摘要:

ANALYST, FEBRUARY 1993, VOL. 318 145 New Look at Analytical Data Through the Gnostical Method TomaS Paukert* and Ivan RubeSka Czech Geological Survey, Malostranske namesti 19, 1 18 21 Prague I , Czechoslovakia Pave1 Kovanic Institute of Information Theory and Automation, Czechoslovak Academ y of Science, Pod vodarenskou v@i 4, 180 00 Prague 8, Czechoslovakia The gnostical theory represents a new, powerful approach towards evaluating data files. This paper describes the application of a gnostical analyser which may help i n finding outliers, testing the homogeneity of sets of data and classifying individual data. As an example, its use for ascertaining the recommended values for reference materials is demonstrated by means of homogeneous and heterogeneous sets of data. Keywords: Robust estimator; gnostical method; recommended value; reference material New analytical procedures may best be verified by analysing certified reference materials (CRMs) with well established concentration values for the constituents to be determined.Apart from this, CRMs are also important for quality assurance and quality control in analytical laboratories. The determination of the true analyte concentration in a CRM from experimental data is therefore an important task toward which the efforts of many workers have been directed. The list of papers dealing with this problem and using statistical methods is fairly extensive. The data to be treated, however, do not always represent a homogenous population. In such instances, robust estimators less susceptible to non-homogeneities in experimental data have been applied by many workers,l-3 but the outcome has often been dubious, if not frustrating.4.5 In this paper we describe the application of a new non-statistical ‘gnostical’ method.The gnostical theory (GT) is an axiomatic-deductive mathematical theory. It has been developed as an alternative to mathematical statistics for the treatment of data containing uncertainty, having a statistical model that is not known, or data with a statistical characterization that does not ad- equately describe the essence of phenomena. The GT may also be successfully applied to data for which, for various reasons, a limited number of observations or measurements are available, i.e., the quality of information is poor or the data are influenced by the contribution of a rare, but strong disturbance of an undefined character. During the development of GT programming, five different classes of gnostical programs have been distinguished.One of them, the gnostical analysis discussed in this work, may be applied for in-depth analysis of limited data files, robust estimations of expectancy or probability of phenomena, distribution functions of data files and their density, robust cluster analyses, investigation of file homogeneity and of the equivalence or diversity of two or more files, robust estima- tions of location parameters and of scale parameters of individual clusters and classification of individual data accord- ing to the degree of their relevance to individual clusters. Gnostical Theory of Uncertain Data Sound data deserve to be given greater weight than unsound data.However, two problems exist in this connection: (i) how to distinguish the unsound data from the sound data; and (ii) how to optimize the weights to make maximum use of the information. Heuristic approaches cannot ensure either optimization or the universal applicability required by practice. It is well * To whom correspondence should be addressed. known that ‘the most practical tool is a good thecry’. Unfortunately, a good theory of such a complex problem cannot be both simple and directly acceptable by way of some ‘common sense’ considerations. In contrast to other theories involving uncertainty, the GT of data files is based on an axiomatic theory of an individual uncertain datum and on a data composition axiom.The axioms of this theory have a simple algebraic nature. To illustrate the functions of the GT, first the main equations will be described. Consider the ith real-valued A, (an ‘additive’ datum) together with its ‘multiplicative’ equivalent, zr = exp(A,) (1) having a strictly positive value. For a positive real scale parameters, a real variable z > 0 and for a sample of N data, define the auxiliary quantities q,(z,s) = (2,/2)2’.* (2) (3) (4) for use in the calculation of N ‘fidelities’: fi(z,s) = 2/[1/q,(z,.9 + q,(z,s)l hr(z,s) = [l/q,(z,s) - qr(z,s)l/[~/qr(z,s) + q,(z,s)I and ‘irrelevancies’: Within the framework of the GT, irrelevance plays the role of the distance between z and 2, the fidelity being the weight of the datum 2,.Introduce the arithmetic mean N f(z,s> = c f i ( z , s Y N ( 5 ) r = l of the fidelities and define the symbol h(z,s) for the irrelevances analogously. Let w(z,s) be the function of weight defined by the relationship W ( z , s ) = Cf(z,s)l2 + [h(z,s)I2 (6) The distribution function generated by the individual datum 2, is then L,(z,s) = [I + hi(~,s)]/2 (7) having the density At least two theoretical results of the GT are immediately applicable to the data files provided by analytical chemistry, i. e . , the two types of data distribution functions (DDF), global (GDF) and local (LDF) . These functions play a role analogous to the probability distribution functions. The field of applica- tion of the gnostical distribution functions is, however, much broader, as these functions do not rely on some statistical146 ANALYST, FEBRUARY 1993, VOL.118 assumptions or probabilistic concepts. They characterize the data patterns and expectations of the subject deduced from their particular shape. For a weak influence of the uncertainty of the data, i.e., minor data errors, the two gnostical DDFs differ to only a negligible extent. However, their behaviour may be different for major data errors. This is a desirable feature that makes these gnostical DDFs very useful for analyses of totally different types of data files. The LDF L(z,s) is simply the arithmetic mean of the distribution functions [eqn. (7)] of individual data: N (9) The GDF G(z,s) is obtained using the function of weight w N [eqn. (611: G(z,s) = Ir(z,s)Iw(z,.~) (10) 1 = 1 The LDF is a relatively universal instrument which can be applied even to data files containing non-homogeneities such as individual subclusters.The derivative of the LDF, called the data density function, has a multi-modal form, in which each mode corresponds to a subcluster. Being a monotonous function for an arbitrary data file, the LDF can always be determined. In the special case of the reasonability of a statistical interpretation of data, the LDF [eqn. (9)] can be an asymptotically consistent kernel estimate of the Parzen type6 of probability distribution function. In such a case, the GT is used as a source of a theoretically justified kernel which generates remarkably clear and smooth density curves, even for small data files. In the much more general case of data files that do not allow statistical interpretation, the equation of the DDF is still valid as a continuous model of the distribution of the expectation that a ‘new’ datum of the same nature as the ‘old’ data will have a certain value.The LDF is ‘locally robust’ in the sense that its local form, corresponding to a subinterval of a data range, does not influence its form in another subinterval. This is due to the steep descent of the gnostical kernel [eqn. (S)]. Unlike the LDF, the GDF has theoretical justification only €or special data files of homogeneous type. Such files should have a unimodal data density function. For a non- homogeneous data file, the GDF may lose the fundamental feature of a distribution function, its monotony.7 This fact makes it possible to perform an efficient test of the homogeneity of the data files.The limited flexibility of the GDF permits the estimation of the proper scale parameter which characterizes the spread of the data. Most unique, however, is the GDF, which is globally robust in the sense of the low sensitivity of its shape with respect to the ‘outliers’ and also to all the other ‘peripheral’ subclusters of the data. This leads to a highly reliable prognosis of rare events (of values of the GDF for very small or very large quantiles). Such tasks often appear in practice in connection with random quality controls, studies of lifetimes, etc. This type of gnostical distribution function has no known statistical analogy. The LDF and GDF differ substantially in their dependence on the scale parameter s.Let F(N) be the ‘empirical’ distribution function of the data file. The function F(N) has the known form of an irregular staircase. The LDF of the same sample can be made to approach the F(N) as close as required, choosing a sufficiently small positive value for the s parameter. In contrast, the maximum distance of the GDF has a minimum for a ‘best’ s, which can be recognized as a robust estimate of the scale parameters. Hence the GDF is as close as possible to the F(W. The choice of s determines the resolution power of individual clusters of data files. An overall survey of the GT, with a detailed description of the mathematics involved, can be found elsewhere.8 Application of the GT in Analytical Chemistry In routine practice, the evaluation of data is usually carried out by statistical calculations.The accuracy of an analytical procedure may best be verified by analysing CRMs with well established ‘recommended’ values (RV). For rock CRMs, the RVs are usually derived from round-robin tests with the participation of many laboratories. If the central values (location parameters PL) for particular elements derived by statistical evaluation of the data are mostly concordant, the assignment job is relatively ‘easy’ and the RV may be established from the various central values. If the PL are discordant, a decision has to be made as to the suitability of the analytical methods employed for the particular concentration levels. Some elements in the Periodic Table are notoriously troublesome for quantitative determination, which may be demonstrated by repeated analyses.Differences of as much as 100% or more are sometimes encountered, especially when determining element contents at ultra-trace levels. To demonstrate the possibilities of the new non-statistical mathematical method discussed, we used data from the 1987 Compilation Report on the Ailsa Craig Granite, AC-E.9 The AC-E reference material, prepared with great care, was distributed to 128 laboratories in 29 countries. From the submitted data, RVs for many elements were successfully established. Some elements were determined by a limited number of laboratories only and for some elements discrepan- cies in the results were evident. Two elements were selected as examples to demonstrate the potential of the GT for evalua- tion of results; first, the results for europium, which represent good concordance of reported data, and second, results for cobalt, which have a significantly heterogeneous file of data.Homogeneous Data File The results reported for europium are a good example of a homogeneous data file. This element was determined in 41 laboratories. All the results are given in Table 1. Table 1 Results for Eu (ppm). For explanation of abbreviations used, see text Laboratory Value Procedure Laboratory Value Procedure Laboratory Value Procedure Laboratory Value Procedure 1 2 3 4 5 6 7 8 9 10 1.39 1.4 1.6 1.7 1.8 1.8 1.86 1.87 1.9 1.9 EMN CSP CSP CSF CSP EMN EMN EMN CSP EMN 11 12 13 14 15 16 17 18 19 20 1.9 1.92 1.92 1.93 1.94 1.94 1.94 1.96 1.98 1.99 EMN CSP EMN CSM CSP EMN EMN EMN EMN EMN 21 22 23 24 25 26 27 28 29 30 2 2 2 2 2 2 2 2.03 2.04 2.05 CSP CSP CSP CSP CSP EMN EMN EMN EMN ASM 31 32 33 34 35 36 37 38 39 40 41 2.07 2.09 2.1 2.1 2.16 2.23 2.3 2.3 2.4 2.68 3.1 EMN CSP CSP EMN EMN EMN BSM CSP CSP EMN ASMANALYST, FEBRUARY 1993, VOL.118 - - - - - 147 1.0 0.8 0.6 - ir ru 0.4 0.2 - 0 Each method is designated by a three-letter code, the first letter indicating the method of sample preparation and the last two the method of determination.9 For sample preparation, A = acid decomposition, B = fusion with fluxes, C = dissolution + separation, D = mixture with buffers and E = simple physical conditioning; for determination, AA = atomic absorption spectrometry (AAS), FX = X-ray fluorescence spectrometry (XRF), SM = mass spectrometry, MN = nuclear methods, SF = flame photometry and SP = direct reading atomic emission spectrometry (AES) .For the first evaluation of data files the data were treated by the GT using the GDF. The resulting distribution function, presented in Fig. 1, may help in investigating the file homogeneity. This type of estimate is based on the a priori assumption that the tested data file is homogeneous. It is robust with respect to peripheral data (outliers). In Fig. 1 both the distribution function and the density function of the data file are shown. The dotted lines indicate the degree for goodness of fit of the distribution function estimated by the Kolgomorov-Smirnov test with plotted intervals for 10, 20, 50, 95 and 99% probability.The central value for europium corresponding to the maximum of the curve was calculated to be 2.007 ppm. 6.946 -6 0 A Log(z) - Fig. 1 Global distribution of the Eu data file. z , Concentration of analyte; P ( z ) , distribution function of the concentration z and dPld(1og z ) , density function of the concentration z . Kolgomorov- Smirnov test: A, 10; B, 20; C, 50; D, 95; and E 99% Log(z1- Fig. 2 95; and E, 99% Local distribution of the Eu data set. A, 10; B, 20; C, SO; D, 9.258 n Log(z) - Fig. 3 data. A, 10; B, 20; C, SO; D, 95; and E, 99% Global distribution of the Eu data file after restriction of 4 Even though the peak in Fig. 1 appears clear, the distribution curve was recalculated using the LDF model, which should distinguish local clusters in a data file.The result, presented in Fig. 2, shows three small maxima, obviously corresponding to the values 1.39, 1.4, 2.68 and 3.1 ppm. These may be interpreted as outliers. After their exclusion, the calculation of the global distribution was repeated and the result is presented in Fig. 3. The central value calculated according to the LDF model is 1.980 k 0.109 ppm. After having eliminated the outliers, it decreases to 1.970 ppm. The value of 0.109 pprn is the standard deviation derived from the GT calculations. In Table 2 this PL is compared with the various mathematical parameters reported in the AC-E c~mpilation.~ The robustness of the GT-derived PL may be demonstrated by the following example. The x, and the gnostical central values (GCVs) were first calculated from the entire number of 41 analyses.Then, the one-step restriction was executed; this means that the values 2.68 and 3.1 ppm were trimmed off and x, and the GCVs were calculated again from the remaining 39 values. The comparison of the calculated results may be seen in Table 3. The GCV evidently changes considerably less than the arithmetic mean. Heterogeneous Data File We selected a data set with non-homogeneous results, at the same concentration level as europium and with a similar number of analyses. The 40-value file for cobalt from the AC-E CRM9 fulfilled these requirements well. The cobalt values have a great spread, ranging from 0.07 to 10 pprn (see Table 4). The outcome of the calculation of the global distribution from these data is presented in Fig.4. The broad maximum clearly reflects the wide range of the results reported. As the congruence between the GDF and the empirical distribution function is poor, the data file cannot be considered as homogeneous. The data set was recalculated once more, using the LDF model, and the result is shown in Fig. 5 . On the local distribution curve, three distinct maxima appear, the location of which, as calculated by the GT, results in three different concentration levels for cobalt, namely 0.16; 1.39 and 4.9 PPm. In Table 5 the central values, as reported for the AC-E in the compilation9 but split according to the method of analysis used, are given. A comparison with the results given by GT immediately suggests that the three maxima correspond to central values of the results by nuclear methods only, by optical spectrometric methods (AAS + AES) and by XRF spectrometry, respectively.It is worth mentioning that the RV chosen was 0.2 pprn and was based on the nuclear methods set of data. Table 2 Mathematical parameters calculated from 41 results for Eu" Parameter N x, M MG xp xgeo x,, xg GCV Value 41 2 2 1.99 2 1.99 2 1.99 1.98 * The derived RV = 2 ppm.9N = number of analyses; M = median; xp = preferred mean calculated after 1 k s elimination; xg = gamma central value; x, = arithmetic mean; M G = Gastwirth median; xgef = geometric mean; x,, = dominant cluster mode; GCV = gnostical central value. For more information about the statistical parameters, see ref. 9. Table 3 Comparison of robustness (for definitions of parameters see Table 2) Number of results x, GCV 41 2.007 1.983 39 1.962 1.974148 ANALYST.FEBRUARY 1993, VOL. 118 Table 4 Results for Co (ppm). For explanation of abbreviations used, see text Laboratory 1 2 3 4 5 6 7 8 9 10 Value 0.07 0.091 0.118 0.13 0.14 0.153 0.18 0.2 0.2 0.21 Procedure Laboratory Value EMN 11 0.21 EMN 12 0.28 EMN 13 0.5 EMN 14 0.6 EMN 15 1 EMN 16 1 ASP 17 1 EMN 18 1.15 EMN 19 1.4 EMN 20 1.4 Procedure EMN EMN AAA EFX AAA ASP EFX EMN AAA EMN Laboratory 21 22 23 24 25 26 27 28 29 30 Value Procedure 1.65 EMN 2 AAA 2 ASP 2 ASP 2 BAA 2 EFX 3 AAA 3 EFX 4 EFX 4.7 DFX Laboratory Value Procedure 31 4.9 DFX 32 5 EFX 33 6 ASP 34 6 BFX 3s 6 EFX 36 7 DFX 37 7 EFX 38 9.5 EFX 39 10 AAA 40 10 EFX 1 .o 0.8 0.6 2 0.4 0.2 0 0 Fig. 4 95; and E, 99% Global distribution of the Co data file.A, 10; B, 20; C, 50; D, 0.271 1 .o 0.8 - A ru 0.6 - I 0 2 z Is1 - I 0.4 TI 0.2 0 0 Fig. 5 and E, 99% Local distribution of Co data file. A, 10; B, 20; C, 50; D, 95; Table 5 Mathematical parameters for the analyses of Co. For explanation of abbreviations used, see text Method used Calculated parameter N Xa M MG XP Xgeo Xcrn xg MN AA SP 14 7 5 0.43 2.84 2.24 0.2 2 2 0.19 1.82 1.7 0.16 1.65 1.3 0.24 1.88 1.34 0.19 1.38 - 0.19 1.73 1.61 FX 5.05 4.95 4.98 5.29 4.01 4.87 4.96 14 GCV 0.16-1.39-4.90 Total 40 2.69 1.53 1.69 1.45 1.1s 0.17 1.4 From the analytical point of view, it seems surprising that for a frequently determined element such as cobalt, such large discrepancies in analytical data may occur. One should nevertheless bear in mind that the AC-E was specially prepared as a CRM for the rare earth elements and the analysis results for cobalt are only a useful by-product of no particular interest.As 0.2 ppm is an unusually low cobalt content in rocks, it is probably well below the concentration range for which the instruments are routinely calibrated. The readings by optical spectrometry and XRF were plausibly evaluated simply by extrapolation to lower concentration. The greater error by XRF probably reflects the fact that 0.2 ppm of cobalt is closer to the limit of detection by XRF than by optical spectrometry. However, the most remarkable outcome, in our view, is that the gnostical analyser applied to the entire set of data did provide three different results which, in addition, are in fairly good concordance with the statistical analysis applied sepa- rately to the results by different analytical methods.This demonstrates the possibilities of the GT when applied to non-homogeneous sets of data in discerning separate data files. Conclusions We have tried to draw attention to a new, powerful tool for treating analytical data provided by the gnostical theory. This is demonstrated by applications of the gnostical analyser to the data from a collaborative study9 and for deriving recom- mended values for the CRM rock AC-E. Although the GT cannot provide RVs from insufficient data, it can reveal their ‘heterogeneity’. For such sets the GT is highly sensitive and may distinguish the independent files even without any additional information. With respect to PL calculation, it also exhibits high robustness, thus providing a theoretically based, practical substitute for empirically derived robust estimators. The GT is still being developed and its applications are very promising. The program system entitled ‘interactive analyzer GA2’ has been adapted for use on an IBM PC. Readers interested in the GT should contact P. Kovanic. References 1 Ellis, P. J . , and Steele, T. W., Geostand. Newsl., 1982, 2, 207. 2 Lister, B., Geostand. Newsl., 1984, 7, 171. 3 Abbey, S., Geostand. Newsl., 1988, 9, 241. 4 Abbey, S . , paper presented at Geoanalysis 90, Huntsville, Canada, 1990. 5 Abbey, S . , Chem. Geol., 1992, 95, 123. 6 Parzen, E., Ann. Math. Stat., 1962, 35, 1065. 7 Baran. R. H.. Automatica, 1988.24, 283. 8 Kovanic, P., Automarica, 1986, 22, 657. 9 Govindaraju, K., Geostand. News/., 1987, 11, 203. Paper 2f02281 H Received May 1, 1992 Accepted August 5, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800145

