST, SEPTEMBER 1993, VOL.118 I NH2 pH <9.0 4e- 6H + narrower than in the absence of Triton X-100. It is evident that the electrode process in very alkaline solution is coupled with stronger adsorption than that at pH values <9.0. 1 1 NH2 NH2 NH2 pH >10.0 4e- 14*+ NH=CH- Ph- O-(CH2)5-O- Ph-CH =NH + 2NH3 1161 A.c Polarography and Cyclic Voltainmetry The phase-selective a.c. polarographic behaviour of pentami- dine isethionate was studied in the pH range 7.0-9.0 for phase angles of 90 and 0". Although at lower concentrations only one peak can be observed, a smaller peak is coupled with the main a.c. peak when the polarograms are recorded at 0" at the higher concentration of 1.0 X 10-4 moll-1 (Fig.4, curve B). The d.c. polarogram is sho I for comparison. The main peak is much smaller than the corresponding d.c. limiting current at all pH values, as is shown in Table 1. The peak potential ( E p ) is distinctly more negative than Et and the waves are very broad, emphasizing the presence of irreversibility and/or an adsorption phenomenon .20 However, this small peak becomes very well-defined at about -1.52 V, when the polarograms are recorded at 90", which confirms that an adsorption phenom- enon is coupled with the electron-transfer reaction. The adsorption of the product of electron transfer was charac- terized by the appearance of a pre-peak in the cyclic voltammogram recorded at a low scan rate, as is shown in Fig. 4 (curve D). A comparison of the peak heights obtained at 0 and 90" at different pH values is shown in Table 1.Analysis shows that the capacitive current increases markedly above pH 10.0. Concomitantly, the faradaic peak decreases in size, and it is evident that the electrode process is more significantly influenced by adsorption in very alkaline solutions. At pH 12.0, the a.c. polarograms recorded at 0" do not show appreciable faradaic current, but at 90" one well-defined capacitive peak at - 1.54 V can be observed, and this increases with concentration, as is shown in Fig. 8. This behaviour suggests that adsorption of the reactant might also be contributing to the process. This is confirmed by cyclic voltammograms obtained at pH 12.0, which show pre- and post-peaks in relation to the main peak at -1.60 V (see Fig.8, curve C). In addition, multi-scan cyclic voltammograms show the absence of current on the second scan (Fig. 8), indicating that the reduction mechanism could involve spontaneous adsorption of the drug in this pH range. Over-all Characteristics of the Reduction of Pentamidine Isethionate The foregoing experiments show that adsorption phenomena strongly affect the reduction mechanism of pentamidine isethionate. The presence of a pre-wave under the determined conditions demonstrates that the reduction of the species is more strongly affected by adsorption of the electrode product at pH values <9.0. In contrast, although product adsorption cannot be neglected, in strongly alkaline solutions the preponderant form of pentamidine isethionate probably adsorbs spontaneously on mercury and provokes maxima and other distortions in the wave.Under exceptional experimental conditions, when adsorp- tion complications were minimized, it was possible to observe a break in the Eh versus pH and E, versus pH relationships at about pH 10.0 and a clear diminution of the peak height. This could indicate the presence of an acid-base equilibrium with at pK, of about 10.0 or possibly the hydrolysis of the amidine group. Because hydrolysis reactions of amidine compounds gener- ally involve bond-cleavage,21 some quantitative polarographic determinations were carried out to check the stability to hydrolysis of the species which predominate in pentamidine isethionate solutions at pH 8.0 and 12.0. Pentamidine isethionate was found to degrade considerably in 1 mol 1-1 sodium hydroxide solution.However, by polarographing the 320 280 240 Wavelengthlnm Fig. 10 UV spectra of 1 X mol 1-1 pentamidine isethionate at A, pH 8.0; B, in 0.20 moll-' sodium hydroxide; C, at pH 8.0 restored; and D, in 1 mol 1-1 sodium hydroxide same solution at pH 8.0, then at pH 12.0, and then again at pH 8.0, an 85% recovery of the pentamidine isethionate was evident in the final polarograms at pH 8.0. This relative stability to hydrolysis at pH 12 was confirmed by UV spectrophotometry, as illustrated by the spectra in Fig. 10. The absorption spectrum of pentamidine isethionate (I,,, = 260 nm) can be restored fully after addition of acid to a pentamidine isethionate solution in 0.2 mol 1-1 sodium hydroxide (in which I,,, = 245 nm).Hence the preponderant form of pentamidine isethionate in strongly alkaline solution appears to be the species containing the unprotonated amidine group. This would be in agreement with the chemistry of unsubstituted benzamidines, for which the dissociation equi- libria are characterized by protonation of the imino nitrogen, and for which the pK, values22 are about 11.2. Hence, with the knowledge that amidine groups can be reduced in aqueous solutions in a two-electron process,lO it seems probable that four electrons would be involved in the reduction of pentamidine isethionate , which has two amidine groups. This is consistent with the value of an, = 1.38 found by d.c. polarography (a = 0.35 for n = 4) at pH 4 0 . Therefore, as with other aromatic amidine compounds,11$*2 the over-all course of the reduction of pentamidine isethionate probably involves reductive cleavage of the amidine group, which can be represented as follows: Determination of Pentamidine Isethionate A pH of about 8.0-9.0 was chosen as optimum for analytical purposes, because it yielded the highest peak currents and gave the best separation from the background electrolyte reduction.The limiting diffusion currents obtained from d.c. polarograms were found to be a linear function of the pentamidine isethionate concentration only when shorter drop times, e.g., 0.4 s, were utilized. Satisfactory calibration graphs were obtained over the concentration range from 0.5 x 10-4 to 2.0 x 10-4 moll-'. Differential-pulse polarography permitted quantitative measurements to be made at lower concentration ranges, the limit of determination for pentamidine isethionate being1162 ANALYST, SEPTEMBER 1993.VOL. 118 300.0 m 0 3.0 6.0 9.0 Concentration/lO-6 mol I-’ Fig. 11 Effect of the pentamidine isethionate concentration on calibration graphs obtained at pH 8.50 using different drop times: A, 1 s (second peak); B, 2 s about 5 X 10-7 mol 1-1. Differential-pulse polarographic calibration graphs could be obtained over the concentration range from 0.5 x 10-6 to 10 x 10-6 moll-1 at pH 8.5 for the second peak when a drop time of 1 s was utilized (see Fig. 11, curve A). The slope obtained, 2.25 x 107 nA mol-1, and the correlation coefficient , 0.9977, indicate excellent linearity. The height of the first peak shows a slight deviation from linearity for concentrations up to 7 x 10-6 moll-’. At a drop time of 2 s, where only one peak is observed, a linear relationship between peak height and the concentration of pentamidine isethionate can be obtained only over the range from 0.6 x 10-6 to 5.0 x 10-6 moll-1 (see Fig. 11, curve B).At higher concentrations, adsorption effects promote devia- tion of the calibration graphs. Although two peaks are observed at pH 8.50 and at a drop time of 1 s, as discussed earlier, differential-pulse polarograms at different concentrations of pentamidine isethionate in the presence of 0.001% Triton X-100 exhibit only one well- defined and reproducible peak, which is excellent for ana- lytical purposes. There is a linear dependence with increase in concentration above about 1 x 10-6 mol 1-1 pentamidine isethionate, as is shown in Fig.9(b). The uniform enhance- ment of peak height above this concentration in the presence of Triton X-100 is apparent, and hence if a linear calibration graph is constructed it passes through the current axis above the origin. The results obtained indicate that polarographic techniques are suitable for the determination of pentamidine isethionate. Preliminary results have shown that improvements in sensitiv- ity can be achieved by using adsorptive stripping voltammetry in strongly alkaline solution, and that this technique can be used to some extent directly on urine samples. This work will be reported later. The authors thank Fisons Pharmaceuticals plc for providing samples, and for their interest in this project.M. V. B. Z. thanks the FAPESP (Brazil) for financial support, and the University of Araraquara, Brazil, for leave of absence. 