摘要:

ANALYST, MARCH 1993, VOL. 118 249 Isomeric Characterization of Polychlorinated Biphenyls Using Gas Chromatography-Fourier Transform Infrared/Gas Chromatography- Mass Spectrometry Doyle M. Hembree, Jr.,* Norman R. Smyrl," Willard E. Davis and David M. Williams Plant Laboratory, Oak Ridge Y-12 Plant,t Martin Marietta Energy Systems, lnc., Oak Ridge, TN 37831-8189, USA A new technique combining both mass spectrometry and infrared spectroscopy to analyse simultaneously the components from a single gas chromatographic injection has been applied to the quantitative and qualitative characterization (including distinguishing positional isomers) of polychlorinated biphenyl (PCB) mixtures. The sensitivity of vibrational spectroscopy to subtle differences in structure was shown to be highly complementary to gas chromatography-mass spectrometry (GC-MS) for qualitative identification of individual PCB isomers and congeners.A key feature of the infrared apparatus is the provision for low-temperature trapping (approximately 77 K) of the GC effluent for subsequent analysis. This technique produces infrared spectra that resem ble normal room-temperature condensed-phase spectra (as opposed to the gas-phase spectra produced by light-pipe gas chromatography-Fourier transform infrared spectroscopy) and leads to lower detection limits (500 pg for 3,3',4,5-tetrachlorobiphenyl). The GC-MS portion of the instrument provides superior quantitative capabilities with su b-picogram detection limits possible using selective ion monitoring. Keywords: Polychlorinated biphenyl; gas chromatograph y-mass spectrometry; gas chromatography- Fourier transform infrared spectroscopy; isomer identification In the past, mixtures of polychlorinated biphenyls (PCBs) have been widely used in many industrial applications because of their many desirable chemical and physical properties. However, the discovery that some members of this relatively large class of compounds (209 isomers and congeners are possible) are toxic to both humans and other forms of life has led to the virtual elimination of their use.' Efforts to dispose of these environmentally persistent compounds, often using improper disposal methods, have resulted in extensive con- tamination of many of the world's ecosystems including the North Atlantic, which may contain up to 79% of the total environmental inventory of USA-derived PCBs.1 In order to assess properly the impact of the environmental burden of PCBs, it is essential that analytical procedures exhibiting adequate sensitivity and capable of making accurate quali- tative as well as quantitative determinations be developed. In particular, it has been suggested that studies be undertaken to assess the environmental fate and toxic effects of individual PCB congeners.2.3 Complex mixtures of PCBs such as those isolated from typical environmental samples are usually analysed by gas chromatography-mass spectrometry (GC-MS) or GC with an electron-capture detector. Many studies have been performed with capillary column GC to identify and quantify, at least tentatively, PCBs based on relative retention times and response factors.2-7 However, isomer-specific identification of individual PCB compounds using GC-MS or GC is extremely difficult in practice and in most routine analyses results are simply reported as total PCBs, with no attempt made to identify individual compounds.The difficulty in distinguishing between similar chemical species becomes important for classes of compounds, such as PCBs, in which only a few members possess a property of interest, such as biological activity (polycyclic aromatic hydrocarbons and nitrogen het- erocycles are further examples of important classes of compounds where isomer-specific identification is desirable). Vibrational spectroscopy is an extremely sensitive probe of molecular structure and many attempts have been made to exploit this fact for analytical purposes.X,') The use of Fourier transform infrared (FTIR) spectroscopy to detect the effluent from a capillary column gas chromatograph using a gold- coated light-pipe has proved powerful (high picogram detec- t I 0 - I I I I I I c (3, m .- I I I I I I 10 ~ 1 7 12 13 14 15 16 17 Ti me/mi n * Authors to whom correspondence should be addressed.t Operated for the US Department of Energy by Martin Marietta Energy Systems, Inc., under contract DE-AC05-840R21400. Fig. 1 Chromatograms of Aroclor 1242. (a) GC-MS total ion chromatogram. ( h ) IR functional group chromatogram (14OO-165O cm-l region) of Aroclor 1242 from the same GC injection as in (a)250 ANALYST, MARCH 1993, VOL. 118 tion limits for strong infrared absorbers), but the technique is several orders of magnitude less sensitive than GC-MS, and produces gas-phase spectra that are difficult to use with the large, commercial condensed-phase libraries.The recent introduction of low-temperature trapping to capture the effluent from a capillary GC column,10 combined with the use of a highly sensitive HgCdTe (MCT) TR detector, has reduced FTIR detection limits to the range of GC-MS (low picogram levels for strong infrared absorbers) and gives rise to condensed-phase spectra resembling those obtained at room temperature using conventional sampling methods. In this study, an instrument combining the capabilities of both FTIR and MS11-13 to analyse simultaneously the various com- ponents in PCB mixtures as separated by capillary-column GC will be discussed. When isomer-specific identification is required GC-FTIR is shown to be highly complementary to conventional GC-MS.Experimental Apparatus The common sample introduction point for the GC-FTIW GC-MS instrument was a gas chromatograph (Model 5890A, Hewlett-Packard, Palo Alto, CA, USA) equipped with two matched 25 m methyl silicone gum capillary columns (Hew- lett-Packard Model HP-1) connected to a single injector. The matched columns gave a 50 : 50 split ratio (ie., 50% of the injection was routed to the MS system and 50% to the FTIR system). The FTIR system was a Digilab FTS-45 spectrometer (Biorad, Digilab Division, Cambridge, MA, USA) equipped with an accessory (Digilab Tracer) for capturing the GC effluent on a ZnSe plate held at low temperature (approxi- mately 77 K).Mass spectrometric data were collected with a Hewlett-Packard Model 5970B mass selective detector (MSD). Further details of the experimental arrangement are presented elsewhere." Table 1 Identification of the chromatographic peaks in Fig. 1 Peak Rctcntion number time/min* IR identification I 11.05 2 12.01 3 12.95 4 13.31 5 13.88 13.91 6 14.10 - 7 14.21 8A 14.63 8B 14.74 8C 14.83 9 15.04 10 15.06 - 11 - 12A 15.74 12B 15.81 I2C 15.85 2,2'-Dichlorobiphenyl 2,4'-Dichlorobiphcnyl 2,2' ,5-Trichlorobiphenyl 2,2' ,3-Trichlorobiphenyl and another trichloroisomer 2,4',5-Trichlorobiphenyl + 2.4,4'-Trichlorobiphenyl 2' ,3,4-Trichlorobiphenyl 2,3,4'-Trichlorobiphenyl 2,2' .S ,5'-Tetrachlorobiphenyl 2,2' .4,5'-Tetrachlorobiphenyl 2,2' ,4,5-Tctrachlorobiphenyl 2,2' ,3,5'-Tetrachlorobiphcnyl 3,4,4'-Trichlorobiphcnyl and a tctrachlorobiphcnyl (no library spectrum) A tctrachlorobiphcnyl (no library spectrum) 2,4,4' ,5-Tetrachlorobiphcnyl 23' .4' ,5-Tetrachlorobiphenyl 2,3' ,4,4'-Tetrachlorobiphenyl A tetrachlorobiphenyl (no library spectrum) MS identification [hit index] 2,4'-; 2,2'- 3,3'-; 4,4'-; 2,6-; 3.4- [98] 4,4'-; 3,3'- [99] 2,4'-; 2,3- [98] 2,3,4-; 2,4',5-; 2,2',5- [99] 2,4,6-; 2,4',5-; 2,4,5-; 2,2',5- [99] 2,3,4-; 2,3',5-; 2,4',5-; 2,4,4'-; 2,4.5- 1991 2,3',5-; 2,3,4-; 2,4,6-; 2,4',.5-; 2,4,4'- [98] 2,4,6-; 2,3',S- [99] 2.2',5,6'-; 2,2',4,5'-; 2,2'5,5'-; 2,2',3,4-; 2,3,4',6-; 2,3,3',5'-; 2,2',3,4'-; 2.2',6,6'- [99] 2,3,4' ,6-; 2,3,3' ,S I - ; 2,3' ,4' ,6-; 2,3,3',4-; 2,2',4,5'-; 2,2.'5,5'-; 3,3' ,5 ,.5'-; 2,2' ,3,4-; 3,3'4.5'- [99] 2,2',3,4-; 2,3,3',5'-; 2,3',4',6-; 2,3.3',4'-; 2,2',4,5'-; 2,2'5,5'-; 3,3' ,5,5'-; 2,3,4' ,6-; 2,2',6,6'- [99] 2,2',3,4-; 2,3,4', 6-; 2,3,3',5'-; 2,3',4',6-; 2,3,3',4'-; 2,2',5,6'-; 2,2',3,4'-; 2,2',6,6'- [99] 2,2',3,4-; 2,2',3,4- [97]? 2,3,4',6-; 2,3,3',5'-; 3,3',4,5'-; 2,3' ,4' ,6-; 2,3,3' ,4'- ; 3,3' ,5,5'-; 2,2' ,6,6'-; 2,2' ,3,4-; 2,3',4',5- [99] 3,3',4,5'-; 2,3,3',4'-; 2,3,4.5-; 3,3',5,5'-; 2,4,4',6-; 2,3',4',5-; 2,3',4,4'-; 2,2',6,6'-; 2,3',5,5'- [99] 3,3',4,4'-; 2,3',4,4'-; 2,2'6,6'-; 2,2',4,4'-; 2,2',3,4-; 2,3,4',6-; 3,3'.5,5'- [99] 2,3,4',6-; 3.3',4,5'-; 2,3,4,5-; 3,3' ,4,4'-; 2,3' ,4' ,5-; 2,3' ,4,4'-; 2,2'.6,6'-; 2,2',3,4-; 2,3,3',5'- [991 2,3,4',6-; 2,3,3',5'-; 3,3'4,5'-; 2,3,3',4'-; 2,3',5,5'-; 2,3',4',5-; 2.3' ,4,4'-; 2,2' ,6' ,6'-; 2.2',3,4- [99] * Rctcntion timc of thc standard of the PCB identified by IR analysis [under the samc chromatographic conditions as Fig.l(b)]. -f Mass spectrum contains both a 258 and 292 mass fragment. The 258 mass (parent ion of trichlorobiphenyls) is about twice the intensity of the 292 mass.ANALYST, MARCH 1993, VOL. 118 2.51 Reagents One hundred and thirty PCB standards (100 pg ml-1 in hexane) were analysed as received from Ultra Scientific (North Kingstown, RI, USA) in order to build GC-FTIW GC-MS spectral libraries. Standard IR and mass spectra were obtained for 50 ng samples (100 ng injection). Aroclor 1242 (Ultra Scientific), a common PCB mixture containing 42% m/m chlorine, was used to test the capabilities of the instrument for analysing complex mixtures of PCBs.Procedure The gas chromatograph was operated in the splitless mode at a measured flow rate of 0.4 ml min-1 at 250°C. High-purity helium treated to remove water was used as the carrier gas. All of the work described in this study was performed with a temperature programme consisting of a 2 min hold at 80°C followed by a ramp to 250°C at 10°C min-1. The MSD was operated in the scanning mode from 50 to 550 u. Infrared ‘functional group’ chromatograms were generated by moni- toring the total IR intcnsity in the 1400-1650 cm-1 region, where most PCBs have relatively intense absorption bands due to C-C skeletal stretching. Individual PCBs were identi- fied using computer searches of the National Institute of Standards and Technology mass spectrometry library (-41 000 compounds) and the IR library generated from PCB standards (using the Digilab spectral library search routine).Results and Discussion The GC-MS total ion chromatogram and the IR functional group chromatogram for a 1 p,l injection of Aroclor 1242 (1 pgml-1 in hexane) are shown in Fig. l ( a ) and (b), respectively. Identifications of the various chromatographic peaks are given in Table 1. The MS compound identifications 0.04 [- I 0.03 - 0.02 - a, C 0.01 - 0 I I 1 I I I 0.06 0.04 0.02 0 I I I I I I 1 2000 1800 1600 1400 1200 1000 800 Wavenurnberlcrn Fig. 2 IR spectra of two PCB isomers isolated from Aroclor 1242. ( a ) A, Standard spcctrum of 2,4’,S-trichlorobiphenyl; and B, spectrum from lcading edgc of peak 5 in Fig. l(h) (13.85 min).( h ) A, Standard spcctrum of 2,4,4’-trichlorobiphenyl; and B, spectrum from trailing cdgc of peak S in Fig. l(h) (13.99 min) are followed by a ‘hit’ index, with a value of 100 representing a perfect match between the unknown and the standard spectrum (only the compounds with the highest hit index o r indexes are given in Table 1). In most cases, multiple compounds had identical indexes, and the order in which the compounds are listed is not significant; MS identifications that agree with the IR results are listed in bold typeface. Examination of the mass spectrometric results clearly illus- trates the difficulties associated with distinguishing positional isomers. In many cases, the correct isomer, identified by both the IR spectrum and retention time of the standard PCB, is not among the compounds with the highest MS hit index (chromatographic peaks 4, 6, 7, 8C, 9, 10, 12A and 12B).In contrast to the MS results, when complete chromato- graphic separation was achieved, the IR spectral search provided unequivocal identification of the unknown PCB , except in cases where a standard spectrum was not available. The TR spectra obtained near the centre of chromatographic peaks containing co-eluting compounds are a superposition of the spectra of all of the compounds giving rise to the peak. Under such conditions, a spectral search usually provides no compound identification. However, by taking spectra from the leading and trailing edges of the peak, relatively ‘clean’ spectra of the first and last components comprising the peak can be obtained.As an example, the initial IR results for peak 5 in Fig. 1 gave no plausible identification of the compound(s) giving rise to the peak, while MS analysis indicated that five trichlorobiphenyl isomers were equally likely (i. e., these f 11 543 I 1800 1600 1400 1200 1000 800 Wavenurnber/crn Fig. 3 with the GC-FTIR Tracer unit (1800-700 cm-l) IR spectrum of 2.5 ng of 3,3’,4,5-tctrachlorobiphcnyl obtained r, h I I I I I 1600 1500 1400 1300 Wave n u rn be r/cm 1 Fig. 4 Signal-to-noise enhancement from FTIR signal averaging. A , IR spcctrum of 1 ng of 3,3’,4,S-tctrachlorobiphenyl (200 scans at 8 cm-1 resolution); and B, 1R spectrum of thc same 3,3’,4,5-tetra- chlorobiphenyl sample as in A obtained ‘on-thc-fly’ (4 scan5 at 8 cm-I resolution)252 ANALYST, MARCH 1993, VOL.118 isomers have essentially identical mass spectra). The 1R spectra obtained from the leading and trailing edges of peak 5 are shown in Fig. 2, and are compared with standard spectra of the identified compounds. Even though HP-1 methyl silicone gum GC stationary phase was used in this study, the elution order of the 2,4’,5- and 2,4,4’-trichlorobiphenyl isomers is consistent with relative retention times reported for SE-54.2 The quantitative response of the GC-FTIR/GC-MS instru- ment was investigated using 3,3’ ,4,5-tetrachlorobiphenyl as a model compound. The IR spectrum of this compound (25 ng) is shown in Fig. 3. The ability to enhance the signal-to-noise (S/N) ratio of smaller amounts of compounds obtained with the Tracer GC-FTIR unit is demonstrated in Fig.4. The bottom spectrum was obtained ‘on-the-fly’ (4 scans at 8 cm-1 resolution) for 1 ng of 3,3’,4,5-tetrachlorobiphenyl. After completion of the GC run, the deposition plate was returned to the spot where the bottom curve was obtained. Two hundred scans at 8 cm-1 resolution were signal averaged to obtain the top spectrum in Fig. 4. Three spectral features at 1543, 1435 and 1366 cm-1 are easily observed in the top spectrum but are not discernible in the bottom spectrum. Straight line regression calibrations for both instruments were computed for 3,3’,4,5-tetrachlorobiphenyl from 1 to 50 ng. The response for both instruments is linear over most of the range. The coefficients of determination (R2) were 0.987 for the GC-MS data and 0.975 for the GC-FTTR calibration (three measurements were averaged for each point on the calibration lines).The detection limit for GC-MS was significantly lower than for FTIR because PCBs are relatively weak IR absorbers, in general, and this tetrachloroisomer is weaker than most other PCRs [a (absorptivity) = 1138 cm-1 measured experimentally for the 1543 cm-1 band]. Based on an S/N of 2, the detection limit for GC-MS was 200 pg using the total ion peak height and 2 pg using selective ion monitoring of the 292 mass fragment, while a detection limit of 500 pg was obtained using the peak height of the 1543 cm-1 band in the FTIR spectrum. Conclusions The ‘double-hyphenated’ technique of GC-FTIR/GC-MS provides, for the first time, a convenient means of obtaining condensed-phase IR spectra (at high sensitivity and similar to conventional spectra obtained at room temperature) simul- taneously with the corresponding mass spectrum.Because the frequencies of the normal modes of vibration of very similar molecules, even positional isomers, are usually different, TR spectroscopy using low-temperature trapping is a powerful qualitative complement to conventional mass spectrometric detection of the effluent from a capillary column gas chroma- tograph. The authors thank R. E. Carroll for preparation of the manuscript. References 1 2 3 4 5 6 7 8 9 10 11 12 13 The National Research Council, Committee on the Assessment of PCBs in the Environment, Polychlorinated Biphenyls. National Academy of Sciences, Washington, DC, 1979. Mullin, M. D., Pochini, C. M., McCrindlc, S . , Romkcs, M . , Safe, S. H., and Safe, L. M., Environ. Sci. Technol.. 1984, 18. 468. Capel, P. D., Rapaport, R. A., Eisenreich, S. J.. and Looney, B. B., Chernosphere, 1985, 14, 439. Cooper, S. D., Moseley, M. A., and Pelliuari, E. D., Anal. Chem., 1985,57,2469. Steichen, R. J., Tucker, R. G., and Mechon, E., J . Chroma- togr., 1982, 236, 113. Bush, B . , Murphy, M. J., Connor, S . , Snow, J., and Barnard, E., J. Chromatogr. Sci., 1985, 23, 509. Hasan, M. N., and Jurs, P. C., Anal. Chem., 1988, 60, 978. Schneider, J. F., Reedy, G. T., and Ettinger, D. G.,J. Cltroma- togr. Sci., 1985, 23, 49. Hembrce, D. M., Garrison, A. A., Crocombe, R. A . , Yoklcy, R. A.. Wehry, E. L., and Mamantov, G., Anal. Chem., 1981 53, 1783. Bourne, S., Hacfner, A. M., Norton, K. L., and Griffiths, P. R., Anal. Chem., 1990, 62, 2448. Smyrl, N. R., Hembree, D. M., Jr., Davis, W. E . , Williams, D. M., and Vance, J. C., Appl. Spectrosc., 1992, 46, 277. Wilkins, C. L., Science, 1983, 222, 291. Cooper, J. R., Bowater, I. C., and Wilkins, C. L., Anal. Chem.. 1986, 58, 2791. Paper 2102153F Received April 27, 1992 Accepted October 8, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800249