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1993, VOL. 118 149 Determination of Free Toluene Diisocyanates in Flexible Polyurethane Foams Using Negative Chemical-ionization Mass Spectrometry Kazik Jedrzejczak and Virindar S. Gaind Occupational Health Laboratory, Ontario Ministry of Labour, 101 Resources Road, Weston, Ontario, Canada M9P 3TI A simple and sensitive method was developed for the determination and confirmation of trace amounts of toluene diisocyanate (TDI) residues in a polyurethane foam. Foam samples were extracted with methanol in a mini-Soxhlet apparatus after addition of deuteriated toluene-2,4-diyl dimethylcarbamate as an internal standard. The extracts were concentrated, derivatized with pentafluoropropionic anhydride at 70 "C and analysed by capillary column gas chromatography with use of selected-ion monitoring in the negative-ion chemical-ionization mode.The limit of detection was 10 ng g-1 for a 1.0 g sample. The recovery was above 79% and reproducibility assessments indicated good accuracy and precision over the concentration range of 20-500 ng g-1. The method was applied to the determination of free TDls in fresh foam samples. Keywords: Toluene diisocyanate; polyurethane foam; gas chromatograph y-mass spectrometry; free isocyanate assay Toluene diisocyanate (TDI) is one of the main components in the manufacture of polyurethane foams and coatings. The TDI-based products would be expected to contain residual traces of the free unreacted isocyanate as a result of the incomplete reaction or of the use of non-stoichiometric amounts.1 A number of methods have been published for the determination of the monomeric TDI in the workplace environment.These include spectrophotometry,2,3 gas chro- matography (GC)4-7 and high-performance liquid chromato- graphy.8-11 However, there is no reliable procedure available for the determination of monomeric TDI residues in the final products such as flexible polyurethane foams. The analysis for monomeric isocyanates in flexible poly- urethane foams presents a significant challenge because of the possibility of workers' exposure when handling foams with residual TDI, for which the American Conference of Govern- mental Industrial Hygienists has adopted a threshold limit value-time weighted average of 0.04 mg m-3. The Guidance Note EH 40 from the Health and Safety Executive, UK, gives 0.02 mg m-3 as the long-term exposure limit €or all isocyanates.The literature on this subject is scanty and reports few methods for such analyses. The titrimetric,l2 spectrophotometric~3 and infrared14 analytical procedures are not sufficiently sensitive or specific for the determination of low concentrations of TDI monomers in foam samples. Conte and Cossils reported a quantitative method involving extraction of flexible urethane foams with anhydrous 1,2-di- chlorobenzene and the GC determination of free TDI with use of a flame-ionization detector. Rastogil6 developed a more sensitive procedure in which TDI monomers were determined as their urea derivatives, formed by the reaction with 9-(methylaminomethy1)anthracene. 16 This method has been applied to the determination of TDI monomers in various types of chemical products containing polyurethanes, such as adhesives, sealing waxes and surface coatings, and was not applicable to flexible foams.This paper describes a highly sensitive and specific analy- tical procedure for the determination and confirmation of trace amounts of TDI residues in polyurethane foams by means of negative-ion chemical-ionization (NICI) mass spec- trometry (MS) and selected-ion monitoring (SIM). It is based on the methanol extraction of foam samples after addition of deuteriated toluene-2,4-diyl dimethylcarbarnate (2,4-TDC- d,) as an internal standard. The extract is derivatized with pentafluoropropionic anhydride (PFPA) reagent in conjunc- tion with pyridine as a catalyst.The pentafluoropropyl derivatives of the two carbamates are determined by capillary GC-MS using SIM in the NICI mode. This technique provides the requisite sensitivity and specificity for determining trace amounts of TDI monomers present in flexible urethane foams, and the identity of the analyte is unequivocally confirmed by its retention time and by the relative molecular mass of the fragment ions. The method was applied to the determination of residual TDI monomers in fresh foam samples. Experimental Reagents The TDI (technical grade SO%), PFPA, potassium di- hydrogen phosphate, toluene, pyridine and methanol were obtained from Aldrich (Milwaukee, WI, USA) and were of analytical-reagent grade. All other reagents and solvents were of the highest purity grade commercially available.Methanol-d4 was purchased from MSD Isotopes (Pointe Claire, Quebec, Canada) and the isotopic composition was a minimum 99.5% of deuterium atoms. Phosphate buffer (1.0 mol 1-1, p H 7). Prepared by dissolving 136 g of potassium dihydrogen phosphate in de-ionized water and adjusting the solution with 40% potas- sium hydroxide to pH 7.0. Apparatus Mass spectra were obtained on a Hewlett-Packard (HP) Model 5989A MS Engine gas chromatograph-mass spec- trometer (Avondale, PA, USA), equipped with an HP 59940 MS ChemSystem and an HP 7673A automatic sampler. An HP 5890A gas chromatograph, equipped with a thermionic (nitrogen-phosphorus) specific detector (TSD), a Model 3396A integrator and a Model 7672A autosampler, was used for GC analysis. Wheaton mini-Soxhlet extractors were purchased from Aldrich.Standard Preparation The toluene-2,4-diyl dimethylcarbamate (2,4-TDC) standard was prepared as follows: 1.0 g of TDI was added dropwise, with stirring, to 20.0 ml of anhydrous methanol. The solution was warmed in a water-bath at 60 "C for 5 min. After cooling in an ice-bath, crystals of 2,4-TDC appeared. The urethane formed was purified by recrystallization twice from carbon tetrachloride. The hot solution was filtered to remove the less-soluble ureas, which could form owing to traces of1 so ANALYST, FEBRUARY 1993, VOL. 118 moisture. The pure solid product had a m.p. 172-175 "C and its electron-impact mass spectra (EI-MS) showed ions at mlz 238 (M+) and mlz 206 (M-CH30H)+, while its positive chemical-ionization MS (PCI-MS), using methane as a reac- tant gas, showed ions at mlz 239 (M + 1)'.Other characteris- tic peaks at mlz 267 (M + 29)+ and mlz 279 (M + 41)+ were also discernible. The infrared spectrum showed a double band at 1697 and 1720 cm-1, which is characteristic of the carbonyl group in carbamates. The standard solution was prepared by dissolving 100 mg of the purified 2,4-TDC in 100 ml of toluene. It was stored at 4 "C and used as a stock standard. Working standards in the range 50-500 pg 1-1 were prepared each day by diluting the stock solution with toluene. The authentic PFP-carbamate derivative was synthesized as follows: 1 g of the 2,4-TDC was treated with 2.6 g of PFPA in 30 ml of toluene at room temperature overnight. The molar ratio of carbamate to PFPA was 1 : 2 .The solvent was evaporated under vacuum and the solid was separated by filtration and recrystallized from methanol by precipitation with ice-water to obtain 1.6 g (77% yield) of the PFP- carbamate. The synthesized derivative was used as a calibra- tion standard to optimize the derivatization reaction condi- tions. Synthesis of Internal Standard (IS) The deuteriated 2,4-TDC-d8 IS was prepared by adding dropwise, with stirring, 0.5 of TDI to 10 ml of anhydrous methanol-d4. The solution was warmed in a water-bath at 60-70°C for 5 min. The deuteriated IS was purified by recrystallization from carbon tetrachloride. The product had a m.p. 180-183°C and its EI-MS showed ions at rnlz 246 (M+) and mlz 210 (M-CD30D)+ and its PCI-MS at mlz 247 (M + l ) + , mlz 275 (M + 29)+ and mlz 287 (M + 41)+.The purity of the synthesized IS was determined by GC-MS-SIM in the EI mode by using the molecular ions at mlz 238 and 246 for 2,4-TDC-do and 2,4-TDC-d8, respectively. The data showed that the deuteriated IS contained less than 1.5% of the corresponding non-deuteriated compound. The integrity of the deuteriated IS after the extraction and derivatization procedure was established by NICI GC-MS of the corresponding PFP-carbamate derivatives through moni- toring of the ions at mlz 383 for PFP-carbamate-do and mlz 389 for PFP-carbamate-d6. The deuteriated standard con- tained less than 3% of the corresponding non-deuteriated species, showing that the conditions for the extraction and heating during the derivatization procedure did not affect the deuterium distribution of the IS.The IS solution was prepared by dissolving 10 mg of purified 2,4-TDC-d8 in 100 ml of methanol. It was stored at 4°C and used as a stock standard. Sample Collection A foam sample (approximately 1.0-2.0 g) was cut into small pieces at the manufacturing plant, submerged immediately into a glass jar containing 20 ml of methanol and dispatched to the laboratory, where the pieces were stored at 4°C prior to analysis. Preparation of Samples and Derivatization The foam sample, together with the remaining solvent, was placed in the mini-extractor. Methanol (20 ml) and 20 p1 of the IS solution, having a concentration of 10 Lig ml-1, were added. After 4 h, the extract was transferred quantitatively into a 50 ml round-bottomed flask.The solvent was evaporated under vacuum and the foam was dried in the ambient air and weighed accurately. The dry residue from the extract was dissolved in 1.0 ml of toluene, and 20 pl of anhydrous pyridine and 50 p1 of PFPA were added. The mixture was heated at 70 "C for 1 h and extracted with 1.0 ml of the phosphate buffer (pH 7.0). The organic layer was transferred into a 1.8 ml vial and dried over anhydrous sodium sulfate. An aliquot of 2 pl was injected into the GC-MS system, which was operated in the NICI mode. Optimization of the Conditions for Derivatization The PFP-carbamate derivatization was optimized as a func- tion of the reaction time. The PFPA (50 pl) was added to 1 ml of a solution of 2,4-TDC (20 pg ml-1) in toluene, and the derivatized product was injected into a GC-TSD system at intervals of 10, 20, 40, 60 and 90 min.The peak areas of the PFP-TDC formed were measured and compared with the peak areas obtained with authentic standards to calculate the percentage yield. GC-MS Procedure The fused-silica capillary column (DB-5,30 m x 0.32 mm i.d., 1.0 pm film thickness, J & W Scientific, Rancho Cordova, CA, USA) was linked directly to the ion source of an HP mass spectrometer set up in the NICI mode with methane as the reagent gas. The transfer line was maintained at 250 "C and the ion source was at 150 "C. The electron multiplier was operated in the range 1500-2000 V to maintain adequate sensitivity. The ion source was tuned each day for maximum sensitivity, using ions at mlz 414, 452, 557 obtained from perfluoro- tributylamine used as a reference compound.The acquisition rate for mass-spectra scans over the range 200-850 u was 1 s cycle-'. The GC conditions were as follows: the injector temperature was 250 "C, and the oven temperature was kept at 80°C for 1 min and then programmed from 80 to 250°C at 10 "C min-1. Helium was used as the carrier gas at a flow rate of 30 ml min-1 and the split-vent activation time was 0.5 min. The injection volume was 2 p1. The peak areas of the ions at mlz 389 for the internal standard and at mlz 383 for the PFP-carbamate derivative, were measured and their ratio was used for calculating the results. Extraction Efficiency A blank polyurethane foam was prepared by cutting foam into small pieces and extracting these with methanol for 24 h to remove any residual TDI or other interfering contaminants.The dry blank foam samples were weighed accurately, treated with 50, 100 and 200 ng g-1 of 2,4-TDI in toluene solution, soaked immediately in 25 ml of methanol and then extracted for 4 h. The same solution was used for the preparation of calibration standards by adding known amounts of TDI to 1.0 ml of methanol. The derivatization procedure was used for both the samples and standards. Results and Discussion Derivatization Reaction The isocyanate moiety reacts readily with molecules capable of a hydrogen exchange, such as alcohols, resulting in the formation of esters of carbamic acids, referred to frequently as 'carbamate' or 'urethane'. Free TDI reacts instantaneously with excess methanol to yield the corresponding 2,4-TDC.Carbamates have been determined as derivatives of the intact molecule as well as derivatives of the products resulting from hydrolysis. 17 For the determination of free isocyanates in flexible foams, pentafluoropropylation of the -NHR group in the carbamate was carried out with PFPA. A stable PFPANALYST, FEBRUARY 1993. VOL. 118 25 000 151 1383 - ( a ) derivative was formed that was suitable for trace analysis by GC-MS: 0 R-N=C=O + CH30H -+ R-NH-C-OCH3 (1) 0 0 0 0 R-NH-C-OCH3 + CZFs-C -0-C-CzFS -+ R-N-C-OCH3 II I1 II ll I C2Fs-C /I 0 + C~FS-COOH (2) The presence of two pentafluoropropyl groups in the deriva- tive enhanced the NICI sensitivity. Addition of pyridine as a catalyst resulted in a quantitative recovery.The best solvent for the formation of the PFP derivative of 2,4-TDC was selected after comparison of the peak heights of the derivative formed when using different solvents, while keeping all other reaction conditions the same. Of all the solvents tested, i.e., 2,2,4-trimethylpentane, benzene, chloro- benzene and toluene, the last showed the highest yield. The derivatization reaction was quantitative and complete after 1 h. Mass-spectral Characteristics of the Derivatives The NICI mode in SIM is often more suitable than the PCI or EI mode for the quantitative analysis for organic species in complex matrices.18 The NICI mass spectra of the PFP derivative of the IS and 2,4-TDC are shown in Fig. 1. The NICI spectrum for PFP-TDC [Fig. l(a)] furnished primarily an ion at m l z 383, which represents (M-C2FSCO)-.Other discernible ions were too small to be useful for analytical 20 000 15 000 10 000 I v) 4- ' E 5000 3 2 e .- 2 0 - a, m 2 18000 16000 Q 14000 12 000 10 000 8000 6000 4000 2000 0 3 a 350 I , , I , " : , 300 320 340 360 380 389 400 420 440 460 480 500 rnlz Fig. 1 Methane NICI mass spectra of the pentafluoropropyl derivative of (a) 2,4-TDC and ( b ) the internal standard at an ion source temperature of 150 "C purposes. The NICI spectrum of PFP-TDC-dc, (IS) [Fig. 1(6)] yielded a significant peak at mlz 389 corresponding to the loss of the C2F5CO- group from the molecular anion. Conse- quently, two ions, i.e., at m/z 389 for the IS derivative and at mlz 383 for the PFP-carbamate derivative, were monitored. The peak-area ratio of m/z 383 to 389 (IS) was calculated to obtain the calibration graph for TDI.Typical selected-ion chromatograms are shown in Fig. 2. As commercially avail- able TDI usually contains 80% of the 2,4- and 20% of the 2,6-diisocyanate isomer, both derivatives were present and well separated. Retention times for the 2,6- and 2,4-TDI derivatives were 13.43 and 13.65 min, respectively. The sum of the areas for both isomers were used for the calculation. No interfering peaks were observed. Precision and Linearity Calibration graphs for PFP-carbamate derivatives were linear from 20 to 500 ng g-l with a correlation coefficient of 0.9954. Foam samples with TDI values exceeding 500 ng g-1 were diluted and re-analysed. The over-all precision and accuracy of the method was tested by analysing prepared foam samples spiked at three concentrations.The results, summarized in Table 1, display an acceptable level of accuracy and precision, with relative standard deviations ranging between 4.3 and 8.5%. The least amount of TDI that could be detected after the above derivatization reaction was 10 ng g-' for a 1 g sample (signal-to-noise ratio = 3). Recovery The efficiency of the extraction procedure was ascertained by comparing the peak areas obtained after processing the 'blank' foam samples, spiked with known amounts of TDI, 13 14 15 Time/min Fig. 2 Selected-ion chromatograms obtained by NICI GC-MS from a polyurethane foam sample containing 225 ng g- of TDI (at rnlz 383) and 200 ng g-1 of the internal standard (at rnlz 389) after the derivatization procedure Table 1 Precision for free TDI in foam samples Concentrationlng g- I Parameter 20 so 100 Mean 18.50 47.20 89.40 Standard deviation 1.579 2.974 3.827 Relative standard Number of replicates 6 6 6 deviation (%) 8.50 6.30 4.30152 ANALYST, FEBRUARY 1993, VOL.118 Table 2 Recovery of spiked TDI from polyurethane foam samples Fortification level/ Recovery* ng g-' ("/.I RSDi (%) 50 66 7.86 100 82 6.64 200 88 4.35 * Mean values. j. RSD = relative standard deviation ( n = 6). with the peak areas for the corresponding standards. The recoveries of TDI from spiked foam 'blanks' ranged from 66 to 88% at 50-200 ng 8-1 levels and averaged 79%, as shown in Table 2. The extraction was complete after 4 h. As the IS was added to the matrix and showed stability under the extraction and derivatization conditions, no correction was needed for the extraction efficiency in the proposed method.Analysis of Fresh Foam Samples The above analytical procedure was applied to the analysis for free TDI in freshly poured polyurethane samples, which were analysed after storage for 1 and 24 h. The TDI content of the fresh samples ranged from 1.2 to 2.4 yg 8-1 when analysed after 1 h. The same foam samples did not show a detectable concentration of free TDI after storage for 24 h. Conclusions A simple, highly sensitive and specific GC-MS method has been developed for the determination of free TDI in flexible polyurethane foams. The free isocyanates were converted into stable carbamates with methanol and further derivatized with PFPA reagent.The conditions for derivatization with PFPA were optimized by varying reaction temperature, reaction time and solvent. With this method, interference from other foam constituents is eliminated and a reliable and fairly precise tool for the determination of trace levels of TDI in foam samples is provided. The authors thank Dr. M. A. Nazar, Chief Scientist, Occupational Health Laboratory, for helpful suggestions and discussions during the preparation of this manuscript. Special thanks are due to Dr. W. S. Wu, Research Scientist, Occupational Health Laboratory, for discussions and sugges- tions during the course of this work. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Ulrich, H., in Kirk-Othmer Encyclopedia of Chemical Technol- ogy, ed. Grayson, M., and Eckroth, D., Wiley-Interscience, New York, 3rd edn., 1982, vol. 23, pp. 576-607. Marcali, K., Anal. Chem., 1957, 29, 552. Meddle, D. W., and Wood, R., Analyst, 1970, 95, 402. Purnell, C. J . , and Walker, R. F., Analyst, 1985, 110, 893. Skarping, G., Sango, C., and Smith, B. E. F., J . Chromatogr., 1981, 208, 313. Dalene, M., Mathiasson, L., Skarping, G., Sango, C., and Sandstrom, J. F., J . Chromatogr., 1988, 435, 469. Brorson, T., Sango, C., and Skarping, G., Znt. J . Environ. Anal. Chem., 1990, 38, 399. Keller, J., and Sandridge, R. L., Anal. Chem., 1979, 51, 1868. Hardy, H. L., and Walker, R. F., Analyst, 1979, 104, 890. Bagon, D. A., Warwick, C. J . , and Brown, R. H., Am. Ind. Hyg. Assoc. J . , 1984, 45, 39. Wu, W. S . , and Gaind, V. S . , Analyst, 1992, 117, 9. Demianczuk, D., Prace Inst. Przem. Org.., 1973, 4, 55. Kubitz, K. A., Anal. Chem., 1957, 29, 814. Tompa, A. S . , Anal. Chem., 1972, 44, 1056. Conte, A., and Cossi, G., J. Chromatogr., 1981, 213, 162. Rastogi, S. C., Chromatographia, 1989, 28, 15. Cochrane, W. P., J . Chromatogr. Sci., 1979, 17, 124. Stilwell, W. G., Bryant, M. S . , and Wishnok, J. S . , Biomed. Environ. Mass Spectrom., 1987, 14, 221. Paper 2103993A Received July 27, 1992 Accepted September 15, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800149