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 References Grout, R. J., in The Chemistry of Amidines and Imidates, ed. Patai, S . , Wiley, Bristol, 1975, vol. 1, pp. 255-277. Goal, K. L., and Campoli-Richards, D. M., Drugs, 1987, 33, 242. Martindale: The Extra Pharmacopoeia, ed. Reynolds, J. E. F., The Pharmaceutical Press, London, 29th edn., 1989, pp. 675-677. Mount, D. L., Miles, J. W., and Churchill, F. C., J. Assoc. Off. Anal. Chem., 1986,69, 624. Waalkes, T. P., and De Vita, V. T., J. Lab. Clin. Med., 1970, 75, 873. Dickinson, C. M., Naviland, T. R., and Churchill, C., J. Chromatogr., 1985,345,91. Lin, J. M. H., Shi, R. J., and Li, E. T., J. Liq. Chromatogr., 1986, 9, 2035. Dusci, L. J., Hackett, L. P., Forbes, A. M., and Ilett, K. F., Ther. Drug. Monit., 1987, 9,422. Ericsson, O., and Rais, M., Ther. Drug Monit., 1990, 12, 362. Jaworski, J. S., and Kaalinowski, M. K., in The Chemistry of Amidines and Imidates, eds. Patai, S . , and Rappoport, Z., Wiley, Bristol, 1991, pp. 789-846. Sevcik, J., Acta Univ. Palacki. Olornuc., Fac. Rerum Nut., 1977, 53, 37. Kane, P. O., Fresenius’ 2. Anal. Chem., 1960, 173, 50. Birke, R. L., Kim, M., and Strassfeld, M., Anal. Chem., 1981, 53, 852. Meites, L., and Israel, Y., J. Am. Chem. SOC., 1961, 83,4903. Rodriguez-Monge, L. M., Munoz, E., Avila, J. L., and Camacho, L., Anal. Chern., 1988, 60,2269. Munoz, E., Avila, J. L., and Camacho, L., Anal. Chem., 1991, 63, 1574. Maestre, M. S., Munoz, E., Avila, J. L., and Camacho, L., Electrochim. Acta, 1992,37, 1129. Flanagan, J. B., Takahashi, K., and Anson, F. C., J. Electroanal. Chem., 1977, 81, 261. Pizeta, I., Lovric, M., Zelic, M., and Branica, M., J. Electroanal. Chem., 1991, 318, 25. Bond, A. M., Modern Polarographic Methods in Analytical Chemistry, Marcel Dekker, New York, 1980, pp. 84-166 and 288-316. Wolfe, R. H., in The Chemistry of Amidines and Imidates, ed. Patai, S . , Wiley, Bristol, 1975, vol. 1, ch. 8, pp. 349-365. Smith, J. A., and Taylor, H., J. Chem. SOC. B, 1969, 66. Paper 3100855J Received February 12, 1993 Accepted May 6, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801157

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1163 Cathodic Stripping Voltammetric Determination of Pentamidine lsethionate at a Hanging Mercury Drop Electrode M. Valnice B. Zanoni and Arnold G. Fogg" Chem is try Department, L oug h bo roug h- University of Tech n olog y, L oug h borough, Leices te rs h ire, U K LEI I 3TU Differential-pulse cathodic stripping voltammetry was used for the determination of trace amounts of pentamidine isethionate at a hanging mercury drop electrode using its reduction peak at -1.57 V in 0.2 moll-' sodium hydroxide. The optimum accumulation potential and accumulation time were -1.1 V and up to 180 s, respectively. Linear calibration graphs were obtained up to 1 x 10-6 mol 1-1: the limit of detection was calculated to be 3.0 x 10-10 moll-1. The effect of various components of urine on the voltammetric response was studied, and albumin, creatinine and uric acid caused interference in the method.The direct determination of the drug (>I x 10-7 mol 1-1) in urine can be effected using a high dilution of the sample. Keywords : Pen tarn idin e iseth ion ate de te rm in ation; cathodic stripping volta mrn e try Pentamidine isethionate [4,4'-(pentamethy1enedioxy)diben- zamidine bis(2-hydroxyethanesulfonate)], an aromatic diami- dine derivative, is an antiprotozoal agent used in the treatment of pneumocystis carinii pneumonia. 1-3 Modern applications of this drug, and the various methods of determining it, have been discussed previously.4 In the earlier paper4 the electro- chemical reduction at mercury electrodes and the differential- pulse polarographic determination of the drug were reported.A cathodic wave is observed at pH >7.0, and the reduction potential (- 1.60 V versus Ag-AgCl) is independent of pH to just below pH 10 (pK, of pentamidine isethionate =lo). Below pH 10, product adsorption is apparent, whereas above pH 10 the determinand adsorbs strongly. For the determina- tion of pentamidine isethionate down to about 1 X 10-6 moll-1, a pH of 8-9 is recommended using Triton X-100 as a maximum suppressor. The aim of the present work was to investigate the possibility of increasing the sensitivity and rapidity of the voltammetric determination of pentamidine isethionate by using cathodic stripping voltammetry, based on the adsorptive accumulation of the drug on the surface of a hanging mercury drop electrode (HMDE) above pH 10.Experimental Apparatus For voltammetric measurements a Metrohm E612 voltammet- ric scanner and E611 voltammetric detector were used with a Houston Instruments Model 2000 x-y recorder. A Metrohm 663 VA stand was used in the HMDE mode. The three- electrode system was completed by means of a glassy carbon auxiliary electrode and an Ag-AgC1 reference electrode. All potentials given are relative to this Ag-AgC1 (3 moll-' KCl) electrode. Reagents All chemicals were of analytical-reagent grade. De-mineral- ized water was obtained from a LiquiPure system. Pentamidine isethionate stock standard solutions (1 X 10-4-4 x 10-4 mol 1-1) were prepared from the dried pure substance (kindly supplied by Fisons Pharmaceuticals) in de-mineralized water.All solutions were freshly prepared daily and more dilute solutions were prepared from the stock solutions as required. ~~ * To whom correspondence should be addressed. Procedures The general procedure adopted for obtaining cathodic strip- ping voltammograms was as follows. An aliquot (20 ml) of sodium hydroxide solution of the required concentration was placed in a clean, dry voltammetric cell and the required volume of standard pentamidine isethionate was added by means of a micropipette. The stirrer was switched on, and the solution was purged for 15 min. An accumulation potential was applied to the working electrode while the solution was stirred continuously. After accumulation for 15 s, a negative- going differential-pulse scan was initiated, the resulting voltammograms being recorded.The operational parameters used were: pulse amplitude, 50 mV; scan rate, 10 mV s-1; and pulse interval, 1 s. Results and Discussion Preliminary studies in Britton-Robinson (B-R) buffers showed a single stripping peak for 1 X 10-5 mol 1-1 pentamidine isethionate at pH values >10.0 with an