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, MARCH 1993, VOL. 118 253 Addition and Measurement of Water in Carbon Dioxide Mobile Phase for Supercritical Fluid Chromatography Dongjin Pyo and Doweon Ju Department of Chemistry, Kang we on National University, Kang weon -do, Korea A method for the addition of water to supercritical C02 is described. Carbon dioxide, the most widely used mobile phase in supercritical fluid chromatography, is a relatively non-polar fluid, and hence the addition of small amounts of polar modifiers could be necessary to migrate polar solutes. In this work, supercritical C02 is delivered from the pump to a p-Porasil column that is saturated with water. After passing through the p-Porasil column, supercritical C02 is changed to a new mobile phase with different polarity, and it is possible to separate polar samples by using this new mobile phase.The amount of water dissolved in supercritical C02 is measured by an amperometric microsensor, which is prepared from a thin film of perfluorosulfonate ionomer. Keywords: Supercritical carbon dioxide fluid; modifier; water; mobile phase Many developments have been reported since the first use of supercritical fluids as chromatographic mobile phases in 1962.1 Especially in the last few years, supercritical fluid chromatography (SFC) has progressed from a laboratory curiosity to a viable analytical technique for solving many otherwise intractable problems. However, the most com- monly used mobile phases in SFC are all relatively non-polar fluids. Carbon dioxide, the most widely used fluid, is no more polar than hexanel33 even at high densities.Solute polarity should be between that of the stationary phase and the mobile phase in order to effect a straightforward separation. Few real-world samples contain only non-polar solutes, so a major objective of research into SFC has been directed towards increasing the range of solute polarity that can be handled by the technique. To bring the SFC technique into routine use, mobile phases that are more polar than the commonly used COl are necessary. The solvent strength of supercritical C02, even at high density, is not sufficient for the elution of polar solutes. Polar mobile phases such as NH3 exhibit useful properties,j but a more practical way to extend the range of compounds separable by SFC is the use of mixed mobile phases. The addition of modifiers (generally organic solvents) to super- critical CO? changes the polarity of the mobile phase and also leads to a de-activation of the column packing material.5 In capillary SFC, most separations are carried out with pure COz because of its compatibility with a flame-ionization detector (FID); indeed, except for formic acid and water the addition of any common modifier precludes the use of an FID.6 Modifiers are essential in packed-column SFC for the elution of polar compounds7 and are extensively used.Several papers have reported the influence of modifiers on peak shape, selectivity and retention time in capillary and packed column SFC.s.7.8 In this work, a new mobile phase with enhanced polarity was prepared by passing supercritical C 0 2 fluid through a p-Porasil column saturated with water.It is shown that some polar samples can be separated with this new mobile phase in packed-column SFC. Experimental A CCS (Computer Chemical Systems, Avondale, PA, USA) Model 5000 supercritical fluid chromatograph was used with a 100 X 2 mm packed column (Nucleosil diol). This system was equipped with a C14W loop injector (Valco Instruments, Houston, TX, USA) and an FID. Supercritical fluid chroma- tography grade C 0 2 (Scott Specialty Gases, Plumsteadville, PA, USA) was used as a basic mobile phase. Experimental conditions for SFC separations were as follows: supereri tical C 0 2 at 150 "C, pressure programmed from 27.56 to 34.45 MPa (4000-5000 Ib in-2) at 0.28 MPa (40 lb in-2) min-l, FID detection at 300 "C, 10 cm3 min-1 restrictor flow at 10.34 MPa (1500 Ib in-2).For the addition of water to supercritical COl, a p-Porasil column, which is manufactured for normal-phase high-performance liquid chromatography by Waters (Milford, MA, USA), was used, its functional group being a hydroxy group (-OH). The p-Porasil column was saturated with water by means of a Model 5560 reciprocating pump (Varian, Palo Alto, CA, USA) and placed between the pump and the injector. In order to measure the amount of water dissolved in the supercritical fluid, an amperometric microsensor was used, which was prepared from perfluorosulfonate (PFSI) polymer.9 A constant-current power supply (0.1 pA, Sungun, Seoul, Korea) was used to measure the voltage drop across the sensor.The sensor output was recorded on a strip-chart recorder (Knauer, Berlin, Germany). Results and Discussion When dealing with the use of modifiers, it should be mentioned that some problems always arise. First, a binary mixture of eluents can contaminate the instrument. In particular, modifiers remaining in a pump can cause corrosion of the pump and be slowly eluted during the next run. This can affect the time required to achieve chemical equilibrium for the subsequent separations. Second, many modifiers can diffuse in the laboratory and contaminate the laboratory air. To overcome these problems, we designed the system that is shown in Fig. 1. A polar modifier (water) is added to the pressurized carbon dioxide fluid after the pump, and hence no modifier remains in the pump.Supercritical C 0 2 is delivered from the pump to the p-Porasil column, which is saturated with water. When supercritical C 0 2 passes through the Restrictor FID n \n p-Porasil r column I I Oven U cop Fig. 1 modifier to thc supercritical fluid mobile phase Schematic diagram of the apparatus used for adding a polar254 ANALYST, MARCH 1993, VOL. 118 p-Porasil column, H20, retained on the -OH groups of p-Porasil by hydrogen bonding, can dissolve in the pressurized supercritical fluids. With this method, non-polar supercritical CO? can have the characteristics of a polar mobile phase because it can absorb polar solvent (H20). Therefore, after passing through the p-Porasil column, supercritical C02 is changed to a new mobile phase with different polarity, and it is possible to separate polar samples using this new mobile phase.An experiment to separate some polar samples (insecticides and fungicides) with this new mobile phase was performed. Figs. 2 and 3 are chromatograms for mixtures of insecticides and fungicides obtained using a mixed mobile phase (super- critical COz-water). In contrast to the experiment in which only COZ was used as mobile phase, excellent separations were obtained. When only C02 was used for these samples, unseparated and very broad peaks were observed in about 25 min. The addition of a small amount of water to supercritical CO2 reduced the retention and improved the peak shapes. The phenomena are in accord with the results reported by Blilie and Greibrokk.5 The structure of each peak is shown in Table 1.To measure the amount of water dissolved in supercritical CO2, a polymer film9 (i. e., a film of PFSI ionomer), which has a high affinity for water, was used. When the PFST film was in contact with two electrodes and a constant current flowed through the film, the water that partitioned into the film from the surrounding environment was electrolytically decom- posed. The change of voltage across two electrodes was used as a measure of the water content of the environment surrounding the sensor. Fig. 4 shows a cross-section of the modifier (water) sensor used in this work. A platinum wire was wrapped with PFSI thin film and another platinum wire was wound in a coil over the assembly. The sensor constructed in this way was placed in a plastic tube, and the entire modifier-measuring device was assembled together as shown in Fig.5. A constant-current (d.c.) source was used to supply a current of about 0.1 pA to the sensor. ‘The resistance of the PFSI film was changed according to the water content of the supercritical COz fluid, A I I 0 5.0 10.0 Time/m in I I I 1 27.56 28.94 30.32 Pressure/M Pa Fig. 2 Chromatogram of a mixture of insecticides and fungicides. A, Thiolix; B, kitazine; and C, captan Table 1 Structures of peaks in the chromatograms Chromatogram Peak Commercial name Chemical name Structure Fig. 2 A Thiolix (insecticide) 1,4,5,6,7,7-Hexachloronorborn-5-ene- 2.3-dimeth an01 sulfite CI B Kitazine (fungicide) C Captan (fungicide) Fig. 3 A Hinosan (fungicide) S-Benzyl-0, 0-diisopropyl phosphorothioate N-(Trichloromethylthio)cyclohex-4-ene- 1 -2-dicarboximide 0-Ethyl-S, S-diphenyl dithiophosphate g=J-Jo>N-s-cc13 0 (i? EtO- P - S-Ph \ S-Ph S I EtO B Parathion (insecticide) O,O-Diethyl-0-4-nitrophenyl phosphorothioate EtO - !- 0 NO^ C DDVP (insecticide) 2,2-Dichlorovinyl dimethylphosphateANALYST. MARCH 1993.VOL. 118 255 t - m C U m .- 0 5.0 10.0 Time/min 27.56 28.94 30.32 Pressu re/M Pa Chromatogram of a mixture of insecticides and fungicides. A, Fig. 3 Hinosan; €3, parathion; and C. DDVP 1 I PFSI Pt wire Fig. 4 Cross-section of the modifier sensor m Sample gas In chYsi Air in FM S P DC I I 1 - c out [PPCI Fig. 5 Schematic diagram of the device used to mcasure water content: P, pump: V, solenoid valve; S, sensor; FM, flow meter; MP, magnesium perchlorate; R, recorder; M, multimeter; DC.12 Vpower supply; PPC, programmable process controller; Ch, Si, charcoal, silica gel; and CS, current source and the voltage difference between the two platinum wires was recorded and measured. Fig. 6 shows chromatograms for the injection of ( a ) air saturated with water, and (b) supercritical C 0 2 fluid saturated with water, through a p-Porasil column at different time intervals. Air saturated with water was generated by passing air, with bubbling, through water contained in two sequential bottles, and was injected directly into the sensor through a three-way solenoid valve (Radio Shack, Fort Worth, TX, USA). The determination of water in COT after passage through the p-Porasil column was performed on-line, i.e. , supercritical C 0 2 fluid was injected through a C14W loop- injector, and the sensor was placed directly after the p-Porasil column.t E rn a, L Y m a, Q a, > m a, n .- .- +d - 1 0 s 7 8 I h 5 s Y S * Time - I1 Fig. 6 Relative peak height at different time intervals (flow rate: 0.5 dm3 min-1, temperature: 20 "C). ( a ) For air saturated with water and ( 6 ) for carbon dioxide fluid saturated with water 50 40 E E 2 30 L rn a, L .- 20 a, a 10 0 20 40 60 80 100 Relative humidity (%) Fig. 7 Sensor response for various relative humidity levels (18.8, 37.1, 47.2, 58.3. 70.4 and 100% relative humidity). A, 1 0 ; B, 15; C, 20; and D, 25 "C By using the data (Fig. 7) o n the correlation of peak height with percentage relative humidity (RH) at different tempera- tures, the water content of supercritical C02 fluid, after passage through the water-saturated p-Porasil column, could be determined.From the data in Fig. 6 ( b ) , the amount of water dissolved in supercritical C02 was measured as 54% RH or 3.00 x 10-3 g dm-3 at a pressure of 25.33 MPa (250 atm). The amount of water in the mobile phase remained constant for about 1 h, but after that decreased exponentially. Therefore, after three or four runs the column must be refilled with water for practical use. Sulfuric acid solutions of known compositions10 were used to generate the standard RH streams. This investigation was supported by a grant from the Korea Science and Engineering Foundation. 1 2 3 4 5 6 7 8 9 10 References Klesper, E., Corwin, A. H., and Turner, D. A., J. Org. Chem., 1962, 27, 700. Hyatt, J . A . , J . Org. Chem., 1984, 49, 5097. Yonker, C. R., Frye, S. L., Lalkwarf, D. R., and Smith, R. D.. J . Phys. Chem., 1986, 90, 3022. Kuei, J. C., Markides, K. E . , and Lee, M. L., J . High Resofui. Chromatogr., 1987, 10, 257. Blilie, A. L., and Greibrokk, T., Anal. Chem., 1985,57,2239. Wright, B. W., and Smith, R. D., J . Chromaiogr., 1986, 355, 367. Schmidt, S., Blomberg, L. G., and Campbell, E. R., Chromato- graphia, 1988. 25, 775. Yonker, C. R., and Smith, R. D., J . Chrornatogr., 1986, 361, 25. Huang, H., and Dasgupta, P. K . , Anal. Chem., 1990,62,1935. CRC Handbook of Chemistry and Physics, ed. Weast, R. C., CRC Press, Boca Raton, FL, 70th edn., 1989, pp. E-43 s. Paper 2104825F Received September 8, 1992 Accepted November 16, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800253