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1993, VOL. 118 153 Determination of Non-ionic Surfactants in Waste Water by Direct Extraction With Fourier Transform Infrared Spectroscopic Detection B. E. Andrew BHP Sheet & Coil Products Division, Private Bag 7, Hastings 3975, Australia Non-ionic surfactants in waste water can be extracted and concentrated in dichloromethane using a simple salting-out procedure. The characteristic C-0-C stretch vibration at 11 10 cm-1 is then utilized to measure surfactant concentration a t levels ranging from 0.1 to 50 mg 1-1. Relative standard deviations vary from 15% at the 10 mg 1-1 level for a single-step extraction, to 5% a t the 1 mg 1-1 level for multiple extraction followed by evaporation and reconstitution. Soaps are the only major interference found for the technique.Comparison of this procedure with the commonly used cobaltothiocyanate method indicates that the combination of direct extraction and Fourier transform infrared detection is more accurate in determining non-ionic surfactants in aqueous solution from the viewpoint of recovery. Various levels of response were observed when analysing samples of different waste waters. The procedure is significantly more rapid when using the single-step extraction and is easily calibrated. Keywords: Non-ionic surfactant; direct extraction; Fourier transform infrared spectroscopy Various procedures have been described for the determi- nation of non-ionic surfactants in waste water and treated effluents. Otsuki and Shiraishi' have determined trace levels of alkylphenol ethoxylates in water using reversed-phase liquid chromatography and mass spectrometric detection.Ventura et a1.2 described the determination of ethoxylate surfactants and their brominated derivatives by gas chroma- tography-fast atom bombardment mass spectrometry. Flow injection has been used by Leon-Gonzalez et al.3 for Triton- type surfactants based on hypsochromic shifts and hyper- chromic effects of trace amounts of such substances on Alizarin Fluorine Blue. Such procedures, however, may not be available to process control laboratories, and hence the determination of non-ionic surfactants in waste water and treated effluents is most commonly carried out using the colorimetric procedure known as the cobaltothiocyanate (CTAS) method." This procedure is time consuming and may be subject to severe interference from other organic compounds which are not non-ionic surfactants.Ion-exchange procedures to remove ionic surfactant interferences would only be employed as a last resort owing to the lengthy analysis time required. In addition, the salting-out steps which form part of complex extraction procedures are inherently messy because full saturation of the aqueous phase is usually specified. The content of surfactant is commonly referenced against a standard non-ionic surfactant material, but the CTAS method is insensitive to non-ionic surfactants containing less than 5 mol of ethoxylate (EO). Ethoxylate compounds form the major proportion of commercial non-ionic surfactants. These materials are clearly identifiable on the basis of their infrared (IR) spectra, with the broad C-0-C stretch vibration at 1110 cm-1 being the most distinctive absorbance band.Another property common to non-ionic surfactants is their solubility in chlorinated solvents. It was therefore considered that combination of these proper- ties warranted investigation as an analytical procedure for determining non-ionic surfactants in waste water and treated effluent. Preliminary tests indicated that dichloromethane was a suitable solvent for non-ionic surfactants and is IR trans- parent at 1110 cm-1. Experimental Equipment A Perkin-Elmer Model 1710 Fourier transform infrared (FTIR) spectrometer was used in conjunction with a Specac fixed-path AgCl cell of 2 mm pathlength. Extraction separat- ing funnels were equipped with Teflon stopcocks. Reagents and Standards Analytical-reagent grade dichloromethane* and sodium chloride were used.Other reagents used were reagent-grade phenol, poly- (ethylene glycol) ( M , = 400), cetyltrimethylammonium bro- mide, sodium dodecyl benzenesulfonate, pentane- 1 ,S-diol and sodium lauryl sulfate. Technical-grade sodium lignin sulfonate and sodium oleate were also used. Recovery studies of commercial non-ionic surfactants were based on preparation of the surfactants in aqueous solution on an as-received basis. Surfactant purity was generally quoted by the manufacturer in the range 95-99%. Non-ionic surfactant stock solution was prepared by dissol- ving Nonidet P40 (phenol ethoxylate) (Merck, formerly BDH) in dichloromethane at 5 g 1-1.This solution was then diluted in dichloromethane on a volume-volume basis for calibration solutions of 0.1, 0.2, 0.4 and 0.8 g 1-1. Procedure A 300 ml aliquot of sample was placed in a SO0 ml separating funnel. Concentrated hydrochloric acid (5 ml) and sodium chloride (40 g) were added and shaken to dissolve. Extraction into dichloromethane (10 ml) was carried out. The dichloro- methane phase was filtered into a glass screw-cap vial and refrigerated if measurements were not made within 4 h. The sample or standard solutions were transferred into a 2 mm pathlength sodium chloride or silver chloride cell and the IR absorbance was measured at 1110 and 990 cm-1. The absorbance at 1110 cm-1 was corrected for the absorbance at 990 cm-1 and any absorbance due to the dichloromethane or trace contamination from reagents or filters.It has been observed that some real-life extracts may generate absorbance * If segregated from other solvents, dichloromethane wastes from this analytical procedure can be retained for redistillation until a suitable volume (1-2 1) has been accumulated. The waste solvent can be redistilled over 50 ml of 50% sodium hydroxide solution. Approxi- mately 80% of the waste solvent volume should be collected for reuse, the balance being distilled back into the waste container until water begins to distil over. The redistilled solvent is equivalent in purity to fresh solvent for use in this analytical procedure.154 320 - 240 8 2 8 - ’ 160 U 80 0 ANALYST, FEBKUAKY 1993, VOL. 118 - - - - - maxima which vary slightly from the specified 1110 cm-! wavenumber. Improved precision at lower levels was obtained by extract- ing 600ml of sample three times with 10 ml aliquots of dichloromethane, using adjusted proportions of acid and salt.The combined extracts were evaporated to dryness and the residue was redissolved in 2 ml of dichloromethane. Results and Discussion The advent of commercial FTIR systems and their associated data acquisition software has significantly expanded the sensitivity of analysis of species in solution. Concentration of non-ionic surfactant into small volumes of dichloromethane from a larger aqueous sample volume brings the mg 1-* levels present in the sample up to g 1-1 levels in dichloromethane. A linear calibration can be obtained in the mg 1-1 range either by direct dissolution of the appropriate amount of standard surfactant in dichloromethane, or by extraction from an aqueous solution and concentration into dichloromethane solution.No significant difference is observed between the two modes of calibration, provided that the standard surfac- tant is prepared from a freshly opened container. Initial tests indicated differences in calibration, which were traced to the hygroscopic nature of the surfactant. Teric LA8 (fatty alcohol ethoxylate) was chosen as the principal compound of study, because this type of non-ionic surfactant is commonly used in emulsions and cleaners at the local worksite. Using the direct extraction procedure at the 10mg1-1 level in water, a mean recovery of 110% was obtained with a standard deviation of 1.2 mg 1-I (n = 12).When spiked into a mill water sample containing 8.6 mg 1-I of unknown non-ionic surfactants, a recovery of 105% was obtained for a 5 mg 1-1 spike of Teric LA8. Alternatively, the evaporative procedure for higher precision was evaluated using a sample containing 1 mg 1-1 of Teric LA8. A mean value of 1.00 mg 1-1 (100% recovery) was found, with a standard deviation of 0.05 mg 1-1 ( n = 6). It would therefore appear that the technique is suitable for rapid analysis of process water systems in the 2-50 mg 1-1 range, or, by use of a more time-consuming procedure, for lower levels of non-ionic surfactant, in the range 0.1-2 mg 1-l. Because the measurement principle is based on a specific IR absorbance after extraction, the interference generated by other co-extracted species which could plausibly be present in the waste or treated effluent stream requires examination. Fig, 1 demonstrates that no interference is generated by anionic species such as alkyl sulfates or sulfonates.Crisps has noted that anionic surfactants interfere significantly in the CTAS method. Interference from glycols only becomes 1 2 3 4 5 6 7 I nte rfe rence Fig. 1 Effect of interferences on recovery: 10mg1-l addition of interferent to 10 mg 1-1 Teric LA8. 1, Linear alkyl sulfate: 2, dodecyl benzenesulfonate; 3, phenol; 4, sodium oleate: 5, lignin sulfonate; 6, poly(ethy1ene glycol) ( M , = 400); and 7, cetyltrimethylammonium bromide. 0, Interference and ., Teric LA8 + spike measurable with long-chain materials. The most significant interference is generated by soaps; a 10 mg 1-1 addition of sodium oleate to 10 mg 1-1 Teric LA8 increased recovery to 310%.Of itself, a blank addition of 10 mg 1-1 sodium oleate showed recovery at 1110 cm-1 of 170%. Unfortunately, as Fig. 2 demonstrates, attempts to remove this interference by various pre-treatments are not very successful. It should be noted, however, that the presence of interfering soaps can be readily detected by the presence of the C=O stretch absor- bance band at 1710 cm-1. Extraction using a neutral salt solution in preference to the acidic salt extraction tends to decrease the level of interference from soaps. In the CTAS procedure, it is normal practice to ensure that solutions are fully saturated with salt prior to extraction with benzene.This is inconvenient as the salt takes time to dissolve, and also tends to obstruct or jam the stopcock of a separating funnel. Studies using various salt levels have indicated that the level of salt used in the dichloromethane procedure is sufficient to ensure full extraction of all the compounds studied, removing the inconveniences associated with full salt saturation. In assessing the effectiveness of this procedure in compari- son with the commonly used CTAS method, a number of non-ionic surfactants were studied for recovery by both procedures, as shown in Table 1. For most commercial non-iont s surfactants higher recoveries are obtained with the IR procedure than with the CTAS procedure. Some materials (e.g. , coconut ethanolamide), which are not classifiable as non-ionics on the basis of the presence of an alkoxylate group, show a positive response to the CTAS procedure.Conversely, one recognized non-ionic surfactant (Genapol PF10) ex- hibited zero recovery with the CTAS method. It therefore appears reasonable to state that the 1R method offers improved accuracy on the basis of more meaningful recoveries from a range of commercial non-ionic surfactants, when compared with recoveries obtained with the CTAS method. The procedure has been used on pilot process samples where non-ionic surfactants co-exist with small amounts of emulsified oil (20-30 mg 1-1). At these levels, spiked recoveries of non-ionic surfactant indicate that interference is minimal, At higher levels ( 0 . 5 1 g 1-1 oil) a preliminary acid-splitting procedure is necessary, followed by pre-extrac- tion with hexane to remove the bulk of mineral oil present.Some non-ionic surfactants may be removed in the hexane phase. For samples with reasonable freedom from suspended or emulsified solids, such as cooling water or treated effluent, spiked recoveries indicate acceptable precision and accuracy with respect to the standard surfactant compound used. The speed of the single-step extraction procedure increases the utility of the test in process control situations. A comparison of the CTAS method and this procedure has been made with samples of waste water from various sources, 1 2 3 Extraction procedure 4 Fig. 2 Effect of attempted soap interference removal on recovery. 1, Standard procedure: 2, hexane pre-extract; 3.anion exchange: and 4, neutral salt extract. 0, soap blank; and ., 10 mg 1-1 Teric LA8ANALYST, FEBRUARY 1993, VOL. 118 155 Table 1 Comparison of method recoveries (based on recovery from 10 mg 1-1 aqueous solution of the surfactant, Nonidet P40 standard) Trade name Brij 35 Crillon CDY Genapol PFlO Hercules AR 150 Miranol L2M-57 Nonidet P40 Nonidet P42 Teric 18 M10 Teric 127 Teric C12 Teric CME7 Teric GN5 Teric GN9 Teric LA8 Teric N5 Teric OF6 Teric T2 Teric X11 Triton CF32 Triton X-100 Description Polyoxyethylene lauryl ether Coconut ethanolamide Ethoxylate/propoxylate Polyethylene glycol ester Disodium dicarboxylated Octyl phenol ethoxylate Octyl phenol ethoxylate Octadecylamine ethoxylate Fatty acid ethoxylate Castor oil ethoxylate Coconut monoethanolamide Nonyl phenol ethoxylate Nonyl phenol ethoxylate Fatty alcohol ethoxylate Alkyl phenol ethoxylate Oleic acid ethoxylate Tall oil ethoxylate Octyl phenol ethoxylate Amine polyglycol condensate Octyl phenol ethoxylate imidazoline ethoxylate Recovery (% ) CTAS FTIR 50 105 20 0 0 190 60 95 35 0 100 110 25 30 15 85 20 45 10 110 20 65 45 95 65 110 45 110 165 40 35 70 55 140 45 150 15 125 50 140 as shown in Fig.3. Results for each source represent the average of five samples taken once per week. It is evident that in several waste streams the levels of non-ionic surfactants determined by the CTAS method are significantly lower than by the extraction-FTIR method; none of the waste streams measured contained detectable soap/fatty acid interferences.It was noted that samples from the low flow aerated sewage plant required filtration and a higher level of salt saturation to achieve separation of the dichloromethane phase; this aqueous material was also more highly coloured. The data appear to conform with earlier comparisons (Table 1) of the relative recoveries of commercial surfactants. Conclusions Extraction of ethoxylate-type non-ionic surfactants, and subsequent measurement using FTIR, is competitive with the current CTAS method in terms of analysis speed and precision. It has been demonstrated that the recovery of a range of commercial non-ionic surfactants by the FTIR method (and hence accuracy) is better than recovery by the CTAS method. By sacrificing some analytical precision, the 1 2 Process [I 4 Fig.3 CTAS and extraction FTIR of various waste waters. 1, Steel mill process cooling water; 2, tertiary treatment plant effluent; 3 , high flow aerated sewage effluent; and 4, low flow aerated sewage effluent. 0, CTAS method; and U, extraction FTIR speed of the FTIR procedure can be accelerated to enable implementation of routine process monitoring of plant streams and treated effluent. Although interferences have not been extensively evaluated, it would appear that most interferences in the CTAS procedure have no influence on the proposed FTIR method; interference from soaps in FTIR measurement is readily detected. Study of four types of waste water has shown that the level of non-ionic surfactants present may be significantly underestimated by the CTAS procedure, assuming that positive interferences have been excluded from the FTIR method. The author thanks the management of BHP Steel, Sheet & Coil Products Division, for permission to publish this paper; also A. Gaetano, who determined comparative recoveries of non-ionic surfactants by the CTAS method. References 1 Otsuki, A., and Shiraishi, H., Anal. Chem., 1979, 51, 2329. 2 Ventura, F., Figueras, A., Caixach, J., Espadaler, I., Romero, J., Guardiola, J., and Rivera, J., Water Res., 1988, 22, 1211. 3 Leon-GonzBlez, M. E., Santos-Delgado, M. J., and Polo-Diez, L. M., Analyst, 1990, 115, 609. 4 Standard Methods for the Examination of Water and Wastewater, American Public Health Association, American Water Works Association, Water Pollution Control Federation, 17th edn., Crisp, P. T., in Nonionic Surfactants, Chemical Analysis, ed. Crisp, P. T., Marcel Dekker, New York, 1987, ch. 3 . 1989, pp. 5-64. 5 Paper 21042486 Received August 6, 1992 Accepted October 26, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800153

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1993, VOL. 118 157 Evaluation of Thermal Lens Detection in Flow Injection With High Background Absorbance A. Chartier, C. G. Fox and J. Georges" Laboratoire des Sciences Analytiques, U.A. 435, Universite Claude-Bernard-L yon I, 43 boulevard du I I Novembre 1918, F69622 Villeurbanne Cedex, France This work describes the use of thermal lens spectrometry as a detection technique for the determination of terbium with 4-(2-pyridylazo)resorcinol by flow injection. The results, obtained for aqueous and aqueous methanolic solutions, are compared with those previously obtained with a conventional spectrophotometric detector. They are discussed with respect to the dependence of the thermal lens effect on the flow, the solvent composition and the background absorbance. Keywords: Now injection; thermal lens detection; influence of the background absorbance; solvent effect Thermal lens spectrometry, which is based on the formation of a refractive-index gradient in a sample following the absorp- tion of a Gaussian beam, is known as a sensitive method for determining weak absorbances in liquids or gases.This aspect has been well described in many papers and reviews.1-6 Thermal lens spectrometry is considered to be a sensitive detection technique for flow injection7-1(' and liquid chromato- graphy. 11-17 In particular, the combination of thermal lens detection and flow injection should be mutually beneficial; the small volumes used in analysis and the need for sensitivity in detection are well matched to the characteristics of laser-based methods.Nevertheless, thermal lens detection in flowing systems requires some considerations on the additional effect of the flow on the thermal lens signal. This effect has been discussed previously and has been shown to depend on the chopping frequency, the flow rate, the geometry of the flow cell and the configuration of flow direction with respect to beam propagation. 18 For this purpose, a dual-beam configuration with continu- ous-wave (CW) laser excitation is generally used. The excitation beam is modulated with a chopper and the probe beam is de-modulated with a lock-in amplifier, which improves the sensitivity and the precision of the measurement. Following the onset of illumination in a chopper cycle, the thermal lens signal increases to a steady state that is reached only when the rate of heat input from the excitation laser is equal to the rate of heat conduction out of the illuminated region.The temporal response of the thermal lens signal is expressed in a simplified form by S ( t ) = 2.3 E A [l + (tc/t)]-l where A is the decadic absorbance, and E is the enhancement factor that represents the sensitivity of thermal lens spec- trometry relative to Beer's law; E is given by P( dnld 7') E=- 1.91hk where P i s the power of the pump beam, h is the wavelength of the probe laser, dn/dTis the refractive-index gradient, and k is the thermal conductivity of the sample. As shown in eqn. (l), the formation of the thermal lens depends on a characteristic signal rise time constant, t,, which represents the response time of the medium to heat input: 0 2 p C, A 1 tc = ( 3 ) 4 K where cr) is the radius of the excitation beam into the sample, and p, C, and k are the density, the specific heat and the * To whom correspondcncc should bc addressed.thermal conductivity of the solvent, respectively. The steady- state conditions are achieved for times greater than 10-20 tc. Previous studies's20 have shown that the flow of the analyte through the flow cell could lead to a significant decrease in the thermal lens signal. This occurs especially when the residence time of the heated sample in the cell is short with respect to the characteristic time constant t, or to the chopping period. Flow effects are generally divided into two components: bulk flow and mixing. Bulk flow includes axial and transverse flows and results in a movement of the heated sample in the flow direction, parallel to the axis of the cell.Mixing, caused by turbulence within the cell, has the effect of mixing the heated material with the surrounding solution in a section of the flow. The influence of the flow can be reduced by increasing the sampling frequency, but to the detriment of the signal amplitude. Tt has been shown18 that, with a cylindrical channel cell with flow parallel to the beam axis, the signal is mainly degraded by turbulence and mixing. However, this effect is small and the thermal lens signal is only degraded to a small extent at flow rates suitable for most of the applications in liquid chromato- graphy or flow injection. The purpose of this work was to study the thermal lens signal in a flow injection system where the absorbance of one of the reagents contributed to a high background and was the main source of noise.Thermal lens detection was applied to the determination of terbium with 4-(2-pyridylazo)resorcinol (PAR) in water and a water-methanol mixture. The limits of detection were compared with those obtained by conventional spectrophotometric detection and with those expected from the theory. Different parameters such as the chopping frequency, the role of the medium, and the influence of the background on the signal intensity were studied. Experimental Thermal Lens Detection The double-beam experimental set-up is depicted in Fig. 1. The excitation radiation was provided by an air-cooled argon ion laser (Ion Laser Technology, Model 5490A) operated in the single line mode with a maximum power of 40 mW at 514 nm. For a better signal-to-noise ratio, the laser was used at half its maximum output power, i.e., 20 mW.The beam diameter was 0.65 mm and the divergence (0) was 0.95 mrad. A half-wave mica retardation plate (Melles Griot, 02 WRM 023) was used to allow the plane of polarization of the laser to be rotated so that the excitation beam was mainly transmitted by the polarizing beam splitter and not reflected. The CW laser beam was then modulated by a variable-frequency optical chopper (Scitec, Model 300). The probe beam was158 ANALYST, FEBRUARY 1993, VOL. 118 m-r”. Helium _I_ L1 A I Fm ......................... . . . . . . . . . 1 . . . . . . .. . . . . \ . . . . . ,......... . . . . . . . . . . 1 . ............ . . . . . . . . . . . 0 1-q p D I RP Ch Lz BS FC Scope =i Fig. 1 Schematic diagram of the experimental set-up. M = Mirror; LI, L2 = focusing lenses: RP = retardation plate; Ch = chopper: BS = beam splitter; FC = flow cell; F = filter: and PD = photodiode supplied by a 6.2 mW helium-neon laser (Optilas, Model H7500P) at 632.8 nm. The excitation and the probe beams were focused independently by two lenses (L2 = 125 mm; L1 = 80 mm). Both beams were then recombined and made co-linear by means of a polarizing beam splitter held on an x-y-z translator. In order to avoid saturation of the photo- diode, the laser power was lowered by rotating the laser head. A mode mismatched configuration of the pump and probe beams was used.With this arrangement, the sensitivity of the method depends on the positioning of the pump and probe beam waists with respect to the sample cell and also on the relative diameter of both beams.21 The optimal position, ensuring maximized sensitivity together with linear calibration graphs, was obtained experimentally, with the cell positioned near the waist of the excitation beam, by translating lens 1 (probe beam) and monitoring the thermal lens signal. The sample cell was a 12 mm3, 1.4 mm i.d. tubular cell with a pathlength of 8 mm (Kratos 2900-0146). The cell axis and, therefore, the sample stream were parallel to the laser beam. After the flow cell, the excitation beam was blocked by a band-pass interference filter. The sample cell and interference filter were tilted slightly to avoid interference effects resulting from back-reflection.The thermal lens signal was detected through a 1 mm pinhole with a silicon photodiode operated in the photovoltaic mode. Its output was passed via a 10 kQ resistor to a lock-in amplifier (Brookdeal, Model 9503). The reference signal for the lock-in amplifier was generated within the chopper head. The time constant and the de-modulator mode of the lock-in amplifier, together with the chopper frequency, were varied in a series of experiments in order to find the best compromise between the signal- to-noise ratio and the peak resolution. The combination used was a frequency of 40 Hz and a time constant of 1 s. Spectrophotometric Detection The conventional spectrophotometric detection was carried out with a high-performance liquid chromatographic detector (Shimadzu SPD6A) operating in the visible region and equipped with a 8 rnm3,l mm i.d.cylindrical channel flow cell with a pathlength of 10 mm. Flow Injection Apparatus The dual-line flow injection system has been previously optimized22 for complex formation of lanthanides with PAR. The two carriers (NaCI-HCI and PAR) were pumped from two reservoirs by a peristaltic pump. The Teflon, low-pres- sure, six-port injection valve (Bioblock) was used to introduce a known amount of terbium into the NaCI-HCI stream. The valve was equipped with a 40 mm3 sample loop. The two carrier streams were mixed at a Teflon T-piece with an angle of intersection of 30°, which provided the best mixing in terms of peak shape and background noise.The output of the 1-piece led directly to a 1 m (0.3 mm i.d.) knitted Teflon tube reactor. The extremity of the reactor was connected to the flow cell. The experiments were carried out at a flow rate of 0.730 cm3 min- I . The PAR-Tb complex was detected at 514 nm. Reagents The 7.5 x rnol dm-3 PAR (Merck) carrier solution was prepared each day by diluting a stock solution ( 5 x 10-3 rnol dm-3 in distilled, de-ionized water) with buffer solution at pH 9.7. This buffer was prepared from 200 cm3 of 2.5 x 10-2 rnol dm-3 sodium tetraborate (Merck), together with a sufficient volume of 1 x 10-1 rnol dm-3 NaOH (Prolabo) to obtain the required pH. The other carrier solution was an equal volume mixture of 2 X 10-2 rnol dm-3 NaCl (Prolabo) and 2 X 10-3 rnol dm-3 HCI (Riedel-de-Haen).Both these solutions were prepared with distilled, de-ionized water. For the experiments carried out with water-methanol mixtures, the pH was adjusted to 9.5 for the aqueous buffer solution, and methanol was then added to both carrier solutions. The carrier solutions were filtered, then de-gassed in an ultrasonic bath for 5 min. The stock solution (30 cm3) of terbium, at a concentration of 1 X 10-3 mol dm-3, was prepared from the chloride (Aldrich) using the carrier. A concentration range of from 1 X to 1 X 10-7 rnol dm-3 lanthanide was prepared by successive dilutions of the stock solution with the carrier. Stock solutions of cobalt, with an absorbance of 0.1, were prepared from cobalt nitrate (Prol- abo) in water or in water-methanol mixtures.Appropriate absorbance values were then obtained by successive dilutions. Methods The results were obtained by recording the average peak height when four samples of each concentration were injected. It was noted that, when the valve was turned to inject, there was a momentary flow perturbation, leading to a small change in the mixing of the PAR reagent at the T-piece. As PAR absorbs strongly at the detection wavelength (514 nm), even a minute change in PAR concentration induced some variation in the background. A blank of NaCI-HCl was therefore injected and the signal height was subtracted from that obtained when the sample was injected. The limits of detection (LOD) were defined as the sample concentration yielding a peak height twice (i.e., 100) the background noise.For static experiments, the LOD was that generally defined, i.e., three times (30) the standard deviation of the back- ground. The dependence of the enhancement factor on solvent composition and on chopping frequency was studied in static and flowing conditions using samples of cobalt nitrate. The flowing conditions were exactly the same as for the determina- tion of terbium, the injected sample and the reagent carrier being replaced by cobalt nitrate in the same medium. The performance of the detection method in static and flowing conditions was compared considering that, in flowing condi- tions with a dual-line system, the sample concentration in the detection cell is smaller than that injected. Sample dilution ( x 1.7) at the T-piece and sample dispersion ( x 1.5) within the flow system between the injection valve and the flow cell contributed to an over-all decrease of the thermal lens signal of more than 60% with respect to that obtained in static conditions. Therefore, in order to compare our results in static and flowing conditions properly, thc dctcction limits, expressed in absorbance, were measured at the top of the flow-injected sample peaks.Results and Discussion The calibration graphs for terbium in water and water- methanol are shown in Fig. 2. The working range extendsANALYST, FEBRUARY 1993, VOI,. 118 m 0 -0.5 4.5 5.5 6.5 7.5 - Log ([Tb3 + l/mol dm - 3 ) Calibration graphs for injcctcd terbium, in water (A) and in Fig. 2 water-methanol (B) with thermal lens dctcction ~~~~ ~ - ~ Table 1 Detection limits for terbium, with PAR, in water and water-mcthanol (40 + 60 v h ) .Values, expressed as absorbance, correspond to the absorbance at the top of thc peak and for the same optical pathlength (8 mm) Convcntional Thermal Icnh dctcction spectrophotometry22 Conccn- Conccn- tration/ t r a t i o n/ Carrier mol dm-? Absorbancc mol dm Absorbance Water 8.5 x 10-8 8.7 x 10-4 5 x 10 8 8.5 x 10-4 Methanol-water 2.4 x 1 0 - 7 5.6 x 10-4 1.2 x 10-7 6.5 x 10-4 beyond 1 x 10-5 mol dm-3, but only the linear part is considered in the figure. The dynamic range lies between 1 X 10-5 and 1 x 10-7 mol dm-3 and 1 x 10-5 and 2.5 X 10-7 mol dm-3 in water and water-methanol, respectively. The limits of detection expressed in concentration and by absor- bance are reported in Table 1 and compared with those previously obtained by means of a visible-range spectrophoto- metric detector.The results obtained in water show that the sensitivity of the detection is not improved by using the thermal lens method. This result is not completely surprising and can be explained by thc following reasons. First, the advantage of improved sensitivity can be outweighed by the disadvantage of the thermal lens method being more susceptible to variations in the experimental sct-up. These originate from irregularities in the beam profile and beam pointing stability, perturbations in the flow rate, and vibrations, small bubbles or suspended particles in the tlow stream; these factors contribute to an increase in the background noise.Second, the thermal lens signal could be degraded by the flow that removes the heatcd material from the probed region, inducing a loss in sensitivity. Finally, it is well known23 that the sensitivity of thermal lens spectrometry depcnds on the physical and optical properties of the solvent. With regard to these properties, water is a poor solvent for the generation o f a thermal lens. Initial optimization of the apparatus was performed on the background signal, using the same carrier solutions, but without any sample injection, by alignment of both beams, by adjustment of the optical components and of the photodiode, and by reduction of the vibrations and flow fluctuations. Once the optimum conditions, yielding the best signal-to-noise ratio, were fulfilled, the second and third points were considered. The dependence of the thermal lens signal on the flow rate is shown in Fig.3.18 These results were obtained with the same experimental configuration and the same flow cell. At low flow rates, and up to a critical value that depends on the 1 .o - 0.8 0 S m .- 0.6 > .- CI - 2 0.4 1 1 2 3 4 5 Flow rate/cm3 min-’ 0 Fig. 3 Dcpcndcnce of the thermal lens signal on the flow rate at constant frcqucncy. Chopping frcqucncy: A. 10; B, 20; and C, 80 Hz loo I 8o I 0 20 40 60 80 100 Methanol (% v/v) Fig. 4 Plot of the thermal lens signal amplitude VCYSUS thc volumc fraction of methanol in water-methanol solvent. 0. Flowing condi- tions; and 0, static conditions chopping frequency, the signal remains constant. At high flow rates, the signal decreases significantly, especially when the chopping frequency is low.It has been demonstrated18 that the decrease in signal intensity is not caused by the axial bulk flow, but mainly by turbulence within the cylindrical channel. However, except at the lowest chopping frequency, the flow rate used in this work should not affect the thermal lens signal. Although the sensitivity was greatly reduced, using a chopping frequency of 40 Hz eliminated any flow effect and allowed a good compromise between noise, signal amplitude and peak resolution. Attempts were then made to improve the sensitivity of the method by using aqueous organic or organic media. Prelimi- nary studies showed that it was possible to use water- methanol mixtures to form a complex of terbium with PAR.The maximum volume fraction of methanol allowed without degrading complex formation too much was set at 60% v/v. The calibration graph obtained with this medium was as good as that obtained with pure water with almost the same dynamic range (Fig.2), but did not result in a better detection limit. Reasons for this are a slight decrease in the absorbance of the PAR-Tb complex, an increase in the background noise and, possibly, an enhancement factor for water-methanol with respect to water lower than the theoretical one. In order to determine the enhancement factor experienced in our work, the behaviour of the thermal lens signal in water-methanol mixtures has been considered. Results obtained with 1 x 10-2 mol dm-3 cobalt nitrate are shown in Fig.4. As expected, on changing from aqueous to aqueous methanolic solutions, the thermal lens signal increases signifi- cantly. However, the change in sensitivity with volume fraction of methanol is not linear and is lower than that expected. Using eqn. (2) with the values of the refractive160 ANALYST, FEBRUARY 1993. VOL. 118 Table 2 Comparison of dctection limits and noisc (both exprcsscd as absorbance) in water and in water-methanol (40 + 60 v/v) with increasing background absorbance. Detection limits correspond to absorbance at the top of the peak. Back- ground Dctcction Carrier Conditions absorbance limit Noise Water Static 0 1.4 x 10-4 2.7 x 10-4 Water-methanol Static 0 3.8 x 10-5 7 x 1 0 - ~ 0.022 3.8 x 10-4 1.9 x 10-4 0.039 4.4 x 10-4 2.2 x 10-4 0.062 4.9 x 2.4 x Flowing 0 4.3 x 10-4 2.1 x 10-4 Flowing 0 1.8 x 10-J 9 x 10 5 index gradient and thermal conductivity for both solvents allows the calculation of the sensitivity enhancement of methanol with respect to water. The ratios of dnldT to k are 1.31 x 10-4 and 1.94 x 10-3 W-1 m for water and methanol, respectively, yielding a theoretical enhancement factor of 14.8 as against 7.5 for the experimental value at 40 Hz.This result was explained by a different value of the characteristic time constant t, [eqn. (3)] in both solvents. Owing to a lower thermal conductivity, methanol has a tc value about 1.5 times that of water, so that the time necessary to reach the steady-state signal is much longer in methanol than in water. When chopped illumination of the sample was used, time evolution of the thermal lens signal varied when the solvent composition was changed, yielding a reduced signal in the solvent where t, is greater.At a chopping frequency of 7 Hz, the sensitivity enhancement was slightly improved (8.7 instead of 7.5), but was still lower than the theoretical value. The latter value was approached only under continuous illumina- tion where the thermal lens reached its steady state whatever the value oft,. Owing to this fact, together with the non-linear dependence of the sensitivity on the volume fraction of methanol, use of water-methanol (40 + 60 v/v) resulted in a sensitivity only 3.5 times greater than that obtained in water. This poor improvement, together with less favourable condi- tions for complex formation in the organic medium, led to a greater value for the detection limit in the methanolic mixture (Table 1).Finally, the effects of background absorbance on the dynamic range, the background noise and the detection limits have been considered. The PAR carrier was replaced by a cobalt solution, and the background absorbances were varied by changing the cobalt concentration in the reagent carrier. Thermal lens detection was checked, in the lower concentra- tion range, with injected samples of cobalt nitrate and increasing values of the background absorbance. The highest value considered corresponded to the background absorbance (0.062) of the PAR in the reagent stream. The calibration graphs exhibited good linearity and the slope remained unchanged whatever the background absorbance.These results indicate that no saturation of the thermal lens was encountered and that the peak height did not depend on the background absorbance. On the contrary, the use of a reagent carrier absorbing at the wavelength of the analytical measure- ment contributes to an increase in the noise of the back- ground, which corroborates previous results obtained with static samples and shows the contribution of sample absor- bance to the relative standard deviation.24J5 It was deter- mined that the background noise originated mainly from the regular fluctuations in PAR concentration when the reagent stream mixes with the NaCI-HCI carrier at the T-piece. As shown iri Table 2, on changing from static to flowing conditions, the background noise did not increase signifi- cantly, when the background absorbance equalled zero.As the background absorbance was increased to 0.022, the noise increased 2-fold and the LOD value increased by the same ratio. Further increase of the background increased the noise and reduced the LOD to only a small extent. Moreover, the detection limit obtained with Tb-PAR was consistent with that achieved with cobalt nitrate when the same background absorbance was used. Conclusions The use of a reagent absorbing at the detection wavclength is not the main limiting factor in the application of thermal lens detection in flow injection. The limited performance of our system, mainly in the water-methanol mixture, results rather from the combination of several drawbacks, including a less fdVOUrabk reaction between PAR and the analyte, an increased noise and a reduced enhancement of the thermal lens effect by the organic solvent.1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 References Dovichi, N. J., CKC Crit. Rev. Anal. Chern.. 1987, 17, 357. Harris, T. D., Anal. Chem., 1982, 54, 741A. Harris, T. D., and Dovichi, N. J . , Anal. Clzenz., 1980,52,695A. Fang. H. L.. and Swofford, R. L., in Ultrasensitive Laser Spectroscopy, cd. Kligcr, 0. S . , Academic Press, Ncw York, 1983, p. 176. Harris, J . M., in Analytical Application,\ of LaJerJ, ed. Piepmeier, E. H., Wilcy-lntcrscience, New York, 1986, p. 541. Georges. J . , and Mermet, J. M., Analusis, 1988, 16, 203. Lcach, R. A., and Harris, J. M., Anal. Chim. Acra, 1984, 164, 91. Lcach. R. A . , and Harris, J. M., Anal. Clzem., 1984, 56, 2801. Jansen, K. L., and Harris, J . M., Anal. Chem., 198.5, 57, 2434. Yang. Y., and Hairrell, R. E., Anal. Chem., 1984, 56, 3002. Nolan, T. G., Hart, B. K., and Dovichi. N. J . , Anal. Chem., 1985, 57, 2703. Leach, R. A . , and Harris. J . M., J. Chromatogr., 1981,218, 15. Buffett, C. E., and Morris, M. D . , Anal. Chem., 1982,54,1824. Buffett, C. E., and Morris, M. D.. Anal. Chem., 1983,55,376. Scpaniak, M. I)., Vargo, J. D., Kettler, C . N., and Maskarinec, M. P., Anal. Chem., 1984, 56, 1252. Yang, Y., and Ho, T. V., Appl. Spectrosc., 1987. 41, 583. Nolan, 1’. G.. and Dovichi, N . J., Anal. Chem., 1987, 59. 2803. Georges. J., and Mermet, J . M., AnulyJt, 1989, 114, 541. Weimer, W. A., and Dovichi, N. J.. Anal. Chem., 1985, 57, 2436. Dovichi. N. J., and Harris, J . M., Anal. Chem., 1981, 53, 689. Berthoud, T., Delorme, N., and Mauchicn, P., Anal. Chem.. 1985, 57, 1216. Chartier, A., €ox, C. G., and Gcorges, J . , AnuluAis, 1992, 20, 269. Mori,. K.. Imasaka. T., and Ishibashi, N.. Anal. Chem.. 1982, 54, 2034. Doviehi, N. J.. and Harris, J . M., Anal. Chem.. 1981, 53, 106. Lowe, R. D., and Snook, R . D., Anal. Chim. Acta, 1991, 250, 95. Paper 210 I 794 F Received April 6, I992 Accepted May 28, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800157