accumu- lation time of 180 s. The peak height is much greater than that obtained with differential-pulse polarography . It was obser- ved that the peak height decreased gradually with decreasing pH and the peak disappeared completely at about pH 9.5. Although it was initially considered that this adsorption at high pH might have been due to hydrolysis of the pentamidine isethionate and the peak obtained be due to reduction of the adsorbed hydrolysis product, this was shown not to be the case.The adsorption and reduction are of the neutral pentamidine isethionate molecule. It was considered that this adsorption could be used as an effective preconcentration step prior to the voltammetric determination of pentamidine isethionate in very alkaline solutions in which the differential-pulse polarographic peaks had previously been shown to be very small.4 Cyclic voltammetric studies showed that pentamidine ise- thionate is rapidly accumulated on an HMDE from a stirred solution. Typical cyclic voltammograms of a 1 X 10-6 moll-' solution of pentamidine isethionate in 0.2 mol 1-1 sodium hydroxide solution are shown in Fig. 1. A narrow cathodic peak, corresponding to an irreversible process, can be seen at - 1.58 V.The peak is absent in the subsequent scan, indicating rapid desorption from the electrode (curve B). The height of the peak was shown to be directly proportional to the scan rate within the range 1CL100 mV s-1 and the peak potential shifts linearly to more negative potentials when the scan rate is increased, indicating that the reduction is that of an adsorbed species .51164 L I I I ANALYST, SEPTEMBER 1993, VOL. 118 Comparison of the voltammetric responses obtained with the linear scan and differential-pulse waveforms showed that the use of the pulse technique improves the sensitivity. Differential-pulse adsorptive stripping voltammograms of 5.8 x 10-7 mol 1-1 pentamidine isethionate solutions in 0.2 mol 1-1 sodium hydroxide using Accumulation times of 0 and 180 s are shown in Fig.2. Well-defined stripping peaks were observed with peak potentials at -1.60 V and peak half- widths of 31 mV. Voltammograms obtained without previous accumulation exhibit a much smaller peak. Several parameters directly affect the voltammetric response. The peak current is linearly related to the pulse amplitude between 20 and 80 mV: a value of 50 mV was chosen as optimum as there is a loss of resolution at higher values. The peak current increased with the drop size on the voltammetric stand, and the larger drop size (nominally 0.4 mm2) was chosen as the optimum. Forced convection (i.e., stirring) during the accumulation step also affected the resulting stripping peak current. Tests were carried out between stirring positions 0 and 6 on the voltammetric stand.The peak current increases with increasing stirring speeds but becomes approximately constant at and above stirring speed 3 (approximately 1500 rev min-I), and this speed was chosen as giving the best results. The effect of accumulation potential (Eacc) on the stripping current at potentials between -0.1 and - 1.5 V for solutions of 3.7 x 10-7 rnol 1-1 pentamidine isethionate in 0.2 rnol 1-1 t ._ I 0 I 1 I -1.10 -1.30 -1.50 EN Fig. 1 Repetitive cyclic voltammograms of 1 X mol 1-1 pentamidine isethionate in 0.2 moll-1 NaOH after stirring for 120 s at -1.1 V (accumulation potential). Scan rate, 70 mV s-l. A, First scan and B, second scan t ._ A 130 nA NaOH using an accumulation time (tact) of 180 s is shown in Fig.3. The peak height decreases when accumulation poten- tials more negative than -1.4 V are used and the reduction potential of pentamidine isethionate is approached. The current is maximum for accumulation between -1.0 and -1.2 V, but a decreased signal (decreased adsorption) is observed at potentials less negative than about -0.8 V. A potential of -1.1 V was adopted as the optimum accumulation potential. In order to investigate the effect of basicity on the peak height and peak potential, the differential-pulse adsorptive stripping voltammetric response was studied for 5.8 x 10-7 rnol 1-1 pentamidine isethionate in solutions of different concentration of sodium hydroxide (see Fig. 4). The peak height increased linearly with sodium hydroxide concentration in the range from 5 x 10-3 to 0.4 rnol 1-1, when the accumulation time and accumulation potential were kept constant at 180 s and -1.1 V, respectively. Also, the peak potential shifts linearly towards more negative values (with a slope of 80 mV per pH unit).The effect of accumulation time for 5.8 x 10-7 rnol 1-1 pentamidine isethionate at several sodium hydroxide concentrations is shown in Fig. 5. It was verified that the occurrence of a measurable peak at low concentrations of sodium hydroxide required a longer accu- mulation time. Similarly, for the same accumulation time, the current reaches a plateau because saturation of the electrode surface occurs more rapidly at higher sodium hydroxide concentrations. For analytical purposes a sodium hydroxide concentration of 0.2 mol 1-1 was therefore chosen.At this concentration a large peak is obtained. No significant hydrolysis of the pentamidine isethionate appears to occur at this concentration t -EN Fig. 3 Influence of accumulation potential on stripping peak current for 3.7 x 10-7 mol 1-1 pentamidine isethionate in 0.2 mol 1-1 NaOH with tact = 180 s \, I I \ I 0- 1.10 0.10 0.60 - EaccN Fig. 4 Differential-pulse adsorptive stripping voltammograms of 5.8 x 10-7 moll-1 pentamidine isethionate in A, 0.005; B, 0.02; and C, 0.1 moll-' NaOH. tact = 180 s; EaCc = -1.1 VANALYST, SEPTEMBER 1993, VOL. 118 1165 of sodium hydroxide, and the peak height is stable with time. The effect of accumulation time on the peak heights observed for 3.7 x 10-7 and 1.0 x 10-6 moll-1 pentamidine isethionate solutions showed that there was a rectilinear relationship up to accumulation times of about 210 and 60 s, respectively.Above this time saturation of the mercury drop was observed. Hence the choice of accumulation time depends on the range of analyte concentration being determined. The peak height is linearly dependent on the pentamidine isethionate concentration. As shown in Fig. 6, linear calibra- tion graphs can be obtained from voltammograms recorded at accumulation times between 30 and 180 s from 1 X 10-7 to 1 x 10-6 rnol 1-1 pentamidine isethionate in 0.2 mol 1-1 sodium hydroxide solution. An accumulation time of 180 s and a sodium hydroxide concentration of 0.2 moll-1 are recommen- ded as optimum conditions for the determination of the drug up to 1 x 10-7 moll-1.The reproducibility of the method was determined by successive measurements of ten solutions of 5.0 X 10-7 mol 1-1 pentamidine isethionate in 0.2 mol 1-1 sodium hydroxide. Relative standard deviations of 4.4% were obtained with preconcentration times of 180 s. The limit of detection was calculated to be 3.0 x 10-10 mol 1-1 with an accumulation time between 5 and 10 min. At lower concentra- tions ( 4 . 