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, MARCH 1993. VOL. 118 257 Oxazole-based Tagging Reagents for Analysis of Secondary Amines and Thiols by Liquid Chromatography With Fluorescence Detection Toshimasa Toyo'oka, Hitesh P. Chokshi, Robert G. Carlson, Richard S. Givens and Susan M. Lunte* Center for Bioanalytical Research, University of Kansas, 2095 Constant Avenue, Lawrence, KS 66047, USA The reactions of three fluorescent tagging reagents, 2-chloro-4,5-diphenyloxazole (DICLOX), 2-fluoro-4,5- diphenyloxazole (DIFOX) and 2-chloro-4,5-bis(p-~,,N-dimethylaminosulfonylphenyl)oxazole (SAOX-CI), with thiols and amines are reported. Emission maxima for the diphenyloxazole (DIOX) and SAOX derivatives of amines were 420 nm (Aex 320 nm) and 485 nm (Aex 360 nm), respectively. The emission wavelengths for the DIOX- and SAOX-thiols are 390 n m (Aex 310 nm) and 425 nm (Aex 330 nm), respectively. In all cases, the derivatives exhibited strong fluorescence whereas the reagents themselves exhibited only weak fluores- cence.The labelled derivatives are very stable, less than 5% decomposition occurs after heating at 60 "C for 2 h. Fluorescence intensities of the amine derivatives were higher in neutral and alkaline than in acidic solutions and were virtually independent of solvent polarity. The thiol derivatives exhibited fluorescence intensities that were relatively constant under all conditions studied. The relative reaction rate toward both thiols and amines was DIFOX > SAOX-CI > DICLOX. The reaction of proline with DIFOX was complete after 60 min at room temperature at pH 9.3.However, the yield with SAOX-CI was only 70% at 60 "C after 3 h, and only a small amount of proline could be derivatized with DICLOX (less than 3%). Thiols, on the other hand, reacted relatively rapidly with SAOX-CI. Therefore, SAOX-CI was used for the determination of thiols and DIFOX was employed for amines in all subsequent studies. Detection limits (signal-to-noise ratio = 2) for authentic DIOX-amines ranged from 3.7 t o 28.4 fmol, and SAOX-thiols ranged from 1.2 t o I .9 fmol. Keywords: Derivatizing reagents; secondary amine; liquid chromatography; fluorescence; amino acid Thiols and amines are important constituents of most living organisms. Several fluorogenic reagents have been employed for the determination of secondary amines, the most common being 5-dimethylaminonaph thalene- 1-sulfonyl chloride (dansyl-CI), 9-fluorenylmethylchloroformate (FMOC) , and 3-khloro-7-nitrobenzofurazan (NBD-Cl) and its fluoro ana- logue (NBD-F).1-8 Likewise, a number of different types of fluorescent tagging reagents, including N-(iodoacetyl- aminoethy1)-5-naphthylamine-I-sulfonic acid,"'() N-dansyl- aziridine, 1 1 N-substituted maleimide, 12-14 bimanesls--'s and halogeno-benzoxadiazoles, 1'1-24 have been designed for thiols and are used as pre- and/or post-column derivatization reagents. Many of these reagents suffer from incomplete reactions, highly fluorescent hydrolysis products, solvent dependency of fluorescence or reagent toxicity. In this paper, a series of halogeno-diaryloxazoles that function as fluorescent labelling reagents for secondary amines and thiols is reported.The reaction of the reagents 2-fluoro-4,s-diphenyloxazole (DIFOX), 2-chloro-4,S- diphenyloxazole (DICLOX) and 2-chloro-4,5-bis(p-N, N- dimethylaminosulfony1phenyl)oxazole (SAOX-Cl) (Fig. 1) with these analytes is described. The spectroscopic charac- teristics, stability and chromatographic behaviour of the labelled compounds are also given. Experimental Materials L-Alanine, L-proline, trans-4-hydroxy-~-proline, L-prolyl-L- leucine, L-prolyl-glycyl-glycine, L-prolyl-L-leucyl-glycine amide, reduced glutathione (GSH), N-acetylcysteine and 2-mercaptopropionylgl ycine (MPG) were purchased from Sigma (St. Louis, MO, USA). Sodium tetraborate decahy- drate was obtained from Fluka (Ronkonkoma, NY, USA). Disodium dihydrogen ethylenediaminetetraacetate (Na2EDTA), sodium phosphate dibasic heptahydrate, sodium phosphate monobasic monohydrate, sodium carbon- ate and hydrochloric acid (35%) obtained from Fisher (Fair Lawn, NJ, USA) were also used.Orthophosphoric acid (85%) and acetonitrile were of HPLC (high-performance liquid chromatography) grade (Fisher). Water used was purificd by a NANOpure I1 system (Sybron/Barnstead, Boston, MA, USA). All other chemicals were of analytical-reagent grade and were used without further purification. 4,5-Diphenyl-2- 3H-oxazolone (DlFOX-OH) and DICLOX were prepared by the method of Gompper and Herlinger.2s Apparatus Two LC-GA pumps (Shimadzu, Columbia, MD, USA) were employed. Gradient elution was controlled by an SCL-6A system controller (Shimadzu) and the data were collected and analysed by a C-R3A Chromatopac (Shimadzu).All samples for chromatographic analysis were injected onto the column with a SIL-GA Auto Injector (Shimadzu). The column used in these studies was a 5 pm Supelco LC-8 column (250 x 4.6 mm i.d.; Supelco, Bellefonte, PA, USA). A Shimadzu SPD-6AV ultraviolet/visible (UV/VlS) detector and a Shimadzu RF-535 fluorescence detector were used in tandem to monitor the column eluent. All mobile phase solutions were sonicated for 15 min prior to use to remove air bubbles. The flow rate was 1 .O ml min-1. . X + NU: DICLOX: R = H, X = CI DIFOX: R = H, X = F SAOX: R = SOzN(CH,),, X = CI Nu: RSH, RNH2, R2NH * To whom correspondcncc should be addrcsscd. Fig. 1 Structures of 2-fluoro-4,s-diphenyloxazolc (DIFOX) and 2-chloro-4,5-bis(j-N, N-dimcthylaminosulfonylphcny1)oxazolc (SAOX-Cl) and thcir rcactions with thiols and secondary amincs258 ANALYST, MARCH 1993, VOL.118 Melting-points (m.p.s.) were determined on a Thomas- Hoover capillary melting-point apparatus and are uncorrec- ted. Infrared (IR) spectra were determined on a Beckman (Fullerton, CA, USA) Acculab 3 grating spectrophotometer or an IBM IW32 Fourier transform infrared (FTIR) spec- trometer. Proton nuclear magnetic resonance ('H ,NMR; 6,) and 13C NMR ( 6 ~ ) spectra were recorded on a Varian (Sunnyvale, CA, USA) XL300 or Bruker (Billerica, MA, USA) AM-500 spectrometer. Chemical shifts are reported in ppm ( 6 ) relative to Me& (6 = 0.00) as an internal standard. The multiplicity of 13C NMR signals were determined by single frequency off-resonance decoupling experiments or distortionless enhancement by polarization transfer.The 1gF NMR (6,) spectra were recorded on the Varian XL300 and the chemical shifts are reported in ppm (6) relative to cx,cx,cx-trifluorotoluene (C6HsCF3; 6 = 63.7 ppm) as an internal standard. The UVNIS absorption spectra were recorded on a Hewlett-Packard (Palo Alto, CA, USA) 8450A diode array spectrophotometer. Fluorescence emission spec- tra were obtained on an Aminco-Bowman spectrofluorimeter using the 1P 28 photomultiplier with a 1 cm quartz cell. Wavelengths are reported in nanometres and are uncorrected. Widths at half-height ( WlI2) in the fluorescence emission spectra are reported in cm-1. Mass spectra were acquired on a Ribermag quadrupole R-10-10, a Varian MAT CH-5 or a VG-ZAB (VG, Danvers, MA, USA) high-resolution mass spectrometer.The thin-layer chromatography analyses were carried out on Analtech silica gel plates (250 pm thickness) with a fluorescence indicator. Synthesis of DIFOX An 18-crown-6-acetonitrile complex was prepared using the method of Goekl and Cram.26 To a solution of DICLOX (1.62 g, 6.35 mmol) in acetonitrile (60 ml), anhydrous KF (3.0 g) and the 18-crown-6-acetonitrile complex (1.5 g) were added. The resultant mixture was heated under reflux for 24 h. The reaction mixture was cooled and filtered, the filtrate was concentrated, and hexane (2 x 50 ml) added to the residue obtained. The resultant mixture was stirred for 10 min and then filtered rapidly.Upon concentration, the hexane filtrate gave a yellow oil, which was distilled (b.p. 130 "C at 0.02 mmHg) to afford 0.85 g (56%) of product. 6H (CDC13) 7.63 (m, ArH), 6.55 (m, ArH), 7.37 (m, ArH); aF (CDC13) -98.2; rnlz (%) 239 (M+, loo), 192 (18), 165 (64), 105 (4), 89 (28), 77 (28) (Calc. for CISHIOFNO: M+, 239.0746. Found M , 239.0748). Synthesis of SAOX-Cl To a suspension of the oxazolone in POC13 (30 ml, 323 mmol) at 0°C triethylamine (0.61 ml, 4.44 mmol) was added drop- wise. The reaction mixture was then heated at 100 "C for 7 h, and the excess P0Cl3 removed on a rotary evaporator. The residue obtained was dissolved in methylene chloride, and the organic layer washed with cold saturated sodium hydrogen carbonate. The methylene chloride layer was separated, dried (over MgSO4) and concentrated to give a yellow solid (1.92 g, 92%), which was chromatographed on silica gel (100 g).Elution with methylene chloride and 10% EtOAc afforded SAOX-Cl as a white solid (1.45 g, 70%); m.p. 222-224 "C; tiH (CDC13) 7.82 (6 H), 7.74 (2 H), 2.77 (6 H, s ) , 2.76 (6 H s); 6c- (CDC13) 147.6, 147.3,137.2,136.8,136.4, 134.9,131.1,128.4, 128.4, 126.3, 126.9, 37.92, 37.90; IR(CHC13) 3020, 2970, 2820,1600,1510,1455,1400, 1350,1160,1090,1050,955,840 cm-I; mlz (YO) 469 (M+, 30), 362 (39), 313 (19), 253 (21), 190 (27), 163 (27), 92 (20) (Calc. for C19H20N30sS2Cl: M+, 469.0532. Found: M , 469.0537). Synthesis of 4,5-Bis@-N,N-dimethylaminosulfonylphenyl)- oxazolone (SAOX-OH) Chlorosulfonic acid (60 ml) was carefully added to DTFOX- OH (7.3 g, 0.03 mmol) at 0 "C.The reaction mixture was heated at 55-60 "C for 4 h, then cooled and added dropwise to ice-water (500 g). The pale-yellow solid was filtered, washed with water (4 1) and 2.0 g (2.22 mmol) of the crude mixture were added to dry benzene (100 ml) and concentrated to dryness to remove traces of water in the sample. hH ([2H6]-Me2SO) 7.87 (1 H , d), 7.82 (1 H, d), 7.74 (1 H, d), 7.61 1.82 (3 H , s, CH,); SC ([2H6]-Me2SO) 153.9, 135.7, 133.9, 133.0, 131.8, 131.5, 129.0, 128.3 (two carbons), 125.4, 123.5, 37.6, 37.5; IR (KBr) 1725, 1590, 1550, 1320 cm-1; rnlz (YO) 451 (M+, 26), 344 ( 5 ) , 235 (7), 211 (6), 185 (7), 164 (lo), 104 (21), 103 (24), 76 (25), 44 (100) (Calc. for C19H21N306S2: M+, 451.0870. Found: M , 451.0873). (1 H , d), 3.70 (1 H, dd, S-CH-H), 3.65 (1 H , dd, S-CH-H), Synthesis of Diphenyloxazole (DIOX) Derivatives of Proline To a mixture of proline (40 mg, 0.348 mmol) and triethyl- amine (90.6 ml) in benzene (25 ml) was added DIFOX (70 mg, 0.293 mmol) in hexane (1 ml). The reaction mixture was stirred at room temperature for 16 h, diluted with water (SO ml), and extracted with benzene (100 ml) and then with methylene chloride (100 ml).The aqueous layer was made acidic with concentrated hydrochloric acid and extracted with diethyl ether (200 ml). The ether layer was dried over MgS04 and concentrated to afford the adduct product as a white solid (48 mg, 49%). aH (CDC13) 7.62 (2 H, m, ArH), 7.45 (2 H, m, ArH), 7.30 (6 H , m, ArH), 4.58 (1 H, dd, N-CH-CO); 3.71 (2 H, m, N-CH2), 1.9-2.5 (4 H, m); bC (CDC13) 172.5, 158.7, 139.8, 133.4, 131.4, 128.7, 128.6, 128.5, 128.4, 128.3, 127.7, 2900, 1740, 1625, 1510, 1490, 1455, 1420, 1250, 1225, 1030, 700 cm-1; mlz (%) 334 (M+, 27), 290 (70), 289 (22), 288 (34), 262 ( 3 9 , 247 (23), 235 (13), 178 (16), 165 (40), 158 (25), 131 (21), 106 (6), 105 (65), 104 (40), 103 (25), 77 (loo), 70 (70), 44 (79) (Calc.for C2,,HlXN203: M+, 334.1316. Found: M , 334.1305). 125.5, 61.3, 48.1, 29.6, 24.2; IR (CHC13) 3400-2300, 3000, Synthesis of SAOX Derivatives of Proline To a solution of proline (115 mg, 1 mmol) in 0.1 mol 1-1 sodium carbonate (100 ml) was added SAOX (96 mg, 0.2 mmol) in acetonitrile (100 ml). The reaction mixture was stirred at room temperature for 3 h, concentrated to remove acetonitrile, and the aqueous solution extracted with ethyl acetate (3 X 50 ml).The aqueous solution was made acidic with 1 mol 1-1 HCI and extracted with ethyl acetate (3 x 50 ml). The ethyl acetate layers obtained from extraction with aqueous HCl were combined, washed with water, dried (over MgS04) and concentrated to afford SAOX-proline as a yellow solid (60 mg, 60%); m.p. 241-243 "C. SH ([2H6]-Me2SO) 7.82 (4 H , br s), 7.70 (4 H , dd), 4.49 (1 H , br d , N-CH-COOH), 3.67 (2 H , m, N-CH2), 2.65 (6 H , s), 2.62 (6 H , s), 1.9-2.45 (4 H, m); 6C ([2H6]-Me2SO) 173.3, 158.5, 138.2, 136.7, 136.5, 134.3, 132.9, 132.4, 128.4, 128.3, 128.0, 124.8, 59.8, 47.7, 37.5,30.2,23.7; IR (KBr) 3420,2080-2880, 1725,1640, 1590, 1455, 1400, 1340, 1160,950,840, 750,700 cm-1; mlz (%) 548 (M+, 1) 504 (4), 476 (l), 461 (l), 396 (l), 342 (1) , 288 (2) , 212 (7), 190 ( 3 ) , 165 (3) (Calc.for C24H2XN407S2: M+, 548.1399. Found: M , 548.1393). Both DIOX-alanine and SAOX-alan- ine were also synthesized and fully characterized. Synthesis of DIOX-N-acetylcysteine To a solution of N-acetylcysteine (163 mg, 1 mmol) in 0.1 mol 1-1 sodium carbonate (100 ml) was added DIFOX (152 mg, 0.64 mmol) in acetonitrile (100 ml) and the mixture stirred at room temperature for 2 h. The reaction mixture wasANALYST, MAKCH 1993, VOL. I18 259 concentrated to remove acetonitrile and the remaining aqueous solution extracted with ethyl acetate (3 x 50 ml). The aqueous layer was thcn made acidic with 1 rnol 1-1 HCI and extracted with ethyl acetate (3 x 50 ml). The ethyl acetate layers obtained from extraction with aqueous HCI were combined, washed with water, dried (over MgS04) and concentrated to afford DIOX-N-acetylcysteine as a white solid (170 mg, 70%); m.p.131-133 "C; bH ([2H6]-MeS04) 8.46 (1 H, d, NH), 7.56 (4 H, m, ArH), 7.43 (m, 6H), 4.66 (1 H, m, (3 H, s, CH,); aC ([2H6]-MeS04) 171.5, 169.4, 158.1, 146.7, 135.7, 131.4, 129.0 (two carbons), 128.7, 128.5, 127.8, 127.4, 126.2,5 1.7,33.3,22.3; IR (KBr) 3400-2600,1725,1650,1500, 1445, 1205, 765, 695 cm-1; UV/VIS (acetonitrile) A,,, (log E) 298 (4.18); fluorescence emission (acetonitrile) h,,, (Wl12) 236 (31), 165 (20), 147 (11), 130 (15), 121 (15), 106 (6), 105 (63), 104 (49, 103 (19), 77 (51) (Calc. for CI7Hl3NO3S: M+, 328.0987. Found: M , 382.0999). CH), 3.70 (1 H, dd, S-CH-H), 3.65 (1 H, dd, S-CH-H), 1.82 385 (3900); mlz (Yo) 382 (M+, l ) , 278 ( 5 ) , 266 (l), 253 (30), Synthesis of SAOX-N-acetylcysteine To a solution of N-acetylcysteine (163 mg, 1 mmol) in 0.1 mol 1-I sodium carbonate (100 ml) was added SAOX-C1 (96 mg, 0.02 mmol) in acetonitrile (100 ml).The reaction mixture was stirred at room temperature for 2 h, then concentrated to remove acetonitrile, and the aqueous solution extracted with ethyl acetate (3 x 50 ml). The aqueous solution was made acidic with 1 rnol 1-1 HCl and extracted with ethyl acetate (3 X SO ml). The ethyl acetate layers obtained from extraction with aqueous HCl were combined, washed with water, dried (over MgS04) and concentrated to afford a yellow solid (101 mg, 85%); m.p. 179-181 "C; aH ([2H6]-MeS04) 8.46 (1 H, d, NH), 7.86 (4H, m, ArH), 7.82 (4H, m), 4.68 (1 H, ddd, CH), 3.79 (1 H, dd, S-CH-H), 3 S O (1 H, dd, S-CH-H), 2.65 (3 H, S , CH3), 2.64 (3 H, d, CH3), 1.82 (3 H, S , CH3); ([2H6]-MeS04) 171.3, 169.4, 159.9, 146.2, 136.2,135.2,134.8, 134.7, 131.3, 128.29, 128.26, 128.12, 126.7, 51.4, 37.5, 33.3, 22.2; IR (KBr) 3400,3070,2600,1730,1650,1600,1500,1335, 1155, 1045, 950, 840, 760, 750, 640 cm-l; UVNIS (acetoni- trile) A,,, (log E) 327 (4.25); fluorescence emission (acetoni- trile) h,,, (W1/2) 385 (3900); m/z (YO) 467 (M+, 16), 253 (3), 212 ( 3 ) , 165 ( 3 ) , 149 (l), 87 (21), 43 (100) (Calc.for C19HZ1NOSS3: M+, 467.0642. Found: M , 467.0633). Stock Solutions Borate buffer (0.1 moll-1; pH 9.3) was prepared by dissolving 19.07 g of sodium tetraborate decahydrate in 1 1 of water containing 2 mmol 1-i Na2EDTA.Both DIOX (-N-acetyl- cysteine, alanine, proline and -OH) and SAOX derivatives (-N-acetylcysteine, -alanine, -proline and -OH) were dis- solved in CH3CN (1 mmol I-' each). Stock solutions (1 mmol 1 - 1 ) of thiols or amines were prepared in water containing 1 mmol l-1 Na2EDTA. All stock solutions, except the borate buffer, were stored in a refrigerator at 4 "C. Effect of Solvent Polarity and pH on the Fluorescence Intensity of the Authentic Derivatives Authentic DIOX or SAOX derivatives ( 5 pmol 1-1) were dissolved in solutions with pH varying from 2 to 12 (0.05 rnol 1-1 Britton-Robinson buffer). The Britton-Robinson buffer was prepared by mixing proportional amounts of Solutions A and B to achieve the desired pH; where Solution A consists of 4.5 g of 85% H3P04, 2.4 g of acetic acid (AcOH) and 2.47 g of boric acid (H3B03) diluted to 1 1, and Solution B is 0.2 rnol 1-1 NaOH (8.0 g 1-1).The fluorescence emission spectra were recorded in a 1 cm quartz cell on an Aminco- Bowman spectrofluorimeter (Model SPF 4-8940 SP) equipped with an IP 28 photomultiplier tube. The emission spectra were recorded with an excitation wavelength of 313 nm. Fluores- cence efficiencies were obtained by comparison with 2,4- diphenyloxazole (Aldrich, Milwaukee, W1, USA) as standard (Qfl = 1.0). The Qfl value for a given derivative was calculated according to the equation Qfl/@"f = Zfl(1 - lo-A')//f,r(l - lo-") where Qfi and Q f l r are the fluorescence quantum efficiencies for the given oxazole derivative and the standard, A and A' are the absorbances of the derivative and the standard solutions, and Zf, and are the areas under the emission curves of the derivative and the standard, respectively.Before the fluores- cence emission spectra were recorded, the UV/VTS spectra of the samples were obtained using a diode array spectropho- tometer. For fluorescence efficiency measurements, the con- centrations of the solutions were adjusted so that the absorbances were less than 0.1, in order to minimize the error due to inner filter effects. Quantum efficiencies were measured in 100% acetonitrile and in a mixture of 50% acetonitrile and 50% 0.05 rnol 1-1 phosphate buffer, pH 7. Determination of Reaction Rate Constants for Thiols and Amines Several 1.5 ml vials containing 0.25 ml of 1 mmol 1-1 DIFOX, DICLOX or SAOX-CI in CH3CN mixed with an equal volume of 10 pmol 1-1 analyte in 0.05 rnol I- sodium tetraborate (pH 9.3) containing 1 mmol 1 - 1 Na2EDTA were prepared.The reaction solutions of DJFOX and DICLOX were kept at room temperature (20-25 "C); SAOX-CI was heated at 60 "C. All solutions were protected from light. At fixed time intervals, a vial was removed from the water-bath and cooled in ice- water, an equal volume of 1 rnol 1-1 HCl-CH3CN (1 + 1) was added to quench the reaction and the aliquot monitored by HPLC with both UV (at 210 nm) and fluorescence detection to determine the progress of the reaction. A reagent blank without analyte was treated in a similar manner. The yield of the derivatization reaction at each point in time was calculated by comparing the peak areas with those of known amounts of the authentic derivative.The pseudo-first-order rate constants were calculated from these data. Stability of DIOX and SAOX Derivatives Several vials of each derivative ( 5 pmol 1 - 1 ) in 0.1 rnol 1 - 1 sodium tetraborate (pH 9.3) containing 2 mmol l-1 Na2EDTA were heated in a water-bath at 60 "C for 2 h . After fixed time intervals, a vial was removed and an aliquot of the solution injected onto the column. The extent of decomposition was determined from the ratio of the peak areas for the derivative before and after heating. HPLC Separation and Detection of Authentic Derivatives Stock solutions of DTOX and SAOX derivatives were diluted to suitable concentrations (0.01-1.0 pmol 1-l). An aliquot of the respective stock solutions was separated by HPLC.Detection limits [signal-to-noise ratio (S/N) of 21 of these derivatives at suitable wavelengths were calculated from the difference between peak height of the derivative and noise level. HPLC With Fluorescence Detection of Secondary Amino Acids and N-Terminal-proline Peptides Derivatized with DIFOX A 0.25 ml volume of DIFOX (1 mmol l-1) in CH3CN and 0.25 ml of a solution of