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1993, VOL. 118 161 Protamine-coated Silica Gel as Packing Material for High-performance - - Liquid Chromatography of Carbohydrates Shuji Yamauchi Central Research Laboratory, SS Pharmaceutical Co. Ltd., I 143 Nanpeidai, Narita-shi, Chiba 286, Japan Noriyuki Nimura and Toshio Kinoshita School of Pharmaceutical Sciences, Kitasato University, 9- I Shirokane-5, Minato-ku, Tokyo 108, Japan Protamine-coated silica gel support for partition chromatography was found to be readily prepared by passing the protein through a silica gel layer. Application of the support to high-performance liquid chromatography provided versatile separation of mono- and oligosaccharides, and sugar alcohols. The support could be operated a t room temperature and gave excellent recovery for the reducing monosac- charides such as mannose and ribose for which a conventional aminoalkyl-bonded phase column gave poor recovery.Keywords: Protamine-coated silica gel column; sugar separation; monosaccharide chromatography; oligosaccharide chromatography; sugar recovery improvement In recent years, the physiological and pathological importance of carbohydrates has been increasingly emphasized. Separa- tion of carbohydrates, usually found as complex mixtures in bio-organisms, by high-performance liquid chromatography (HPLC) has been widely used. Amine phase columns rep- resented by the aminoalkyl-bonded silica gel columns have been used for this purpose.1" Partition chromatography of mono- and oligosaccharides on these columns has afforded high resolution. However, these columns gave poor recovery for some reducing sugars such as mannose and ribose since these sugars react with the primary amino group of the stationary phase to give N-glycosides or their Amadori rearrangement products .7-9 The amide columns developed recently showed both high resolution and recovery of carbo- hydrates.10-12 However, they should be operated at higher temperatures, around 80 "C, in order to avoid the separation of sugar anomers, which make the chromatogram too complex.In the present study, we found that protamine, a naturally occurring, strongly basic protein, composed mainly of argi- nine, binds readily to silica gel. This property was applied to the preparation of protamine-coated silica gel support for HPLC. The protamine-coated silica gel column (PCS column) should be suitable for the chromatography of carbohydrates as its guanidyl group is unlikely to react with reducing sugars.The preparation of the PCS column and its application to the partition chromatography of carbohydrates are described. Experimental Reagents and Materials Protamine, free base grade IV, from salmon was obtained from Sigma (St. Louis, MO, USA). Mono- and oligosac- charide standards were obtained from Wako Pure Chemicals (Osaka, Japan), Seikagaku (Tokyo, Japan) and Tokyokasei (Tokyo, Japan). Control serum (Precipath U) was purchased from Boehringer Mannheim Yamanouchi (Tokyo, Japan). Materials to be analysed were dissolved in 50% v/v aceto- nitrile. Methanol, anhydrous ethanol and acetonitrile (all of HPLC grade) were obtained from Wako Pure Chemicals.De-ionized and distilled water was used throughout. All other chemicals and solvents were of analytical-reagent grade or of the highest grade commercially available. A 5 pm porous silica gel (Nagel, Duren, Germany) packed into a 150 x 4.6 mm i.d. stainless-steel tube was purchased from Chemco (Osaka, Japan), Preparation of PCS Column A pre-packed silica gel column (150 x 4.6 mm i.d.) was washed with anhydrous ethanol. A 0.1% m/v protamine solution in 50% v/v methanol was passed through the column at a flow rate of 0.5 ml min-1 for 15 h to give a protarnine- coated column. The column was then washed with 50% v/v methanol at a flow rate of 0.5 ml min-1 for 1 h and was conditioned with 80% v/v acetonitrile-water at a flow rate of 1 ml min-1 for 24 h .The amount of protamine adsorbed on the silica gel surface was determined by measuring the absorbances, at 284 nm, of the protamine solutions, before and after elution. Apparatus The liquid chromatographic system consisted of a Model 655-12 pump (Hitachi, Tokyo, Japan), a Model 7125 injector (Rheodyne, Cotati, CA, USA) and an RI-2 refractive index detector (Japan Analytical Industies, Tokyo, Japan). The column temperature was controlled with Taitec CL-80 Coolnit and a Taitec thermominder Jr-1000 (Saitama, Japan). Sample Pre-treatment To one bottle of the control serum was added 2.5 ml of water and the mixture was mechanically shaken with an Iwaki KM shaker (Model VS, Iwaki, Tokyo, Japan) for 10 min. To 1.0 ml of the solution was added 9 ml of chloroform.The mixture was shaken as described above. After centrifugation at 8000g for 10 min, 20 p1 of the supernatant was submitted to the assay. Results and Discussion The PCS column was readily prepared by eluting protamine free-base solution through a silica gel column. One ml of the silica gel adsorbed 82.3 f 2.4 mg ( n = 7) of protamine and the protamine layer was stable under ordinary separation condi- tions for carbohydrates. These facts indicate that the basic guanidyl groups of protamine molecule interact strongly with acidic silica gel support. Figs. 1-4 show the chromatographic profiles of mono- and oligosaccharides and sugar alcohols separated on the PCS column. These figures demonstrate that the PCS column recognizes the fine configurational differ- ences between the saccharides. Fig.1 shows the chromato- gram of ribose, arabinose, xylose, fructose, mannose, glucose and galactose, obtained by eluting with 85% v/v acetonitrile. This concentration of acetonitrile was found to be optimum162 ANALYST. FEBRUARY 1993, VOL. 118 , 0 5 10 15 Time/mi n Fig. 1 Separation of monosaccharides on a 1.50 X 4.6 mm i.d. PCS column. Mobile phase, acetonitrile-water (8.5 + 15); flow rate, 1.0 ml min-I; column temperature, 30 "C; detection, refractive index detector at ~ 6 4 ; and 200 pg of each sample was injected. Peaks: S, solvent; A, ribose: B, arabinose; C, xylose; D, fructose; E, mannose; F, glucose; and G, galactose 0 5 10 15 20 25 Ti rne/m in Fig. 2 Separation of dissaccharides. Mobile phase, acetonitrile- water (80 + 20).Further conditions are the same as described under Fig. 1. Peaks: S, solvent: A, sucrose; B, maltose; C, lactose; and D, isomaltose for the separation of reducing monosaccharides. Fig. 2 shows the separation of the disaccharide mixture containing sucrose, maltose, lactose and isomaltose. Glucose and maltooligosac- charides were clearly separated as shown in Fig. 3. Fig. 4 demonstrates that the PCS column is also applicable to the separation of sugar alcohols such as glycerin, xylitol, sorbitol and inositol. Glucosamine and N-acetyl glucosamine were tested for separation. They could be separated but an unsymmetrical peak was observed for glucosamine. This may be ascribable to the strong interaction between the amino group of the sugar and the guanidyl group of protamine.Addition of protamine in,the mobile phase improved the shape of the glucosamine peak. Uronic acids and sialic acid were also examined but they were too strongly retained on the column and hardly eluted from the PCS column. The PCS column showed excellent separation of carbo- hydrates comparable with those of the aminoalkyl-bonded phase. The elution order of saccharides from the PCS column 0 5 10 15 20 25 Ti me/mi n B Fig. 3 Separation of glucose and maltooligosaccharides. Mobile phase, acetonitrile-water (70 + 30). Further conditions are the same as described under Fig. 1. Peaks: S, solvent; A, glucose; B, maltose; C, maltotriose; D, maltotetraose; E, maltopentaose; and F, mal- tohexaose ^ I B I I I I 0 5 10 15 20 Time/mi n Fig. 4 Separation of sugar alcohols.Mobile phase, acetonitrile- water (80 + 20). Further conditions are the same as described under Fig. 1. Peaks: S, solvent; A, glycerin; B, xylitol; C, sorbitol; and D, inositol was almost the same as those from amine phase columns such as the aminoalkyl-bonded silica gel column. The retention mechanism of the PCS column, therefore, seems to be based on a partition mode similar to those of the conventional amine phase columns.13-~6 This may be due to the highly hydrophilic nature of guanidyl groups of the protamine molecule which is similar to those of the amino groups on the aminoalkyl- bonded support. Silica gel packings coated with other proteins may provide a different mode of separation. Although conventional aminoalkyl-bonded columns are known to separate these saccharides, they often give poor recovery for some aldoses such as mannose or ribose.7-9 Table 1 compares the recovery of reducing carbohydrates separated on the PCS column with those separated on a conventional aminoalkyl-bonded column.The recoveries (% )ANALYST, FEBRUARY 1993, VOL. 118 163 Table 1 Comparison of recovery of reducing carbohydrates on a PCS-column and a conventional aminoalkyl phase column (NH2 column) Recovery (%) Analyte PCS column* NH2 column? Ribose 86 31 Arabinose 91 36 Mannose 99 32 Glucose 100 90 Sucrose 100 100 * Eluent: acetonitrile-water (80 + 20). t Eluent: acetonitrile-tetrapropylammonium hydroxide solution (75 + 25) (PH 10.0).17 C I I I I I - m 03 v) .- 0 5 10 15 20 25 I I I I I I I I I I I 0 5 10 15 20 25 Ti me/rni n Fig.5 Chromato rams of monosaccharides se arated on a PCS column kept at tu) 30, (b) 10, (c) 4 and Pd) 0 "C. Eluent, acetonitrile-water (80 + 20); and flow rate, 0.7 ml min-l. Further conditions are the same as described under Fig. 1. Peaks: S, solvent; A , ribose; B, mannose; and C, galactose were normalized to that of sucrose. The PCS column showed excellent recovery for almost all sugars tested, while the aminoalkyl-bonded column gave poor recovery for several sugars. This may be ascribable to the fact that the guanidyl group of the PCS column is very different from that of the primary amine group of the aminoalkyl-bonded column. The excellent recovery of the PCS column suggests that the guanidyl group does not react with reducing carbohydrates whereas the amino group of the aminoalkyl-bonded support reacts readily with 20 I I C - B - - I - Y Y Y U - r - - u Y A 5 I I I I I I 0 12 24 36 48 72 96 120 I nj ect io n ti me/h Fig.6 water (80 + 20). Further conditions are the same as described undc Fig. 1. A , mannose; B, glucose; C, sucrose; and D, maltose Reproducibility of retention time. Mobile phase, acetonitrile- S (bl S 0 5 10 15 0 5 10 15 Tirne/rn in Fig. 7 Chromatograms of ( a ) glucose standard solution and (b) control serum on a PCS column. Detection, refractive index detector at ~ 3 2 . Further conditions are the same as described under Fig. 1 the carbonyl groups of the carbohydrates to give glycosyl- amines. On the other hand, amide-type columns, such as TSK gel Amide-80, have recently been claimed to give excellent recovery for the aldoses that show poor recovery on the aminoalkyl-bonded columns.However, anomers of reducing carbohydrates were separated on the amide-type columns and this phenomenon made the chromatogram much too compli- cated and interfered with the separation of sugars with different configuration. Although anomers are known to give a single peak on the amide column by eluting with an alkaline mobile phase, such treatment may degrade the column. Otherwise, the amide column should be operated at higher temperature (about 80 "C) to obtain a single peak for anomers. Fig. 5 shows that the PCS column gives a single peak for a reducing carbohydrate over a wide range of temperature, These aldoses were separated in the range from 0 to 30 "C without resolution of anomers.Anomers were not resolved even at 0 "C, though broadening of the peaks was observed164 ANALYST, FEBRUARY 1993. VOL. 118 €or some sugars. These data substantiate the fact that the PCS column is suitable for operation at room temperature using mobile phases at neutral pH. This advantage may be a result of the basicity of the guanidyl group of protamine which accelerates the mutarotation to make the anomers indistinguishable. Fig. 6 shows the reproducibility of retention times for mannose, glucose, sucrose and maltose. When the PCS column was continuously eluted with 80% v/v acetonitrile at a flow rate of 1.0 ml min-1, the retention times were fairly reproducible. The retention time decreased by about 5% after elution for 5 d but the resolution was maintained.The reduction of the retention time seemed to be due to the loss of protamine from the silica gel surface. The retention time was restored by re-coating the column with the protamine solution described in the Experimental section. Fig. 7 shows the chromatogram of a supernatant obtained by the deproteinization of a control serum. Clear separation and excellent recovery (95%, n = 5 ) were observed. This column is expected to be applicable to the analysis of various biological fluids. The calibration curves of ribose, arabinose, mannose, glucose and sucrose showed excellent linearity for concentra- tions from 50 to 250 pg and passed through the origin. The detection limits of the above samples, using the RI detector were 5 pg (signal-to-noise ratio = 2). This detection limit is comparable to that of RI detection applied to the carbohy- drates eluted from the other amine phase columns.18 Conclusions The PCS column is readily prepared and provides excellent separation and recovery of various carbohydrates.The column is expected to be utilized in the fields of biochemistry and clinical chemistry. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 References Gherezghiher, T., Koss, M. C., Nordquist, R. E., and Wilkinson, Ch. P., J. Chromatogr., 1987, 9, 413. Verzele, M. Simoens, G., and van Damme, F., Chromato- gruphia, 1987, 23, 292. Engelhardt, H., and Orth, P., J. Liq. Chromatogr., 1987, 10, 1999. Buszewski, B . , and Lodkowski, J . Liq. Chromatogr., 1991, 14, 1185. Gotsick, J. T., and Benson, R. F., J . Liq. Chromatogr., 1991, 14, 1887. Erler, V., and Heublein, G., J. Chromatogr., 1991, 588, 340. Verzele, M., Simoens, G., and van Damme, F., Chromato- graphia, 1987, 23, 292. Hicks. K. B., Adv. Carbohydr. Chem. Biochem., 1988,46, 17. Hirata, N., Tamura, Y., Kasai, M., Yanagihara, Y., and Noguchi, K., J . Chromatogr., 1992, 592, 93. Koizumi, K., Utamura, T., Kubota, Y., and Hizukuri, S., J . Chromatogr., 1987, 409, 396. Nomura, Y., Agric. Biol. Chem., 1989, 53, 3313. Fujii, Y., Ikeda, Y., Okamoto, I., and Yamazaki, M., J. Chromatogr., 1990, 508, 241. D’Amboise, M., Noel, D., and Hamai, T., Carbohydr. Res., 1980, 79, 1. Verhaar, L. A. T., and Kuster, B. F. M., J . Chromatogr., 1982, 234, 57. Nikolov, Z . L., and Reilly, P. J., J. Chromatogr., 1985, 325, 287. Blanken, W. M., Bergh, M. L. E., Koeppen, P. L., and Eijnden, D. H., Anal. Biochem., 1985, 145, 322. Kinoshita, T., unpublished results. Clement, A., Young, D., and Brechet, C., J . Liq. Chromatogr., 1992, 15, 805. Paper 2104631 H Received August 28, 1992 Accepted November 2, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800161