0 x 10-9 mol 1-1) the peak is present but it was sometimes difficult to measure owing to the supporting electrolyte discharge. However, it is possible to obtain a linear concentration dependence for this peak within the concentra- tion range from 5.0 x 10-9 to 1 x 10-8 mol 1-1 with an accumulation time of 5 min in stirred solutions. In order to investigate the possibility of applying cathodic stripping voltammetry to the determination of pentamidine 400.0 300.0 a = 200.0 .3 100.0 0 16 I 100 200 300 taccls Fig. 5 Effect of accumulation time on peak current for pentamidine isethionate (5.8 X lo-’ rnol 1-l) obtained for various concentrations ofNaOH:A,0.005;B,0.1;C,0.2;andD,0.4mol1-1.E,,,= -1.1V 400.0 r - 7 7 q = 200.0 I 100.0 B A 0 5.0 10.0 15.0 20.0 Concentration/mol 1-1 Fig.6 Calibration graphs for pentamidine isethionate obtained for various accumulation times in 0.2 moll-1 NaOH: A, 30; B, 60; C, 120; and D, 180 s isethionate in urine samples, a study was carried out of the influence of various urine components on the stripping peak currents of the drug. The following urine constituents were selected: albumin, uric acid, urea, glucose, creatinine, potas- sium, chloride and phosphate.It was demonstrated that there is no significant interference from phosphate, chloride and potassium in amounts at least four times higher than those usually found in urine. Fig. 7 shows the effect of the addition of increasing amounts of each urine constituent on the peak current. Albumin and uric acid, even at the low levels at which they are usually present in urine, lower the signal by 60 and 8O%, respectively, and the peak disappears completely in the presence of 50 mg ml-1 of these components. Glucose has virtually no effect on the adsorption and reduction of pentamidine isethionate. Urea and creatinine at concentra- tions above 60 mg ml-1 lowered the voltammetric response by 60%.The possibility of determining pentamidine isethionate in human urine was investigated. It was found that the direct determination of pentamidine isethionate in urine is possible by employing a high dilution of the sample with the supporting electrolyte. Urine was spiked to a concentration of 2.2 x 10-5 moll-1 in pentamidine isethionate and 0.50 ml aliquots of this sample were added to 20 ml of 0.2 moll-’ sodium hydroxide in the cell. The voltammogram was recorded after accumulating for 180 s at - 1.1 V. The peak potential was slightly shifted to a less negative potential but the peak height was not affected by the presence of urine. Quantification of the urine content of the drug was accomplished by two standard additions. The recovery obtained was 70% and typical voltammograms are shown in Fig.8(b). However, the peaks obtained for urine samples containing lower concentrations of pentamidine isethionate were partially suppressed. In addition, it was shown that direct dilution of 1 ml of urine, spiked to a concentration of 7 X 10-7 moll-’ with pentamidine isethion- ate by adding it to 19 ml of 0.2 mol 1-1 sodium hydroxide solution in the voltammetric cell, lowered the voltammetric response by 5% in relation to that obtained in the absence of urine [see Fig. 8(a)]. Further dilutions involving more than 4 ml of urine to 16 ml of sodium hydroxide lead to a depression of the pentamidine isethionate peak of 70%. Taking into account the above results, the direct determina- tion of pentamidine isethionate in human urine is possible, but only under exceptional conditions.It is not possible to determine pentamidine isethionate in urine samples contain- ing very low concentrations of this drug or by using a higher ratio of urine in relation to the supporting electrolyte, probably owing to the effective inhibition of the preconcentra- tion process on the electrode. Further, some natural constitu- ents of the urine give rise to peaks close to that of pentamidine 100.0 80.0 - s .-! .> 60.0 0 - 40.0 20.0 0 20.0 40.0 60.0 Conce n t rat ion/pg m I - 1 Fig. 7 Effects of various urine components on stripping peak current for 5.8 X 10-7 moll-’ pentamidine isethionate in 0.2 moll-1 NaOH, with an accumulation time of 180 s at -1.1 V. A, Albumin; B, uric acid; C, urea; D, creatinine; and E, glucose1166 ANALYST, SEPTEMBER 1993, VOL.118 9, I -1.20 -1.40 -1.60 -1.30 -1.50 -1.70 EN Fig. 8 ( a ) Differential-pulse adsorptive stripping voltammograms of 7 x 10-7 moll-1 pentamidine isethionate in 0.2 moll-' NaOH in the presence of urine diluted: A, without urine; B, 1 + 19 ml in urine-electrolyte; and C, 5 + 15 ml in urine-electrolyte. tacc = 120 s; E,, = - 1.1 V; scan rate = 10 mV s-1. (b) Stripping curves obtained using the standard additions method. X, 0.5 ml of urine spiked with pentamidine isethionate in 20 ml of 0.2 moll-1 NaOH; Y, addition of 0.2 ml of a 2 x 10-5 mol 1-1 standard pentamidine isethionate solution; and Z, further addition of 0.2 ml of standard solution isethionate [see Fig. 8(a)]. Hence a urine sample clean-up procedure would be necessary prior to the determination of pentamidine isethionate. Previous workers677 have used solid- phase extraction with adsorptive stripping voltammetry, and this might prove a fruitful route. The authors thank Fisons Pharmaceuticals plc for providing samples of pentamidine isethionate, and for their interest in this work. M. V. B. Z. thanks FAPESP (Brazil) for financial support and the Instituto de Quimica (UNESP), Araraquara, Brazil, for leave of absence. References 1 Martindale: The Extra Pharmacopoeia, ed. Reynolds, J. E. F.. The Pharmaceutical Press, London, 29th edn., 1989, pp. 675-677. 2 Pearson, R. D., and Hewlett, E. L., Ann. Intern. Med., 1985, 103,782. 3 Sands, M., Rev. Infect. Dis., 1985,7, 625. 4 Zanoni, M. V. B., and Fogg, A. G., Analyst, 1993, 118, 1157. 5 Laviron, E., J. Electroanal. Chem., 1974,52, 35. 6 Hernandez, L., Zapardiel, A., Lopes, J. A. P., and Bermejo, E., Talanta, 1988,35,287. 7 Telting-Diaz, Miranda Ordieres, A. J., Costa Garcia, A., Tuii6n Blanco, P., Diamond, D., and Smyth, M. R., Analyst, 1990, 115, 1215. Paper 3100992 K Received February 19, I993 Accepted May 6, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801163

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1993, VOL. 118 1167 Direct Determination of Ethanol in all Types of Alcoholic Beverages by Near-infrared Derivative Spectrometry Maximo Gallignani, Salvador Garrigues and Miguel de la Guardia* Department of Analytical Chemistry, University of Valencia, 50 Dr. Moliner St. 46100, Burjassot (Valencia), Spain A rapid and accurate method has been developed for the direct determination of ethanol in all types of alcoholic beverages. The method, which does not require any sample treatment (except for simple de-gassing for beer samples or dilution with distilled water for spirits), is based on the use of the first derivative of the near-infrared absorbance spectrum. Measurements, carried out between the 1680 nm peak and the 1703 nm valley, provide a typical calibration graph [dA/dh = 0.00002 + 0.01 27c (where c is the concentration of ethanol in % v/v)] for a concentration range up to 25% v/v with a regression coefficient of 0.999992 and a limit of detection (for k = 3) of 0.1 % v/v.The interference of sugars in the determination of ethanol can be seen by the presence of bands in the derivative spectrum in the range between 1300 and 1800 nm and can be corrected. Results comparable to those found by a reference pycnometric procedure or by gas chromatography were obtained when the proposed method was applied to the determination of ethanol in beer, white and red wines, whisky, gin and rum samples and also in sweet wines and fruit liqueurs. Keywords: Near-infrared derivative spectrometry; ethanol determination; alcoholic beverage The determination of ethanol in alcoholic beverages is a very important problem for social and economic reasons, particu- larly in relation to the taxes imposed on alcohol in different