a mixture of secondary amino acids and proline-containing peptides (10 pmol 1- 1 each of L-proline, hydroxy-L-proline, 1,-prolyl-glycyl-glycine and L-prolyl-L-leu- cine) in 0.05 moll-1 sodium tetraborate (pH 9.3) containing 1 mmol 1-1 Na,EDTA was added to a vial (1.5 ml volume). The260 ANALYST, MARCH 1993, VOL. 118 vial was capped and kept at room temperature protected from light for 1 h.Then 0.5 ml of 1 mol I-' HCI-CH3CN (1 + 1) was added to the solution to quench the reaction. The acidic solution was diluted to the desired concentration with H20-CH3CN (1 + l ) , and an aliquot injected onto the column. The DIFOX derivatives were separated under isocratic conditions using an eluent of 0.05 moll-1 phosphate (PH 7)-CHnCN (7 + 3). Results and Discussion Reaction Rate of Amines With DIFOX, DICLOX or SAOX-Cl The reactions of DIFOX, DICLOX and SAOX-CI with proline were compared. The time course for the formation of the labelled derivative was determined by HPLC with fluorescence detection. The production of the labelled proline by reaction with DIFOX at room temperature (RT) in 0.05 mol 1-1 sodium tetraborate (pH 9.3)-CH3CN (1 + 1) reached a maximum after 60 min and remained constant for 2 h. In contrast, DICLOX reacted with less than 3% of the proline even after 3 h under the same reaction conditions.The reaction of proline with SAOX-Cl at KT was extremely slow, with a yield of only about 5% after 4 h. However, when the latter reaction was carried out at 60 "C for 3 h the yield was increased to 70%. Pseudo-first-order rate constants of 2.26 x 10-1 and 5.63 x 10-2 min- 1 , respectively, were determined from reactions of 0.5 mmol 1-1 of the reagent (DIFOX at RT or SAOX-C1 at 60 "C) and 5 pmol 1-1 proline at pH 9.3. The reaction of alanine with DIFOX at RT was slower, 1.14 X 10-2 min-1, and was not complete even after 5 h. When the reaction was carried out in 0.1 mol I-' Na2C03 (pH 11.5), the yield of DIOX-alanine was even lower.This low yield is due in part to the competing hydrolysis of DIFOX. The use of elevated temperatures was not suitable for the derivatization reaction with DIFOX, as both the number and amount of side products increase with rising temperature. Therefore, the rec- ommended reaction conditions for DIFOX are RT at pH 9.3 in a solution containing 1 mmol 1-1 Na2EDTA. No side products other than the hydrolysis product were detected in the reaction of alanine with SAOX-CI at 60 "C over a 5 h reaction time. The reactivity of halogeno-diaryloxazoles toward nucleo- philes is affected by the electron-withdrawing substituent on the aromatic ring (Fig. l), as has been shown for halogeno- benzoxadiazoles.7-~() 2-Chloro-4,5-bis(p-nitrophenyloxazole), a dinitro derivative of DICLOX, reacts instantly with proline at RT.However, the derivative exhibits no fluorescence in polar solvents such as acetonitrile or methanol. Based on results with proline, the reactivity of p-substituted diarylox- azoles follows the order: NO2 > S02NMe2 > H. Furthermore, DIFOX is more reactive than its chloro analogue based on a comparison of DTFOX with DICLOX. Reaction of Thiols With DIFOX or SAOX-Cl The reactivity of DIFOX and SAOX-Cl toward N-acetyl- cysteine, which was selected as a representative of biological thiols, was also investigated in aqueous media (pH 9.3). The yields with SAOX-CI increased with the reaction time and reached a maximum at 60 "C after 90 min. The time-yield profile of DIFOX at RT was almost superimposable with that of SAOX-CI at 60 "C.The pseudo-first-order rate constants were measured by reacting 0.5 mmol 1-1 reagent (DTFOX or SAOX-Cl) and 5 pmol 1-1 of N-acetylcysteine at pH 9.3. The rate constants for N-acetylcysteine with both reagents were comparable: 7.90 x 10-2 min-1 (DIFOX at RT) and 1.04 x 10-1 min-1 (SAOX-Cl at 60 "C). Therefore, either reagent can be used for thiol analysis by HPLC. As the reaction of SAOX-CI with amines is slower than that with DIFOX under the same derivatization conditions, SAOX-CI appears to provide better selectivity for the determination of thiols. Effect of Solvent Polarity and pH on the Fluorescence Characteristics of DIOX and SAOX Derivatives of Amines In contrast to NBD and dansyl derivatives, oxazole derivatives show little change in fluorescence quantum efficiencies with increasing solvent polarity (Table 1 shows the values obtained under aqueous conditions and Table 2 gives these values under non-aqueous conditions).The +fl value for proline changes from 0.39 to 0.51 in going from 100% acetonitrile to 50% acetonitrile. The ++, value for the SAOX derivatives decreases slightly in fluorescence as the solvent becomes more polar (0.45-0.37). This is in contrast to NBD-hydroxyproline, which exhibits a 45-fold decrease in +fl . Likewise, dansyl derivatives exhibit low values for Qfl under aqueous conditions (Table 1). To obtain the optimum pH and suitable detection wavelengths for the determination of amines, the correlation between pH and fluorescence intensity was investigated with authentic derivatives.As shown in Fig. 2, DIFOX derivatives of proline exhibited higher fluorescence emission intensity than alanine derivatives at all pH values tested. For all DIOX-amino acid derivatives, the fluorescence emission intensities were higher in neutral and alkaline solution (pH 5-10) than in acidic solution. Protonation of the nitrogen in amino acid derivatives (DIOX-amino acids) might be occur- ring at pH values lower than 3, leading to a decrease in the fluorescence intensity, as indicated in Fig. 2. The fluorescence emission intensities of SAOX-amino acids were about twice as high as those of DIOX-amino acids. The fluorescence intensities of DIOX-OH and SAOX-OH were relatively low at all pH values when compared with the amino acid derivatives.The emission intensity of DIOX-OH was slightly higher, whereas that of SAOX-OH was slightly lower in alkaline solution (pH 10 and 12). Table 1 Fluorescence properties of oxazole derivatives under aqueous conditions Derivative D I OX-pro1 i ne* DIOX-CI* DIOX-OH* SAOX-prolinet NBD-h ydrox y- I ,-pro1 i ne$ DNS-gl ycine? SAOX-OHt L,&m (log E ) 320 (4.22) 286 (4.20) 301 (4.19) 361 (4.33) 500 (3.10) 334 (4.35) 355 (0.73) h,,,,/nm (Wu2/cm - 1 ) 420 (4300) 365 (4500) 410 (4200) 440 (4 1 00) 536 430 475 (3500) @tl Ref. 0.5 1 - 0.07 - 0.0s - 0.37 - 0.05 - 0.010 6 0.06 10 * In acetonitrile-phosphate buffer ( I + I ) (0.05 rnol 1-1. pH 7.0). Relative to Qfl = 1 .O for 2,4-diphenyloxazole in cyclohexane at he, = 313 nm.-t In H20, pH 9. Relative to Qfl = 0.55 for quinine hydrogen sulfate (in H2SOJ) at he, = 366 nm. $ In 0.01 mol 1-1 Tris-CI- buffer, pH 7. Relative to Qfl = 0.93 for fluorescein in 0.01 moll-' NaOH. Table 2 Fluorescence properties of oxazole derivatives under non- aqueous conditions h,,,/nm h,,,/nm Derivative (log E ) ( Wl/z/cm-l) @tl Ref. DIFOX-proline* 320(4.22) 420(4300) 0.39 - NBD-hydroxy-r.-prolinet 470 (2.2) 526 0.24 6 336 (0.77) SAOX-proline* 361 (4.33) 475 (3500) 0.45 - * In acetonitrile. Relative to @fl = 1.0 for 2,4-diphenyloxazole. t In methanol-HCI. Relative to Qtl = 0.55 for quinine hydrogen $ In methanol. Relative to Qtl = 0.93 for fluorescein. DNS-tryptophan$ 335 533 0.37 10 sulfate.ANALYST, MARCH 1993. VOL. 118 26 1 W 4- K O 2 4 6 8 10 12 14 PH Fig.2 Effect of pH on the fluorescence intensities of authentic DIOX derivatives: A, DIOX-prolinc; €3, DIOX-alanine; and C, DIOX-OH. Authcntic derivative ( 5 pmol 1-1 each) was dissolved in 0.05 mol 1-1 Britton-Robinson buffer (pH 2-12). The fluorescence intensities were mcasured at the maximum wavelength v) 4- .- 5 1200 ? .- z 1000 + - 5 800 .- v) C a, 4- ,E 600 a, C 9 400 2 2 200 W m a, ._ c. - K O - @---+--; A A </C - 2 4 6 8 10 12 14 PH Fig. 3 Effect of pH on the tluorescence intensities of authentic derivatives: A, DIOX-N-acetylcysteine; B, SAOX-N-acctylcysteinc; C. DIOX-OH; and D, SAOX-OH. Authentic dcrivative ( 5 pmol 1 - 1 each) was dissolved in 0.05 mol I-' Britton-Robinson buffer (pH 2-12). The fluorescence intensities were measurcd at the maximum wavelength The wavelengths for the emission maxima of DIOX-proline (420-433 nm), DIOX-alanine (422-427 nm), SAOX-proline (483493 nm) and SAOX-alanine (485-490 nm) were not greatly affected by increasing the pH above 7 (less than a 13 nm bathochromic shift).However, a 20 nm red shift for the excitation wavelength was observed in going from pH 2 to 7. In contrast, the emission maxima of the hydrolysis products DIOX-OH and SAOX-OH showed a much greater depen- dence on pH and exhibited a bathochromic shift (45-80 nm) with increasing pH. Effect of pH on the Fluorescence Characteristics of Thiol Derivatives The fluorescence emission intensities of DIOX-N-acetyl- cysteine and SAOX-N-acetylcysteine remain constant over a wide pH range. The intensities of SAOX-N-acetylcysteine were about 5 times higher than those of DIOX-N-acetyl- cysteine at all pH values tested (Fig.3). The shift in the maximum wavelength of the thiol derivatives was negligible at all pH values (Table 1). Thiol determination with these reagents is possible over a wide pH range, as the derivatives show a higher fluorescence intensity and negligible shift of the excitation and emission wavelengths at all pH values. The relatively low fluorescence intensities of the hydrolysis pro- t - m t m 0 .- 0 10 20 Time/min 0 10 20 30 40 50 60 Time/min Fig. 4 HPLC separation of authentic DIOX derivatives. ( a ) Eluent, 0.1 rnol 1-' H3P04-CH3CN ( I + 1); and ( h ) eluent, 0.05 mol I-' phosphate (pH 7.0)-CH3CN (7 + 3). Peak 1 , DIOX-OH (0.21 pmol); peak 2, DIOX-alaninc (0.16 pmol); and peak 3 , DIOX-proline (0.19 pmol).Fluoresccncc dctcction, 420 nm (Iex 320 nm). Other HPLC conditions are given in the Experimental section ducts, DIOX-OH and SAOX-OH, over the entire pH range is an advantage of these reagents because of the low interference of the side products from derivatization. Stability of DIOX and SAOX Derivatives The stability of DTOX-proline, SAOX-proline, DIOX-N- acetylcysteine and SAOX-N-acetylcysteine was examined under various conditions. The derivatives exhibited less than 5% decomposition over a period of 2 h in 0.1 moll-' sodium tetraborate (pH 9.3) containing 2 mmol 1 - 1 Na2EDTA at 60 "C. Comparable results were observed in 100% acetonitrile solution at 60 "C. No decomposition was seen at pH 1 and RT for any sample tested.The oxazole derivatives have been found to be insensitive to visible (room) light, but will decompose after prolonged exposure to UV radiation .27 HPLC Separation and Detection of Authentic Derivatives The HPLC separation of a mixture of authentic amine derivatives was carried out using a reversed-phase octyl-silica gel column (Supelco LC-8). At acidic pH, the hydrolysis products SAOX-OH or DIOX-OH elute before the amino acids [Fig. 4(a)]. The elution order at pH 7.0 was opposite to that in acidic solution [Fig. 4(b)] with DIFOX-OH eluting at approximately 33 and 66 min, respectively.262 ANALYST, MARCH 1993, VOL. 118 Table 3 Retention timcs (tR) and detection limits of authentic DIOX and SAOX dcrivatives Derivativc Eluent* D1 OX-pro1 i ne% 1 3 DIOX-alaninct 1 3 DIOX-OH1 1 3 S AOX-prolincS I 2 SAOX-alaninc$ 1 2 S AOX- 0 H ?: 1 2 Detection limit r,lmin (S/N = 2)/fmol 8.6 5.8 11.5 8.7 11.6 10.9 10.1 14.6 56.6 313 8.6 77.') 8.8 5.3 16.Y 5.3 10.4 9.0 13.8 7.3 32.7 53 1 11.9 126 * Eluent composition: 1,0.05 rnol I-' phosphate (pH 7.0)-CH3CN (7 + 3); 2, 0 .1 rnol 1-1 H3P04-CH3CN (6 + 4); and 3, 0.1 rnol 1-1 H3POJ-CH3CN ( I + 1). t A,, = 320; A, = 420 nm. 3 A,, = 360; A,,,, = 485 nm. t - m C 13) m .- 1 I I I I I I 0 10 20 30 40 50 60 Time/m in Fig. 5 HPLC separation of proline-containing pcptidcs derivatized with DIFOX. Peak 1, hydroxy-L-proline (99 fmol); peak 2, 1.-proline (130 fmol); peak 3, L-prolyl-glycyl-glycine (110 fmol); peak 4, L-prolyl-1.-leucine (96 fmol); and peak 5 , DIOX-OH. Chromato- graphic conditions: 0.05 moll-1 phosphate (pH 7.0)-CH3CN (7 + 3) The detection limits (S/N = 2) calculated from peak heights are listed in Table 3.The detection limits for authentic DIOX-alanine and DIOX-proline were 5.8-14.6 fmol; those for SAOX-alanine and SAOX-proline were 5.3-9.0 fmol. In contrast, the detection limits of the hydrolysis products were extremely high (DIOX-OH, 77.9-313 fmol; SAOX-OH, 126-531 fmol) compared with amino acid derivatives. The low yield (70%) under derivatization conditions with SAOX-Cl makes DIFOX the more useful reagent for the determination of secondary amines. Detection limits under typical deriva- tization conditions for proline were 100 fmol at an S/N of 2.28 The HPLC separation of authentic N-acetylcysteine deriva- tives and hydrolysis products was carried out using a reversed- phase octyl-silica gel column (Supelco LC-8).The hydrolysis products in acidic eluents (18.4 as against 17.2 min and 22.1 as against 19.6 min) overlapped the peaks of the N-acetylcysteine derivatives. In contrast, the complete separation was accom- plished at a neutral pH [0.05 rnol 1-1 phosphate (pH The detection limits for authentic SAOX-N-acetylcysteine and DIOX-N-acetylcysteine in a neutral eluent were 1.2 and 3.6 fmol, respectively. Those in acidic eluent were 1.5 fmol 7 .O)-CH,CN] . t - m C 13) m .- 0 10 20 30 Time/min 0 10 20 30 40 Ti me/m i n Fig. 6 HPLC separation of thiols derivatizcd with SAOX-CI. ( a ) Eluent, 0.1 mol 1-1 H3P04-CH3CN (65 + 35); and ( 6 ) eluent, 0.05 mol 1-1 phosphate (pH 7.0)-CH3CN (7 + 3).Pcak 1, GSH (SO fmol); peak 2, N-acetylcysteine (SO fmol); peak 3 . MPG (59 fmol); and peak 4, SAOX-OH. Column, LC-8 (250 X 4.6 mm i.d., 5 pm) at RT; fluorescence detection, 425 nm (A,, 330 nm); and flow rate, 1.0 ml min-' (SAOX-N-acetylcystei ne) and 6.2 fmol (DIOX-N-ace tyl- cysteine) (Table 2). The detection limits of SAOX-N-acetyl- cysteine were about 3 times lower than those of DIOX-N- acetylcysteine in both eluents. The detection limits using conventional fluorescence detection are lower than for other derivatization methods for thiols described previously.5 Based on reaction speed, selectivity of the reagents toward thiols, and the detection limits of thiol derivatives, SAOX-CI is a more effective tagging reagent than DIFOX for the determi- nation of thiols.HPLC Separation of Proline Peptides Derivatized With DIFOX The separation of a mixture of secondary amino acids (L-proline and hydroxy-L-proline) and a mixture of short- chain peptides (L-prolyl-L-leucine, L-prolyl-glycyl-glycine and L-prolyl-L-leucyl-glycine arnide) was studied by isocratic elution using a reversed-phase HPLC column (LC-8). Aceto- nitrile-phosphate buffer (pH 7.0) was selected as the eluent, as the compounds give a higher fluorescence intensity in neutral solution (pH 7.0) than in acidic solution. Furthermore, the capacity factor of the hydrolysis compound (DIOX-OH) of DIFOX is larger in neutral solution than in acidic solutionANALYST, MARCH 1993. VOI,. 118 263 and its fluorescence intensity at pH 7.0 is lower. Under these conditions, the DIOX-OH would not be expected to interfere in the separation of the target amines.The HPLC separation of DIOX-hydroxy-L-proline, DIOX- L-proline and DIOX-peptides by isocratic elution with 0.05 moll-’ phosphate buffer (pH 7.0)-CH3CN (7 + 3) is shown in Fig. 5 . The elution order of the DIOX derivatives was hydroxy-r>-proline, I,-proline, L-prolyl-glycyl-glycine and L-prolyl-L-leucine. The detection limits (S/N = 2) were 3.7, 12.6, 7.4 and 28.4 fmol, respectively. With UV detection at 210 nm using the same effluent, DIOX-OH eluted at 58 min (Fig. 5 ) . HPLC Separation of Thiols Derivatized With SAOX-CI The separation of some thiol compounds labelled with SAOX-CI was investigated by isocratic elution using a reversed-phase column (LC-8). As shown in Fig. 6, complete separation of three biological thiols was obtained in acidic [O. 1 mol 1-1 H3P04-CH3CN (6.5 + 3.5)] and neutral [0.0.5 mol 1-1 phosphate (pH 7.0)-CH3CN (7 + 3)]; SAOX-OH in neutral and acidic eluents was eluted at 32 and 19 min, respectively.I n acidic eluent, SAOX-OH interfered with the separation of the derivatives of cysteine and homocysteine. The peak areas are almost the same for each thiol derivative. The SAOX-CI itself did not elute under these isocratic conditions. The detection limits (S/N = 2) for GSH, N-acetylcysteine and MPG in neutral eluents were 1.4, 1.3 and 1.4 fmol, respectively. I n acidic eluent, the extrapolated detection limits (S/N = 2) were 1.2 (GSH), 1 ..5 (N-acetylcysteine) and 1.9 fmol (MPG). Conclusions The purpose of this paper is to report the use of DICLOX, DIFOX and SAOX-CI for the determination of secondary amines and thiols.DIFOX was found to be the most useful reagent of the three tested for detection of secondary amines. The reaction rate with primary amines is an order of magnitude slower than that with secondary amines, leading to increased selectivity. In addition, the excitation maxima for the secondary amino acid derivatives (320 nm) are well suited to He-Cd laser excitation using the 32.5 nm line. Owing to its excellent sensitivity and its selectivity for thiols, SAOX is ideal for the determination of thiols. Examples of the use of these reagents for the analysis of biological samples will be reported elsewhere. This work was supported in part by funds from Kansas Technology Enterprise Corporation and Oread Laboratories and an equipment grant from Shimadzu Corporation USA.I 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 References Imai, K., and Toyo’oka, T.. in Design and Choice of Suitable Labelling Reagents for Liquid Chromatography, cds. Frei, R. W., and Zeck, IS., J. Chromatogr. Lib., 1988, vol. 39A, Elsevier, Amsterdam, pp. 209. Chen, R. F., Arch. Biochem. Biophys.. 1967, 120, 609. Einarsson, S., Josefson, B., and Lagerkvist, S . , J . Chrornatogr., 1983,282, 609. Ghosh, P. B . , and Whitehouse, M. W., Biochem. J . , 1968,108, 155. Imai, K., and Watanabe, Y., Anal. Chim. Acta, 1981, 130,377. Ahnoff, M., Grundevik, I., Arfwidsson. A.. Fonselius. J . . and Persson, B.-A., Anal. Chem., 1981, 53, 485. Watanabe, Y . . and Imai, K., J. Chromatogr., 1982, 239, 723. Andrew, J . L., Ghosh. P., Ternai, B., and Whitehouse, M. W., Arch. Biochem. Biophyb., 1982, 214, 386. Wu, C. W., and Stryer, L., Proc. Natl. Acad. Sci. USA, 1972, 69, 1104. Fricdman, F. K., Chang, M. Y.. and Bcychok, S . , J . Biol. Chem., 1978, 253, 2368. Johnson, J. D . , and Schwartz. A., .I. Biol. Chem., 1978, 253, 5243. Kanaoka, Y.. Yakugaku Zasshi, 1980. 100, 973. Weltman, J . K., Szaro, R. P., Frackelton, A . R., Bunting. J . R., and Cathou, R. E . , J . B i d . Chem., 1973, 248. 3173. Kagedal, B . , and Kallberg, M., J . Chromutogr., 1982,229,409. Kosowcr, N. S . , Kosowcr, E. M., Ncwton, G . L., and Ranney, H. M., Proc. Natl. Acad. Sci. USA, 1979, 76, 3382. Kosower, N. S . , Newton. G . L.. Kosower, E. M., and Ranney, H. M., Biochim. Biophys. Acta, 1980, 622, 201. Newton, G. L., Dorian, R . , and Fahey, R. C., Anal. Riochem.. 1981, 114, 383. Toyo’oka, T., and Imai, K., Analyst. 1984, 109, 1003. Toyo’oka. T., and Imai, K., J. Chromatogr., 1983. 282, 495. Imai, K., and Toyo’oka, T., Methods Enzymol., 1987, 143, 67. Toyo’oka, T., and Imai, K., Anal. Chem., 1984, 56, 2461. Toyo’oka, T., and Imai, K., Anal. Chern., 1985, 57, 1931. Toyo’oka, T . , Suzuki, T., Saito. Y . , Uzu, S., and Imai, K . , Analyst, 1989, 114, 413. Carlson, R. G., Chokshi, H . P., Givens, R. S . , and Toyo‘oka. T., J. Org. Chem., submitted for publication. Gompper, R., and Hcrlinger, H.. Chem. Ber., 1956, 89, 2816. Goekl, G . W., and Cram, D. J . , J . Org. Chern., 1974,39,2445. Carlson, R. G . , unpublished data. Lunte, S. M., and Wilson, M., unpublished data. Paper 2/05 75 1 D Accepted October 30, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800257