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1993, VOL. 118 165 High-performance Liquid Chromatographic Determination of 5Hydroxyindoles by Post-column Fluorescence Derivatization Junichi Ishida, Ryuji lizuka and Masatoshi Yamaguchi" Faculty of Pharmaceutical Sciences, Fukuoka University, Nanakuma, Johnan-ku, Fukuoka 814-01, Japan A selective and sensitive high-performance liquid chromatographic method with post-column fluorescence detection has been developed for the determination of 5-hydroxyindoles. Five biogenic 5-hydroxyindoles (5-hydroxytryptop ha n, 5-hyd roxyt rypta mi ne, 5-hyd roxyi ndol-3-ylacetic acid, 5- hyd roxytryptop hol and N-acetyl-5-hydroxytryptamine) were separated by isocratic elution on a reversed-phase column, TSKgel ODS-80Tm, and then converted into fluorescent derivatives by post-column reaction with benzylamine and potassium hexacyanoferrate(l1i) reagents.The detection limits (signal-to-noise ratio = 3) for the indoles were in the range 140-470 fmol per injection volume (I00 pl). The method was applied to the determination of 5-hydroxyindol-3-ylacetic acid in human urine, with direct sample injection. Keywords: 5-Hydroxyindole and 5-h ydroxyindol-3- ylacetic acid; post-column fluorescence reaction; high-performance liquid chromatography; benzylamine; human urine 5-Hydroxyindoles are metabolites of tryptophan and play physiologically important roles in the human body. 5-Hydroxytryptamine (5-HT) , one of the active metabolites of tryptophan, occurs in the central nervous system and blood platelets. As intestinal carcinoid tumours release large amounts of 5-HT, higher concentrations of 5-hydroxyindol-3- ylacetic acid (5-HIAA), the major metabolite of 5-HT, are found in urine from the patients.l.2 Further, increased levels of 5-HT have been implicated in certain mental disorders such as schizophrenia and migraine."." Therefore, the determina- tion of these compounds in biological samples is useful for the elucidation of tryptophan metabolism and for the diagnosis of the above-men tioned disorders. Various methods using high-performance liquid chromato- graphy (HPLC) have been developed for the determination of 5-hydroxyindoles with native fluorescence5-8 and by electro- chemical detection.+-1 1 However, these detection methods are not always specific for 5-hydroxyindoles.Therefore, suitable pre-treatments such as liquid--liquid5-~~10 and solid-phase7.9 extractions have been required for the determination of the compounds in biological samples, as the crude sample contains many fluorescent and electrochemically active com- pounds that otherwise would interfere with the analysis.Recently, new HPLC methods,'*-'." involving column-switch- ing techniques, have been introduced for the purpose of injecting biological samples directly onto a column. Neverthe- less, these techniques need special equipment for the analysis. In previous work,15 we have reported that several aromatic methylamines such as benzylamine and 3,4-dimethoxybenzyl- amine react selectively and sensitively with 5-hydroxyindoles in weakly alkaline medium in the presence of potassium hexacyanoferrate(ii1); these derivatives yielded intense fluorescence in alkaline medium (Fig.1). In this study, benzylamine was found to be most effective as a post-column fluorescence derivatization reagent for 5-hydroxyindoles. We have developed a selective and sensi- tive HPLC method for the determination of 5-hydroxyindoles, based on the separation of the indoles by reversed-phase chromatography followed by an on-line post-column derivati- zation with benzylamine. Furthermore, the method was applied to the determination of 5-HIAA in human urine, with direct sample injection. 5-Hydroxytryptophol (5-HTOL) was used as an internal standard (IS). * To whom correspondence should be addressed. Experimental Chemicals and Solutions De-ionized and distilled water, purified with a Milli-Q I1 (Millipore, Milford, MA, USA) system, was used for all aqueous solutions.5-Hydroxytryptophan (5-OH-Trp), 5-HT, 5-HIAA, 5-HTOL and N-acetyl-5-hydroxytryptamine (N-Ac- 5-HT) were purchased from Sigma (St. Louis, MO, USA). Standard solutions of these compounds were prepared in water as recommended and kept frozen (-20°C) in amber coloured test-tubes. Other chemicals were of the highest purity available and were used as received. The reagent solution for post-column derivatization was prepared as follows. Benzylamine (or 3,4-dimethoxybenzylamine) and potassium hexacyanoferrate(n1) were dissolved in a mixture of acetonitrile and 25 mmol 1-1 borate buffer (pH 10.0) (1 + 1 v/v) to yield 20 and 3.0 mmol 1-1, respectively, at the final concentrations.Chromatographic System Fig. 2 shows a schematic diagram of the HPLC post-column fluorescence system. Chromatography was performed with an L-6000 high-performance liquid chromatograph (Hitachi, Tokyo, Japan) equipped with a Rheodyne (Cotati, CA, USA) 7125 syringe-loading sample-injector valve (100 pl loop). 5-Hydroxyindoles were separated on a reversed-phase column, TSKgel ODs-80Tm (150 x 4.6 mm i.d.; particle size, 5 pm; Tosoh, Tokyo, Japan), by isocratic elution with a mixture of acetonitrile and 10 mmol 1-1 acetate buffer (pH 4.7) ( 5 + 95 v/v) as mobile phase. The flow rate of the mobile phase was 1.0 ml min-1. The column temperature was ambient (18-25 "C). The eluate from the HPLC column was mixed with the reagent solution, delivered by a 980 PU pump (Jasco, Tokyo, Japan) using a T-type mixing device.The flow rate of the reagent solution was 0.5 ml min-1. The mixture was then passed through a reaction coil (7 m x 0.5 mm i.d., PTFE tube, GL Sciences, Tokyo, Japan) immersed in a 70 "C water-bath. After the fluorescence reaction, the mixture was passed through a cooling coil (0.5 m x 0.5 mm i.d., PTFE tube) immersed in ice-water. The fluorescence intensity from each 5-hydroxyindole in the final eluate was monitored at an excitation wavelength of 345 nm and an emission wavelength of 481 nm by a Shimadzu RF-530 spectrofluorimeter (Kyoto, Japan) equipped with a 12 PI flow cell. Chromatograms were recorded with a TOA FBR-251A recorder (TOA Elec- tronics., Tokyo, Japan).166 PI ANALYST, FEBRUAKY 1993, VOL. 118 Column Q H H 5-Hydroxyindoles Benzylam i ne Fluorescent derivative Fig.1 Fluorescence reaction of 5-hydroxyindoles with benzylamine Fig. 2 Schematic flow diagram of the HPLC-fluorescence system. P1, HPLC pump for mobile phase (Hitachi L-6000); Pz, HPLC pump for reagent solution (Jasco 980 PU); I, injection valve (Rheodyne 7125,100 pl); G, guard column (TSKgel ODS-80Tm); column, TSKgel ODS-80Tm (150 x 4.6 mm i.d.); M, reaction coil (PTFE, 7 m x 0.5 mm i.d.); C, cooling coil (PTFE, 0.5 m x 0.5 mm i.d.); D. fluorescence detector (Shimadzu RF 530); E, mobile phase; and R , tluorogenic reagent solution. Flow rate: E, 1 .O; R, 0.5 ml min-l Uncorrected fluorescence spectra were measured by means of a Hitachi 650-60 spectrofluorimeter with silica cells (10 x 10 mm); spectral bandwidths of 5 nm were used for both the excitation and emission monochromators.Procedure for the Determination of 5-HIAA in Human Urine Urine samples (24 h) from healthy volunteers and a patient were collected in glass bottles each containing 10 ml of 25% (15.6 mol 1-1) hydrochloric acid. Aliquots of 1 ml were stored at -20 "C prior to analysis. The urine was diluted 20 times with water before analysis and passed through a disposable filter (0.45 pm, 13 mm i.d., cellulose acetate). To 100 pl of the diluted urine were added 100 PI of 1.5 nmol ml-1 5-HTOL. The mixture was kept in an amber coloured test-tube with ice-cooling for stabilizing the analyte and IS. An aliquot (100 pl) of the mixture was analysed using the HPLC system. For the establishment of the calibration graph, 100 pl of 5-HTOL solution were replaced with an IS solution containing 5-HIAA (0.5 pmol-2.5 nmol).The net peak-height ratios of 5-HIAA and IS were plotted against the spiked 5-HIAA concentrations. Results and Discussion Selection of Benzylamine as a Post-column Derivatization Reagent In previous work,lS eight aromatic methylamines were screened as fluorogenic reagents for 5-hydroxyindoles by a manual method. Consequently, benzylamine and 3,4- dimethoxybenzylamine were found to yield the largest signal-to-noise ratio and the highest fluorescence intensity, respectively. In this study, these two compounds were examined as post-column reagents. Although 3,4-dimethoxy- benzylamine afforded peak heights about twice as high as those of benzylamine for all the indoles tested, the back- ground fluorescence from the former compound was about 50 times higher in intensity than that from the latter. Therefore, benzylamine was selected as the post-column derivatization reagent, with its extremely low blank fluorescence, for the determination of 5-hydroxyindoles. HPLC Conditions The best separation of five 5-hydroxyindoles was achieved on a reversed-phase column, TSKgel ODS-80Tm7 with aceto- nitrile-10 mmol 1-1 acetate buffer (pH 4.7) (5 + 95 v/v) as mobile phase.Fig. 3 shows a typical chromatogram obtained with a standard mixture of the compounds. These compounds could be separated within 40 min; the retention times (min) were 5.5 (5-OH-Trp), 8.0 (5-HT), 13.0 (5-HIAA), 23.0 (5-HTOL) and 39.0 (N-Ac-5-HT). The eluates from all the peaks had fluorescence excitation and emission maxima around 345 and 480 nm, respectively; the fluorescence excitation and emission spectra were almost identical with those observed using the manual method.15 Post-column Derivatization Conditions The concentrations of benzylamine and potassium hexacyano- ferrate(iii) had an effect on fluorescence development.Concentrations of benzylamine in the range 15-30 mmol 1- 1 in the reagent solution provided maximum and constant peaks from 5hydroxyindoles; a 20 mmol l-1 solution was used in the recommended procedure. Potassium hexacyanoferrate(1ii) at 2.G5.0 mmol 1-1 in the reagent solution afforded maximum peaks heights (Fig. 4); a concentration of 3.0 mmol 1-1 was adopted for the preparation of the reagent.The fluorescence reaction occurred effectively around pH 9 and the resulting derivatives fluoresced intensely around pH 11.15 Hence, the fluorogenic reagent solution was prepared in alkaline borate buffer. All the indoles tested afforded almostANALYST, FEBRUARY 1993, VOL. 118 167 t v) C 0 0 v) 2 L c 0 a, 4- n 1 - 2 3 A L r 1 I I I 0 10 20 30 40 50 Time/m in Fig. 3 Chromatogram of a standard mixture of five 5-hydroxy- indoles. A portion (100 PI) of a standard solution (SO pmol each on-column) was injected into the HPLC system. Peaks: 1 = S-OH-Trp; 2 = 5-HT; 3 = 5-HIAA; 4 = 5-HTOL; and 5 = IV-Ac-5-HT. HPLC conditions as described under Experimental I 1 I I I , 0 1 2 3 4 5 K3Fe(CN)dmmol 1-1 Fig. 4 the fluorescence derivatization. Curves 1-5 as for Fig. 3 Effect of potassium hexacyanoferrate(il1) concentration on maximum and constant peak heights with buffer solutions at pH values between 9 and 10.5; 25 mmol 1 - 1 borate buffer (pH 10.0) was selected for the recommended procedure.The pH of the final reaction mixture was almost 10. Organic solvents such as dimethyl sulfoxide and acetonitrile were required for dissolving benzylamine and accelerating the fluorescence reaction. Maximum peak heights were obtained over the concentration range 40-60% v/v in the reagent solution for both the solvents. However, the background noise width from dimethyl sulfoxide was about twice as high as that from acetonitrile. One reason for its high background noise might be that the mixing of the mobile phase and reagent solution was not adequate because of the high viscosity of dimethyl sulfoxide.Therefore, aqueous 50% acetonitile was used for the preparation of the reagent solution. The fluorescence reaction with benzylamine proceeded fairly rapidly, even at 0°C. However, an elevated reaction 100 h v) c 3 4- .- 2 2 4- .- a 50 - r' 0) Q) c Y m a .- 0 I I I I I 20 40 60 80 100 Tem peratu re/"C Fig. 5 tion. Curves 1-5 as for Fig. 3 Effect of reaction temperature on the fluoresccncc derivatiza- I I 1 1 I 2 4 6 8 10 Reaction coil length/m Fig. 6 tion. Curves 1-5 as for Fig. 3 Effect of reaction coil lcngth on the fluorescence derivatiza- temperature was required for development of the fluores- cence during the post-column reaction for all the compounds tested, as shown in Fig. 5; a 70 "C water-bath was selected.Fig. 6 shows the effect of reaction coil length on the fluorescence development; a 7 m reaction coil was adopted to yield the highest peaks. The post-column reaction time was estimated to be approximately 1 min. A linear relationship was observed between the peak height and the amount of each 5-hydroxyindole in the injection volume (100 pl) up to at least 2.5 nmol. The linear correlation coefficients were above 0.998 for all the compounds. The lower limits of detection a t a signal-to-noise ratio of three wcre 140, 180, 300, 300 and 470 fmol for 5-OH-Trp7 5-HT, 5-HTAA, 5-HTOL and N-Ac-5-HT, respectively. Determination of 5-HIAA in Human Urine The major metabolite of 5-HT in urine is 5-HIAA, and the determination is important in the diagnosis and therapy follow-up of carcinoid tumours. Hence, the above-mentioned168 ANALYST, FEBRUARY 1993, VOL.118 1 2 t v) C 0 n 2 L c, 0 a, c, 0" I I I 1 I I 0 5 10 15 20 25 Ti me/mi n Fig. 7 Chromatogram of the urine sample obtained from a hcalthy subject. A portion (100 plj of urine diluted 20-fold with water was treated as described under Experimental. Peaks: 1 = 5-HIAA (33.31 ymol d-l); 2 = IS; and 3 = endogenous fluorescent compounds t a, r: 0 n 2 L- c, O 0, c, 0" 4 3 2 I I I I I 0 5 10 15 20 25 Ti me/m i n Fig. 8 Chromatogram of the urine sample obtained from a carcinoid patient. A portion (100 plj of urine diluted 20-fold with water was treated as described under Experimental. Peaks: 1 = 5-HIAA (76.94 pmol d-lj; 2 = IS; 3 = endogenous fluorescent compounds; and 4 = endogenous compound, fluoresced by the post-column reaction.Detector sensitivity was half that in Fig. 7 method using post-column HPLC was applied to the deter- mination of urinary 5-HIAA. Typical chromatograms obtained with human urine from a healthy subject and a carcinoid patient are shown in Figs. 7 and 8, respectively. Both 5-HIAA (peak 1) and the IS (peak 2) were detected with baseline separation. The peak for 5-HIAA was identified as the fluorescent derivative of 5-HIAA on the basis of the retention time compared with that of the standard compound and by co-chromatography of the standard and a biological sample with 2-8% acetonitrile as mobile phase. When the analysis was performed without benzylamine andlor potassium hexacyanoferrate(iii), only peaks 1, 2 and 4 disappeared from the chromatograms. These results suggest I I I 1 I I 0 5 10 15 20 25 30 Ti me/m i n Fig.9 Chromatogram of S-HIAA in human urine detected by native fluorescence. The same diluted urine sample as in Fig. 7 was subjected to post-column HPLC. The HPLC conditions were the same as described under Expcrimcntal except that the reagent solution was replaccd by acctonitrilc and 25 mmol I-' borate buffer (pH 10.0) (1 + 1 v/v). Peaks: as in Fig. 7 Table 1 Urinary excretion (24 h) of 5-HIAA from healthy volunteers and a carcinoid patient 5-HI A A/ Hcalthy subject Age Sex pmol d-1 1 2 3 4 5 6 7 8 9 10 Mean Standard deviation 21 21 22 23 23 26 33 21 22 23 Male Male Male Male Male Male Male Female Female Female 41.04 38.62 41.36 34.69 33.31 33.57 46.87 43.19 33.30 26.20 37.22 6.09 Carcinoid patient 6 Male 76.94 that peaks 1 and 2 are the fluorescent derivatives of 5-HIAA and IS, respectively.Peak 4 in the chromatogram (Fig. 8), obtained with the carcinoid patient urine, was observed only when the reaction was performed with benzylamine and potassium hexacyanoferrate(iI1). This peak might be due to an endogenous 5-hydroxyindole. The detailed peak component, however, has remained unknown. No other 5-hydroxyindoles (SOH-Trp, 5-HT and N-Ac-5-HT) were detected in either normal or carcinoid patient urine. Peaks 3 were observed in the chromatogram, even when the reaction was performed without benzylamine and/or hexacyanoferrate(ii1). These peaks are endogenous fluorescent compounds in human urine. In order to evaluate the selectivity and sensitivity of the method, the same urine sample as in Fig.7 was analysed using native fluorescence (hex 275 nm, A, 330 nm) of 5-HIAA at the same detector sensitivity; the analysis was carried out by using acetonitrile-25 mmol 1-1 borate buffer (pH 10.0) (1 + 1 v/v) instead of the fluorogenic reagent solution (Fig. 9). Although the small peaks for 5-HTAA and the IS were observed in the chromatogram, their peaks were partially overlapped by other endogenous fluorescent compounds. Furthermore, the peak heights for the endogenous compounds varied with individual urine samples. Hence, the proposed post-column derivatiza- tion method is highly selective and sensitive for 5-hydroxy-ANALYST, FEBRUARY 1993, VOL. 118 169 indoles compared with that based on detection by native fluorescence.For a precise determination of 5-HIAA in urine, 5-HTOL was used as an IS; 5-HTOL is excreted in urine as free and conjugated forms, and the urinary levels are increased in patients with carcinoid tumours.16 In addition, the enhanced levels are observed in urine after ingestion of alcohols and serotonin-rich foods. lh-18 However, the urinary levels of free 5-HTOL are very low, even in the urine referred to above, and are approximately 1% or less than that of 5-HIAA.1618 Actually, no peak was observed at the retention time for 5-HTOL in either of the chromatograms, when the urine not spiked with 5-HTOL was treated as in the procedure. Therefore, 5-HTOL can be used as an IS for the determina- tion of 5-HIAA in urine.The calibration graph for the IS method was linear up to at least 500 nmol ml-1 in urine in the relationship between the ratios of the peak heights of 5-HIAA to that of the IS and the amounts of 5-HIAA added to urine. The linear correlation coefficient ( r ) of the calibration graph was 0.997. The recovery of 5-HIAA added to 100 pl of the diluted urine was 102.9 -t 3.3% (mean k standard deviation, n = 5 ) . The within-day precision of the method was established by repeated determinations (n = 8) on human urine (33.31 pmol d-1); the relative standard deviation was 2.5%. The lower limit of detection for 5-HIAA was 6 pmol ml-1 in diluted urine at a signal-to-noise ratio of three. This sensitivity is approximately 5-50 times higher than that of HPLC with native f l u o r e ~ c e n c e ~ ~ 8 ~ 1 3 ~ ~ ~ and approximately three times higher than that of HPLC with electrochemical detection." The concentrations of 5-HIAA in human urine from healthy subjects (n = 10) and from a carcinoid patient were measured by this method (Table 1).The 5-HIAA concentrations in normal urine were similar to those obtained by other workers; the reported 5-HIAA concentrations were between 2 and 50 pmol d-1.538-l'.'3 On the other hand, the concentration of 5-HIAA for a carcinoid patient was about two to three times higher than those obtained for healthy subjects. In conclusion, the proposed post-column HPLC method permits the highly selective and sensitive determination of 5-hydroxyindoles and can be applied to the determination of 5-HIAA in human urine without prior sample purification.The method requires a small portion of urine ( 5 PI), and, therefore, should be useful for biological investigations where only a small amount of sample is available. Moreover, the simplicity and reproducibility of the method are sufficient for clinical-laboratory screening. This work was partly supported by a Grant-in-Aid for Encouragement of Young Scientists (04771915) from the Ministry of Education, Science and Culture, Japan. The authors are grateful to M. Isokane for his skillful assistance and to Y. Mibuchi, Japan Medical Laboratory, Kyushu (Fukuoka, Japan), for the supply of urine samples from a patient. 1 2 3 4 5 6 7 8 9 10 I1 12 13 14 15 16 17 18 References Emson, P. C., Gilbert, R. F. T., Martensson, H . , and Nobin, A . , Cancer, 1984, 54. 715. Udenfriend, S . , Titus, E . , and Weissback, H . , J . Riol. Chem., 195.5, 216, 499. Tagari, P. C., Boullin, D. J., and Davics, C., Clin. Chem. (Winston-Sulem, N . C.), 1984. 30, 131. Garver, D. L . , and Davis, J . M., Life Sci., 1979. 24, 383. Rosano, T. G . , Meola, J . M.. and Swift. T. A . , Clin. Chem. ( Winston-Salem. N . C.), 1982, 28, 207. Skrinska, V., and Hahn, S . , J . Chrornatogr.. 1984, 311, 380. Wielders, J. P. M.. and Mink, J . K.. J. Chromatogr.. 1984,310, 379. Gironi. A . , Seghicri. G., Niccolai. M.. and Mammini. P . , Clin. Chem. (Winston-Salem, N . C.), 1988, 34, 2504. Parker, N. C., Levtzow, C. B . , Wright, P. W., Woodard, L. L., and Chapman, J . F., Clin. Chem. (Winston-Salem, N.C.), 1986, 32, 1476. Richard, D. A . , and Titheradge, A . C.. Biomed. Chromatogr., 1987, 2, 115. Helander, A . , Beck, O., Wennbcrg, M., Wikstrom, T.. and Jacobson, G., Anal. Biochem., 1991, 196, 170. Ncbinger, P . , and Koel, M., J. Chrornatogr., 1988. 427. 326. Koel, M., and Ncbingcr, P . , Biomecl. Chrornatogr., 1989, 3, 114. Koel. M., and Nebinger, P . , 1. Chromatogr., 1989, 495, 263. Tshida, J . , Yamaguchi, M., and Nakamura, M., Analyst, 1991, 116, 301. Davis, V. E . , Brown, H., Huff, J . A . , and Cashaw, J . A., J. Clin. Invest., 1966, 45, 1000. Davis, V. E., Brown, H., Huff, J . A . , and Cashaw, J . A., J. Lab. Clin. Med., 1967, 69, 132. Fcldstein, A., Hoagland, H., Freeman, H., and Williamson, O., Life Sci., 1967. 6, 53. Paper 2103703C Received July 13, 1992 Accepted October 13, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800165