countries.This type of analysis is carried out in many laboratories, not only by producers but also in state and customs laboratories. Hence attempts have been made to simplify the methods available for the determination of ethanol and to develop new strategies that can be used with low-cost instrumentation and without complex sample treat- ment procedures. The current official methods for the determination of ethanol are based on physical measurements, carried out after a previous distillation of the sample in order to separate the ethanol.1-4 However, chemical methods can be used, based on the oxidation of ethanol to acetic acid with potassium dichromate and the determination of the excess oxidizing agent by using Fe" or by iodimetry.5 Instrumental methods can also be employed for the determination of ethanol, and gas chromatography,6.7 liquid chromatography,8.9 potenfiometrylo and nuclear magnetic resonance spectroscopy11 ~2 have been proposed.In recent years the enzymic determination of ethanol has undergone considerable development and a series of enzymic methods, based on different techniques, which can be carried out in batch or flow injection mode, have been proposed.13-16 Infrared (IR) spectrometry, both in the mid and the near range, provides interesting possibilities for the determination of ethanol. In the mid-IR range, ethanol has been determined using cells of an appropriate material, to avoid damage from water, and with a small thickness, in order to reduce the high absorbance of IR radiation by ethanol and water solutions.17 In order to carry out these determinations a series of mathematical treatments, such as the use of orthogonal functions17-19 or derivative spectrometry,17,20,21 have been proposed.The near-infrared (NIR) offers the possibility of determin- ing ethanol in the presence of water by using ordinary glass cells, without the need for short pathlengths. A series of methods based on the use of NIR reflectance analysis have been applied to the determination of ethanol in bee1-22-2~ and wines.25 Transmission NIR spectrometry has been employed * To whom correspondence should be addressed. for the determination of ethanol in beers26 and molasses.27 Also, methods employing optical fibres can be used for the monitoring of fermentation processes28 and for the analysis of wines.29 However, in general, methods based on NIR have been applied to the determination of ethanol in only one type of alcoholic beverage, viz., beers or wines, and there appears to be no general procedure which permits the direct determi- nation of ethanol in all types of alcoholic beverages, even in samples with a high sugar content.Hence, the aim of the present work was to develop a suitable methodology for this purpose. Experimental Apparatus A Perkin-Elmer Lambda 9 double-beam ultraviolet/visible/ near-infrared (UVNISINIR) spectrometer, equipped with a tungsten-halogen lamp source and a PbS detector was used to carry out absorbance measurements.The spectrometer covers a wavelength range from 185 to 3200 nm and is equipped with a monochromator with a 260 lines per mm grating to enable it to operate in the NIR region. An Epson EL2 computer, with a PECSS software package for UV/VLS/NIR (from Perkin- Elmer, release 4.01), controls the spectrometer functions and permits absorbance measurements to be carried out and the corresponding derivative to be calculated. Glass cells, with 1 cm and 1 mm pathlengths, were employed to carry out the direct determination of ethanol in alcoholic beverages. For the chromatographic determination of ethanol, a Perkin-Elmer Model Sigma 3 gas chromatograph with a flame-ionization detector was employed; it was equipped with a 2 m x 2 mm i.d.glass column with 15% m/m Carbowax on Chromosorb (80-100 mesh). For the determination of ethanol by the pycnometric procedure a Gibertini Hydromatic hydrostatic balance was employed. General Procedure NIR derivative spectrometry For the NIR derivative spectrometric determination of ethanol in alcoholic beverages two alternative procedures were carried out, depending on the sugar content of the samples to be analysed. However, in both instances the only1168 ANALYST, SEPTEMBER 1993, VOL. 118 sample preparation required was the de-gassing of the beer samples and a 10 + 25 dilution of the spirit samples. For samples with a low sugar content, the direct measurement of dA/dh between the peak at 1680 nm and the valley at 1703 nm and the interpolation of these .values in the corresponding calibration graph, obtained from ethanol standards diluted with water, provides accurate results.For samples with a high sugar content, it is necessary to ' carry out two measurements in the NIR derivative mode, one using the 1 cm pathlength cell between 1680 and 1703 nm and the other using the 1 mm cell at 1400 nm. From these experimental values, and from the derivative values found, under the same conditions, for standards of ethanol and sugar (taking into account the nature of the sugar present in the sample, viz., natural sugars or added saccharose), the true concentration of ethanol can be obtained by solving the corresponding system of two equations with two unknowns.This system permits an estimation of the sugar content in alcoholic beverages to be obtained. Gas chromatography In order to check the accuracy of the proposed method, ethanol was determined in a series of samples of whisky in the laboratory of the Valencia Customs (SOIVRE) following a direct procedure which consisted of diluting the sample 1 + 9 with distilled water and injecting 1 p1 of the sample into a gas chromatograph using a nitrogen carrier gas flow rate of 2&30 ml min-1, an oven temperature of 105 "C and injector and detector temperatures of 250 "C. A 3.85% v/v standard, prepared from analytical-reagent grade (96% v/v) ethanol, was used as a reference. Pycnometric procedure Other reference values were obtained in the Laboratorio Agrario de la Conselleria de Agricultura of Burjassot using a pycnometric procedure based on the distillation of 250 ml of wine in an alkaline medium obtained from CaO and the measurement of the density of the distillate.Results and Discussion NIR Spectrum of Ethanol in Water Solutions The NIR spectrum of absolute ethanol in the wavelength range from 1100 to 1800 nm [Fig. l(a)] provides, in general, absorbance values lower than those obtained for pure water. However, in a narrow range, from 1670 to 1740 nm, the spectrum of ethanol has three well-defined bands and a shoulder, which could be clearly identified in the presence of water [see Fig. l(b)]. On the other hand, using water as a reference, positive absorbance peaks can be obtained for ethanol in the above-mentioned range [Fig.l(c)]. By increasing the pathlength of the cell from 1 mm to 1 cm, the sensitivity of the NIR spectrometric determination of ethanol also increases and adequate sensitivity values for the determination of ethanol in alcoholic beverages can be obtained. Fig. 2 shows the NIR spectra of ethanol standards of different concentrations and the spectrum of a wine sample containing 12.5% v/v ethanol. From these spectra, it can be concluded that the direct determination of ethanol in alcoholic beverages by NIR spectrometry presents serious problems in the establishment of the spectral baseline, owing to the matrix. However, it can be shown that the absorbance maximum at 1693 nm is well-defined, in both standards and samples, and that it offers a series of possibilities for NIR analysis.Derivative NIR Spectra of Ethanol The first-derivative spectra of ethanol can be employed to correct the matrix interference in the NIR analysis of alcoholic beverages. 