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST. MARCH 1993. VOL. 118 265 Low Level Determination of Formaldehyde in Water by High-performance Liquid Chromatography Evangel0 Cotsaris and Brenton C. Nicholson Australian Centre for Water Quality Research, Private Mail Bag, Salisbury, South Australia, 5108 Low levels of formaldehyde in water were determined by derivatization with 2,4-dinitrophenylhydrazine a t an optimized pH (1.5-2.5), solid-phase extraction with c18 adsorption cartridges and analysis by reversed-phase high-performance liquid chromatography with ultraviolet detection. A novel procedure for the removal of formaldehyde present as an impurity in blank water was responsible for lowering the detection limit to 0.1 pg 1-1 for a 200 ml sample. A strong anion-exchange resin, in the hydrogen sulfite form, was used to adsorb formaldehyde and any other aldehyde impurities present in blank water.The use of c18 adsorption cartridges also minimized background effects. The recovery of C1-C3 aldehydes spiked into purified blank water was 83-93% with a relative standard deviation of 1.4-6.4%. Keywords: Formaldehyde; aldehydes; solid-phase extraction; high-performance liquid chromatography; 2,4-dinitrophen ylh ydrazine Formaldehyde is an increasingly important environmental pollutant, with evidence of its adverse effects on health becoming more apparent. 1-3 Formaldehyde enters the water environment mainly as a result of human activities, major sources being the discharge of trade waste effluents into waterways and the ozonation of water and waste waters. Existing methods for the determination of formaldehyde in water usually involve derivatization with 2,4-dinitrophenylhy- drazine (2,4-DNPH) followed by solvent extraction and analysis by high-performance liquid chromatography (HPLC).4-h These procedures are relatively easy to carry out and give reasonable sensitivity using ultraviolet (UV) detec- tion.However, there are very few reliable techniques re- ported for the determination of the low concentrations of formaldehyde usually found in natural waters. Most analytical methods for the determination of formaldehyde are limited in their sensitivity as a result of high blank responses. Trouble- some background peaks are difficult to eliminate because reagents and solvents contain trace amounts of formaldehyde as impurities. Some workers have attempted to decrease blank responses and hence lower detection limits by replacing solvent extraction with solid-phase extraction techniques.7.8 Ogawa and Fritz7 found that small columns packed with zeolite ZSM-5 were able to concentrate most low molecular mass aldehydes and ketones except for formaldehyde where recoveries were poor (1%).Takami et al.8 demonstrated that moderately sulfonated cation-exchange resins could indeed concentrate formaldehyde present at microgram per litre levels from drinking water. However, these techniques are not particularly convenient as the solid-phase extraction car- tridges require tedious preparation with custom-made pack- ings. Furthermore, substantial background peaks are still observed at trace levels and attempts to reproduce the performance of some of these solid-phase packings by other workers have been unsuccessful .y Formaldehyde present as an impurity in blank water may also contribute to high blank responses.Some workers have overcome the high blank formaldehyde response by utilizing unchlorinated bore water for the preparation of blank water.6 However, the widespread occurrence of formaldehyde in the environment makes it virtually impossible to determine trace concentrations. To date, no attempt has been made to lower the detection limit by removing formaldehyde contamination from water that is used for the preparation of blank solutions. In this study, a novel procedure for the removal of formaldehyde present as an impurity in blank water was used to lower the detection limits and to improve sensitivity.The method is based on the formation of the 2,4-DNPH derivative of formaldehyde at an optimized pH (1 S), solid-phase extraction with CI8 adsorption cartridges, and analysis by reversed-phase HPLC with UV detection at 365 nm. Experimental Reagents Aldehydes and ketones were obtained from commercial sources. Apart from formaldehyde, all carbonyl compounds were distilled under nitrogen prior to use. Owing to trimer formation the purification of the aldehydes is essential if reliable results are to be obtained. Formaldehyde solution, about 38% m/v, was assayed by the sulfide-iodimetric method10 and was used without further purification. Hexane and dichloromethane were obtained from Mallinckrodt (nan- ograde) and further purified by extraction with 40% m/v sodium hydrogen sulfite solution.Acetonitrile High-performance liquid chromatography grade acetonitrile (Ajax Chemicals, Australia) was used. Blank water Blank water was prepared by the purification of distilled water to remove any traces of aldehyde impurities. A glass column (25 x 1 cm) was packed with a strongly basic anion-exchange Dianion PA 318 resin (Mitsubishi Chemical Industries, Japan). The column was pre-treated by eluting with 250 ml of 1 moll-1 hydrochloric acid, 250 ml of distilled water, 250 ml of saturated sodium hydrogen sulfite solution and rinsed with a furhter 200 ml of distilled water. Blank water was prepared by passing distilled water through the pre-treated column at a rate of 2 bed volumes per hour.At least 10 1 of distilled water with a total contaminant aldehyde concentration of up to 10 pg 1-1 can be purified before any significant break- through occurs. 2,4- Dinitrophenylhydrazine Aldehyde impurities were removed from 2,4-DNPH by a procedure adopted from van Hoof et a1.5 The 2,4-DNPH (100 mg) , previously extracted five times with hexane-dichloro- methane (70 + 30) and recrystallized twice from acetonitrile, was dissolved in 10 ml of 10 mol 1-1 hydrochloric acid and 90 ml of blank water. The resultant solution was extracted three times with hexane (50 ml) to remove trace impurities and stored under the same solvent. The purified reagent solution is stable for at least one week, but should be re-extracted with266 ANALYST, MARCH 1993, VOL. 118 fresh hexane prior to use if the lowest detection limit is required.2,4- Dinitrophenylh ydraz one standards Carbonyl hydrazone derivatives were prepared by standard procedures" and purified by recrystallization from ethanol. Stock solutions (1 g 1-1) of each derivative were prepared in acetonitrile. A mixed standard solution (1 mg 1 - 1 ) was prepared from the stock solutions in 50% v/v acetonitrile- water. The mixed standard should be freshly prepared prior to use. Derivatization A 4.0 ml aliquot of 2 moll-' 2,4-DNPH was added to a 200 ml water sample in a calibrated glass flask and the contents shaken for approximately 0.5 min to ensure good mixing. The pH is controlled using the 2 moll- 1 2,4-DNPH solution so that addition of the reagent to the sample results in a pH of 1.5-2.5.Water samples with very high alkalinities should be checked for pH and adjusted accordingly to pH 1.5-2.5 with 1 mol I-l HCI. After allowing 1 h for derivatization to proceed to completion, the derivative was extracted to prevent decompo- sition and further reaction. Extraction Procedure The 2,4-DNPH derivatives were extracted using a 3 ml C18 (100 mg) Bond-Elut cartridge (Varian) preconditioned with 20 ml of acetonitrile followed by 40 ml of blank water. The derivatives were eluted with 1 ml of acetonitrile followed by 1 ml of blank water, the eluates were combined in a 2 ml calibrated flask and made up to the mark with water. Typically, 100 pl injection volumes were employed for HPLC determination. Apparatus The HPLC determinations were performed with a Waters Associates liquid chromatograph equipped with two Model 501 solvent delivery systems, Model 680 automated solvent gradient controller, Model 444 UV detector set at a wavelength of 365 nm and a Model U6K injector.The column was a C18 reversed-phase Brownlee (Spheri-5 RP18), 5 pm particle size, 25 cm X 7 mm semi-preparative column. Chromatographic Conditions The mobile phase used was an acetonitrile-water mixture (65 + 35) at a flow rate of 1.5 ml min-1. Results and Discussion Optimization of pH for Maximum Recoveries The yields of 2,4-DNPH derivatives as a function of pH for a constant reaction time are shown in Fig. 1. Maximum recoveries of the C1-Cs aldehyde derivatives were obtained in the pH range 1.5-2.5. The acid concentration of the reagent was chosen to produce a pH at the lower end of the range in distilled water (pH l S ) , to allow for slight increases in pH that may occur with water samples that have a high alkalinity.For example, when 2 moll-' of 2,4-DNPH were added to a water sample with a high alkalinity of 450 mg 1-1, the resultant pH was 1.95 and still within the pH range for maximum recoveries. Recoveries of aldehyde derivatives are dependent on two factors: ( a ) the extraction efficiency of the hydrazone deriva- tives on CIS adsorption cartridges; and (6) the reaction between 2,4-DNPH and the aldehyde. Both factors are known to be pH dependent. At low pH (<1.5) it was postulated that a proportion of the aldehyde derivatives were in the ionic form loo 80 i 8 40 a 2o t I I I I I I I 0 1 2 3 4 5 PH Fig.1 Effect of reaction pH on the percentage yield of formal- dehyde, acetaldehyde and propionaldehyde 2,4-DNPH derivatives. Reaction time, 1 h. 0, Formaldehyde; X, acetaldehyde; and U, propionaldehyde 100 90 80 70 1 60 8 12 50 40 a, 30 20 10 I X I I 1 I I 0 1 2 3 4 5 Reaction time/h Fig. 2 Rate of derivatization of C,-C3 aldehydes with 2,4-DNPH at pH 1.5 versus time. 0, Formaldehyde; X , acetaldehyde; and U, propionaldehyde and were not adsorbed onto C18 cartridges. At higher pH (>2.5) the derivatives should be in the molecular form and would be expected to adsorb onto the C18 cartridges via a hydrophobic interaction. However, there must be another mechanism in operation because the yields decreased with an increase in pH. This is presumably due to a decrease in the rate of the reaction.The reaction proceeds by a multi-step mechanism where the rate-limiting step involves the addition of 2,4-DNPH reagent to the protonated carbonyl moiety. At varying pH there are competing effects between the availabil- ity of the 2,4-DNPH reagent and the reactivity of the carbonyl group, hence the rate passes through a maximum which is characteristic of the basicity of 2,4-DNPH.12 Bicking et al. 13 observed similar results when investigating the effect of pH on the reaction of 2,4-DNPH with formaldehyde and acetal- dehyde. Consequently, there are two mechanisms operating that are dependent on pH, the adsorption of hydrazone derivative onto the CIS adsorption cartridge and the reaction between 2,4-DNPH and the aldehyde.We were unable to determine the mechanism that was the most critical. The mild acidic reaction conditions (pH 1.5-2.5) are preferred to the more common highly acidic conditions (pH 0.5) employed in conventional derivatization.4-6 Under highly acidic conditions a false positive formaldehyde result may be obtained, as numerous environmentally occurring compounds have been shown to generate formaldehyde under a variety of conditions.14 In addition, the recovery of the aldehyde 2,4-DNPH derivatives by solid-phase extraction at pH 0.5 is poorer in comparison with that at pH 1.5-2.5.ANALYST, MARCH 1993, VOL. 118 267 Optimization of Reaction Time The derivatization reaction for C1-C3 aldehydes was studied as a function of time at the optimum pH and found to reach equilibrium after 0.5 h as shown in Fig.2. Reaction times longer than 1 h resulted in decreased yields; in particular the formaldehyde derivative did not appear to be stable for long periods. Blank Responses and Detection Limit A strong anion-exchange resin in the hydrogen sulfite form was used for the adsorption of trace concentrations of aldehydes present in blank water. Aldehydes and methyl ketones react with the hydrogen sulfite ions to form ionic a-hydroxyalkanesulfonates at the exchange site. Several workers have reported that aldehydes and methyl ketones are retained by an anion-exchange column in the hydrogen sulfite form.15-17 Although no investigation was undertaken to determine the most suitable resin, the work by Williams and Strauss'7 demonstrated that a porous-type resin instead of a gel-type anion exchanger was the most appropriate.The resin (Dianion PA 318) evaluated was found to produce very low blank responses when used under conditions described in the experiment. t - (0 C Is: v) .- 0 5 10 15 I n Z ( 6) n 5 I I I I 0 5 10 15 Retention timelmin Fig. 3 (a) HPLC trace of 2,4-DNPH derivatives of carbonyl standards, formaldehyde, acetaldehyde, acctone and propional- dehyde, equivalent to 10 pg 1-1 concentration each. (b) Chromato- gram of blank water. Scale: X 0.06 a.u.f.s. in both instances The detection limit, defined as twice the size of the background peaks for formaldehyde, was 0.1 pg 1-1 for a 200 ml water sample. The blank is an order of magnitude lower than for most other published methods. Fig.3 shows a typical liquid chromatogram of the 2,4-DNPH derivatives of some low molecular mass carbonyl compounds and the blank test with 200 ml of water. The low detection limit is attributable to the removal of formaldehyde present in the reagents and blank water. Commercially available 2,4-DNPH contains substantial amounts of aldehydes that produce interferences in trace analysis. These impurities can by eliminated by purifying the reagent by extraction and recrystallization as described. However, formaldehyde is also introduced as a contaminant in the blank water. Without purification of the blank water by resin adsorption, a detection limit of below 1 pg 1-1 is not attainable. Another factor contributing to the low blank responses was the use of commercially available C18 adsorption cartridges for solid-phase extraction to replace the conventional liquid- liquid extraction techniques and other solid-phase packings.Solid-phase extraction with CI8 cartridges minimizes the use of solvents and accordingly reduces the level of aldehyde contamination from this source. t - m C tJ, tn .- 0 5 10 15 0 5 10 15 Retention tirnehin Fig. 4 Typical chromatograms of some low level aldehyde analyses. ( a ) Distilled water and (b) Milli-Q water. Scale: x 0.06 a.u.f.s. in both instances268 Table 1 Recovery of aldehydes from spiked blank water* Amount Amount Recovery-! Compound spiked/pg found/pg (YO) SD$ (Yo) Formaldeh ydc 50 46.0 92.0 6.4 Propionaldehyde 50 44.2 88.4 1.4 Acetaldehyde 50 46.5 93.0 2.3 * A 200 ml volumc of blank water was used for recovery tests.7 Averagc of three runs. 'i: Standard deviation. Table 2 Aldehyde concentrations of various samples Formaldehyde/ Acetaldehyde/ Propion- Samplc pg I-' pg 1-1 aldehyde/pgl-l Distilled water* 7.3 6.8 ND1- Milli-Q water+ 3.3 11.9 ND Rainwater 0.7 0.1 ND Borewater 0.6 0.2 ND plant in which formaldehyde was used as a membrane preservative. * Prior to distillation. water was obtained from reverse osmosis .I- No peak detected. $ Water purified through a Milli-Q system. Recoveries The recovery of CI-C3 aldehydes was determined by repeated spikes on 200 ml of purified blank water at the SO pg 1-1 level. The results shown in Table 1 indicate that recoveries of derivatives were 8843% with a 1.4-6.4% relative standard deviation. Application The method was found to be suitable for the determination of trace concentrations of aldehydes in a variety of waters.Results for the trace analysis of a number of samples are reported in Table 2. Typical chromatograms are shown in Fig. 4. The application of this improved method to a range of waters has demonstrated the widespread occurrence of formaldehyde and to a lesser extent other simple aldehydes and ketones. Accurate determination of trace concentrations of aldehydes in water is now possible. ANALYST, MARCH 1993, VOL. 118 The proposed method has advantages over conventional methods in terms of rapid and simple extraction procedures, mild derivatization conditions, high analytical sensitivity and low background effects. We thank G. Skouroumounis for valuable discussions and M.Ayling for technical assistance during the development of this method and the Engineering and Water Supply Department for supporting this work. 1 2 3 4 5 6 7 8 9 10 I 1 12 13 14 15 16 17 References Hileman, B., Environ. Sci. Technol., 1982, 16, 543. U.S. Federal Panel On Formaldehyde, Environ. Health Perspect., 1982, 44, 139. Hileman. B., Environ. Sci. Technol., 1984, 18, 218. Fung, K., and Grosjean, D., Anal. Chem., 1981, 53. 168. van Hoof, F., Wittocx, A., van Buggenhout, E., and Janssens, J., Anal. Chim. Acta, 1985, 169, 419. Whittle, P. J., and Rennie, P. J., Analyst, 1988. 113, 665. Ogawa, I., and Fritz, J. S . , J. Chromatogr., 1985, 329, 81. Takami, K . , Kuwata, K., Sugimae, A . , and Nakamoto, M., Anal. Chem., 1985, 57, 243. I'omkins, B. A., McMahon, J. M., Caldwell, W. M . , and Wilson, D. L., J . Assoc. Off. Anal. Chem., 1989, 72, 835. American Public Health Association, Standard Methods for the Examination of Water and Wastewater, American Public Health Association, American Water Works Association, Water Pollu- tion Control Federation, Washington, DC, 16th edn., 1985, pp. 479-480. Shrincr, R. L., Fuson, R. C., and Curtin, D. Y., The Systematic Identification of Organic Compounds, Wiley, New York, 5th cdn., 1965. Hine, J., Physical Organic Chemistry, McGraw-Hill, New York, 1956, ch. 8 and 11. Bicking, M. K. L., Cooke, W. M., Kawahara, F. K., and Longbottom, L. E., J . Chromatogr., 1988, 455, 310. Sawicki, E.. and Sawicki, C. R., Aldehydes-Photometric AnalyJis, Academic Press, London, 1975, vol. 1. Samuelson. O., Ion Exchangers in Analytical Chemistry, Wiley, New York, 1953, ch. 16. DuVal, D. L., Rogers, M., and Fritz, J. S . , Anal. Chrm., 1985, 57, 1583. Williams, P. J . , and Strauss. C. R., 1. Sci. Food Agric., 1978,29, 527. Paper 2105678J Received October 26, 1992 Accepted November 30, I992