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1993, VOL. 118 17 1 Determination of Patulin by Reversed-phase High-performance Liquid Chromatography With Extraction by Diphasic Dialysis Javier Prieta, Miguel A. Moreno, Javier Bayo, Susana Diaz, Guillermo Suarez and Lucas Dominguez" Departamento Patologia Animal I, Facultad de Veterinaria, Universidad Complutense, 28040 Madrid, Spain Ramon Canela and Vicente Sanchis Department de Tecnologia d'Aliments, Escola Tecnica Superior d'Enginyeria Agraria, 25006 L leida, Spain A simple and economical method has been developed for the determination of patulin in apple juice. The sample is extracted with ethyl acetate in a diphasic dialysis system, and the extract is cleaned up by elution from a Sep-Pak cartridge. Patulin is detected and determined by reversed-phase high-performance liquid chromatography using a Novapak CI8 column and an ultraviolet detector.The lower detection limit is 1 pg 1-1 and the recovery is 85% at the 20 pg 1-1 level. Keywords: Patulin; reversed-phase high-performance liquid chromatography; diphasic dialysis; apple juice Patulin is a mycotoxin of low relative molecular mass (154). It is the a,P-unsaturated lactone 4-hydroxy-4H-furo[3,2- c]pyran-2(6H)-one,l currently known to be produced by approximately 60 species of moulds belonging to over 30 genera, mainly Aspergillus and Penicillium.2 Patulin is toxic to a wide range of biological systems including bacteria, fungi, plants, protozoa and animals. In humans, its toxicity is mild: administered orally, it produces gastric irritation with nausea and vomiting; applied topically, it produces cutaneous inflammation.It is mutagenic and teratogenic.3.4 Long-term experiments have revealed the production of tumours in rats injected subcutaneously,5 but Osswald et a1.6 found no carcinogenicity when it was adminis- tered orally to rats and mice. Patulin has mainly been found in apples and apple products (juice, juice concentrate, jam, compote, confectionery and cider) and occasionally it has been determinated in silages, cereals, flour, bread, cakes and cheese and in pears, apricots, peaches, plums and grapes, as well as in products derived from these fruits.7 Tolerance levels have been established for patulin in various foods and drinks, usually apples and apple juices. The World Health Organization (WHO) has established 50 pg 1-1 as the recommended limit in apple juice.8 In this laboratory, a semi-permeable membrane technology called diphasic dialysis has been developed,' for the extraction of low relative molecular mass compounds from different substrates. The diphasic dialysis membrane procedure has previously been applied to the extraction of patulin from apple juice, using thin-layer chromatography (TLC) as a detection system.However, the results obtained were semiquantitative; the limit of detection was only 50 pg 1-1 and the recovery was 65% at the 100 pg 1-1 level.10 We have now modified and improved the latter method by using Sep-Pak cartridges as the clean-up procedure and reversed-phase high-performance liquid chromatography (RP-HPLC) as the detection system.Experimental Instrumental The following liquid chromatography equipment was used: A Rheodyne (Cotati, CA, USA) injector, a Model 510 pump, a 150 X 3.9 mm stainless-steel reversed-phase column packed with Novapak CI8, a Model 486 ultraviolet detector, a Maxima 820 software and a Waters System Interface Module (Waters, Milford, MA, USA). * To whom correspondence should be addressed. Extractions were carried out using a controlled-environ- ment incubator shaker, Model G25 (New Brunswick Scien- tific, Edison, NJ, USA). Materials A patulin standard in crystalline form was obtained from Sigma (St. Louis, MO, USA). Working solutions of patulin at 50 and 5 pg 1-1 were prepared with acetonitrile, then kept refrigerated (4 "C) and calibrated by ultraviolet absorption.11 Experiments were performed using Visking dialysis tubing of regenerated cellulose, size 20/32 (Cat. No. 44110, Serva, Feinbiochemical , Heidelberg, Germany) .9 The reagents used were chloroform, ethyl acetate and anhydrous sodium sulfate (all analytical-reagent grade) and water and tetrahydrofuran (THF) (HPLC grade). Extraction Procedure Apple juice was spiked with 20 pg 1-1 of patulin. Unspiked apple juice samples were used as blanks. Ethyl acetate was placed in a hydrated dialysis tube (100 ml of solvent in 750 mm length, 50 ml in 450 mm, 25 ml in 300 rnm and 10 ml in 250 mm length). The samples (10, 25 and 50 ml of apple juice) and dialysis tube were placed in flasks and extracted by shaking (150 rev min-1). The contents of the dialysis tube were decanted into a separating funnel, and the aqueous layer was discarded.The volume of the organic extract was then measured in a graduated test-tube; thereafter, the extract was placed in a flask and dried with 1 g of anhydrous sodium sulfate. A 5 ml portion of the extract was transferred into a vial and evaporated to dryness. The residue was then dissolved in 1 ml of chloroform. Clean-up The clean-up procedure was carried out using a silica Sep-Pak cartridge (No. 51900, Waters), which was prepared with 5 ml of chloroform. The extract was then injected into the cartridge and elution effected with 1 ml of chloroform, 1 ml of chloroform-ethyl acetate (8 + 2) and 1 ml of chloroform-ethyl acetate (5 + 5), consecutively, All of the resulting eluates were discarded. Thereafter, elution was effected with 2 ml of chloroform-ethyl acetate (2 + 8); this fraction was collected and evaporated to dryness under a gentle stream of air.The residue was then dissolved in 1 ml of water-THF (99 + 1).172 ANALYST, FEBRUARY 1993, VOL. 118 -100 I I I 71.26 71.93 " 0 8 A 10 15 20 25 30 35 40 TemperaturePC Fig. 1 Recovery of patulin in spiked apple juice with extraction by diphasic dialysis. Effect of the extraction temperature on percentage recovery. Experiments were performed in quadruplicate using 50 ml of apple juice (spiked at 20 pg 1-I) and 50 ml of ethyl acetate. Extraction time, 24 h 70.38 7565 72.54 71.3 3 A v v -I 0 I I I I 0 5 10 15 20 25 Time/h v Fig. 2 Recovery of patulin in spiked apple juice with extraction by diphasic dialysis. Effect of the extraction time on percentage recovery. Experiments were performed in duplicate at room temperature using 50 ml of apple juice (spiked at 20 pg 1-I) and SO ml of ethyl acetate Table 1 Recovery of patulin in spiked apple juice with extraction by diphasic dialysis.Results with sample-to-solvent volume ratio of 1 : 1" Volume of Volume of Patulin apple ethyl Patulin detected/ Recovery RSDt 10 10 0.2 0.12 63.64 3.98 25 25 0.5 0.34 69.63 1.74 50 50 1 .o 0.71 71.30 4.18 100 100 2.0 1.37 68.67 2.76 * Extraction overnight at room temperature. All experiments were t RSD = relative standard deviation. juice/ml acetate/ml addedlpg pg ("/I (Yo) performed at least in duplicate. HPLC Analysis The water-THF (99 + 1) mixture was used as mobile phase for the HPLC system.The flow rate was adjusted to 1.0 ml min-l and detection was carried out at 275 nm. Aliquots (20 p1) of standard solutions with different concentrations of patulin were injected into the HPLC system. The calibration curve * was obtained by plotting peak area versus concentration. The response for patulin was linear, at least between 0.05 and 5.0 ng. Sample solutions (20 pl) in THF-water (99 + 1) were injected into the HPLC system in duplicate. The concentration of patulin in the sample was determined directly from the calibration curve by means of the Maxima 820 software. Results and Discussion Factors taken into account when the method was being developed were temperature, extraction time and sample-to- solvent volume ratio. Table 2 Recovery of patulin in spiked apple juice with extraction by diphasic dialysis.Results with sample-to-solvent volume ratio of 1 : 2" Volume of Volume of Patulin apple ethyl Patulin detected/ Recovery RSDt 25 50 0.5 0.42 85.66 2.05 so 100 1.0 0.68 80.75 15.08 * Extraction overnight at room temperature. All experiments were t RSD = relative standard deviation. juice/ml acetate/ml added/pg pg (Yo 1 (Yo) performed at least in duplicate. 3.40 3.20 3.00 2.80 2.60 I I I I I 1.00 2.00 3.00 4.00 5.00 3.60 3.50 > 3.40 0 - 1 0, v, 3.30 .- 3.20 3.10 1.00 2.00 3.00 4.00 5.00 3.60 3.40 3.20 3.00 2.80 I I I I I 1.00 2.00 3.00 4.00 5.00 Time/min Fig. 3 High-performance liquid chromatograms of (a) ap le juice blank (unspiked); ( b ) apple juice spiked with 1 pg 1-l; and &) apple juice spiked with 20 pg I-'.The calibration was obtained using peak area and not peak height. The measured values were determined using the Maxima 820 software using the calibration as follows: c = 0.04944 + 178 x x R. Where c = concentration and R = peak response in volts The effect of the extraction temperature (15, 26 and 37 "C) on percentage recovery was tested, but no differences were observed (Fig. 1). Extraction-time tests were carried out for periods of 1,3, 5 , 8 , 16 and 24 h, recoveries obtained for 5 , 8 , 16 and 24 h being higher than 70% (Fig. 2). The results of experiments performed with a sample-to- solvent volume ratio of 1 : 1 v/v are presented in Table 1; the highest recoveries were obtained with 25 and 50 ml of appleANALYST. FEBRUARY 1993, VOL. 118 173 juice.As shown in Table 2, the use of a sample-to-solvent volume ratio of 1 : 2 led to an increase in recovery. The best result was obtained by using 25 ml of apple juice and 50 ml of ethyl acetate, the recovery being slightly higher than 85%. The mobile phase and column used in the HPLC determina- tion were selected to yield the maximum separation of patulin from interfering substances in apple juices. Under the stated conditions, as little as 0.01 ng of patulin standard could be detected. The limit of detection of the technique was 1 pg 1-1 of patulin in apple juice. Fig. 3 shows typical chromatograms for unspiked and spiked apple juice samples. The most widely used quantitative tools for patulin determi- nation are TLC and HPLC, although TLC methods are less sensitive (20 pg 1-1; Association of Official Analytical Chemists Official Method)ll and provide semiquantitative results.The separation by HPLC leads to an improvement in sensitivity: Ware et al. have detected 11 pg 1-1;12 Moller and Josefsson, 5 pg l-1;13 and Stray, 1 pg l-1.14 However, in every instance, the extraction and purification procedures are similar to those used for TLC. Most of these methods involve tedious extraction and require large volumes of solvents. The method for the determination of patulin in apple juice reported in this paper has several advantages over other techniques. The procedure is very simple, involves much less work on the part of the operator and saves large volumes of organic solvents, as only 60 ml of solvent are required compared with the 500 ml used when following some conventional methods.11312.14 The clean-up stage is performed using Sep-Pak cartridges instead of having to employ a conventional chromatographic column; hence, again time and solvents are saved. In summary, the optimum conditions for the method were determined to be 25 ml of apple juice, 50 ml of ethyl acetate and extraction overnight at room temperature. Under these conditions, the recovery was higher than 85% and the sensitivity was sufficient to allow detection of 1 pg 1-1 levels thereby allowing a safety margin for the determination of the patulin levels in apple juice to abide by the levels recom- mended by the WHO. Also, since the patulin content in commercial apple juice is usually lower than 50 pg 1-1, the method is suitable for monitoring purposes in the production of apple juice.This work was supported by the Spanish Project 90/0785 (Fondo de Investigaciones Sanitarias de la Seguridad Social). 1 2 3 4 5 6 7 8 9 10 11 12 13 14 References Scott, P. M., in Mycotoxins, ed. Purchase, I. F. H., Elsevier, Amsterdam, 1974, pp. 383-403. Steiman, R., Seigle-Murandi, F., Sage, L., and Krivobok, S., Mycopathologia, 1989, 105, 129. Ciegler, A., Beckwith, A. C., and Jackson, L. K., Environ. Microbiol., 1976, 31, 664. Matthiaschk, G., and Korte, A., Mycotoxin Res., 1986, 2, 89. Dickens, F., and Jones, H. E. H., Br. J. Cancer, 1961, 15, 8.5. Osswald, H., Frank, H. K., Komitowski, D., and Winter, H., Food Cosmet. Toxicol., 1978, 16, 243. Harrison, M. A., J. Food Safety, 1989, 9, 147. Jelinek, C. F., Pohland, A. E., and Wood, G. E., J. Assoc. Off. Anal. Chem., 1989, 72, 227. Dominguez, L., Blanco, J. L., Moreno, M. A., Diaz, S . , Prieta, J., Camara, J. M., Bayo, J., and Suarez, G., J. Assoc. Off. Anal. Chem., 1992, 75, 8.54. Prieta, J., Moreno, J. L., Blanco, J. L., SuBrez, G., and Dominguez, L., J. Food Prot., 1992, 55. AOAC Official Methods of Analysis, ed. Helrich, K., Associa- tion of Official Analytical Chemists, Washington, DC, 15th edn., 1990, pp. 1209-1210. Ware, G. M., Thorpe, C. W., and Pohland, A. E., J. Assoc. Off. Anal. Chem., 1974, 57, 1111. Moller, T. E., and Josefsson, E., J. Assoc. Off. Anal. Chem., 1980, 63, 1055. Stray, H., J . Assoc. Off. Anal. Chem., 1978, 61, 129. Paper 2103698C Received July 13, 1992 Accepted October 16, I992