1.50 1 .oo 0.50 0 I I 1100 1300 1500 1700 (b) 0.50 0.150 0.075 0 -0.075 -0.150 1 I I 1640 1680 1720 Wavelengthlnm 1760 Fig. 1 NIR spectra of absolute ethanol and water: (a) and (b) using CC4 as a reference; (c using water as a reference. Ethanol (-), water ( - . . - a ) and CC14 2---). In all instances a 1 mm pathlength cell was used I I 0.20 Q) c lu e 8 0.10 2 0 I I I I 1640 1680 1720 1760 Wavelengthlnm Fig. 2 NIR spectra of ethanol standards (-) containing 10,12 and 15% v/v ethanol and the spectrum of a wine sample with 12.5% v/v ethanol (---). Pathlength: 1 cm. Reference: water (.--.-)ANALYST, SEPTEMBER 1993, VOL.118 0.300 U 1169 - (b) Fig. 3 shows the first-derivative NIR spectrum of absolute ethanol and that of pure water, using CC14 as reference and 1 mm pathlength cells, and, as can be seen, ethanol can be determined without interference from water in the wavelength range from 1677 to 1698 nm. On the other hand, using water as a reference [see Fig. 3(b)], .well-defined peaks can be obtained. NIR Derivative Spectrometric Determination of Ethanol By using 1 cm pathlength cells and carrying out the peak-to- valley measurements between 1680 and 1703 nm, ethanol can be determined in alcoholic beverages by NIR derivative spectrometry with a dynamic range between 0 and 25% v/v. A typical calibration graph, under these conditions, corresponds to dA/dh = 0.00002 + 0.0127~ (where c is the concentration of ethanol in 70 v/v) with a regression coefficient (Y) = 0.999992 (see Fig 4).Under the above conditions the matrix interfer- ence found for the NIR determination of ethanol in wine, using absorbance spectra, can be avoided, as can be seen in Fig. 5. Analytical Figures of Merit of the NIR Derivative Spectrometric Determination of Ethanol Table 1 summarizes the main figures of merit for the determination of ethanol using both 1 mm and 1 cm pathlength cells and, in the latter instance, carrying out the measurements from peak to valley and from zero to valley. As can be seen, the limit of detection obtained under the optimum sensitivity conditions (for k = 3, probability level 99.86%) corresponds to 0.1% v/v, and so the methodology developed could be applied to the analysis of all types of alcoholic beverages from beer to wines and spirits.The repeatability of the first-derivative NIR spectrometric determination of ethanol was established from the relative standard deviation of five independent measurements of a sample containing 10% v/v ethanol (for a 1 cm pathlength cell) 0.300 Pa' I 0.150 - x -0.150 w - I I 1640 1680 1720 1760 Wavelengthlnm Fig. 3 First derivative spectra of absolute ethanol (-) and pure water (...-.), obtained with a pathlength of 1 mm and using (a) CC4 as a reference (---) and (b) water as a reference. Pathlength: 1 mm and 25% v/v ethanol (for a 1 mm pathlength cell). A good indication of the precision of the NIR derivative measure- ments can be obtained from the fact that a 10.0% v/v solution of ethanol provides a typical absorbance value of 0.1264 k 0.0003 and that a 10.25% v/v solution of ethanol provides a value of 0.1293 _+ 0.0003, carrying out the measurements between 1680 and 1703 nm and using a 1 cm pathlength cell.Analysis of Real Samples Ethanol was determined in real samples of alcoholic beverages and, as can be seen in Table 2, values comparable to those reported on the bottles were found for beer, red and white wine, rum, gin, brandy, whisky, vodka and fruit liqueurs. In all instances, analyses were carried out without any sample pre-treatment except for the beer sample for which a previous de-gassing was carried out by sonication in an ultrasonic water-bath, and for spirits, for which a previous dilution of 10 + 15 with distilled water is recommended in order to obtain experimental derivative values in the middle part of the linear range of the calibration graph prepared by using a 1 cm pathlength cell.The regression between values found by the proposed method and those reported (Fig. 6) provides the equation y = 0.10 + 0.9942~~ with Y = 0.99988, which demonstrates that the method can be applied in all the concentration ranges studied without requiring a blank correction (the intercept is statistic- ally equal to zero) or presenting constant relative errors (the slope is statistically equal to l).30 0.30 0.15 x 9 U 0 -0.15 0.30 -1 I I 1640 1680 1720 Wavelengthlnm 1760 Fig. 4 Calibration graph obtained for the determination of ethanol.(a) Spectra of ethanol standards (from 5 to 25% v/v) (-) using water (.-..-) as a reference. Pathlength: 1 cm. (b) Calibration graph obtained by measuring dA/dh between 1680 and 1703 nm (0) and at 1703 nm (A) 0.120 0.080 4 0.040 -0 0 s -0.040 -0.080 1640 1680 1720 Wavelengthlnm 1760 Fig. 5 NIR derivative spectra of ethanol standards containing 10, 12 and 15% v/v ethanol (-) and a wine sample (---) containing 12.5% v/v ethanol. Pathlength: 1 cm. Reference: water (-...-)1170 ANALYST, SEPTEMBER 1993, VOL. 118 Table 1 Figures of merit for the first-derivative NIR spectrometric determination of ethanol Cell pathlength 1 mm 1 cm Parameter (1677,-1698,)* nm (1680,-1703,)* nm (1703,)" nm Calibration equation: a0 + bc (c in % v/v) 0.0001 + 0.00262~ -0.0000~ + 0.0127~ -0.00007 + 0.0043~ Regression coefficient ( r ) 0.99996 0.999992 0.99986 Sensitivity+ 0.0262 0.0127 0.0043 Typical signals 0.0655 f 0.0005 0.1260 f 0.0004 0.0430 +_ 0.0020 LODT (% V/V) 0.5 0.1 0.35 Dynamic range (% v/v) 0.5-50 0.1-25 0.35-30 Repeatability* (YO) 0.8 0.3 4.5 * The subscripts p and v indicate peak and valley, respectively.t Sensitivity expressed in absorbance c-1 cm-1; where c is in % v/v. * Repeatability: determined by the relative standard deviation of five independent measurements of a sample containing 10% v/v ethanol for a 1 5 Typical derivative signal obtained for a 10% v/v ethanol content for a 1 cm pathlength cell and 25% v/v ethanol for a 1 mm pathlength 7 LOD: Limit of detection for k = 3; probability level 99.86%. cm pathlength cell and 25% v/v ethanol for a 1 mm pathlength cell.cell. Table 2 NIR derivative analysis of real alcoholic beverage samples Sample Beer White wine Gin Dark rum White rum Vodka Brandy Sherry liqueur Whisky Whisky Spirit 1 2 1 2 3 1 2 1 2 3 4 5 1 1 1 1 1 2 1 2 1 Value reported Value found by NIR* (% v/v) (% v/v) 4.50 5.40 11.0 10.0 12.0 40 40 40 40 40 40 40 38 37.5 39 25 40 40 43 43 40 * Mean ? standard deviation (n = 5). 4.47 f 0.08 5.5 f 0.1 11.25 k 0.09 9.9 +- 0.1 12.2 +. 0.1 39.6 k 0.2 40.2 k 0.1 40.30 k 0.09 39.9 f 0.1 40.17 ? 0.09 40.2 f 0.08 39.8 f 0.2 37.7 Ifr 0.2 37.3 f 0.1 38.9 f 0.2 24.96 f 0.07 39.6 f 0.1 40.16 f 0.09 42.4 k 0.1 42.6 ? 