ISSN:0003-2654    

DOI:10.1039/AN9931800265

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, MARCH 1993, VOL. 118 269 High-performance Liquid Chromatographic Detection of Trace A/-Nitrosoamines by Pre-column Derivatization With 4-(2=Phthalimidyl) benzoyl Chloride* Minghui Zheng Department of Chemical Engineering, Tangshan University, Tangshan, Hebei, People‘s Republic of China Chengguang Fut and Hongda Xu Research Centre of Ph ysical and Chemical Analysis, Hebei University, Baoding, Hebei, People’s Republic of China A simple and sensitive method for the detection of N-nitrosoamines is described. N-Nitrosoamines in dichloromethane solution can be denitrosated to secondary amines with a hydrogen bromide-acetic acid mixture, the amines formed then react rapidly with 4-(2-phthalimidyl)benzoyl chloride to give fluorescent amides, which can be separated on an octadecylsilane column with aqueous acetonitrile as eluent.N- N it r oso d i met h y I a m i n e , N-n it rosod i et h y I a m i n e , N- n it rosod i p ro p y I a m i n e , N- n it rosod i b u ty I a m i n e , N- n it ro- sopyrrolidine and N-nitrosopiperidine were used as model compounds to optimize the derivatization and chromatographic conditions. The realtive standard deviations ( n = 7) a t an analyte concentration of 8 x 10-6 mol 1-1 were less than 5%. The detection limits were in the range 0.4-1. 6 pmol per injection. Keywords : N - Nitrosoam in e; 4- (2-p hthalimid yl) benzo yl chloride; flu0 rescence deriva tization; high - performance liquid chromatography The carcinogenicity of N-nitrosodimethylamine was first discovered in 1956 by Magee and Barnes.’ Since then, about 300 N-nitroso compounds have been discovered, approxi- mately 90% of which are known to be potent carcinogens in animals.2 N-Nitrosoamines are formed in air, water, soil and even in the human body when the appropriate amines and nitrite precursors are present.High-performance liquid chromatographic separations of N-nitrosoamines have been reported.”-” Pre-column derivati- zation6-7 or post-column derivatization8.9 in high-performance liquid chromatography are used in order to allow the sensitive and selective detection of N-nitrosoamines. 4-(2-Phthalimidyl)benzoyl chloride (PIB-CI) was used as a fluorescent derivatization reagent for compounds with amino groups. 1 0 However, no systematic studies of derivatization conditions of PIB-CI with amines have been reported and there has been no study of amine derivatives of PlB-C1 using liquid chromatography.In this paper, we develop a simple and sensitive method for the detection of trace N-nitrosoamines. The technique involves the use of high-performance liquid chromatography with fluorescence detection after pre-column reaction. N- Nitrosoamines were denitrosated to secondary amines with denitrosating reagent, the amines formed then react rapidly with PIB-CI to give fluorescent amides, which can be separated on an octadecylsilane column. The derivatization and chromatographic conditions were optimized on the basis of experiments. Experimental Apparatus A Perkin-Elmer series 3 liquid chromatograph equipped with an MPF-44B fluorescence detector, a 5 pm Nucleosil C18 * Presented at the 4th Asian Chemical Congress, Beijing, China, t To whom correspondence should be addressed.1991. column (125 x 4.6 mm i.d.) and a Rheodyne Model 7105 injection valve with a 20 pl sample loop. Reagents All chemicals used were of analytical-reagent grade unless stated otherwise. Doubly distilled water was used throughout. The N-nitrosoamines were synthesized by reaction of secondary amines with nitrites in an acidic medium,2>10 and their structures were confirmed by mass spectrometry. These compounds were dissolved in dichloromethane and the stock solution (2 x 10-4 mol dm-3) was diluted before use. The PIB-CI was prepared as described previously” and was dissolved in acetonitrile to give a 5 x 10-3 mol 1-1 reagent solution. The denitrosation reagent was prepared by diluting 5 ml of 47% m/m aqueous hydrobromic acid (guaranteed-reagent grade) to give a final volume of 26 ml with acetic anhydride.40 50 60 70 CH3CN (% V/V) Fig. 1 Plot of capacity factor ( k ’ ) as a function of ace t oni t rile percentage compo&ion. Curvcs: A, N-nitrosodimcthylamine; B, N-nitrosopyrrolidine; C, N-nitrosodiethylaminc; D, N-nitrosopiperi- dine; E, N-nitrosodipropylamine; F, N-nitrosodibutylamine270 ANALYST, MARCH 1993, VOL. 118 Procedure A 70 pl aliquot of denitrosation reagent was added to 100 pI of test solution (64 pmol-20 nmol per 100 PI) in a test-tube and heated for 5 min in a water-bath at 40 "C. After removal of the solvent under a flow of nitrogen, 100 pl of 0.4 moll-' sodium hydrogen carbonate solution and 100 pl of PlB-CI solution were added.The mixture was allowed to stand at room temperature for about 1 min. A 10 pl aliquot of the final mixture was injected into the high-performance liquid chro- matograph. Results and Discussion Optimization of Conditions for Denitrosation Reaction The N-nitrosoamines are known to undergo cleavage at the N-NO bond in the presence of a hydrogen bromide-acetic A B C 20 16 12 8 4 0 tRlm i n Fig. 2 Chromatogram of PIB-Cl derivatives. Mobile phase, aceto- nitrile-water (48 + 52 v h ) ; flow rate, 0.8 ml min-l; detector wavelength, A,, = 299 nm, A, = 426 nm; injection volume, 10 pl. Peak assignment is same as in Fig. 1. Each lettered peak corresponds to 20 pmol of N-nitrosoamine acid mixture, resulting in the formation of the corresponding secondary amines and the liberation of nitric oxide.2~7~10 According to Drescher and Frank,12 the denitrosation of N-nitroso compounds in dilute dichloromethane solution occurs readily at ambient temperature and the yield of denitrosation products is essentially independent of hydrogen bromide concentration provided that a minimum excess of approximately 103 mol of hydrogen bromide per mol of N-nitroso compound is maintained.On the basis of our experiments, we found this to be true except for N-nitroso- pyrrolidine. The denitrosation of N-nitrosopyrrolidine was completed in 90 rnin under the above conditions. A systematic study of temperature and reaction time was performed with the temperature being varied between 20 and 40°C and the reaction time between 1 and 90 min.The study indicated the optimum conditions to be 5 rnin at 40°C. Fluorescence Derivatization The method of Tsuruta and Kobashi" was modified to the extent that the fluorescence derivatization was facilitated with sodium hydrogen carbonate rather than sodium hydroxide and reaction time was 1 rnin rather than 30 min. The optimum concentrations of sodium hydrogen carbonate and PIB-CI for maximum reaction yield have been investigated. The maxi- mum yield was achieved if the mole ratios of PIB-CI to N-nitrosoamine was >3 and the sodium hydrogen carbonate to N-nitrosoamine was >120. The derivatization of PlB-C1 with the secondary amines, produced by denitrosation, proceeded rapidly and was independent of temperature (0-80 "C), the reaction was complete within 1 rnin even at 0°C.Therefore, the mixture was allowed to stand for 1 min at room temperature. Fluorescence Properties of the Derivatives The fluorescence spectra were measured in aqueous acetonit- rile by stop-flow scanning. The wavelength maxima of fluorescence excitation and emission of the derivatives are 299 and 426 nm, respectively. Chromatographic Conditions The separation of PIB-CI derivatives was studied on a reversed-phase column with aqueous acetonitrile. The depen- dence of retention time on the acetonitrile concentration in the eluent is shown in Fig. 1. The optimum separation was obtained with 48% v/v aqueous acetonitrile. Fig. 2 shows a typical chromatogram obtained with a mixture of six N-nitro- soamines. Performance of Fluorescence Detection Several characteristic of the method are given in Table 1.Interferences Aromatic amines gave no fluorescent products under the proposed derivatization conditions. Alcohols did not interfere Table 1 Regression analysis of calibration graphs and other quantitative data for the N-nitrosoarnines Compound Linear range/ Calibration Correlation nmol ml-1 equation* coefficient N-Nitrosodimethylamine 0.32-200 y = -0.311 + 0.175~ 0.999 N-Nitrosopyrrolidine 0.16-200 y = 0.381 + 0.124~ 0.999 N-nitrosodiethy lamine 0.32-200 y = -0.140 + 0.256~ 0.999 N-Nitrosopiperidine 0.32-200 y = 0.393 + 0.240~ 0.999 N-Nitrosodipropylamine 0.64-200 y = -0.840 + 0.397~ 0.999 N-Nitrosodibutylmine 0.64-200 y = -0.269 + 0.782~ 0.999 * y in cm, x in nmol ml-1. $ Signal-to-noise = 3.Relative standard deviation ( n = 7). [N-nitrosoamine] = 8 x 10-6 mol 1-1. RSDi 2.8 3.9 4.1 3.7 4.9 3.9 (Yo 1 Detection limit$/pmol 0.6 0.4 0.8 0.4 1.6 1.6ANALYST, MARCH 1993, VOL. 118 27 1 with the determination of N-nitrosoamines when present at least up to 0.5 mol 1-1. Aliphatic amines could react with PIB-CI under the above conditions. In order to identify N-nitrosoamines in the sample, a portion of the sample was reacted directly with PIB-Cl in a weak alkaline medium (pH = 8-10> and the amount of aliphatic amines in the sample was determined. Another portion of the sample was treated as mentioned in the text, and the total amount of N-nitrosoam- ines and aliphatic amines were examined. The concentration of the N-nitrosoamines could then be calculated.Conclusion The method described retains simplicity of technique and can be applied to determination of N-nitrosoamines in the environment. In combination with a solid-phase extraction technique13 most of the N-nitrosoamines in aqueous samples can be detected down to required levels. References 1 2 Magee. P. N., and Barnes, J.. J . Cancer, 1956. 10, 114. Xu, H. X., N-Nitroso Compounds in the Environment, Science Press, Bcijing, 1988. 3 4 5 6 7 8 9 10 11 12 13 Issaq, H. J., McConnell, J . H., and Weiss, D. E., J . Liq. Chrornatogr., 1986, 9, 1783. Issaq, H. J., Glennon, M., Weiss, D. E . , Chmany, C. N., and Saavedra, J . E . , J . Liq. Chromatogr., 1986, 9, 2763. Issaq, H. J . , Atamna. I. Z . , Schultg. N . M., Muschik, G. M., and Saavedra, J. E., J . Liq. Chrornatogr., 1989, 12, 771. Wan, Q. H., and Fu, C. G., Cepu, 1986, 4, 238. Wang, Z . , Fu, C. G., and Xu, H. D., J . Chrornatogr., 1992,589, 349. Lee, S. H., and Field, I,. R., J. Chromatogr., 1987, 386, 137. Righezza, M., Murello, M. H., and Siouffi, A. M., J . Chrorna- togr., 1987. 410, 145. Hu, R. M., and Ma, L. S . , Analysis of N-Nitroso Compounds. Science Press, Beijing, 1980. Tsuruta, Y.. and Kobashi, K., Anal. Chim. Acta. 1987, 193, 309. Drescher, G. S . , and Frank, C. W., Anal. Chem., 1978. 50. 2218. Fu, C. G., and Wan, Q. H., Fenxi Huaxue, 1985, 13, 595. Paper 2104093J Received July 30, I992 Accepted October 28, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800269