ISSN:0003-2654    

DOI:10.1039/AN9931800171

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1993. VOL. 118 17.5 Determination of Formic Acid Vapour Using Piezoelectric Crystals With 4-Ethyl-3-thiosemicarbazide and 2,6=Diacetylpyridine Coatings Juan A. Munoz Leyva and Jose L. Hidalgo Hidalgo de Cisneros Department of Analytical Chemistry, Apdo. 40-1 1510 Puerto Real, Cadiz, Spain Daniel Garcia Gomez de Barreda and Antonio J. Fraidias Becerra Department of Marine Sciences and Engineering, University of Cadiz, Spain Two sensors have been suggested for formic acid vapour. These sensors are based on piezoelectric crystals coated, by immersion, with films of 4-ethyl-3-thiosemicarbazide and 2,6-diacetylpyridine. The sensors have been perfected using a new static system comprising an oscillator mounted on a die-plate and placed inside a testing chamber which, by means of a series of auxiliary devices, can reproduce and control environmental conditions.The sensors show their sensitivity to formic acid for at least 60 d. Selectivity is adequate with both coatings. The sensors begin t o respond t o formic acid vapour immediately, reaching a stable frequency after 5 min; average recovery times are about 28 min. For the 4-ethyl-3-thiosemicarbazide sensor, the sensitivity is 29.7 Hz mg-1 I (for an amount of coating corresponding to Afo = 475 Hz) with a correlation coefficient of 0.9997 and a detection limit of 0.17 m g 1-1. For the 2,6-diacetylpyridine sensor, the values are 10.9 Hz mg-1 1-1, 0.987 and 0.33 m g 1-1 respectively (Afo = 824 Hz). The sensors could be used t o detect formic acid vapour in the atmosphere and in working environments i f a suitable preconcentration device were connected. The advantages over existing techniques, i.e., gas chromatography and pulsed ultraviolet photoacoustic spectroscopy, should be evident from their simplicity, reduced cost and capacity for use in situ.Keywords : Formic acid determination; piezoelectric crystal; sensor; 4-eth yl-3-th iosemicarbazide; 2,6-diace t ylp yridin e There is increasing interest in the development of detection systems for toxic gases, and piezoelectric (PZ) crystals are playing a remarkable role in this field. This is due to their special characteristics, which facilitate the construction of portable, simple and cheap detection units. The first work on the application of PZ crystals for gas detection was by King.1 Later a PZ crystal detector for hydrocarbons in the atmosphere was developed.2 The atmosphere is the medium in which research in this field has been most widely applied.Guilbault and co-workers3-10 discovered several applications of coated PZ crystals for the detection of some gases in the atmosphere. Ho etal.11 constructed a portable detector to measure toluene in the environment. Fraser et ~1.12 developed a multi-sensor system to measure pollution in the air. Alder et n1.13 tested many metallic salts as coatings for PZ crystals for detecting hydrocyanic acid in the atmosphere. Mierzwinski and Witkiewiczl4 examined the use of PZ crystals as air pollution detectors, and Alder and McCallum's and McCalluml6 also carried out some remarkable investigations on the applications of PZ crystals to sensors.Formic acid is very toxic when inhaled or swallowed, and is present, mainly as ~ a p o u r , ' ~ in significant amounts in the atmosphere and in acid rain, where it accounts for 0.1-19% of the acidity. 18 It also has numerous industrial applications. Therefore, it is extremely important to have sensor systems for its detection. The existing detection methods use either gas chromato- graphy'g or gas chromatography with a preconcentration stage on a column with adsorbent;19 determination has also been described using pulsed ultraviolet photoacoustic spectroscopy.2O These techniques give a high level of sensitivity but they involve the use of sophisticated instruments. Only two processes using sensors to detect formic acid have been described and both detect formic acid in solution.One uses an optical fibre sensor for continuous measurement in biological systems.21 The Naval Research Laboratory optical waveguide sensor works by the attachment of a colorimetric redox dye indicator film to the outer surface of a hollow cylindrical waveguide, which, when immersed-in an aqueous solution, can reversibly detect various reducing species, reaching a sensitivity to formic acid of 7.5 mmoll-1 (345 mg 1-1). The other sensor described in the literature is based on the property of certain bacteria which produce hydrogen from formic acid, formic acid can be detected using Clostridium hutyricum and a fuel type electrode22 consisting of a platinum anode and a silver peroxide cathode that can be used for measuring hydrogen. The minimum concentration for determination is 10 mg 1-1.In the literature, we found no sensor for the determination of formic acid in its vapour form. This paper describes for the first time two sensors for formic acid vapours, based on AT-cut PZ crystals, coated with 4-ethyl-3-thiosemicarbazide and 2,6-diacetylpyridine films. These sensors combine good selectivity and much greater sensitivity than those suggested for formic acid in solution, although not sufficient for application to the atmosphere or working environments, threshold limit value-time weighted average (TLV-TWA) 0.009 mg 1-1.23 Sensitivity might be improved by the addition of an initial preconcentration stage, cf. the method of Ali et al. for determination of gaseous ammonia using a coated quartz PZ crystal.24 To study and perfect the new sensors we used our own design system25 which is capable of working statically, this is more easily adaptable to measurement in situ, unlike other techniques proposed with other sensors, which work dynamically.Experimental Reagents Formic acid, diphenylamine , 4-ethyl-3-thiosemicarbazide, 2,6-diacetylpyridine, phenylhydrazine, 2,2-bipyridine and reagents used in the study of interferences, were of analytical- reagent grade. The nitrogen was of type N-50 from SEO (Spanish Oxygen Company).176 ANALYST. FEBRUARY 1993, VOL. 118 Materials and Equipment The oscillator circuits employed were made of common commercial electronic components (resistors, capacitors, transistors). The 8.9 MHz, AT-cut (Universal Sensors) PZ crystals, were mounted on HC-25/U type bearings and the rest of the equipment consisted of control and measurement instruments described previously.25 The sensor unit is placed inside a testing chamber, which is capable of reproducing normal atmospheric conditions.Procedures Preparation of the coated crystal The crystal was cleaned by submerging it in ethanol. It was dried in warm air then placed in the testing chamber under vacuum for 5 min, dry nitrogen was injected and after 30 min the frequency was measured. The crystal was removed from the chamber and coated by immersion for 5 s or more according to the amount of coating desired, in a solution of the appropriate reagent (0.190 g of diphenylamine in 10 ml of ethanol; 0.100 g of 4-ethyl-3-thiosemicarbazide in 10 ml of ethanol; 0.100 g of 2,6-diacetylpyridine in 10 ml of acetone; 0.085 g of phenylhydrazine in 9 ml of ethanol; and 0.100 g of 2,2-bipyridine in 10 ml of ethanol).The coated crystals, dried as described, are protected for storage with a metallic capsule and placed in a vacuum desiccator. Measurement of the response to formic acid In a nitrogen or air atmosphere (in the testing chamber) and in the absence of formic acid vapour the stabilized frequency of the coated crystal was measured and noted. The chamber was evacuated and the formic acid injected. The nitrogen or air atmosphere was re-established (at the same pressure and temperature) and the frequency again measured and noted. The difference, Af, between this measurement and the previous one was calculated.The sensor was restored to its original condition by evacuation and nitrogen admission after this second measurement. The concentration of formic acid vapour was calculated from the previously constructed calibration graph of Af versus the concentration of formic acid. Results and Discussion Selection of Coatings The acid nature of the sample for analysis led us to select basic coatings. Furthermore, in the search for selective chemical sorption processes we considered compounds containing nitrogen atoms, which will probably give the following: -N- -N- I I H+ as described by Nieuwenhuizen and Barendsz.26 With all the coated crystals a reduction in frequency proportional to the concentration of formic acid in the chamber was noted but the reduction varied markedly for different coatings.For a fixed amount, 3.31 mg 1-1, of formic acid in the chamber, the differences in frequency (Hz) of the coated crystals, in the absence and presence of formic acid were: diphenylamine, 8; phenylhydrazine, 2; 2,2’-bipyridine, 7;. 4-ethyl-3-thiosemicarbazide, 60; and 2,6-diacetylpyridine, 40. On the basis of these results the 4-ethyl-3-thiosemicarbazide and 2,6-diacetylpyridine were chosen as coatings. Fig.1 PressurelkPa Effect of pressure on the 4-ethyl-3-thiosemicarbazide sensor 96.0 99.6 103.2 106.8 110.5 PressurelkPa Fig. 2 Effect or pressure on the 2,6-diacetylpyridine sensor Variation of Response With the Amount of Coating and With Time The Afo value is dependent on the amount of coating. Crystals were prepared with three different amounts of coating substance.The values of Afo are listed in Table 1. However, it was found that the frequencies of the coated crystals shift with time. Using the crystals mentioned above, Af, was measured daily for 60 d and the results are also shown in Table 1. Generally, Afo decreased with time owing to a slow loss of coating but after 30-40d Afo may increase owing to crystal contamination. This small frequency variation was taken into account when measuring Afo at a particular time. Variation in the Measuring Conditions The testing chamber reproduces environmental conditions and enables a study of the influence of variations in pressure, temperature or relative humidity on the sensor response to formic acid to be made. The frequencies of the coated crystals (for a given Afo) were first measured without and then with a known amount of formic acid.The atmospheric parameter was then varied and the corresponding frequencies recorded. Pressure Experiments were carried out at a constant temperature of 27°C. The chamber was evacuated and dry nitrogen at different pressures was injected with a fixed amount (10 ~ 1 ) of formic acid. Figs. 1 and 2 show that between 96 and 110.5 kPa, in the presence of formic acid, the frequency variation is 0.413 Hz kPa-1 (0.675 Hz kPa-1 between 96 and 103 kPa) for the 4-ethyl-3-thiosemicarbazide coated crystal (Ah, = 185 Hz), but it is only 0.15 Hz kPa-1 (0.30 Hz kPa-1 between 96 and 103 kPa) for the 2,6-diacetylpyridine coated crystal (Afo = 1077 Hz). In the absence of formic acid and under the same conditions, the differences in frequency are almost unchanged, especially for the 2,6-diacetylpyridine coating.Crystal response under a dry air atmosphere has also been studied. The results are very similar to those obtained under the nitrogen atmosphere.ANALYST, FEBRUARY 1993, VOL. 118 177 Temperature Experiments were carried out under a nitrogen atmosphere at a constant pressure of 105 kPa, injecting a volume of 10 ~1 of formic acid (3.31 mg 1-1) and measuring the recovery times of the sensor. It was found that the effect of temperature, both on frequency and recovery time, is dependent on the PZ mately constant. However, in the presence of formic acid, the frequency increases as the temperature rises, suggesting a decrease in sorption, which is confirmed by the decrease in Af.When the temperature rises, the recovery time initially stays constant, then increases and finally decreases abruptly so that at 40°C it is reduced by one half. crystal coating. For 4-ethyl-3-thiosemicarbazide (crystal with Afo = 185 Hz), crystal frequency both in the absence and presence of formic acid, decreases when the temperature increases, which appears illogical. Although the frequencies differ in both situations, Af decreases as the temperature rises, as shown in Table 2, which means that the sorption onto the coated crystal is reduced. When the temperature rises, the recovery time also increases at first, and then decreases. For 2,6-diacetylpyridine (crystal with Afo = 1077 Hz), and in the absence of formic acid, the crystal frequency fluctuates as the temperature rises, but on average remains approxi- Table 1 Variation of Afo with time AfolHz Time elapsedld 4-Ethyl-3-thiosemicarbazide 1" 1551 771 308 2 1542 682 252 5 1237 614 181 - 10 1181 - 20 1121 - 30 1107 443 186 40 1104 430 190 - - 60 1200 - * Just coated crystals. 2,6-Diacetylpyridine 1113 890 682 1051 890 699 1017 859 697 1008 - 1048 830 - 1073 - 622 1113 863 556 - - 1766 - Table 2 Temperature influence on frequencies of formic acid sensors 2,6-Diacetylpyridine 4-Ethyl-3-thiosemicarbazide TPC AflHz tRlmin A flHz tRlmin 15 91 30 40 20 20 78 30 38 25 25 50 35 31 30 30 54 35 26 35 35 41 20 30 20 40 45 15 29 11 Table 3 Influence of concentration of formic acid on frequency and recovery time 2,6-Diacetylpyridine 4-Ethyl-3-thiosemicarbazide Afo = Afo = Afo = Afo = 824 Hz 687 Hz 475 Hz 185 Hz mg1-l A F t R t Af tR Af t~ Af tR 0 .1 7 0 0 0 0 4 3 3 3 0 . 3 3 8 5 4 5 6 3 7 3 3.31 37 21 23 20 96 20 39 20 8.28 43 24 32 26 114 26 53 26 16.55 47 26 44 27 130 30 68 30 24.82 56 35 58 35 146 35 96 35 33.10 72 35 76 35 179 35 122 35 * AflHz. i tR/min. Humidity In view of the great influence of water on the frequency of some coated crystals,25 a study on the effect of relative humidity (RH) on the oscillation frequency of the sensors tested was carried out. A test-chamber temperature of 25 "C with a vacuum to cause the water to evaporate was used and 5, 10 and 5 0 ~ 1 of water (equivalent respectively to 10, 20 and 100% relative humidity) with 10 p1 of formic acid was injected.Dry air was injected into the chamber until a pressure of 100.8 kPa was reached; the oscillation frequencies of the crystal were measured, and compared with those measured in the absence of formic acid. The frequency of the PZ crystal with a 4-ethyl-3-thiosemi- carbazide coating (Afo = 1177 Hz) shows a slight but uniform reduction between 0 and 100% relative humidity (RH), amounting to 0.2 Hz (unit of RH)-1, both in the absence and presence of formic acid. The difference in frequencies is constant over that range, thus it was concluded that the humidity does not interfere in the determination of formic acid unless the RH variation is very large. The results of the tests for the 2,6-diacetylpyridine coated crystal (Afo = 1042 Hz) were similar to those described above, although Af first decreases when the RH rises to 20% and then from 20 to 100% RH, shows a slight increase. Influence of the Concentration of Formic Acid Calibration of the sensors in the presence of formic acid was carried out in a dry air atmosphere (0% RH) at 25°C and a pressure of 101.3 kPa.For each concentration of formic acid we measured the frequencies, both in the presence and in the absence of the acid. The difference between the frequencies, Af, was plotted against the formic acid concentration. Table3 shows the results obtained for the crystals coated with 4-ethyl-3-thiosemicarbazide and with 2,6-diacetylpyri- dine. It can be seen that the relationship between the frequency variations Af (in the presence and absence of the sample) and the concentration of formic acid, reveals two trends: one between 0.17 and 3.31mg1-1 and the other between 3.31 and 33.1 mg 1-1.The calibration equations for the coated crystals are given in Table 4. Greater sensitivity was obtained with the 4-ethyl-3-thio- semicarbazide coating for Afo = 475 Hz, being 29.7 Hz mg-11. The detection limit (defined as the concentration of sample that produces a three times greater variation in frequency than that produced in the absence of the sample) was 0.17 mg 1-1. Recovery Time (tR) and Variation of Response With Time The time taken for the sensor to recover its original frequency after exposure to formic acid was studied. For the 4-ethyl- 3-thiosemicarbazide coated sensor, the average tR was 29 min and for the 2,6-diacetylpyridine coated crystal, 28 min.Nevertheless, as shown in Table 3, the recovery time is dependent on the formic acid concentration. Table 4 Calibration equations for the coated crystals over two formic acid concentration ranges [HC02H] = 0.17-3.31/mgl-1 [HC02H] = 3.31-33.llmg 1-1 Coating AfdHz Af r A foIHz Af r 4-Et hyl-3-t hiosemicarbazide 2,6-Diace t y lpyridine 475 29.7[HC02H] - 2.4 0.9997 475 2.6[HCOzH] + 88.3 0.9890 185 ll.l[HC02H] + 2.2 0.9984 185 2.8[HCOzH] + 28.2 0.9947 824 10.9[HC02H] + 1.2 0.9870 824 1.10[HC02H] + 32.2 0.9730 687 6.9[HCOzH] + 0.24 0.9930 687 1.74[HC02H] + 16.7 0.9772178 ANALYST, FEBRUARY 1993, VOL. 118 Table 5 Statistical analysis of repeatability data Coating AfJHz (AJ)JHz GAp'Hz (Gaf)m/Hz A JlHz &(%) tRlmin 20.90 k 2.50 2,6-Diacetylpyridine 824 37.10 1.75 0.53 37.10 k 1.21 k3.3 19.54 Ifr.2.29 4-Ethyl-3-thiosemicarbanide 185 37.25 1.85 0.56 37.25 k 1.24 k3.3 Table 6 Study of interferences Interferent Chloroform Diethyl ether Benzene Acetone Acrolein Formaldehyde Diisopropyl ether Isobutyl methyl ketone Isopentyl alcohol Ace tonitrile Carbon tetrachloride Chlorobenzene Toluene Acetic acid Hexane Butyl acetate Dichloromethane Propan-2-01 Tributyl phosphate Hydrochloric acid Ethanol Concentra- tion/mg 1-1 61 29 36 32 34 16 29 32 50 31 63 45 36 43 27 35 54 32 40 17 32 A JD "/HZ 14 0 6 12 0 15 9 10 4 8 11 0 22 0 8 13 0 0 0 31 7 AfEJ-lHZ 14 19 2 12 0 17 0 0 11 5 9 7 0 24 4 21 15 0 0 50 9 * Decrease in frequency for 4-ethyl-3-smithiocarbazide. AJo = t Decrease in frequency for 2,6-diacetylpyridine, Afo = 460 Hz.830 Hz. Repeatability In order to determine the repeatability, eleven measurements were carried out under identical reaction chamber conditions with a formic acid concentration of 3.31 mg 1-1. Statistical treatments (95% probability level) for both sensors are listed in Table 5. Interferences First, the frequency was measured with a known amount of formic acid (1Op1, equivalent to 3.31 mgl-1 within the chamber) and then the frequency was measured with the same amount of formic acid and a 5-20 times greater volume of the interfering substance. The responses of crystals coated with 4-ethyl-3-thiosemicar- bazide and 2,6-diacetylpyridine to 21 interfering species were studied. Conditions of steady pressure (99.6 kPa) and temper- ature (28 "C) were maintained.The concentration of inter- ferents (mgl-1) and the decrease in frequency and for 4-ethyl-3-thiosemicarbazide (AfE) and 2,6-diacetylpyridine (AfD), respectively, are listed in Table 6. From the results it can be concluded that for both coatings the greatest interferences are caused by species of an acid nature (e.g., hydrochloric acid, acetic acid). Formaldehyde also showed significant interference. The authors acknowledge their gratitude to CICYT (Project AMB92-0863) for financial support. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 References King, W. H., Anal. Chem., 1964, 36, 1735. King, W. H . , J. Environ. Sci. Technol., 1970, 4, 1136. Scheide, E. P., and Guilbault, G. G., Anal. Chem., 1972. 44, 1764. Karmarkar, K.H., and Guilbault, G . G.. Anal. Chim. Acta., 1974, 71, 419. Karmarkar, K. H., and Guilbault, G. G., Anal. Chim. Acta, 1975, 75, 111. Hlavay, J., and Guilbault, G . G., Anal. Chem., 1978, 50, 1044. Hlavay, J., and Guilbault, G . G., Anal. Chem., 1978, 50, 965. Karmarkar, K. H . , and Guilbault, G. G.. Environ. Lett., 1975, 10, 237. Webber, L. M., Karmarkar, K. H., and Guilbault, G . G . , Anal. Chim. Acta., 1978, 97, 29. Tomita, Y., Ho, M. H . , and Guilbault, G. G., Anal. Chem., 1979, 51, 1745. Ho, M. H . , Guilbault, G. G., and Reitz, B . , Anal. Chem. 1983, 55, 1830. Fraser, S. M., Edmonds, T. E . , and West, T. S . , Analyst, 1986, 111, 1183. Alder, J. F . , Bentley, A. E., and Drew, P. K., Anal. Chim. Acta, 1986, 182, 123. Mierzwinski, A., and Witkiewicz, Z., Uchr. Powierrza, 1984, 18, 73; Chem. Absrr., 1985, 102, 66570s. Alder, J. F., and McCallum. J. J., Anafysr, 1983, 108, 1169. McCallum. J. J., Analyst, 1989, 114, 1173. Andreae, M. 0.. Talbot, R. W., and Lee, S . , Geophys. Res., D: Atmos., 1987, 92, 6635: Chem. Abstr., 1987, 107, 138084q. Rood, A. P., and Streeter, R. R., Am. Ind. Hyg. Assoc. J . , 1985, 46, 257. Nishiriura, S . , and Esaka, S . , Kyoto-Fu Eisei Kogai Kenkyusho Nempo, 1983. 28, 92; Chem. Ab.str. 1985, 103,41701t. Cvijin, P. V., Gilmore, D . A.. and Atkinson, G. H., Appl. Spectrosc., 1988, 42, 770. Matsunaga, T., Karube, I., and Suzuki, S . , Eiir. J . Appl. Microbiol. Bioeng. 1979, 21, 1845. Chemical Sensors and Microinstrumentation , eds. Murray, R. W., Dessy, R. E., Heineman, W. R., Janata, J., and Seitz, W. R., American Chemical Society. Washington DC, 1989. 1990-1991 TLVs- Valores Limites e Indices Biologicos de la ACGlH, American Industrial Hygiene Association (Spanish section), Generalitat de Valenqia, Valenqia, Spain, 1990. Ali. Z . , Thomas, C. L. P., Alder, J. F., and Marshall, J. B., Analysr, 1992, 117, 899. Mufioz, J. A., Hidalgo, J. L., Fraidias, A., and Garcia, D., Tulanta, submitted for publication. Nieuwenhuizen, M. S . and Barendsz, A. W., Sens. Actuators, 1987,11,45. Paper 21022 78H Received May 1, 1992 Accepted November 2, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800175

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1993, VOL. 118 179 Voltammetric Determination of Total Alkannin Using a Glassy Carbon Electrode Rasamee Chaisuksant and Anastasios Voulgaropoulos* Analytical Chemistry Laboratory, Chemistry Department, Aristotle University of Thessaloniki, 54006 Thessaloniki, Greece Antonios S. Mellidis and Vassilios P. Papageorgiou laboratory of Organic Chemistry, Department of Chemical Engineering, Aristotle University of Thessa Ion iki, 54006 Thessalo n iki, Greece A sensitive and reliable voltammetric method for the determination of total alkannin using a glassy carbon electrode was developed. The method was applied to the determination of total alkannin in the roots of Alkanna tinctoria after hydrolysis of the isolated mixture of pigments consisting of alkannin esters. The detection limit was 0.55 mg I-' and the relative standard deviation 2.29%.For the accuracy of the method the hydrolysis step was crucial. Keywords: Alkannin; voltammetry; glassy carbon electrode; A1 kanna tinctoria The isohexenylnaphthazarins, commonly known as alkannins or shikonins, are lipophilic red pigments. 1 These compounds have been isolated from the roots of various species of the family Boraginaceae.2-6 The structural features common to all these pigments are the naphthazarin moiety and the isohexenyl side-chain [( 1) naphthazarin, (2) alkannin, (3) alkannin isovalerate and (4) alkannin acetate]. They differ in the following respects: (i) their manner of rotating polarized light (chirality); (ii) the position of the alcoholic group in the side-chain; and (iii) the acid moiety which esterifies the alcoholic group.OH 0 OH 0 1 OH 0 \ I 2 : R = H 3: R = CH3CHCH2CO- I CH3 4: R = CH3CO- All isohexenylnaphthazarins are biologically extremely potent compounds, because they have pronounced anti- bacterial,2,7-9 antiturnour'" and wound-healing11312 activities and are regarded as a new class of drugs.13 The biological importance of these compounds, along with the need to determine low levels of the active components for future elucidation of the mechanism of their healing activity, prompted the development of a sensitive and reliable method for their determination as total alkannin. Different methods have been published for the determina- tion of the total content of isohexenylnaphthazarins'4-'7; however, none of these methods give reliable results, as they * To whom correspondence should be addresscd. do not take into account either the tendency of these compounds to polymerize or the different sensitivities of the various alkannin esters to the applied technique.In this work, a sensitive voltammetric method was devel- oped for the determination of isohexenylnapthazarins as total alkannin after appropriate hydrolysis. This method was applied to the determination of total alkannin in the roots of Alkanna tinctoria after suitable sample preparation. Experimental Apparatus A Metrohm E506 Polarecord equipped with an E505 polaro- graphic stand was used for the differential-pulse voltammetric measurements and a Wenking PGS 81 Potentio-Galvano-Scan potentiostat equipped with a Wenking VSG 72 voltage scan generator (Gerhard Bank-Elektronik) and with a Hewlett- Packard 7004B x-y recorder was used for cyclic voltammetry.A GC-Typ-ZU628 glassy carbon electrode (Metrohm) was used as the working electrode; the counter electrode was platinum wire and the reference electrode was silver-silver chloride saturated with potassium chloride. The glassy carbon electrode was polished with aluminium oxide between samples according to the manufacturer's instructions. Reagents All reagents were of analytical-reagent grade. Solutions of 0.01 and 0.1 mol 1-1 tetraethylammonium tetrafluoroborate (TEATFB) in dimethylformamide (DMF) were used as supporting electrolytes for voltammetric measurements. Pure alkannin isovalerate (3) and alkannin acetate (4) were isolated from the roots of A.tinctoria as illustrated in Fig. 1. High-purity alkannin (2) was obtained by alkaline hydrolysis of a mixture of alkannin esters and its further purification by column chromatography on Kieselgel 60 (70-230 mesh) (Merck) with chloroform as eluent. The purity and identity of these compounds were checked by their melting points and spectral data (infrared, proton nuclear magnetic resonance and mass spe~tra).2>~-5 Voltammetric Procedures Nitrogen (99.999%) was used without further purification to de-aerate the samples prior to the voltammetric measure- ments. The nitrogen was purged €or 10 min. The parameters of the differential-pulse mode were: pulse height, -50 mV;180 A. tinctoria ANALYST, FEBRUARY 1993, VOL. 118 ( i ) Soxhlet extraction with hexane ( i i ) Evaporation of the solvent ( i ) Fractional extraction with MeOH ( i i ) Filtration Methanolic filtrate ( i ) Precipitation of alkannins with methanolic CU(CH~COO)~-H~O ( i i ) Filtration fluorescing substances I Fattyand I Suspension into diethyl ether and treatment with diluted HCI (1 + 1) I Evaporation of the solvent Al ka n ni n ,I ( i ) Repeated column r isovalerate (2) acetate (3) Mixture ( i ) Alkaline hydrolysis ( i i ) Acidification ( i i i ) Extraction with CHCI3 (iv) Evaporation of the solvent I I measurement Residue for I Column chromatography +, Fig.1 isolation of pure alkannins Sample preparation for the determination of total alkannin in the roots of Alkanna tinctoria and analytical procedures for the pulse duration, 56 ms; scan rate, 8.3 mV s-I; initial potential, ( E i , ) = 0.0 V; and final potential, (E,) = -1.4 V.For cyclic voltammetry, the concentration of the supporting electrolyte was 0.1 mol 1-1 TEATFB in DMF and the scan rates used were 50 and.100 mV s-1. Sample Preparation A 200 g amount of powdered dried roots of A . tinctoria was subje\cted to the procedure shown in Fig. 1. About 70 mg of the product, mainly consisting of a mixture of alkannin esters and small amounts of naturally occurring polymeric pigments, was hydrolysed by addition of 30 ml of 2% sodium hydroxide solution at room temperature with continuous stirring in the absence of light. The progress of saponification was controlled by thin-layer chromatography (TLC) and it was found that the alkaline hydrolysis was completed after 8.5 h.The TLC was performed on Kieselgel 60FZs4 plates (Merck) with benzene-chloroform- acetone (50 + 50 + 1) as the developing solvent. It should be noted that longer hydrolysis times result in polymerization of alkannin and must be avoided. After saponification, the solution was cooled in an ice-bath and acidified by addition of 4 ml of sulfuric acid (9.85%). The acidified solution was extracted with chloroform (3 X 100 ml) rather than with diethyl ether, so that the final residue contained the least possible amount of free fatty acids so as not to affect the voltammetric measurements. The combined extracts were washed with distilled water (2 x 100 ml). The chloroform layer was dried over anhydrous sodium sulfate and the solvent removed in vacuo (below 50 "C).About 12 mg of the deep red residue were dissolved in 25 ml of DMF and the sample was ready for measurement.ANALYST. FEBRUARY 1993. VOL. 118 L' 181 A I -1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 E N versus Ag-AgCI Fig. 2 Differential-pulse voltammograms of A, naphthazarin (3.32 x 10W5 rnol I-l); B, alkannin (3.35 x 10-5 rnol 1-l); C, alkannin acetate (3.17 x 10-5 mol 1-1); D, alkannin isolvalerate (3.29 x 10-5 mol 1-l); and E, blank in DMF using 0.01 rnol I-* TEATFB as the supporting electrolyte and glassy carbon as the working electrode. Scan rate, 8.3 mV s-1 L 3 u -2.0 -1.2 -0.4 +0.2 E N versus Ag-AgCI Fig. 3 Cyclic voltammograms of (a) naphthazarin (2 x 10F4 moll-1) and ( b ) alkannin (2 X 10-4 rnol 1-1) for scan rates of 50 mV s-1 (solid line) and 100 mV s-1 (broken line), using 0.1 mol I-' TEATFB as the supporting electrolyte in DMF Results and Discussion It has been established that in solvents of low proton availability, quinones undergo two one-electron polaro- graphic reductions .I8319 Q + e- $ Q- Q- + e- s Q2- The first reduction process is reversible and the second can be reversible or irreversible.The cause of irreversibility of the second reduction process has been attributed to proton addition from the solvent. The electrochemical characteristics of some isohexenyl- naphthazarins are shown in Fig. 2. Both reduction processes are reversible for naphthazarin (1) and alkannin (2), as shown by the same calculated for various pulse magnitudes and for different polarities from the equation E, = E1/2 -(AE/2), where AE is taken with sign.20J1 Thus, for naphthazarin (E112)1 = - 0.28 V and (El/& = - 1.06 V and for alkannin (E1/2)1 = - 0.34 V and (E1/2)2 = - 1.15 V, using pulse heights of -50, -25, +25 and +50 mV.Additionally, the cyclic voltammograms of these two compounds shown in Fig. 3 confirm the reversibility of the above processes, as the differences in the cathodic and anodic peak potentials are the same for different scan rates. However, the second reduction process is irreversible for alkannin isovalerate (3) and alkannin acetate (4), showing the strong influence of the esterified hydroxyl group. This is not clear from Fig. 2 owing to the background interference at about -0.8 V, overlapping with the peaks of the alkannin esters at a concentration of 2 x 10-4 rnol 1-1.However, for concentrations an order of magnitude higher (1 x 10-3 moll-1) the cyclic voltammograms of these two esters produce higher peaks than the background interference, indicating an irreversible process, as is clearly shown in Fig. 4. The presence of the side-chain in alkannin shifts the peak potentials in comparison with naphthazarin, whereas the esterification of alkannin shifts the peak potentials to less negative values. Moreover, there are significant differences in the magnitude of the peak current. Hence the direct determination of these derivatives expressed as total alkannin leads to erroneous results. The sample preparation is crucial for the reproducibility and accuracy of the method, because polymerization is possible at any stage.The polymerization is strongly influenced by the extraction heating time, the waiting time of the extracts, the concentration of sodium hydroxide, the hydrolysis time and the presence of light. I I I -2.0 -1.2 -0.4 +0.2 E N versus Ag-AgCI Fig. 4 Cyclic voltammograms of A, alkannin acetate (1 x 10-3 rnol I-l) and B, alkannin isovalerate (1 X rnol I-l), using 0.1 rnol 1-l TEATFB as the supporting electrolyte in DMF. Scan rate, SO mV s-1182 ANALYST, FEBRUARY 1993, VOL. I18 A Ac I Table 1 Replicate measurements of total alkannin content in the mixture of pigments of Alkanna tinctoria Total alkannin Sample Mass of Total alkannin in the mixture of No. pigmentstg foundlg pigments (%) 1 0.0694 0.0386 55.6 2 0.0702 0.0408 58.1 3 0.0694 0.0408 58.8 4 0.0757 0.0433 57.1 5 0.0731 0.0429 58.7 X = 57.770, s = 1.32%.RSD = 2.29%. Table 2 Recovery of alkannin added to the mixture of pigments of Alkanna tinctoria before hydrolysis -1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 EN versus Ag-AgCI Fig. 5 Differential-pulse voltammograms of pigments from the roots of Alkunna tinctoria in DMF, using 0.01 mol 1-1 TEATFB as the supporting electolyte. A, Blank; B, before hydrolysis, 10.7 ppm of pigments; and C, after hydrolysis. Scan rate, 8.3 mV s- 1 Sample Mass of alkannin Mass of alkannin No. addedlg foundlg Recovery (%) 1 0.0317 0.0300 94.64 2 0.0407 0.0395 97.05 3 0.0420 0.0452 107.61 4 0.0436 0.0454 104.13 5 0.0435 0.0386 88.74 The polymerization may occur from the unsaturated side- chain, but is mainly via the phenol-phenol coupling effect.Thus, in the simplest case of polymerization, dimer products of the probable structure 5 are formed. OH 0 OCOR 5 For comparison, in Fig. 5 the differential-pulse voltammo- grams of pigments before and after hydrolysis are given. After hydrolysis, the voltammograms become typical of those for pure alkannin [Fig. 2(curve R)]. For the calibration graph the peak potential of the signal used was E, = - 0.32 V versus Ag-AgC1. The calibration graph is described by the equation y = [5.85 k 15.29 (nA)] + [68.29 k 0.91 (nA mg-l)]x Five replicate measurements were performed to determine the relative standard deviation (RSD), as shown in Table 1. Additionally, to check the recovery of the method, known amounts of alkannin were added to the mixture of pigments prior to saponification; the results are given in Table 2 and show satisfactory efficiency. However, the reproducibility in Table 2 is not as good as that in Table 1, probably owing to the different calibrants used.1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 References Thomson, R. H., Naturally Occurring Quinones, Academic Press, New York, 1971. Shukla, Y. N., Tandon, J.S., Bhakuni, D. C . , and Dhar, M. M., Experientia . 1969, 25, 357. Kyogoku, K., Terayama, H., Tachi, Y., Suzuki, T., and Komatsu, M., Shoyakugaku Zasshi, 1973, 27, 24. Papageorgiou, V.P., and Digenis, G. A., Planta Med., 1980,39, 81. Papageorgiou, V. P., Mellidis, A. S . , and Sagredos, A. N., Chem. Chron., 1980, 9, 57. Mellidis, A. S . , and Papageorgiou, V.P., J . Nat. Prod., 1987, 50, 618. Tanaka. Y., and Odani, T., Yakugaku Zasshi, 1972,92,525. Tanaka, Y., Odani, T., and Kanaya, T., Shoyakugaku Zasshi, 1974. 28, 173. Papageorgiou, V. P., Winkler, A., Sagredos, A. N., and Digenis, G. A., Planta Med., 1979,35, 56. Gupta, S. K., and Mathur, I. S . , Indian J. Cancer, 1973, 9, 50. Papageorgiou. V. P., Experientia, 1978, 34, 1499. Papageorgiou, V. P., US Pat., 4282250, 1981. Papageorgiou, V. P., Planta Med., 1980, 38, 193. Van Damme, J . G., and De Neve. R. E., J . Pharm. Sci., 1979, 68, 16. Fedoreev, S. A., Krivoshchekova, G. E., Denisenko, V. A., Gorovoi, P. G., and Maksimov, 0. B . , Chem. Nat. Cornpd. (USSR), 1979, 15, 546. Tsukada, M., Fukui, H., Habara, C., and Tabata, M . , Shoyakugaku Zasshi, 1983, 27, 299. Papageorgiou, V , P., Liakopoulou-Kyriakides, M., and Papadakis, Ch.. Flavour Fragrance J., 1985, 1, 21. Peover, M. E., J . Chem. Soc., 1962,4540. Peover, M. E., Trans. Faraday Soc., 1962.58, 1656. Birke, R. L., Anal. Chem., 1978, 50, 1489. Dillard, J. W., O’Deu, J. J., and Osteryoung, R. A., Anal. Chem., 1979, 51, 125. Paper 2103399B Received June 29, 1992 Accepted October 2, I992