0.1 40.2 f 0.1 Three whisky samples were analysed in this laboratory, using the proposed method, and by an independent labora- tory, using a gas chromatographic method, and, as can be seen in Table 3, the values obtained by both procedures are comparable and of the same order as the values reported on the bottles.Interference of Sugars in the NIR Derivative Spectrometric Determination of Ethanol For some samples, with a high content of sugars, and also for liqueurs with added sugar, the NIR derivative spectrometric determination of ethanol, following the procedure described above leads to large errors; for example, for a sweet white wine containing 15% v/v ethanol, a value of 16.43% v/v ethanol was found, and for peppermint liqueur samples containing 25 and 30% v/v ethanol, the values found were 26.30 and 32.4% v/v, respectively. Fig. 7(a) shows, as an example, the NIR derivative spectra of two standards containing 15 and 16% v/v ethanol, a sweet wine sample, with a 15% v/v ethanol content, and a peppermint sample diluted in order to obtain the same 50.0 40.0 - 3 30.0 a m - ; 20.0 C 3 0 LL 10.0 0 10.0 20.0 30.0 40.0 50.0 Reported value (% v/v) Fig.6 Regression between values found for the determination of ethanol by NIR derivative spectrometry and those reported (---). Theoretical line with a slope of 1 and an intercept of 0 (- ) Table 3 Comparison between the NIR derivative spectrometric method for the determination of ethanol and gas chromatography Ethanol content Ethanol content found by NIR content chromatography spectrometry* Stated ethanol found by gas derivative Sample (Yo v/v) (Yo v/v) (Yo v/v) Whisky 1 40 40.36 40.16 k 0.09 2 40 39.22 39.6 k 0.1 Whisky 1 43 42.0 42.4 f 0.1 * Mean k standard deviation (n = 5).concentration of ethanol. As can be seen, the measurements between 1680 and 1703 nm are strongly affected by the matrix. For both types of alcoholic beverages, the valley at 1650 nm is deeper than previously observed. The different types of sugars present in each sample (saccharose is added to the peppermint and the equivalent of a binary mixture of fructose and glucose to the sweet wine) cause different changes in the spectra. Hence for samples with added saccharose, the peak at 1680 nm and the valley at 1703 nm provide a bathochromic shift of 7 nm, but the position of the bands is not affected by other sugars, such as glucose or fructose. The spectra, recorded between 1300 and 1800 nm, using a 1 mm pathlength cell, [see Fig.7(b) and (c)] show the changes caused by the presence of sugars, as compared with ethanolANALYST, SEPTEMBER 1993, VOL. 118 1171 1640 1680 1720 1760 0.10 0 x 5 -0.10 -0.20 0.03 0 x U 3 -0.03 U -0.06 -0.09 1 I Sugarlg I-' I 1300 1500 1700 Wavelengthh m Fig. 8 (a) Effect of a 1 + 1 glucose-fructose mixture on the NIR derivative spectrum of a 10% v/v ethanol standard solution. Sugar content: 0(---), 30, 60, 90, 120, 150, 180 and 210 g 1-l (-). Pathlength: 1 mm. Reference: water (..---) (b) Variation of the absolute value of dA/dh, measured at 1400 nm, as a function of the sugar concentration Table 4 Effect of the sugar concentration on the first-derivative NIR measurement at 1400 nm by using a 1 mm pathlength cell Sugar [(dAldh)] = a0 + bc (c in g I-')* r Fructose - o .m 4 + 2.04 x 10-4 c 0.9989 Glucose o.oooo5 + 2.01 x 10-4 c 0.9991 0.9992 Saccharose o.0001 + 1.98 x 10-4c 0.9989 0 and 210 g 1-1. Fructose-Glucose? - o . m 6 + 2.02 x 10-4 c * In all instances the sugar concentration range studied was between + 1 + 1 mixture of glucose and fructose. 1300 1500 1700 0.06 h A I 0.040 1 standards. It can be seen that the smoothed first-derivative NIR spectra of samples containing both ethanol and sugars exhibit a common valley at 1400 nm, which can be used for the determination of the concentration of both compounds in the same sample. On the other hand, in the 1650-1750 nm range, the use of a 1 mm pathlength cell demonstrates that the maximum observed at 1680 nm when a 1 cm pathlength cell is used, is actually the sum of two bands at 1677 and 1689 nm and that the different types of sugars present in the sample modify either the first or the second band.A systematic study of the interference of saccharose, fructose and glucose (from 0 to 210 g 1-1 in each instance) in the NIR derivative spectrometric determination of ethanol was carried out. In addition the effect of 1 + 1 glucose- fructose mixture on this determination was also studied. Fig. 8(a) shows, as an example, the effect of the concentration of a glucose-fructose mixture on the spectrum of a 10% v/v 0 x -0.040 U -0.080 -0.120 1300 1500 1700 Wavelengthlnm Fig 9 Calibration obtained for the NIR derivative spectrometric determination of ethanol by using a 1 mm pathlength cell. (-) Measurement at the 1400 nm band.Reference: water (.-.--). (Inset gives regression line) ethanol standard, from which a linear calibration graph of dAldh for the valley at 1400 nm versus the sugar concentration can be prepared [Fig. 8(b)]. This calibration presents the typical expression I(dAldh)1400nmI = - O.oooO1 + 2 x lo-% (c in g 1-I), with r = 0.9996, which is similar for all the different sugars tested (see Table 4). The above-mentioned band at about 1400 nm can be related to the ethanol concentration and hence a typical regression line for ethanol, l(dA/dh)1400nmI = O.oo00, i 4.45 X 1O-k (where c is the ethanol concentration in % v/v), with r = 0.9991, can be obtained (see Fig. 9). From these two regression lines it is possible to obtain the corresponding coefficients for an equation which relates the first-derivative (dA/dh) values at 1400 nm (using a 1 mm1172 ANALYST, SEPTEMBER 1993, VOL.118 Table 5 Effect of the sugar concentration on the first-derivative NIR measurement between 1680 and 1703 nm using a 1 cm pathlength cell Sugar [(dAldh)] = a0 + bc (c in g 1-1)* r Glucose -o.oooo,.+ 2.0 x 10-5 c 0.9981 Fructose o.oooi + 1.61 x 10-4c 0.9997 Glucose-Fructose? o.oooi + 9.1 x 10-5 c 0.9992 Saccharose o.oooo7 + 5.8 x 10-5 c 0.9987 0 and 210 g I-'. * In all instances the sugar concentration range studied was between t 1 + 1 mixture of glucose and fructose. Table 6 NIR derivative spectrometric determination of ethanol in samples with a high content of sugars Reported NIR(,)* NIR(,)t NIR(,)* Sample value (Yo v/v) (% v/v) (% v/v) (YO v/v) Whitewine 1 15 16.4 15.4 14.95 k 0.09 Peppermint 1 25 27.3 25.4 24.8f0.1 Redwine 1 15.1 16.8 16.0 15.2k0.1 2 17.0 18.6 17.6 16.97 k 0.08 3 17.9 18.9 18.3 18.00 k 0.09 * NIR(,): Direct determination carried out by measuring dAldh between the 1680 nm peak and the 1703 nm valley.t NIR(,): Direct determination carried out by measuring dAldh at the 1703 nm valley. * NIR(,): Corrected values obtained by carrying out dA/dh measurements with 1 mm and 1 cm pathlength cells and taking into account the parameters corresponding to the contribution of the sugar to the NIR measurements. Results given are mean k standard deviation. Table 7 Comparison between results found by a pycnometric procedure and by NIR derivative spectrometry for the determination of ethanol in wine samples with a high content of sugar Pycnometric NIR(,,)* NIR(,)t NIR(,)* Sample value (% v/v) (% v/v) (% v/v) (% v/v) Redwine 1 9.13 9.87 9.3 9.05k0.07 2 13.28 13.9 13.5 13.39 f 0.09 3 12.72 12.9 12.8 12.60 k 0.08 Whitewine 1 14.82 16.2 15.3 14.90k0.1 2 11.98 13.2 12.3 12.05 k 0.09 3 12.21 12.6 12.3 12.10+0.1 * NIR(,,-,): Direct determination carried out by measuring dAldh between the 1680 nm peak and the 1703 nm valley.t NIR,,): Direct determination carried out by measuring dAldh at the 1703 nm valley. * NIR(,): Corrected values obtained by carrying out dAldh measurements with 1 mm and 1 cm pathlength cells and taking into account the parameters corresponding to the contribution of the sugar to the NIR measurements. Results given are mean k standard deviation.pathlength cell) with the concentrations of ethanol and sugar in the sample, i.e., I(dAldh)14mnm( = 0.00445cethanol + 2 X By using a 1 cm pathlength cell, a relationship can also be found between the derivative values, measured between 1680 and 1703 nm, and the concentrations of ethanol or sugars, from which a series of regression lines can be obtained. Table 5 summarizes the data obtained and, as can be seen, the parameters of these equations depend on the type of sugar; hence for the analysis of alcoholic beverages the equations to be used must be selected taking into account the nature of the sugars present in the sample. Hence, for the determination of ethanol in wines the following equation, which includes the coefficients for ethanol and for a mixture of fructose and glucose, must be employed: (dA/dh)1680-1703 nrn - 0.0127cethanol + 9.1 X 10-5 cSugar, whereas for the determina- tion of ethanol in peppermint (which also contains added saccharose) , the following equation is recommended: (dAl l@-4~sugar.