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, MARCH 1993. VOL. 118 273 Ion-exclusion Chromatographic Determination of Hydrogen Carbonate in Natural Waters Using Unmodified Silica Gel and Conductimetric Detection Michio Zenki, Tomiko Nabekura and Atsushi Kobayashi Department of Chemistry, Faculty of Science, Oka yama University of Science, Ridai-cho, Oka yama 700, Japan Tadashi lwachido College of Liberal Arts, Oka yama University, Tsushima, Oka yama 700, Japan Yasuaki Shimoishi School of Health Sciences, Oka yama University, Shikata, Oka yama 700, Japan A simple and rapid ion-exclusion chromatographic method for the determination of hydrogen carbonate has been developed. Unmodified silica gel was used for the separation column instead of H+-form cation- exchange resin. By elution with water and monitoring with a conductimetric detector, excellent separation can be achieved.The chromatographic conditions for the separation of hydrogen carbonate, pore and particle sizes of packings and the separation mechanism are discussed. For the determination of hydrogen carbonate, borate buffer solution (pH 7.3) was added in order t o keep the pH of the sample solution constant. The calibration graph was found t o be linear in the range 2-20 pg ml-1 of hydrogen carbonate. Common anions and cations such as chloride, nitrate, sulfate, sodium and calcium do not interfere. The method has been applied t o the determination of hydrogen carbonate in river and lake waters. Keywords: lon-exclusion chromatography; hydrogen carbonate; unmodified silica gel; conductimetric detection The simple and accurate determination of inorganic carbon species (e.g., carbon dioxide, hydrogen carbonate and carbo- nate) is an important requirement in water purification, environmental and biological research.Ion chromatography (1C)l with conductimetric detection is a useful tool for the determination of common anions, such as fluoride, chloride, bromide, nitrate, nitrite, phosphate and sulfate. The combination of a low-capacity ion-exchange column and low-conductivity eluents allows the determination of such ions at sub-ppm levels. However, the detection of hydrogen carbonate at trace levels is difficult, because it is a very weak acid (pK,, = 6.34) and exists ordinarily as anions in basic solution. In addition, the resolution of hydrogen carbonate seems to be incomplete, and the peak for hydrogen carbonate overlaps sometimes with peaks for other com- pounds.Also, sodium hydrogen carbonate is used frequently as a component of the eluent for the separation of several anions .2,3 Detection of hydrogen carbonate using a spectrophoto- metric detector is also difficult because of its weak absorption above 210 nm. Indirect ultraviolet (UV) detection44 was therefore carried out, using organic acids as eluents. However, hydrogen carbonate is similar in its retention behaviour to that of other common anions, e.g., chloride. On the other hand, ion-exclusion chromatography (IEC), developed by Wheaton and Bauman,7 is a convenient method for the separation of non-ionic species from ionic species. Kreling and DeZwaan8 and Tanaka and Fritz9 reported the determination of hydrogen carbonate by IEC using a cation- exchange resin (H+-form, sulfonated polystyrene-divinylben- zene copolymer) and water as the eluent. This is the only reliable method that has been confirmed.Silica gel is known to act as an ion exchanger.10.11 Smith and Pietrzyk12 have revealed that many inorganic cations can be separated on an unmodified silica gel column by IC. We have also reported some results regarding the chromatographic separations of alkali and alkaline-earth metal cations on unmodified silica ge1.13.14 The pK, value of the silanol group is reported to be 7.115 and it is assumed that the silanol group of the hydrated silica gel surface dissociates to a certain extent in slightly acidic or neutral solutions.Therefore, ion exclusion can be expected to take place in a similar manner to that occurring on a sulfonated cation-exchange resin. To date, no studies on the separation and determination of hydrogen carbonate with unmodified silica gel have been reported. The purpose of this work was to demonstrate the use of an unmodified silica gel column for the separation of hydrogen carbonate by IEC. The method has been applied to the determination of hydrogen carbonate in natural waters. Experimental Apparatus The high-performance liquid chromatographic system con- sisted of an LC-6A pump, a CDD-6A conductivity detector, a CTO-6AS column oven (all from Shimadzu, Kyoto, Japan) and a Rheodyne 7125 loop injector (Cotati, CA, USA). A Shimadzu Model C-R6A printer-plotter integrator was used to record the signal response.A 250 x 4 mm i.d. stainless-steel column filled with Develosil 30-5 (Nomura Kagaku, Seto, Japan) was used. A packed column of TSK SCX (150 X 4 mm i.d.) was purchased from Tosoh (Tokyo, Japan) and used for comparative purposes. The column temperature was main- tained at 35 "C. The flow rate of the mobile phase was fixed at 0.8 ml min-1, and the injected sample size was 100 PI. Reagents All the reagents used were of analytical-reagent grade. Water was obtained from a Millipore (Milford, MA, USA) Milli-Q water purification system (18 MQ). Stock standard solution (1000 pg ml-1). Prepared by dissolving 1.377 k 0.005 g of sodium hydrogen carbonate (Wako Pure Chemicals, Osaka, Japan) in 1 1 of water. Buffer solution (PH = 7.3).Prepared by dissolving 3.81 k 0.01 g of sodium tetraborate in approximately 400 ml of water. The pH was adjusted to 7.3 k 0.05 with boric acid, and the solution was diluted to 500 ml with water.274 ANALYST, MARCH 1993, VOL. 118 Sample Preparation All the samples were passed through a 0.45 pm Millipore filter, and the filtrates were mixed with equal volumes of borate buffer solution and injected onto the column. Results and Discussion C hromatograms A typical chromatogram, obtained with a silica-gel column (Develosil 30-5), distilled water as the mobile phase and conductimetric detection, is shown in Fig. l(a). Under the same conditions, some different types of packing, such as cation- and anion-exchange resins and octadecylsilane, were tested. Fig. l(h) depicts a chromatogram obtained with a TSK SCX column, which was used by Tanaka and Fritz.9 It is interesting that both chromatograms are very similar, even though the properties and characteristics of the column packings are different from each other.The SCX column is a sulfonated polystyrene-divinylbenzene copolymer with a high cation-exchange capacity (4.2 mequiv- 1 g), while Develosil 30-5 is a porous, spherical silica gel (pore size 3 nm; particle size 5 pm; surface area 650 m2 g-1). The surfaces of silica gels and their ion-exchange properties have been discussed exten- sively by Unger.1" Chromatograms obtained with the silica gel columns are explained by assuming that the ion exclusion by the silanol group takes place in a manner similar to that of the sulfonic acid group.With other columns tested (LiCrosorb RP-18, Develosil ODs-5 and TSKgel TC Anion PW), the separation and resolution of hydrogen carbonate were incom- plete. Pore and Particle Size The effects of the pore and particle size of the silica gels on the separation of hydrogen carbonate have been investigated. Nine packings, having pore and particle sizes of 3,5 and 10 nm and 5 , 7 and 10 pm, respectively, were packed into the stainless-steel column (250 x 4 mm i.d.). With decreasing pore and particle size of silica gels, the surface area increases, i.e., the number of ion-exchange sites increases and hence the separation and resolution by the ion-exchange reaction can be affected significantly. In fact, the expected results were obtained for the separation and resolution of analytes in previous papers.1"14 In this work, however, it was found that the retention behaviour (capacity factor) of hydrogen carbon- t - 0 C 0, in .- 2 i, 0 10 0 10 Timehi n Fig.1 Chromatograms of hydrogen carbonate. (a) Develosil 30-5 (250 x 4 mm i.d.). ( h ) TSK SCX (150 x 6 mm i.d.). Peak 1, dip peak; and peak 2, HC03- (20 pg ml- I ) ate was almost independent of the pore and particle sizes of the packings. Only the shapes of the peaks were sharper with decreasing particle size. Fig. 2 shows the relationship between the particle size of the packings and the theoretical plate numbers (n). Silica gels with a smaller pore size (3 nm) and a smaller particle size ( 5 pm) are suitable for obtaining well-resolved, sharper peaks. Column Length and Enhancement Column The dependence of the separation of hydrogen carbonate on the column length was investigated.Three types of column, 50, 150 and 250 mm long (each 4 mm i.d.), were used and compared. In IEC a considerable amount of the resin is essential to achieve a reasonable separation.9 Good resolution was obtained in about 5.2 min on the 250 X 4 mm i.d. column. The effect of ion-exchange enhancement9 was investigated by inserting a second column of silica gel after the separation column. No enhancement was observed, which shows that the mechanism of enhancement differs between the silica gel and the sulfonated polystyrene-divinylbenzene copolymer. More careful consideration is necessary. Mobile Phase In order to determine the effect of pH on retention, mobile phases buffered at pH 4.0, 5.0, 6.0,7.0 and 8.0 with HC1 and NaOH were prepared and investigated.Higher pH values (above pH 8) were not tested as solutions of such pH damage the column. The peak resolution became poorer as the pH of the mobile phase decreased, and below pH 4 the peak for hydrogen carbonate disappeared. This was because of the lack of sensitivity, i.e., no dissociation of hydrogen carbonate. Throughout the experiments, the signal due to carbon dioxide was not observed because of its neutrality. Mobile phases of higher pH (pH > 6.5) were found to produce a large baseline drift because of the increase in background conductivity. Therefore, distilled water was adopted for this work. pH of the Sample Solution The calibration graph was established from injection of different concentrations of the standard hydrogen carbonate solutions onto the column.The signal responses as a function of hydrogen carbonate concentrations up to 20 pg ml-l are shown in Fig. 3(a). As expected, the plots are curved because of the difference in the degree of dissociation as a function of 10 0 2 x 5 C 0 I I I I I 2 4 6 8 1 0 Pore sizehm Fig. 2 Theoretical plate number VCYSUS pore size of packings. Particle size: A, 5 ; B, 7; and C, 10 pmANALYST, MARCH 1993. VOL. 118 275 0 5 10 15 20 Concentration of HC03-/pg ml-l Fig. 3 buffer solution (pH 7 . 3 ) Calibration graph. ( u ) Without buffer solution; and (h) with 2 4 6 8 10 Signal response versus pH of sample solution. Concentration PH Fig. 4 of HC03-: A, 10; and B, 20 pg ml - I hydrogen carbonate concentration.y In order to improve the linearity of the calibration graph, samples were adjusted to constant pH by addition of a buffer solution prior to injection onto the column.Fig. 4 shows the relationship between pH of the sample solutions and signal response (peak height). The peak increased with increasing pH of the sample solution up to 8.5. This is a result of the increase of dissociation of carbonic acid, i.e., hydrogen carbonate is predominant. In fact the molar fraction of hydrogen carbonate, calculated from the pK, values of carbonic acid (pK,, = 6.34, pK,, = 10.36), reaches a maximum at pH 8.5. Unfortunately, it is known that the dissolution of silica gel becomes more serious above pH 8 . As a pH between 7.0 and 7.5 was satisfactory for a quantitative determination, pH 7.3 was adopted.Several types of buffer solution, such as acetate, ammo- nium, borate, phosphate and Good's buffer, were tested. Sodium tetraborate-boric acid buffer was selected because a suitable chromatogram was obtained, which is shown in Fig. 5(a). Boric acid (pK,, = 9.24) is as weak an acid as is hydrogen carbonate and is also expected to be eluted by IEC. It is known that the order of elution depends on the acid-dissociation constant, but the peak of borate appears between the dip peak and the hydrogen carbonate peak (retention time = 3.1 min). Even though the concentration of borate is high (1.0 X 10-2 mol I-'), the strength of the borate signal is weak and therefore adequate. Column Temperature The effects of temperature on columns were investigated.Though the sensitivity (peak height) increased with increasing column temperature, the capacity factor for hydrogen car- bonate decreased gradually. Therefore, the temperature of the column oven was maintained at 35 "C. -- 0 10 0 10 Time/mi n Fig. 5 Chromatograms obtained with borate buffer solution. (a) Standard solution (HC03-, 20 pg ml-I). (b) Rivcr watcr (Takahashi). Pcak 1, dip peak; peak 2. H2B03-; and peak 3, HC03- Interferences Possible sources of interference in natural waters, such as chloride, nitrate, sulfate, silicate, phosphate, ammonium, sodium, potassium, magnesium, calcium, aluminium and iron, were investigated at concentrations of up to ten times that of hydrogen carbonate. All the ions tested were eluted faster than the borate peak, and caused no interference even at a ratio of interfering species to hydrogen carbonate of 10 : 1.In the analysis of some actual samples by IEC, sometimes the dip peak grew larger and the species that eluted after the dip peak were overlapped or disappeared. In this case, either removal of the matrix or dilution of the sample solution was necessary. However, in thc proposed method, the borate peak was affected to some extent by the dip peak, but not by the hydrogen carbonate peak [Fig. 5 ( h ) ] . Therefore, there is no interference in the determination of hydrogen carbonate. Calibration The calibration graph was linear over the range 2-20 pg mi-1 for hydrogen carbonate. As shown in Fig. 3(b), the calibration plot does not pass through the origin because of the residual carbonate concentration in the borate buffer solution.The correlation coefficient for the calibration graph was 0.9999. When 10 and 20 pg mi-1 standard hydrogen carbonate solutions were injected consecutively the relative standard deviations measured for ten runs were 0.89 and 1.9%, respectively. Application to Natural Waters The method has been applied to the determination of hydrogen carbonate in some environmental waters. River and lake water samples were collected in Okayama Prefecture,276 ~~ Table 1 Determination of hydrogen carbonate in natural waters Amount/pg ml-I Sample" Recovery (dilution) Initial? Added Found? (%) Asahi 22.7 * 0.3 4.0 26.0 * 0.4 97.4 10.0 33.8 t 0.4 103 Takahashi 45.0 k 0.6 4.0 50.6 k 0.5 103 Yoshii 23.5 k 0.4 4.0 27.8 2 0.3 101 10.0 34.6 f 0.5 103 lkenouchi 72.6 k 0.5 10.0 82.6 k 0.5 100 Miyashita 120 2 0.6 10.0 128 k 0.6 98.5 Hikoki 158 f 0.9 10.0 170 k 1.1 101 * Samples were collected on January 28, 1992.t Average of five determinations ? standard deviations River- (1 + a 10.0 47.6k0.6 97.1 Lakc- (1 + S ) (1 + 10) ( I + 10) Other method$ 24-29 45-6 1 24 - - - - - - $ lndirect UV detection6 and gas-diffusion flow injection.16 Japan. These samples were passed through a membrane filter (pore size 0.45 pm; Millipore) and submitted to chromato- graphy as soon as possible. The recovery test was carried out by adding, with a microsyringe, 10 pl of the standard hydrogen carbonate solution to 10 ml of the filtered sample solutions. The results obtained for recovery tests are summarized in Table 1.A chromatogram o f river water samples is shown in Fig. 5(b). A large dip peak due to large amounts of common anions (such as chloride, nitrate and sulfate) was observed. However, the resolution between the dip peak, borate peak ANALYST, MARCH 1993, VOL. 118 and hydrogen carbonate peak was satisfactory. The results of the recovery tests also show that the proposed method could be utilized for practical applications. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1s 16 References Small, H., in Ion Chromatography, ed. Hercules, D., Plenum Press, New York. 1989, p. 149. Hanaoka, Y . . Murayama, T., Muramoto, S . , Matsuurd, T.. and Nanba. A., J . Chromutogr., 1982, 239, 537. Saigne, C . , Kirchner, S . , and Legrand, M., Anal. Chim. Actu, 1987, 203, 1 1 . Brandt, G., and Kettrup, A., Fresenius' Z. Anal. Chem., 1985, 320,485. Brandt, G., Matuschek, G., and Kettrup, A . , Freseniu' 2. Anal. Chem., 1985, 321, 653. Hironaka, T., Oshima, M., and Motomizu, S . , RunJeki Kagaku, 1987, 36, 503. Wheaton, R. M . , and Bauman, W. C., Ind. Eng. Chem., 1953, 45, 228. Kreling, J. R.. and DeZwaan, J . , Anal. ChPm., 1986, 58, 3028. Tanaka, K., and Fritz, J . S . , Anal. Chem.. 1987, 59, 708. Unger, K. K., J . Chromatogr. Lihr., 1979, 16, 130. Dugger, D. L., Stanton, J . H., Irby, B. N., McConnell, B. L., Cummingc, W. W., and Maatman, R. W., J . Phys. ('hem.. 1964, 68, 757. Smith, R. L., and Pietrzyk. D. J., Anal. Chem.. 1984, 56. 610. Iwachido, T., Shinomiya, M.. and Zcnki, M., Anal. Sci., 1990, 6, 277. lwachido, T., Ikeda, T., and Zenki, M., Anal. Sci., 1990,6,593. Hair, M. L., and Hertl, W., J. Phys. Chem., 1970, 74, 91. Kuwaki, T.. T8ci, K., Akiba, M., Oshima, M.. and Motomizu, S . , Bumeki Kagaku, 1987, 36, T132. Paper 2104799C Received September 7, 1992 AcceDted November 6. I992

ISSN:0003-2654    

DOI:10.1039/AN9931800273

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, MARCH 1993. VOL. 118 277 Organic-phase Biosensors for Monitoring Phenol and Hydrogen Peroxide in Pharmaceutical Antibacterial Products Joseph Wang, Yuehe Lin and Liang Chen Department of Chemistry, New Mexico State University, Las Cruces, NM 88003, USA Organic-phase biosensors open new opportunities for assays of challenging pharmaceutical products. Such opportunities are illustrated for the rapid determination of phenol and peroxide antiseptics in different anti-infective formulations. The tyrosinase and peroxidase enzyme electrodes offer reliable quantification of these antibacterial agents following sample dissolution in the organic solvent. The dynamic properties of these enzyme electrodes are exploited for rapid and reproducible flow-injection assays of the pharmaceutical products (relative standard deviation = 1.6-1.9%).Such developments should facilitate rapid quality control testing in the pharmaceutical industry and should be applicable to other therapeutic agents and products. Applicability to cosmetic products containing hydrogen peroxide is also demonstrated. Keywords: Organic-phase biosensor; phenol; pharmaceutical analysis; enzyme electrode The remarkable finding that enzymes can maintain their biocatalytic activity in non-aqueous environments‘ has led to the development of organic-phase biosensors.2 The operation of enzyme electrodes in organic solvents offers several important advantages, including measurements of additional (hydrophobic) substrates, assays of new environments, ex- tended sensor stability or simplified immobilization schemes.The ability to analyse previously inaccessible sample matrices can greatly expand the possibilities for biosensors. For example, recent studies have illustrated the utility of choles- terol oxidase and tyrosinase electrodes for direct ampero- metric assays of butter3 and olive oils,4 respectively. Nu- merous other potential applications of organic-phase enzyme electrodes are expected to be explored in the near future. This paper describes the utility of organic-phase biosensors for challenging pharmaceutical products. Many pharmaceut- ical formulations are not readily dissolved in aqueous media. Hence, their bioassays (using traditional enzyme electrodes) usually require time-consuming sample manipulations, e.g. , solvent extraction. The introduction of organic-phase enzyme electrodes obviates the need for such sample pretreatment and facilitates rapid assays of pharmaceutical products. Such opportunities are illustrated here for the monitoring of phenolic and peroxide antiseptics in anti-infective phar- maceutical formulations.The antibacterial activity of phenols and hydrogen peroxide5.6 has led to their extensive thera- peutic use. Tyrosinase- and peroxidase-based biosensors, known for their effective operation in non-aqueous environ- ments,”7-9 are shown here to be highly suitable for the rapid determination of their corresponding substrates in phar- maceutical products. The reported coupling of these sensors with fast flow-injection operation should be particularly attractive for quality-control and process-monitoring applica- tions in the pharmaceutical industry.Experimental Apparatus Amperometric measurements were performed with a CV-27 voltammograph [ Bioanalytical Systems (BAS)], in connection with an x-y-r recorder (BAS). The 10 ml cell (Model CV-2, BAS) was joined to the enzyme electrode, reference electrode [Ag-AgCI (3 mol 1-1 NaCI), Model RE-1, BAS] and platinum-wire auxiliary electrode through holes in its Teflon cover. A magnetic stirrer and stirring bar facilitated the transport of the substrates. The flow-injection system eon- sisted of a 50 ml syringe/carrier reservoir, held by the syringe pump (Model 341B, Sage), a Rainin Model 5041 sample injection valve (20 PI), interconnecting Teflon tubing and the thin-layer electrochemical detector.Modification of the glassy carbon electrode (Model MF 2012, BAS) was achieved by covering the surface with a 10 pl drop of the mixed enzyme-Eastman- AQ polymer solution. The coating was then allowed to dry with an air gun. The mixed enzyme-Eastman-AQ solutions were prepared by dissolving 2 mg of tyrosinase or 6 mg of horseradish peroxidase in 200 PI of the 1.4% polymer solution. Reagents Tyrosinase (EC 1.14.18.1, 2400 U mg-1) (1 U = 16.67 nkat) and horseradish peroxidase (HRP, EC 1.11.1.7, 90 U mg-I) were received from Sigma. Phenol (Fisher), hydrogen per- oxide, tetraethylammonium p-toluenesulfonate (TEATS), acetonitrile (HPLC grade) and ferrocene (Aldrich) were used as received. The poly(ester-sulfonie acid) polymer (Eastman AQ 55D, 28% dispersion) was received from Eastman Chemical Products; prior to mixing with the enzyme it was diluted 20-fold with de-ionized water.‘Unguentine Plus’ (Mentholatum), ‘Campho-Phenique’ (Sterling Drug), ‘Stat- One, Hydrogen Peroxide Gel’ (Continental Consumer Pro- ducts) and ‘Creme Hair Bleach’ (Del Lab) were purchased from a local drugstore. Samples were dissolved in the following manner: ‘Campho-Phenique’ products, 0.10 g in 5 ml of propan-1-01; ‘Unguentine Plus’, 0.47 g in 10 ml of propan-1-01; ‘Hydrogen Peroxide Gel’, 0.113 g in 10 ml of acetonitrile, and ‘Creme Hair Bleach’, 0.20 g in 5 ml of propan- 1-01. Procedure Experiments were performed (at 25 k 1 “C) in acetonitrile solutions (containing 4% v/v water and 0.05 mol I-’ TEATS) by holding the working electrode at the desired potential and allowing the transient current to decay.Potentials of 0.0 and -0.25 V were used for the quantification of hydrogen peroxide and phenol, respectively. Measurements of hydrogen peroxide were carried out in the presence of 5 x 10-3 mol 1-1 ferrocene. Results and Discussion Several organic-phase electrodes, based on the activity of tyrosinase and peroxidase in non-aqueous media, have been developed in recent years.4.7-10 One promising fabrication method involves the entrapment of these enzymes within278 ANALYST, MARCH 1993, VOL. 118 Eastman-AQ surface coatings (which are stable in various organic media). lo Fig. 1 displays amperometric responses of the tyrosinase-Eastman-AQ coated electrode (in acetonitrile) to the addition of various dissolved pharmaceutical products (A) and to subsequent additions of a phenol standard solution (B-E).The tyrosinase electrode responds very rapidly to these additions, producing steady-state currents within 10- 12 s. Note also the high sensitivity to these micromolar changes in the substrate concentration compared with the absence of response without the enzyme (broken line). The well-defined response of the organic-phase enzyme electrode offers convenient determinations of the phenol antiseptics based on the standard additions method. The resulting standard additions plots for these pharmaceutical formula- tions are displayed in Fig. 2. All plots exhibit a linear dependence (correlation coefficients, 0.999) that permits reliable quantification (following correction for the dilution factor).Phenol levels of 4.5% (a), 0.5% ( b ) and4.5%0 (c) were thus calculated for these samples. Such values are in very good agreement with the labelled values [4.7% (a), 0.50% ( 6 ) and 4.7% (c)]. It should be pointed out that the antiseptic products used in Figs. 1 and 2 are insoluble in water, but can be readily dissolved in propan-1-01. Owing to its strong oxidizing power, hydrogen peroxide is commonly used as an antiseptic in anti-infective formulations, or as a bleaching agent in cosmetic products. The peroxidase- Eastman- AQ organic-phase biosensor can facilitate the quan- tification of hydrogen peroxide in such challenging products. Fig. 3 shows current-time recordings obtained at the peroxi- dase electrode for additions of dissolved ‘Hydrogen Peroxide Gel’ pharmaceutical (a) and ‘Creme Hair Bleach’ cosmetic ( b ) products, and for successive 1 x 10-5 mol 1 - 1 additions of t 4- C 2 3 0 _ldB _____------- Time - Fig.1 Amperometric response to additions of dissolved phar- maceutical products (A), followed by that to successive additions of 1 x 10-5 mol 1-1 phenol (B-E). Samples, ‘Campho-Phenique Cold Sore Gel’ (u), ‘Ungucntine Plus Pain Relieving Cream’ (b), and ‘Campho-Phenique Pain Relieving Antiseptic’ (c). Sample dilutions (in cell) of 1000- [(a) and (c)] and 200-fold ( b ) . Operating potential, -0.25 V; medium, acetronitrile-water (96 + 4% v/v) containing 0.05 mol I-’ TEATS. The corresponding response without the enzyme is shown by the broken line hydrogen peroxide (B-E). In a similar way to its tyrosinase counterpart, the peroxidase electrode exhibits a fast (-20 s) and sensitive response to changes in the substrate concentra- tion, which permits convenient quantification.Note also the absence of response without the enzyme (broken lines). The resulting standard additions plots (not shown) were highly linear (correlation coefficients, 0.999) and yielded hydrogen peroxide values of 3.3% (a) and 8.9% (6). [The labelled value for product (a) is 3.0%, while no value is given for product (b)]. Other commonly used peroxide species (e.g., benzoyl peroxide) can be measured in a similar fashion, based on the reported response of peroxidase electrodes towards organic peroxides. 11 The fast response of the tyrosinase and peroxidase organic- phase electrodes can be exploited for high-speed flow-injec- tion assays, as desired for quality control applications.Such flow analysis is simplified by the organic-phase operation, as the need for on-line sample pre-treatment (e.g., solvent extraction) is eliminated. We have recently reported on the adaptation of organic-phase biosensors for monitoring flowing streams. 10,12 Fig. 4 displays the amperometric response to phenol in the ‘Campho-Phenique Cold Sore Gel’ (F) and ‘Campho-Phenique Pain Relieving Antiseptic’ (G) products, together with peaks for phenol standard solutions of increas- ing concentration [5 x 10-5-25 x 10-5 mol 1-1 (A-E)]. The tyrosinase detector responds very rapidly to dynamic changes in the concentration, characteristic of flow-injection systems. Phenol levels of 4.4% (F) and 4.5% (G) can therefore be calculated, which are in good agreement with the labelled value (4.7%).Various experimental variables affecting the flow-injection/ organic-phase biosensing response were evaluated. Fig. 5 , A shows the effect of the flow rate on the phenol peak. A nearly exponential decrease of the response is observed on increasing the flow rate from 0.4 to 3.4 ml min-1. Apparently, the enzymic reaction requires slower passage of the sample plug to I I I ( a ) 40 - 20 - 01 ” I I 1 20 0 20 40 Concentration/ymol dm-3 Fig. 2 Standard additions plots for phenol quantification in various pharmaceutical products: ‘Campho-Phenique Cold Sore Gel’ (a), ‘Unguentine Plus Pain Relieving Cream (b) and ‘Campho-Phenique Pain Relieving Antiseptic (c).Conditions as in Fig. 1ANALYST, MARCH 1993, VOL. 118 279 t 4- C 2 3 0 150 nA H 1 min (a) / A _________--- ------- Time -. Fig. 3 Ampcrometric response to additions of dissolved products (A). followed by that to successive additions of 1 X rnol 1-1 hydrogen peroxide (B-E). Samples, ‘Hydrogen Peroxide Gel’ ( a ) and ‘Creme Hair Bleach’ ( h ) (dilution factor, 1000 and 1000, respect- ively). Operating potential 0.0 V; medium, acetonitrile-water (96 + 4% v/v) containing 0.05 rnol 1-1 TEATS and 5 x mol I - ’ ferrocenc. The corresponding response without the enzyme is shown by the broken line t + C 2 3 0 7 2 0 nA L- Time - Fig. 4 Flow-injection peaks for phcnol solutions of increasing concentration [S x 10-5-2S x rnol 1-I (A-E)] and for injections of dissolved ‘Campho-Phcniquc Cold Sore Gel’ (F) and ‘Cam- pho-Pheniquc Pain Relieving Antiseptic’ (G) products.Dissolved samples were diluted 66.6-fold in the carrier solution prior to injection. Flow rate. 0.50 ml min-1; carrier + electrolyte, acetonitrile -water (96 + 4% vh), containing 0.05 rnol 1-1 TEATS; operating potential, -0.25 V produce appreciable currents. The effect of the operating potential is shown in Fig. 5 , B. A well-defined sigmoidal hydrodynamic voltammogram (HDV) , with a rapid current increase between -0.05 and -0.25 V, is observed. Such an HDV, expected for the detection of the enzymically produced quinone species, indicates that ohmic drop effects are not substantial in the organic media (provided sufficient elec- trolyte is present).The flow-injection response to phenol in pharmaceutical products is also highly reproducible. Detection peaks for a series of 20 repetitive assays of the ‘Campho-Phenique Pain Relieving Antiseptic’ ( a ) and ‘Campho-Phenique Cold Sore Gel’ (b) samples are displayed in Fig. 6. These prolonged 0 80 2 . w IT L (3 40 0 Flow rate/ml min-’ 2 4 I - -0.2 -0.4 PotentialN Fig. 5 Effect of flow rate (A) and operating potential (B) on the tlow-injection response to a 2 x 10-4 rnol 1-1 phcnol solution. Operating potential (A), -0.25 V; flow rate (B), 0.50 ml min-1. Conditions as in Fig. 4 t c.l 2 3 0 1 ! ’ 5min ’ I Time - Fig. 6 Detection peaks for repetitive injections of dissolved ‘Cam- pho-Phenique Pain Relieving Antiseptic’ (u) and ‘Campho-Phcnique Cold Sore Gel (b) products.Other conditions as in Fig. 4 series yielded relative standard deviations (RSDs) of 1.7% ( a ) and 1.6% (b). Note again the fast response and rapid return to the baseline, as expected for Eastman-QA based organic- phase biosensors. 10 Flow-injection measurements of hydrogen peroxide in the ‘Hydrogen Peroxide Gel’ and the ‘Creme Hair Bleach’ samples are displayed in Fig. 7. Also shown are peaks for hydrogen peroxide solutions of increasing concentration [ 1 x 10-44 X rnol 1-1 (A-D)] (a). The HRP-organic-phase electrode offers a fast, linear and stable flow-injection response [correlation coefficient (a), 0.999; RSDs (n = 20) of 1.9% ( b ) and 1.6% ( c ) ] . Hydrogen peroxide levels of 6.9% ( b ) and 2.8% ( c ) have thus been calculated. In conclusion, organic-phase biosensors offer unique oppor- tunities for the pharmaceutical industry.Although such opportunities are illustrated here for measurements of phe- nolic and peroxide antiseptics, they could easily be extended to bioassays of other therapeutic agents and products. The280 ANALYST, MARCH 1993, VOL. 118 20nA , 6) I 5 min a) D nA Time - Fig. 7 Flow-injection peaks for hydrogen peroxide solutions of increasing concentration [ 1 x 10--r_4 X 10 -4 mol 1-’ (A-D)] ( a ) and for repetitive injections of dissolved ‘Creme Hair Bleach ( h ) and ‘Hydrogen Peroxide Gel’ ( c ) products. Dissolved samples were diluted 500- ( h ) and 50-fold (c) in carrier solution prior to injection. Operating potential, 0.0 V; carrier + electrolyte (96 + 4% v/v) containing 0.05 mol I-’ TEATS and 5 x lo-? mol I-’ ferrocene; flow rate, 0.5 ml min-l speed and simplicity that accrue from the organic-phase biosensing operation should greatly facilitate quality-control testing of pharmaceutical and cosmetic products.In addition, the organic-phase biosensors should be suitable for in situ process monitoring during the production of drugs. The fast-responding flow-injection systems should be particularly attractive for the tasks of quality control and process monitoring in the pharmaceutical industry. Similar improve- ments and opportunities are anticipated for previously inac- cessible and challenging sample matrices from other indus- tries. Helpful discussions with C. Kelly are acknowledged. References 1 2 3 4 5 6 7 8 9 10 11 12 Zaks, A., and Klibanov, A. M., J . Biol. Chem., 1988,263,3194. Saini. S . , Hall, G. F., Downs, M. E., and Turner, A. P. F., Anal. Chim. Acta, 1991, 249, 1. Hall, G. F., and Turner, A. P. F.. Anal. Lett., 1991, 24, 1375. Wang, J., Reviejo, A. J., and Mannino, S., Anal. Lett., 1992, 25, 1399. Merck Manual of Diagnosis and Therapy, ed. Bcrkow, R . , Merck. Rahway. NJ. 14th edn., 1982, p. 2300. Jenkins, G., and Hartung, W., The Chemistry of Organic Medicinal Products, Wilcy, New York. 1950, p. 100. Hall, G. F., Best, D. J., and Turner, A. P. F., Anal. Chim. Acta, 1988, 213, 113. Schubert, F., Saini, S . , and Turner, A. P. F., Anal. Chim. Acta. 1992, 245. 133. Wang, J., Wu, L. H., and Angnes, L., Anal. Chem., 1991, 63, 2993. Wang, J., Lin, Y., and Chen, Q.. Electroanalysis, 1993, 5, 23. Wang, J . , Freiha. B., Naser, N., Romero, E. G., Wollenbcrger, U., Ozsoz, M., and Evans, O., Anal. Chim. Acta, 1991,254,81. Wang, J . , and Lin, Y . , Anal. Chim. Acta, 1993, 271, 53. Paper 2105282 B Received October 2, 1992 Accepted December 2, I992