ISSN:0003-2654    

DOI:10.1039/AN9931800179

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1993, VOL. 118 183 Polarographic Determination of Total Pyrethroid Residues in Stored Cereals Gianfranco Corbini, Cinzia Biondi, Daniela Proietti, Elena Dreassi and Piero Corti Dipartimento Farmaco Chimico Technologico, Universita di Siena, Siena, Italy A new method for determination of the total cyanopyrethroid residue concentration in stored cereals by differential-pulse polarography is proposed. Cyanopyrethroids were previously extracted with a non-polar solvent and purified by solid-phase extraction. The total pyrethroid residue concentrations are determined by differential-pulse polarography of 3-phenoxybenzaldehyde formed quantitatively by rapid and reproducible alkaline alcoholysis of the sample. The method is linear, quantitative and reproducible for maize and wheat analysis in the range 0.1-3.0 mg kg-1.A thin-layer chromatographic method for identification of six widely used cyanopyrethroids in the sample is described. The electrodic reduction process of 3-phenoxybenzal- dehyde is examined and the reduction products isolated and identified. Keywords: Pyrethroid; di fferential-pulse polarograph y; cereal; 3-phenoxybenzaldeh yde; solid-phase extraction Pyrethroids are an interesting class of compounds with insecticidal properties'-9 and are widely used in the storage of foodstuffs. The products currently in use are summarized in Fig. 1 (fourth generation pyrethroids). All are characterized by the structural feature of an ester group with a cyano group in an a-position. This paper describes a polarographic method of assaying total residues of this group of compounds in wheat and maize in storage.The method is based on qualitative analysis by thin-layer chromatography (TLC) for the identification of the specific pyrethroid present in the matrix and on the determina- tion of total pyrethroid concentration by differential-pulse polarography (DPP) of 3-phenoxybenzaldehyde formed quantitatively by rapid and reproducible alkaline alcoholysis of the sample. The method was verified by double analysis of several samples using gas chromatography (GC). 1 0 The electrode reduction process, upon which the analytical method is based, was also studied. Experimental Materials and Methods Methanol, chloroform, acetonitrile and cyclohexane were LiChrosolv (Merck, Darmstadt, Germany).For polaro- graphy, analytical-reagent grade tetraethylammonium bro- mide (Carlo Erba, Milan, Italy), doubly distilled water and triply distilled mercury were used. Reference standards for pyrethroids were obtained from Lab Service Analytica (Bologna, Italy). The reference standard for 3-phenoxy- benzaldehyde was obtained from Aldrich (Steinheim, Germany). For TLC, high-performance (HP) TLC plates (silica gel 60) with a concentrating zone (Merck, Darmstadt, Germany) were used. Equipment The electrochemical equipment used consisted of an EG & G Princeton Applied Research Model 264A polarographic analyser with a Model 303A stationary drop electrode and a Model RE 0073 x-y recorder. The nuclear magnetic resonance (NMR) spectra were recorded with a Bruker Model AC 200 spectrometer. Polarographic Conditions The polarographic conditions are summarized in Table 1.Isolation of Product by Potassium Hydroxide Treatment of C yanopyrethroids About 200 mg of pyrethroid was added to 60 ml of 0.2 mol I-' KOH solution in methanol-water (4 + 1). The solution was saturated with nitrogen before adding the pyrethroid, and was kept under nitrogen for the duration of the alcoholysis. The solution was kept cool and stirred for 1 h. Chromatography then revealed two products, neither of which was the original pyrethroid. Stirring was suspended and the solution extracted three times in 20 ml aliquots with cyclohexane. The cyclohex- ane phase was dried on anhydrous sodium sulfate and vacuum concentrated. When the volume of organic solvent was reduced to 3 ml, 400 mg of silica gel A0 was added and the mixture dried.The gel was transferred to the top of a chromatographic column (30 X 2 cm) filled with silica gel 60. On elution with a mobile phase consisting of cyclohexane- chloroform-methanol (6 + 1 + O . l ) , pure 3-phenoxybenzal- dehyde and the methyl ester of the acid were separated. Sample Analysis Extraction A 20 g sample of flour obtained by milling a wheat or maize grain sample for 10 s was placed in an Allihn tube (10.5 X 2cm; Steroglass, Perugia, Italy) with a porous G-1 septum holding a Sartorius (Gottingen, Germany) No. 13400-20 S fibre-glass filter. The column was eluted with cyclohexane- chloroform (4 + 1) five times, each time 5 ml of the mixture was added. The mobile phase was collected by free fall and then the residue was dried by vacuum aspiration.The total extract thus obtained (25 ml) was subsequently purified. Sample preparation The extract was run through two LC-Si (1 g) Supelclean (Supelco, Bellafonte, PA, USA) columns mounted in series on a Visiprep solid-phase extraction vacuum manifold (Supelco). The columns had first been activated with 18 ml cyclohexane-methanol (4 + 1). The columns were dried by aspiration and treated with two 1.5 ml fractions of mobile phase consisting of cyclohexane- chloroform-methanol (6 + 1 + 0.1); the eluates were combined and dried. The residue was made up with 2 ml of methanol, supplemented with 0.05 mol 1-1 alcoholic solution of KOH (100 PI) and allowed to react for 7 min. After alkaline alcoholysis, the solution was diluted with 7 ml of methanol.Before addition of KOH, 20 PI of the solution was run on an HPTLC concentation zone plate (fixed phase: silica gel 60)184 ANALYST, FEBRUARY 1993, VOL. 118 0 1 Cypermethrin 2 Deltamethrin CN 3 Fenpropathrin 4 Fenvalerate CH(CH3)Z I NHCHCOOCH I CN 0 5 Fluvalinate and eluted three times with cyclohexane-acetic acid (1 + 0.05). After elution, the plates were treated with a methanolic solution of sulfuric acid (5%) and heated in an oven at 80°C for 5 min. Several fluorescent spots of different colours corresponding to the compounds 1-7 (Table 2) were revealed by Wood's lamp (Model UVL-21, Ultra-Violet Products, San Gabriel, CA, USA). Polarographic analysis After centrifugation, 8 ml of the methanol solution were placed in a polarographic cell, supplemented with 2 ml of de-ionized water and 200 mg of tetraethylammonium bro- mide, and analysed by DPP.The calibration enabled the peak height at - 1.4 V to be read, giving the amount of 3-phenoxy- benzaldehyde in the cell, which indicated the amount of pyrethroid in the sample. Results and Discussion Polarography Study None of the pyrethroids analysed was electroactive at the mercury cathode in acid or neutral water-alcohol solution or in non-aqueous solution (acetonitrile). Polarography per- formed in basic, non-alcoholic environments revealed cathode processes that were subject to fast, uncontrolled variation. These reduction processes can be ascribed to hydrolysis of the ester group. Under the water-alcohol solution conditions OH-, CH30H \ 0 II .OH + R-C-0 1 CH3 7 Scheme 1 Cyanopyrethroid alcoholysis 6 Flucynathre 1.5 Fig.1 Fourth generation pyrethroids structural formulae ? Table 1 Polarographic conditions 1.25 Indicator electrode Reference electrode Ag-AgC1 Counter electrode Scanning speed Stationary Hg drop; drop time 1 s 1 mm diameter, 25 mm long platinum wire 2 mV s- for DPP, 5 mV s-1 for DC and 100 mV s- for CV technique 4- CN I 1.0 ' I I 0 5 10 -u P" Impulse amplitude 50 mV De-aeration time 50 min Fig. 2 Half-wave potential curve of 3-phenoxybenzaldehyde as a function of pHANALYST, FEBRUARY 1993, VOL. 118 185 used, the process leads to the formation of the methyl ester of the aliphatic residue and 3-phenoxybenzaldehyde (7), an electroactive chemical species. 11-13 The alcoholysis scheme is therefore as follows (Scheme 1): This is confirmed by NMR and mass spectrometric identifi- cation of compound 7 and the different specific methyl esters of the carboxylic residues of the pyrethroids shown in Fig.1. The electrochemical reaction of compound 7 was studied in methanol-water solution (4 + 1) and consisted of a single process in the pH range 1.5-14. The half-wave potential as a function of pH is shown in Fig. 2. This is an example of an irreversible system in which the electrode process is accom- panied by a proton-transfer reaction, the pH increases with reduction in the rate of supply of hydrogen ions and the electrodic reduction occurs at a more negative potential. Under these experimental conditions the process cannot be used analytically as there is no correlation between the limiting diffusion current and concentration.According to the literature,14 it is reasonable to suppose that this behaviour is due to acid or base catalysis or equilibria, which stabilize the hydrated, non-electroactive forms responsible for the uncon- trolled variation in the concentration of the electroactive aldehyde. Analytical experimental conditions were obtained using tetraethylammonium bromide in methanol-water (4 + 1) as the support electrolyte. Under these conditions, the half-wave potential of the reduction of 3-phenoxybenzaldehyde was -1.40 V with respect to the Ag-AgC1 electrode, the process was controlled by diffusion and the limiting current was linear with respect to concentration. Electrolysis of 3-Phenoxybenzaldehyde 3-Phenoxybenzaldehyde was reduced at fixed potential on a mercury cathode, integrating the Faraday current.The reduction mixture was run on a TLC plate. Two reaction products were found at all times, and the number of electrons in the process was 1.3 after the passage of a change of 75 C and the complete disappearance of the aldehyde. At intermediate times, the number of electrons, calculated on the basis of the polarographic decrease in aldehyde, increased from 1.1 to 1.4. The two final products were separated by column chromato- graphy and identified as 3-phenoxybenzyl alcohol and 1,2-di- (3-phenoxyphenyl)ethane-1,2-diol derived from two parallel processes (see Scheme 2): These results are in accord with literature data on the reduction of the aldehyde group.14 The compounds were identified by mass spectrometry and NMR.In the NMR spectrum of the mixture, the value of the total integral of aliphatic protons, calculated with respect to 18 aromatic protons and corresponding to molecular doubling, gave a composition of the reduction mixture of 75% diol and 25% alcohol. The composition was calculated with a simple linear system of two equations in two unknowns which took into account that the ratio of aliphatic to aromatic protons is double for the alcohol. The presence of two distinct signals with a total integral of two for the aliphatic protons in the spectrum of the pure diol is due to the non-equivalence of protons in the meso- and racemic forms, occurring as a result of the presence of two chiral centres; the hydroxyl signals were also distinct, as for the meso- and racemic tartaric acid.From the ratio of the partial integrals of the aliphatic protons it was calculated that the two forms are in the ratio 60:40, but the predominant form was not determined. For tartaric acid, the meso-form has smaller chemical shifts. Fig. 3 shows the patterns of molar fractions of the aldehyde, alcohol and diol detected during electrolysis: reduction to two electrons was always according to Scheme 2 and the process became more significant when the concentration of aldehyde was low. This is in accord with the increasing trend in the number of electrons calculated at intermediate times, and even the final value calculated (1.3) is very close to the theoretical value (1.2) assuming the one-electron process to account for 7.5% and the two-electron process for 2.5% of the reaction.The concentrations were determined, over the period of electrolysis, by polarographic analysis for the aldehyde and TLC with densitometric detection for the reduction products. + e - 0 &CH- 3-Phenoxybenzyl alcohol Scheme 2 Proposed cathodic reduction process of 3-phenoxybenzaldehyde186 ANALYST, FEBRUARY 1993, VOL. 118 0.8 K .O 0.6 c 2 Lc L 0.4 5 0.2 0 20 40 60 80 Charge/C Fig. 3 Molar fraction variations during controlled potential elec- trolysis. A, 3-Phenoxybenzaldehyde; B, 3-phcnoxybenzyl alcohol; and C, 1,2-di(3-phenoxyphenyI)ethane-1.2-diol 1600 7 1200 - E m x . 5 800 2 e 0, - a" 400 0 400 800 1200 1600 Gas chromatographyhg ml 1 Fig.5 methodologies for cyanopyrethroid dosage of maize ( r = 0.9917) Comparison between polarographic and gas-chromatographic Table 2 Chromatographic parameters for TLC identification of pyret hroids spot elution/ Compound mm 1 Cypermethrin 65 2 Deltamethrin 58 3 Fenpropathrin 52 4 Fenvalerate 45 5 Fluvalinate 41 6 Flucynathre 32 7 3-Phenoxybenzaldehyde 76 Detection Wood's lamp limits coloration (PPb) Light yellow 24 Dcep/bright-yellow 20 Light blue 24 Pink 26 Yellow-orange 20 Peach-pink 29 Ycllow-green 21 1 .o 1.5 1 .o 1.5 -EN Fig. 4 Differential-pulse polarography of ( u ) maize and (b) wheat extracts. A, Pyrethroid-free extracts; B, extract with added fluvali- nate corresponding to 2.8 pg of 3-phenoxybenzaldehyde; and C, extract with added fluvalinate corresponding to 5.6 big of 3-phenoxy- benzaldehydc Analytical Method The most appropriate conditions for alkaline hydrolysis were established first.Operating under the experimental conditions described at three different pyrethroid contents (0.1, 0.2 and 0.3 mg ml-I), five samples of fluvalinate and one of each of the other pyrethroids were analysed. When the mixture was checked by TLC (mobile phase , cyclohexane-chloroform- methanol, 6 + 1 + 0.1; fixed phase, silica gel 60; 50 pl deposits), the pyrethroid spot had completely disappeared after only 7 min, the 3-phenoxybenzaldehyde and alcohol- derived ester spots were constant and there were no secondary spots. In this mobile phase, the six pyrethroids had the same (0.62) or very similar retardation factor (RF) values (fluvali- nate 0.60).This was very useful for subsequent extractions from maize and wheat. Analysis of the alcoholysis solution by DPP showed values for compound 7 of between 99.4 and 100.9% of the theoretical value of 100% hydrolysis in all cases; this gave a mean value of 99.7 & 0.5% and a relative standard deviation (RSD) never exceeding 0.5%. On the basis of these results, we developed a method for the determination of pyrethroids in maize and wheat. Cyanopy- rethroids were previously extracted with a non-polar solvent and purified by solid-phase extraction, the total pyrethroid residue concentration was determined by DPP of 3-phenoxy- benzaldehyde formed quantitatively by rapid and reprodu- cible alkaline alcoholysis of the sample. The abundant oily component of the extracts made purification of the analysis sample necessary.Simple extraction left matrix components that did not give specific voltammetric responses but caused a background pattern which prevented any possible analytical application of the polarographic method. The results were satisfactory when the operating conditions defined in the Experimental section were respected. Fig. 4 shows the polarographic resolution of an extract of maize and one of wheat before and after the addition of appropriate amounts of fluvalinate. The reproducibility of the analytical process from maize and wheat was determined using a pool of samples of graminae not treated with chemical agents but ground and supplemented with appropriate amounts of pyrethroid. In every case the recoveries were between 91.1 and 94.3%, the standard deviation never exceeded 1.3% and the RSD was always below 0.5%.The regression line was constructed directly with 3-phen- oxybenzaldehyde. Plotting current (n A) against concentration (pg per 10 ml of cell solution) in the conccntration range 1-30 pg per 10 ml, the following relations were obtained on the basis of six regression lines determined on the same day and six more determined one per day over a period of six days: (maize) y = (47.80 + 1 . 4 3 ) ~ - (11.44 + 0.47) Y = 0.9987 (wheat) y = (47.01 + 1.61)~ - (10.72 + 0.26) Y = 0.9991 where Y is the correlation coefficient. The two lines were not significantly different (p = 0.05). For the six cyanopyrethroids considered, it follows that the method is useful in the range of concentration 0.1-3.0 mg kg- 1 and that the detection limits are well below the maximum allowable level of pyrethroid (1 mg kg-1) stipulated by ltalian law (Ministerial Decree of June 6, 1985). These results led us to compare our method with GC.10 Proceeding as described for the evaluation of recovery, 30 laboratory samples were prepared ( 5 for each compound) by treating maize and wheat samples with the various pyrethroidsANALYST, FEBRUARY 1993, VOL.118 187 and 1 2 3 with a content ranging randomly from 200 to 1500 ppb of active principle (pyrethroid); every sample of the extracted organic derivate was analysed by GC before submission to polarographic analysis. Fig. 5 shows the distribution of polarographic versus GC data for maize: the two analytical methods are essentially in agreement ( r = 0.9917). Similar results were found for wheat (Y = 0.9875).Discrimination between different compounds was performed by TLC analysis before alkaline alcoholysis: fluorescent spots of different colour, corresponding to the compounds 1-7, were revealed by Wood’s lamp; Table 2 reports satisfactory identification levels for single compounds adequate detection limits. References Lund, A. E., and Narahashi, T. F.. Pharmacol. Exp. Ther., 1981, 219, 464. Pelhate, M., Hue, B., and Satelle, D. B., in Insect Neurobiology and Pesticide Action, Society of Chemical Industry, London, 1980, p. 65. Narahashi. T., and Lund, A. E . , in Insect Neurobiology and Pesticide Action. Society of Chemical Industry, London, 1980, p. 497. 4 5 6 7 8 9 10 11 12 13 14 Narahashi, T., Neuropharmacology of insects, Ciba Foundation Symposium 88, ed. Pitman, London, 1982, p. 291. Gammon, D. W.. Pestic. Sci., 1979, 9, 78. Gammon, D. W., Brown, M. A., and Casida, J. E., Pestic. Biochem., 1981, 15, 181. Van den Bercken, J . , Kroese, A. B. A., and Akkermans, L. M. A., in Neurotoxicology of Insecticides and Phermones, ed. Narahashi, T., Plenum, New York, p. 183. Vijerberg, H. P. M., Van der Zalm, J., and Van der Bercken, J . , Nature (London), 1982, 295, 601. Goodman Gilman, A., and Goodman, L. S . , Le Basi Farma- cologiche della Terapiu, ed. Montanaro. N., Zanichelli, Bologna, 1991, p. 1541. Fitch, W. L., Helisten, C. C., Visser, I. M., and Miller, W. W., in Analytical Methods for Pesticides and Plant Growth Regula- tors, eds. Zweig. G., and Sherma, J., Academic Press, New York, 1984, vol. 13, p. 84. Zuman, P., Barnes, D., and Ryvolova-Kejharova, A., Discuss. Faraday SOC., 1968, 45, 203. Zuman, P., J. Electroanal. Chem., 1977, 75, 523. Sedon, J. H., and Zuman, P., J . Org. Chem., 1976, 41, 1957. Feoktistov, L. G., and Lund, H., in Organic Electrochemistry, eds. Baizer, M. M., and Lund, H., Marcel Dekker, New York, 2nd edn., 1983, p. 315. Paper 2/04044A Received July 28th, 1992 Accepted October 7th, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800183

出版商:RSC    

年代:1993

数据来源: RSC