- dh)168@1703 nm = 0.0127 Cethanol + 5.8 x 10-5 csugar. Hence for the analysis of samples with a high content of sugars the determination of ethanol must be performed by carrying out NIR derivative measurements with 1 mm and 1 cm pathlength cells at 1400 nm and between 1680 and 1703 nm, respectively, and solving the corresponding system of two equations with two unknowns; the ethanol concentration is expresed in YO v/v and the sugar content in g 1-1. As can be seen in the equations summarized in Tables 4 and 5, the sensitivity of the NIR derivative measurements for the determination of sugars is very low and so the method, which is suitable for the determination of ethanol, only permits an estimation of the sugar concentration to be obtained.Values obtained for sugars can be used to correct the concentration of ethanol, but the values must be improved by means of other measurements in order to provide an accurate and precise determination o$ the sugar concentration. Determination of Ethanol in Samples With a High Sugar Content The above-mentioned procedure was employed for the determination of ethanol in different samples with high contents of sugars. The data in Table 6 demonstrate that the proposed method permits the correction of errors obtained by direct measurement of dA/dh and provides results comparable to those reported.Comparison of the NIR Derivative Spectrometric Procedure With Other Methods The official method for the determination of ethanol in alcoholic beverages is based on the measurement of the density of the sample, with a precision of +0.1% v/v ethanol, after a previous distillation of the sample in an alkaline medium. Table 7 shows the results obtained for a series of samples analysed by the pycnometric procedure and also by the proposed method. As can be seen, both procedures provide comparable results, and the precision of the NIR derivative measurements is adequate for carrying out these determina- tions. One procedure, commonly employed in control labora- tories for the determination of ethanol in spirits, is based on the direct analysis of samples, previously diluted with distilled water, by gas chromatography using a specific column for the determination of alcohols.This method is rapid and has been thoroughly validated. Table 3 summarizes the results found for the determination of ethanol in whisky samples by both gas chromatography and NIR derivative spectrometry. As can be seen, both procedures provide comparable results. It can be concluded that the proposed NIR derivative method is very simple and more rapid than other methodolo- gies described in the literature and that it provides accurate, sensitive and reproducible results. One of the major advan- tages of the method is that it can be applied to the direct determination of ethanol in all types of alcoholic beverages.Compared with other published methods, based on Fourier transform IR17-19 and NIR22-25 analysis, the proposed method does not require any complex mathematical treatment of the data and, by using direct derivative spectra, accurate results can be obtained. The precision of the proposed NIR derivative method (k0.lY0 v/v) is better than that reported by Buchanan et aZ.29 (f0.33% v/v) and Cavinato et a1.28 (+0.2% v/v). This could be due to the use of a 1 cm pathlength tranmission cell instead of the fibre optic systems employed in the earlier papers. The proposed method is the only procedure that can be applied to the analysis of samples of very different types without incurring problems related to the matrix. In addition, a suitable methodology for correcting errors derived from the presence of sugars has been developed.ANALYST, SEPTEMBER 1993, VOL.118 1173 M. G. acknowledges a grant from the Agencia Espafiola de Cooperaci6n Internacional to carry out Ph.D. studies and the financial support of Los Andes University and CONICIT (Venezuela). S. G. acknowledges a grant from’the Conselleria de Cultura, Educaci6 i Cihcia de la Generalitat Valenciana to carry out Ph.D. studies. The’authors also thank Cristina Martinez for performing the gas chromatographic analyses and Angel Orozco for carrying out the pycnometric pro- cedure. 1 2 3 4 5 6 7 8 9 10 11 12 References Commission Regulation (EEC) No. 2676/90, Off. J. Eur. Comm., 1990,33, L272. Ministerio de Sanidad y Consumo, Anulisis de Alimentos.Mktodos Oficiales y Recomendados por el Centro de Investigacidn y Control de Calidad, Servicio de Publicaciones del Ministerio de Sanidad y Consumo, Madrid, 1985. Official Methods of Analysis of the Association of Official Analytical Chemists, AOAC, Washington, DC, 1990. Amerine, M. A., and Ought, C. S., Methods of Analysis of Musts and Wines, Wiley, New York, 1980. Pilone, G. J., J. Assoc. Off. Anal. Chem., 1985, 68, 188. Caputi, A. J. and Mooney, D. P., J. Assoc. Off. Anal. Chem., 1983, 66, 1152. Cutaia, A. J., J. Assoc. Off. Anal. Chem., 1984, 67, 192. Iwachido, T., Ishimaruk, K., and Toei, K., Anal.Sci., 1986, 2, 495. Morawski, J., Dincer, A. K., and Ivie, K., Food Technol., 1983, 37,57. Kakabadse, G. J., Lab. Pract., 1990, 39, 51. Guillou, M., and Belanche, M., Connaiss. Vigne Vin., 1989,23, 215. Guillou, M., and Tellier, C., Anal. Chem., 1988, 60, 2182. 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Lazaro, F., Luque de Castro, M. D., and Varcircel, M., Anal. Chim. Acta, 1986, 185, 57. Lazaro, F., Luque de Castro, M. D., and Varcircel, M., Anal. Chem., 1987,59, 1859. Junge, C., J. Assoc. Off. Anal. Chem., 1987, 70, 1089. Worsfold, P. J., RfiZiEka, J., and Hansen, E. H., Analyst, 1981, 106, 1309. Lopez Mahia, P., Simal GBndara, J., and Paseiro Losada, P., Vib. Spectrosc., 1992, 3, 133. Glenn, A. L., J. Pharm. Pharmacol., 1963, 15, Suppl., 123T. Agwu, J. V., and Glenn, A. L., J. Pharm. Pharmacol., 1967,19, 76s. Paseiro Losada, P., and Simal Lozano, J., Anal. Bromatol., 1984, 36, 97. Heisz, O., Labor Praxis, 1989, 13, 402. Criddle, W. J., Parry, K. W., and Jones, T. P., Analyst, 1987, 112, 615. Halsey, S. A., J. Znst. Brew., 1985,91, 306. Halsey, S. A., Anal. Proc., 1986,23, 126. Requejo-Gdmez, A., Tec. Lab., 1983, 8, 911. Coventry, A. G., and Hunston, M. J. Cereal Foods World, 1984, 29,715. Dumoulin, E. D., Azain, B. P., and Guerain, J. T., J. Food Sci., 1987, 52, 626. Cavinato, A. G., Mayes, D. M., Zhihong, G., and Callis, J. B., Anal. Chem., 1990, 62, 1977. Buchanan, B. R., Honigs, D. E., Lee, C. J., and Roth, W., Appl. Spectrosc., 1988,42, 1106. Miller, J. C. and Miller, J. N., Statistics for Analytical Chemistry, Wiley, New York, 1984. Paper 3101564E Received March 18, 1993 Accepted May 14, 1993

ISSN:0003-2654    

DOI:10.1039/AN9931801167

出版商:RSC    

年代:1993

数据来源: RSC