ISSN:0003-2654    

DOI:10.1039/AN9931800277

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, MARCH 1993, VOL. 118 281 Agarose Gel Electrophoresis System for the Separation of Antibiotics used in Animal Agriculture Michael J. Salvatore and llya Feygin Merck & Co., Inc. P.O. Box 2000, Rahway, NJ 07065, USA Stanley E. Katz Department of Biochemistry and Microbiology, Cook College/NJAES, Rutgers-the State University of New Jersey, New Brunswick, NJ 08903-0231, USA "b A novel electrophoresis system using agarose gel has been developed for the separation and as an aid in the classification of antibiotics. This system utilizes Nunc cell factory disposable tissue culture dishes, which serve as bioassay dish and cooling chamber for agarose gel, in a custom designed electrophoresis unit. Tris( hydroxymethyl) methylamine-succinate buffer at pH 6.0 and 8.0 are employed as the electrolyte for electrophoresis.Bioautography was used as the indicator of mobility. Any agar diffusion assay can be modified to use this system. A suggested name for this system is Nunc cell factory agarose gel electrophoresis (NUAGE). Selected antibiotics, representative of the aminoglycoside, @-lactam, macrolide, moenocinol, peptide, polyene, polyether, quinone and tetracycline classes, were separated with this system. Keywords: Agarose gel electrophoresis; animal agriculture; bioautograph y; separation of antibiotics; classification of antibiotics The current official, Association of Official Analytical Chemists (AOAC), methods of analysis rely primarily upon agar diffusion assays for the quantification of antibiotic levels in feeds.' These methods assume that the declared antibiotic is solely responsible for the biological response.These assays also assume a prior knowledge of the constituents of the feed. For example, an assay for erythromycin would be used to analyse feeds assumed to contain erythromycin. The key to error in these assays is the 'assumption' of the identity of the antibiotic prior to its analysis. There is no provision to identify the antibiotic in this type of assay because the results are demonstrated only by zones of inhibition. The AOAC methods do not address the analysis of antibiotics occurring in mixtures. They cannot, per se, detect, separate or differentiate between individual, classes or mix- tures of antibiotics. The design of these assays can completely mask the presence of other antibiotics.This is also true for combinations of antibiotics that act in synergy or as antag- onists with each other; all of which can lead to erroneous and misleading results. It is, therefore, important to develop an antibiotic identifi- cation and classification system. Such a system should provide rapid detection with the simplicity of an agar diffusion assay and a means of detecting separating and identifying individual classes and mixtures of antibiotics. A separation system designed to differentiate between antibiotics can be achieved by various means, some of which include: ( i ) exploitation of some chemical or physical property of the antibiotic molecule (i.e. , charge, size); (ii) selective destruction or inhibition of the activity of the antibiotic ( i .e . , pH change, chemical inactivation); (iii) selective indication of individual antibiotics (i.e. , colour reaction, spectra); and (iv) selectivity by indirect means (i.e. , resistant microorganisms). To date, only separa- tion by chemical or physical property has shown any merit. There have been many classification systems based on chemical and/or physical properties described for the detec- tion, separation and identification of antibiotics. These systems have been described for paper chromatography,2-10 thin-layer chromatography,11-18 paper electrophoresis1Y--25 and gel electrophoresis .*fi-35 All of these systems utilize differential solubility and/or mobility to detect, separate and identify antibiotics. Paper and thin-layer chromatography systems take advan- tage of the antibiotic solubility and differential mobility in a solvent for classification. These systems use multiple transfer steps, solvents and support matrices to separate and charac- terize antibiotics into discernible groups.Low levels of antibiotics have to be concentrated prior to testing. If bioautography is used to indicate mobility and biological activity there is also a drying and antibiotic diffusion step. In addition, complex mixtures can contain constituents (salts, fatty acids, etc.) that interefere usually by causing streaking or yield RF (retardation factor) values that are not reproducible.2-1* These types of systems are tedious and time consuming, ineffective for all antibiotics, insensitive and can expose the antibiotic to harsh conditions.The aforementioned chromatography systems are very useful for organic-soluble antibiotics. Aqueous-soluble anti- biotics, such as tetracyclines, aminoglycosides, etc. , are very difficult to characterize due to their lack of mobility. Paper and gel electrophoresis utilize charge, shape and size to separate and characterize antibiotic molecules. Differential mobilities of antibiotic molecules in an electric field can be used for characterization. Systems have been described that utilize organic solvents, such as chloroform,36 with paper electrophoresis and aqueous solvents, such as phosphate buffer, with agarose gel electrophoresis.6~~~-*~~~2*~24-3~ Paper electrophoresis is more suitable for antibiotic residue analysis than it is for chromatography.It is a simpler and more direct method of analysis. It can handle organic- as well as aqueous-soluble antibiotics for characterization. However, this type of system contains the same insensitivity, over- loading, streaking, drying and antibiotic diffusion problems as chromatography. In addition, electroendoosmosis can further affect the accuracy of the characterization. Gel electrophoresis is, by far, the best of the separation systems. Overloading, and thus sensitivity can be minimized by using thicker support layers without increasing electro- endoosmosis. Streaking can be minimized by decreasing the amount of gel comprising the matrix. Finally, drying and antibiotic diffusion problems are eliminated by overlaying the gel with a biological indicator contained in a similar type of matrix.The gel electrophoresis system presented herein detects,282 ANALYST, MARCH 1993, VOL. 118 separates, classifies and identifies 17 antibiotics currently used in animal agriculture when used in conjunction with the solvent separation system of Salvatore and Katz.23 Experimental Reagents The reagent ingredients were dissolved in de-ioinized water, the pH was adjusted and the total volume made up 11. Tris-succinate buffer, p H 6.0. Tris(hydroxymethy1)methyl- amine (Tris; 1.82 g) and 0.95 g succinic acid. Tris-succinate bu ffer, p H 8.0. Tris( h ydrox y me t h y I)me th yl- amine (3.03 g) and 0.85 g succinic acid. Sterile isotonic saline solution. Sodium chloride (9.0 g). Sterilize for 20 min at 121 "C. 1% Agarose-Tris-succinate, p H 6.0.Agarose (10.0 g) and 1.0 1 of Tris-succinate buffer pH 6.0. Sterilize for 20 min at 121 "C. After sterilization agarose-Tris-succinate was kept at 48 "C until use. 1% Agarose-Tris-succinate, p H 8.0. Agarose (10.0 g) and 1.0 I of Tris-succinate buffer pH 8.0. Sterilize for 20 min at 121 "C. After sterilization agarose-Tris-succinate was kept at 48 "C until use. Culture Media The media described below were re-hydrated in 11 of distilled water, adjusted to the appropriate pH and autoclaved at 121°C for 20min. Media containing agar were kept in a water-bath at 48°C until use. Medium A . Antibiotic medium 3 (53.0g) and 0.1 g man- ganese(ous) sulfate H20 (Fisher Scientific, Springfield, NJ, Medium B. Nutrient broth (8.0 g) and 2.0 g yeast extract, Medium C.Medium B and 15.Og Bacto-Agar, pH 6.95- Medium D. Yeast extract (10.0 g) and 10.0 g anhydrous Medium E. Medium D and 15.0g Bacto-Agar, pH6.55- All biological media were purchased from Difco Labora- USA), pH 6.95-7.05. pH 6.95-7.05. 7.05. dextrose (Fisher Scientific), pH 6.55-6.65. 6.65. tories, Detroit, MI, USA). Antibiotics Antibiotics were chosen because of their use in animal agriculture in the USA. Their preparation is outlined in the AOAC manual. 1 Fosfomycin and L-proline (no antiobiotic activity) were used as indicators of mobility. Fosfomycin was prepared by following the same procedure as that for streptomycin. The following antibiotics were used in these studies: bacitracin, zinc salt; bambermycin; chlortetracycline hydrochloride; erythromycin; fosfomycin, disodium; hygromycin; lincomycin hydrochloride; monensin, sodium salt; neomycin sulfate; novobiocin, sodium salt; nystatin; oleandomycin, phosphate salt; oxytetracycline dihydrate; penicillin G, sodium salt; spectinomycin dihydrochloride; streptomycin sulfate; tylosin tartrate; virginiamycin. All antibiotics except bambermycin, fosfomycin and virginiamy- cin were purchased from Sigma, St.Louis, MO, USA. Bambermycin was obtained from Hoechst Pharmaceuticals, Somerville, NJ, USA, virginiamycin was obtained from Smith Kline and Beckman, Paoli, PA, USA. Fosfomycin was obtained from Merck, P.O. Box 2000, Rahway, NJ, USA. Culture Maintenance and Preparation Bacillus subtilis Bacillus subtilis [American Type Culture Collection (ATCC) 66331 was prepared according to the procedures outlined in the AOAC manual.1 Spore suspensions were sub-divided into 2ml aliquots ("6 x 108 viable spores per ml) and stored at 4°C. Overlays were prepared by dilution of 0.5 ml of spore suspension with 4.5 ml of medium B. Five ml of diluted spore suspension were added per 100 ml of molten medium C. Saccharomyces cerevisiae Saccharomyces cerevisiae (ATCC 9763) was prepared accord- ing to the procedures outlined in the AOAC manual.' A fresh culture was used each day to prepare overlays. For the overlays, a fresh broth culture was diluted with medium D to 70% transmission at 660nm in a Spectronic 20 spectropho- tometer (Fisher Scientific). Four millilitres of this culture was added per 100 ml of molten medium E. Electrophoretic Gels Two hundred ml of agarose-Tris-succinate buffer at either pH 6.0 or 8.0 (see Reagents section) was dispensed onto the top support shelf of a Nunc cell factory [Laboratory Dispos- able Products (Springfield, NJ, USA); Fig 11.Air bubbles were removed by flaming with a bunsen burner and the agarose was allowed to solidify on a levelling table. After solidifying, 950 ml of a 30% solution of antifreeze, coolant and water was dispensed into the bottom chamber of the cell factory. The gel was cooled to 4 "C in a refrigerator and stored until use. Electrophoresis Equivalent volumes (about 750ml) of a solution of Tris- succinate buffer at 4 "C was dispensed into each buffer trough Fig. 1 Side view of a Nunc cell factory containing an agarose gel layer. A, Coolant outlet; B, coolant inlet; C, cooling chamber; D, agarose layer; and E, support shelf ~ ~~ ~- ~~~ ~~ Fig. 2 Corner view of electrophoretic chamber (base) containing Nunc cell factory, paper wicks and wick supports.A, Nunc cell factory; B, paper wicks; C, buffer troughs; D, electrodes; E, cell factory support; F, trough connector; G, trough baffle; H, trough outlet (to pump); I , wick support; J, levelling screw; and K, cover supportANALYST, MARCH 1993, VOL. 118 283 Fig. 3 Top view of cover for electrophoretic chamber with electrode wire leads. A, Vacuum ports; B, inlet and outlet ports: C, air vent controls; and D, electrode wire leads Fig. 4 Corner view of electrophoretic chamber with cover and cooling connector valves. A, Inlet connector valve; B, outlet connector valve; C, purge valve; D, chamber cover; E, cathode wire lead; and F, anode wire lead in the electrophoresis chamber (Fig.2). A circulating pump [Master Flex variable speed pump, Cole Parmer (Chicago, IL, USA)] connected to both buffer troughs was used to recircu- late electrolyte (about 10 rev min-1). A cell factory containing agarose-Tris-succinate gel at the appropriate pH was placed on the gel support between the two buffer troughs. Two glass wick supports, one at each end of the cell factory, were also placed on the gel support. Four mm wells were then punched into the centre of the gel about 2.5 cm apart with a gel punch [Bio-Rad Laboratories (Richmond, CA, USA)]. Twenty microlitres of sample to be assayed was pipetted into each well. To track the migration from the origin and current flow an indicator dye (Bromocresol Green, 10% solution) was pipetted into the centre well for each determination.Paper wicks [15.25 x 14 cm (6 x 54 in) Whatman filter paper No. 31 were soaked in buffer and placed about 1 cm from each end of the gel, on top of the wick support and into the buffer trough. The chamber cover (Fig.3) was placed on top of the electrophoretic chamber that connected the buffer troughs to a computer controlled power supply (Bio-Rad Laboratories). To remove excess condensate, the cover was also connected to a vacuum line via rubber tubing and a liquid trap. The cooling chamber of the cell factory was connected to a refrigerated water circulator (Bio-Rad Laboratories) via an inlet and outlet valve (Fig. 4). The circulator was brought to the appropriate temperature, the air purged from the coolant lines and the feed and return valves opened.The coolant was allowed to circulate until the appropriate gel temperature was achieved. The chamber was connected to the power supply and a constant voltage applied to the gel for the specified time. Bioautography After electrophoresis, the cell factories were removed from the electrophoresis chamber, the coolant was poured out of the bottom chamber and water at 50 "C was dispensed into the bottom chamber. The gel was allowed to reach 48 "C. The cell factory was then placed on a level table and 150ml of the appropriate inoculated medium (see Culture Maintenance and Preparation Section) was poured over the top of the agarose layer and allowed to solidify.The water in the bottom layer was poured out of the bottom chamber, the plate was covered and incubated at the appropriate temperature for 18 h (or overnight). Mobility was demonstrated by measuring the location of the zones of inhibition. The distance from the origin to the beginning of the zone, plus one half the diameter, along with the position of the zone was recorded. Electroendoosmosis Twenty microlitres of a 5 mgml-l aqueous solution of L-proline was pipetted into each well of a cell factory and assayed electrophoretically under the appropriate conditions. A methanol solution of 1% ninhydrin was then pipetted into each well and along each lane of the gel. These gels were warmed for 1 h by filling the bottom chamber of the cell factory with 48°C water.A yellow colour appeared where there was a presence of L-proline. L-Proline was assayed in each position five times and the results were averaged. Results and Discussion Design of Electrophoretic Equipment Gel matrix and support Many types of gel matrices have been described for use in electrophoresis; some of which are better suited than others for specific applications. Polyacrylamide and starch gels are more suitable for protein separations whereas agar and agarose are typically used for the separation of smaller molecules. Therefore, it is imperative to use a gel that has a consistent composition (batch-to-batch) and lacks interfering components. The amount of agarose typically used in these types of gels ranges from 1 to 2% .h,*6-11().24-3().32-35 However, a 1% concen- tration was chosen in this system to minimize any residual electroendoosmotic effects and allow easier spreading.Con- centrations of greater than 1% led to stiffer gels that had a tendency to clump and solidify before a uniform layer could be spread. Concentrations of less than 1% led to softer gels that accumulated water during electrophoresis and the results were not reproducible. The amount of prepared agarose used per plate was 200 ml, which gave a gel 4 mm in thickness. Volumes less than this generally led to clumping and uneven gels. Cooling is one of the major problems in electrophoresis. Uneven cooling can lead to poor separations. Therefore, a means to consistently cool gel layers during electrophoresis is of primary importance.Nunc cell factories, which are usually used for the proliferation of tissue/cell cultures, were selected as cooling chambers. A side view of a cell factory loaded with agarose gel prior to electrophoresis is shown in Fig. 1. It consists of a sealed chamber with inlet (A) and outlet (B) valves and a 1 cm lip on top of the chamber. The top part of the chamber and the top lip can be used as a dish to support the gel layer. It is also level from one end to the other to allow pouring of a gel layer that is flat and of uniform thickness. The chamber normally used for growing tissue culture can be filled with antifreeze and the inlet and outlet valves connected to a chiller/reeirculator, which will circulatc coolant, at constant temperature, to the chamber and ultimately cool the gel layer.Electrophoretic chamber It was necessary to design a chamber in which the electro- phoresis could be performed. This chamber is shown in Fig. 2. It consists of: a shelf (E), which will support the cell factory/agarose gel; and two buffer troughs (C); which contain284 ANALYST, MARCH 1993, VOL. 118 baffles (G) and platinum wire electrodes (D). These troughs were situated at either end of the shelf with enough space to allow the placement of wick supports (I). To eliminate suction effects, a tube (F) was placed between the troughs connecting them and levelling screws (J) were placed on the base. Electrolyte gradients were also eliminated by connecting trough outlets (H) with tygon tubing to a pump circulating electrolyte between troughs. The height of the support shelf and the outside of the electrophoretic chamber were adjusted to accommodate the connection of the troughs and the agarose gel with paper wicks (B).Finally, supports (K) for the cover were placed opposite the electrodes preventing motion. bThe cover for the chamber is seen in Fig. 3. It is designed to cover the distance between the buffer troughs completely. It has two ports (B) that fit over the inlet and outlet that connect the cell factory to the chiller/recirculator. It also has two vacuum ports (A) for the removal of moisture and any gas generated during electrophoresis. These vacuum ports are connected to a vacuum trap via rubber tubing. To control air flow into and ultimately out of the chamber, two air vent controls (C) are placed in the centre of the cover.Finally, two wire leads (D) with female to male connectors were attached to the lid, the female connectors mated exactly with the male connectors on the chamber. Connection to a power pack is made via these leads. Orientation of the poles can be in either direction. It was also necessary to develop a valve system to connect the cell factory to the chiller/recirculator. This was accomp- lished by a loop of tygon tubing with branch points. Each branch point carried coolant to and from the electrophoretic chamber. This tubing was connected (Fig. 4) to an inlet connector valve (A), which had a bypass valve (C) to purge any air before it could enter the coolant chamber of the cell factory. The outlet connector valve (B) was connected to tubing that returned the coolant to the chiller/recirculator.It also has a purge valve (C) to remove unwanted air from the system. Electrophoretic Conditions In order to obtain consistent results from each electrophoretic separation, it was necessary to standardize the system. Each parameter had to be examined to maximize resolution and maintain consistency. Before such parameters were examined, it was necessary to determine the type of electrolyte to be used. This electrolyte would be used in the aqueous as well as the solid phase of the system and would be selected such that any background interference would be minimized. For this reason citrate (divalent cation chelator) and phosphate (cation chelator and antibiotic uptake inhibitor) buffers were eliminated as elec- trolyes .37,38 Smither and Vaughan35 developed a system that utilized Tris-succinate at pH6.0 and 8.0 to separate anti- biotics in agar and agarose gel and which had no background interference.Therefore, Tris-succinate was the electrolyte of choice. It is also important to determine mobility as related to pH as Smither and Vaughan35 have demonstrated. The pH values chosen, 6.0 and 8.0, were sufficiently spread to allow differentiation between the antibiotics and still lie in the range of biological activity. The best means of applying the sample to the gel had to be determined. It was apparent that a well should be used as a reservoir for the sample. This would minimize spreading and the diffusion of the antibiotic into the gel. The well size should be sufficiently small not to affect migration but large enough to hold a concentration of antibiotic that could be detected.A well size of 4 mm was chosen. Sensitivity and reproducibility problems were seen with smaller well sizes. Well sizes of larger than 4 mm resulted in migration and reproducibility problems. Finally, it was necessary to select a marker that could be used to determine maximum migration. The antibiotics fosfomycin and streptomycin were chosen as indicators of mobility. Fosfomycin is one of the fastest migrating anionic antibiotics whereas streptomycin is one of the fastest migrat- ing cationic antibiotics. They were used to standardize electrophoretic conditions. It was also decided to use bioauto- graphy to track the mobility of these anitibiotics.B. subtilis was used as an indicator of the activity and thus mobility. Volumes of agar of less than 150m1, when used as an overlay on the agarose gel, solidified too quickly causing clumping and uneven thicknesses of the agar layer. Time Run time was studied under the premise that this system should be simple and rapid. Thus, the faster the run time the simpler and quicker the assay. The results of various run times at pH6.0 and 8.0 for fosfomycin and streptomycin are summarized in Table 1. Run times of 1.0, 1.5 and 2.0 h were chosen. The results indicate that a 1 h run time was not sufficient for separation. The antibiotics tested attain maxi- mum velocity within 1-1.5 h. With a 1.5 h run time, the average distance migrated increased to 127% over the distance migrated in 1 h.A subsequent increase in run time to 2 h led only to a 3% average increase in the distance migrated. Therefore, the 1 .5 h separation time was chosen as the modest increases in mobility for fosfomycin and streptomycin did not warrant the extra time taken for 2 h. Temperature Temperature is another parameter that must be controlled t o maintain consistency and accuracy. Other workers have indicated that a temperature range of 10 to 20°C should be maintained during electrophoresis.2"3s With modern tem- perature controllers, it is easier to regulatc the system temperature. Therefore, a narrower range of temperature can be maintained, resulting in increased accuracy and repro- ducibility. The results of electrophoresis performed on fosfom ycin and streptomycin at 10, 15 and 20°C are summarized in Table 2.Overall a 1 "C increase in temperature resulted in an increase Table 1 Electrophoretic mobility (values given in cm) of streptomycin and fosfomycin at various timcs. (Conditions: 20 V cm-1, 1 0 "C. Bacillus subtilis was used as the biological indicator of activity) Timc/h Antibiotic 1 .o 1 .5 2.0 Fosfomycin a* 3.2 a7.1 a 73 pH 6 . 0 - Streptomycin ~ $ 2 . 7 c 5.2 c 5.6 Streptomycin C 3.3 c 5.8 C 6.0 pH 8.0- Fosfomycin a4.5 a10.4 a 10.6 * a = Anion. $ c = Cation. Table 2 Electrophoretic mobility (values given in cm) of streptomycin and fosfomycin at various temperatures. (Conditions: 20 V cm-l, 1 .5 h Bacillus subtilis was used as thc biological indicator of activity) Temperature/"C Antibiotic 10 1s 20 Fosfom ycin a* 7.1 a 7.9 a 8.6 pH 6.0- Streptomycin c-t 5.2 c5.7 C 6.3 pH 8 .G Fosfom ycin a10.4 a l l . 1 a 12.3 Streptomycin CS.8 c 6.5 c 7.2 * a = Anion. 7 c = Cation.ANALYST, MARCH 1993, VOL. 118 285 Table 3 Electrophoretic mobility (values given in cm) of streptomycin and fosfomycin at various voltages. (Conditions: 10 "C, 1 .5 h. Bacillus suhtilis was used as the biological indicator of activity. For voltages of 25 and 30 V cm-l paper wicks werc attachcd to the agarose gel) Table 5 Electrophoretic mobility of selected antibiotics. (Conditions: 1.5 h, 20 V cm-1, 10°C. Bacillus subtilis was used as the biological indicator for all antibiotics except nystatin; Sacchavornyces cerevisiae was the biological indicator for nystatin. Mobility of each antibiotic was adjusted for electroendoosmotic drift) VoltageIV cm-1 ~ Antibiotic 20 25 30 Fosfornycin a* 7.I a 10.2 a 13.6 S t r y t ornycjn c-l- 5.2 c 9.6 c 13.9 Fosfomycin a 10.4 ND$ ND Streptomycin cS.8 c 11.0 ND qU.64- pef S.[& * a = Anion. t c = Cation. $ ND = No activity dctccted. Table 4 Electrophoretic mobility (values given in cm) of streptomycin and fosfomycin at 25 V cm-I and various timcs. (Conditions: 10°C. 25 V cm-1. Bacillus suhtilis was used as the biological indicator of activity. For all timcs papcr wicks wcre attachcd to thc agarosc gel) Timclh Antibiotic 0.5 1 .o 1.5 Fosfom ycin a* 5.0 a 7.5 a 10.2 Streptomycin c-l- 4.8 c 7.2 c 9.6 Forfomycin a7.5 a l 1 . l ND$ Streptomycin c5.0 c 7.6 c 11.0 pH 6.(& pH 8.0- * a = Anion. i c = Cation. $ ND = No activity detected.in mobility of 2.1% at pH 6.0 and 1.8% at pH8.0 for fosfomycin. With streptomycin, a 2.1% increase was seen at pH6.0. At pH8.0 there was a 2.4% increase in mobility. However, for streptomycin at both pH 6.0 and 8.0, tailing was observed with increases in temperature. At 1.5 "C a 1.0 cm tail and at 20°C a 2.5 cm tail was observed. The results indicate that although there is enhanced separation at increased temperatures tailing also increases. Therefore, it would be best to run separations at 10°C. Voltage The voltage used per centimetre of gel (the driving force) is another parameter that will affect the mobility and accuracy of the system. It is very important to use a power source that can provide constant current. Lightbown and de RossP and Smither and Vaughan35 suggested votlages of' between 20 and 30 V cm-1.The results of electrophoresis performed on fosfomycin and streptomycin at 20, 25 and 30Vcm-1 are summarized in Table 3 and the results of electrophoresis of fosfomycin and streptomycin at 25 V cm-1 at 0.5, 1.0 and 1.5 h are shown in Table 4. The data indicated that 20 V cm-1 was the maximum voltage that should be used with this system. At voltages of 25 and 30Vcm-1 the paper wicks adhered to the agarose gel making them difficult to remove. It must also be kept in mind that any increase in voltage will result in a quadratic increase in the heat generated. Thus, the lower the voltage the easier it is to regulate cooling. Electroendoosmosis When designing any electrophoretic separation system, it is important to consider the effects that electroendoosmosis will have on the migration of an ion.Typically, electroendoosmo- sis will cause a shift in migration towards the cathode; the Mo bi 1 it ylcm Antibiotic Erythromycin OJemdpmygjn Tylosin Hygromycin b Neomycin Streptomycin Lincom yci n Spectinomycin Chlortctracyclinc Oxytetracycline B acitracin Virginiam ycins Bamberm ycins Moncnsin Novobiocin Nystatin Penicillin G * c = Cation. t a = Anion. Class Macrolide MiwQJjds Macrolide Aminoglycoside Aminoglycoside Aminoglycoside Aminoglycoside-like Aminogl ycosidc-li kc Tetracycline Tetracycline Pcptidc Peptide-like Mocnocinol Pol yet her Quinone Pol yene @-Lactam QH 6.0 c* 2.5 c1.2 cs.2 c 5.5 c 5.2 c 3.0 c 6.0 c 1.6 c 1.3 c 1.6 c 1.3 a 3.8la 2.4 a 1.4 a 1.4 c 0.6 a 2.7 ~ 2 ~ ~ 5 QH 8.0 c3.1 c.1.3 cs.0 c5.5 c 5.8 c 2.6 c 5.0 a l 1.5 a 1.3 a 0.5 a 2.3la 0.8 a 2.3la 1.6 a1.3 a 2.4 c 0.6 a 3.8 $319 extent of which is a property of the support matrix.Electro- endoosmosis was monitored at both pH 6.0 and 8.0 utilizing L-proline as the migrating ion and ninhydrin as the indicator of mobility. These results indicate slight migration (0.2 cm) toward the cathode at both pH6.0 and 8.0. Electroendoos- motic effects were the same at both pHs. Although this migration is negligible, 0.2 cm should be used as a correction factor, when determining the exact mobilities of antibiotics. Correction factors should be determined using the individual apparatus. Electrophoretic Mobility of Antibiotics The results of the electrophoretic mobilities of the 17 antibiotics tested are given in Tables.The macrolide anti- biotics (erythromycin, oleandomycin and tylosin) migrated as cations. Erythromycin and oleandomycin migrated the fur- thest and showed similar mobilities at both pH6.0 and 8.0. These two antibiotics also demonstrated an increase in mobility with an increase in pH. The aminoglycoside class of antibiotics (hygromycin B, neomycin and streptomycin) migrated as cations and have similar mobilities. At pH 6.0, neomycin migrated the furthest of these antibiotics, to 5.5 cm. At pH 8.0, however, strep- tomycin migrated the furthest with an increase in mobility to 5.8 cm, followed by neomycin, which had no enhancement of mobility, and hygromycin B, which showed a decrease of mobility to 5.0 cm. The aminoglycoside-like class of antibiotics (lincomycin and spectrinomycin) also migrated as cations with lincomycin migrating 3.0 cm and spectrinomycin 6.0 cm, at pH 6.0.Both antibiotics demonstrated a decrease in mobility with increas- ing pH. The tetracycline class of antibiotics (chlortetracycline and oxytetracycline) migrated as cations at pH 6.0, with chlor- tetracycline being the fastest of the two antibiotics. At pH 8.0, however, these two antibiotics reversed direction and migrated as anions with chlortetracycline being the faster of the two antibiotics. The peptide and peptide-like antibiotics (bacitracin and virginiamycin) migrated toward the cathode at pH 6.0; ba- citracin migrated the furthest. With a change in pH, both of286 ANALYST, MARCH 1993, VOL.118 Table 6 Selected electrophoretic conditions for the separation of antibiotics. (Conditions were selected that will maximize separation while maintaining efficiency) Fig. 5 Top view of a pH 6.0 agarose gel containing an indicator dye and various antibiotics after electrophoresis and bioautography . ERM, Erythromycin; OLE, oleandomycin; TYL, tylosin; HYG, hygromycin B; NEO, neomycin; and STM, streptomycin. Bacillus subtilis was used as the biological indicator Fig. 6 Top view of a pH 8.0 agarose gel containing an indicator dye and various antibiotics after electrophoresis and bioautography. ERM, Erythromycin; OLE, oleandomycin; TYL, tylosin; HYG, hygromycin B; NEO, neomycin; and STM, streptomycin. Bacillus subtilis was used as the biological indicator these antibiotics acted as anions with virginiamycins migrating the furthest.Virginiamycins at this pH also separates into two components, both of which migrated further than bacitracin. The peptide-like antibiotics (bacitracin and virginiamycins) migrated toward the cathode at pH 6.0; bacitracin migrated the furthest. With a change in pH, both of these antibiotics acted as anions with virginiamycins migrating the furthest. Virginiamycins at this pH also separates into two components, both of which migrated further than bacitracin. Also summarized in Table 5 are the electrophoretic mobilities of the miscellaneous class of antibiotics. Bam- bermycin demonstrated two components at both pH 6.0 and 8.0. Both components migrated toward the anode and demonstrated a decrease in mobility with an increase in pH.Monensin acted as an anion and migrated toward the anode. The distance migrated decreased slightly with an increase in pH. Novobiocin also acted as an anion by migrating to the anode. Unlike monensin, it demonstrated an increase in mobility when the pH was raised to 8.0. Nystatin exhibited some slight cationic activity and did not demonstrate any difference in migration from one pH to another. Penicillin G migrated towards the anode at both pH6.0 and 8.0. It displayed an increase in mobility with an increase in pH. Examples of the typical biological responses seen after electrophoresis and bioautography for the macrolide and aminoglycoside antibiotics at pH6.0 and 8.0 are shown in Parameter Condition Time Temperature Volt age Electrolyte Matrix Matrix thickness Sample volume Sample receptable Overlay thickness 1.5 h 10 “C 20 V cm-1 Tris-succinate; pH 6.0 and 8.0 Agarose; 1.0% 4 mm; 200 ml per dish 40 pl Well: 4 mm 3 mm; 150 ml per dish Figs.5 and 6, respectively. Also demonstrated are the enhancements in zone size with increasing pH. Conclusion The conditions that maximize separation of antibiotics while maintaining efficiency are summarized in Table 6. This system represents an efficient and easy to perform method for the separation and identification of classes of antibiotics and, in some cases, individual antibiotics of the same class. This system eliminates the insensitivity, overloading, streaking and antibiotic diffusion problems associated with chromatography and paper electrophoresis.It has the ability to separate and characterize antibiotics that are usually affected by harsh conditions (drying, solvent exposure, etc.). Agar diffusion assays currently in use for antibiotic testing can be adapted to this system. This system has been used for the quantitative and qualitative analysis of 17 antibiotics used in animal feeds.”) Finally, elements of this system are disposable, Nunc cell factories make gel handling very easy and assay plates can be disposed of like any other disposable bioassay dish. Paper of the Journal Series, New Jersey Agricultural Experi- ment Station, Publication No. D-04-01112, supported in part by state funds. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 References Official Methods of Analysis of the Association of Official Analytical Chemists, ed.William, S., 15th edn., Association of Official Analytical Chemists, Arlington, VA, 1990, sect. 957.23- 975.61. Bctina, V., Nature (London), 1958, 182, 796. Betina, V., J. Chromutogr., 1964, 15, 379. Betina, V., in Methods in Enzymology, cd. Hash, J . H., Academic, New York, 1975, vol. 43, pp. 101-172. Betina, V., and Nemec, P., Nature (London), 1960, 187, 1111. Goodall, R. R., and Levi, A. A., Nature (London), 1946, 158, 675. Macek, K . , in Chromatography: A Laboratory Handbook of Cliromatogruphic and Electrophoretic Methods, ed. Heftmann, E., 3rd edn., Van Nostrand Reinhold, New York, 2975, Wagman, G. H., and Bailey, J. V.. J. Chromutogr., 1969, 41, 263. Wagman, G. H., and Weinstein, M. J . , Chromatography of Antibiotics, Elsevier, New York, 1973, pp.23-405. Wagman, G. H., and Weinstein, M. J . , Chromatography of Antibiotics, 2nd edn., Elsevier, New York, 1984. pp. 23-504. Aszalos, A., Davis, S . , and Frost, D., J. Chromatogr., 1968,37, 487. Aszalos, A., and Frost, D., in Methods in Enzymology, ed. Hash, J . H., Academic, New York, 1968. vol. 18, 172-213. Betina, V., J. Antibiot. Ser. A , 1964, 17, 127. Ikekawa, T., Iwami, F., Akita, E., and Umezawa, H., J. Antibiot. Ser. A , 1963, 16, 56. Ito, Y., Namba, M., Nagahama, N., Yamaguchi, T., and Okuda, T., J. Antibiot., Ser. A , 1964, 17, 218. pp. 139-163.ANALYST, MARCH 1993, VOL. 118 287 16 17 18 19 20 21 22 23 24 25 26 Stahl, E., in Chromutography: A Luboratory Hundbook of Chromatographic and Electrophoretic Methods, ed. Hcftmann, E., 3rd edn., Van Nostrand Reinhold.New York, 1975, pp. 164-188. Synder, L. R., in Chromatography: A Laboratory Hundbook of Chromatographic and Electrophoretic Methods, ed. Hcftmann, E., 3rd cdn., Van Nostrand Reinhold, New York, 1975, pp. 46-76. Salisbury, C. D. C., Rigby. C. E., and Chan, W., J. Agric. Food Chem., 1989, 37, 105. Audubcrt, R.. and dc Mcndc, S . , The Principles of Electro- phoresis, Hutchinson, London, 1959, pp. 1-120. Lederer, M., A n Introduction to Paper Electrophoresis and Related Methods, Elsevier, New York, 1957, pp. 1-193. Maede, K., Yagi, A., Naganawa, H., Kondo, S . , and Umczawa, H., J . Antibiot., 1969,22, 635. Michl, H., in Chromatography: A Luboratory Handbook of Chromatographic and Electrophoresis Methods, cd. Hcftmann, E., 3rd edn., Van Nostrand Reinhold, New York, 1975, Salvatore, M. J., Jr., and Katz, S. E., J . Assoc. Off Anal. Chem., 1988. 71, 1101. Umczawa, H., and Kondo, S . , in Methods in Enzymology, ed. Hash, J. H., Academic, New York, 1975, vol. 43, pp. 279-291. Vondracek. M. ~ in Chromatography: A Laboratory Hundbook of chromatographic and Electrophoretic Methods, ed. Heft- mann, E., 3rd cdn., Van Nostrand Reinhold. New York, 1975, Freidlin, P. J., Avian Dis., 1988, 32, 370. pp. 282-311. pp. 815-840. 27 28 29 30 31 32 33 34 35 36 37 38 39 Giddings, J . C.. in Chromatography A Laboratory Handbook of Chromutographic and Electrophoretic Methods, ed. Heftmann, E . , 3rdedn., Van Nostrand Reinhold, New York, 1975, pp. 27-45. Grynne, B., Acta Pathol. Microbiol. Scand., Sect. B: Micro- biol., 1973, 8, 583. Horng, H., and KO. H., Proc. Natl. Sci. Counc., Kepub. China, Part 2, 1977, 10, 321. Horng, C., Hsich. J., KO, H., Jan, R., and Li, J . , Proc. Natl. Sci. Counc., Repub. China, 1979, 3, 382. Jorgenson, J . W., Anal. Chem.. 1986. 58. 743A. Lightbown, J. W., and de Rossi, P., Analyst, 1965, 90, 89. Lott, A. F.. and Vaughan, D. R., Soc. Appl. Bucteriol. Tech. Ser., 1983, 18, 331. Lott, A. F., Smither, R., and Vaughan, D. R . , J . Assoc. Off Anal. Chem., 1985, 68, 1018. Smithcr, R., and Vaughan, D. R., J . Appl. Bacteriol., 1978,44, 421. Whitaker, J. R. in Paper Chromatography und Electrophoresis, ed. Zweig, G., and Whitaker, J. R.. Academic, New York, Cassidy, P. J . , 1990. personal communication. Zimmcrman, S. B . , 1990, personal communication. Salvatorc, M. J . , and Katz, S . E., J . Assoc. Off. Anal. Chem., in the press. Paper 2103044F Received June 9, 1992 Accepted October 8, 1992 1967, VOI. I , pp. 1-356.

ISSN:0003-2654    

DOI:10.1039/AN9931800281

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, MARCH 1993, VOL. 118 Optical Characteristics of a Ruthenium(i1) Complex Immobilized Silicone Rubber Film for Oxygen Measurement Xiang-Ming Li, Fu-Chang Ruan and Kwok-Yin Wong Department of Applied Biology and Chemical Technology, Hong Kong Polytechnic, Hong Kong 289 I in a The optical characteristics of tris(4,7-diphenyl-l ,lo-phenanthroline)ruthenium(ii) perchlorate immobilized in a silicone rubber film were studied for its application t o the measurement of oxygen. The luminescence intensity and the degree of quenching of the ruthenium complex by oxygen were shown to be affected by the concentration of the complex in the silicone rubber film. The optimum concentration was found t o be about 0.2 mmol dm-3. At this concentration, the silicone rubber film containing the immobilized ruthenium complex emits the highest luminescence intensity and is able to undergo a high degree of quenching of the luminescence by oxygen. The quenching curves for 20 samples containing various concentrations of the ruthenium complex were correlated with high accuracy by using a modified form of the Stern-Volmer equation.The film preparation procedure and the solvent used were found to be critical for performance. The effect of the film thickness on the luminescence intensity and the dynamic response was also studied. Keywords: Optical sensing; oxygen; ruthenium complex; silicone rubber Oxygen concentration is one of the most frequently measured parameters in chemical, biological and biomedical systems. The Winkler titration has been adopted for the measurement of oxygen for many years and is considered, to some extent, to be the standard method.' However, the time-consuming and cumbersome nature of the titration has hindered its applica- tion to process monitoring.The Clark electrode2 was con- sidered to be a breakthrough as regards techniques for the measurement of oxygen. However, it is based on the reduction of oxygen at the cathode and the diffusion-limited passage of oxygen through the membrane. Any factor that can change the diffusion resistance, such as fouling of the membrane or a change in flow conditions in the testing fluid, could cause measurement error.3,4 The development of new techniques for the measurement of oxygen has been a continuous process over the last decade. Techniques for optical sensing of oxygen, based on the quenching by oxygen of the luminescence of various chemicals, such as organic dyes,s polycyclic aromatic hydrocarbonsG9 and transition metal complexes, w-14 have been developed.Of the compounds studied, the transition metal complexes have the distinct advantages of a long excited-state lifetime and strong absorption in the visible region. Recent developments in optical oxygen sensors have been reviewed. 15 One of the advantages of using luminescent chemicals for optical sensing of oxygen is that the mechanism does not consume oxygen. Such a feature makes the technique immune to the interferences caused by changes in resistance during diffusion of oxygen into the probe. Another advantage of the optical sensing technique is that it can be miniaturized, as a consequence of the development of miniaturized optical fibre techniques, for in vivo measurement.The optical sensing technique also provides a means for application to the remote sensing of oxygen. The luminescent chemical can be placed at the location where the monitoring of oxygen i s required, and the light source and light sensor can be placed in the visible range without physical contact with the luminescent chemical. A pressure-sensitive paint using platinum-porphyrin devel- oped by Kavandi et aZ.16 is one example of remote sensing. This paint can be used for continuous mapping of the surface pressure on aircraft or other aerodynamic surfaces. The underlying principle of the paint is that the change in pressure results in a change in oxygen level in the paint and also in the luminescence intensity of the platinum-porphyrin dissolved in a silicone resin.In order to prevent the possibility of a reaction occurring between the luminescent chemical and the species in the measuring environment , the luminescent chemical should usually be immobilized in a matrix. Various studiesl4.17 have shown that silicone rubber is a suitable matrix for the immobilization of luminescent transition metal complexes. The luminescence of the complex in the silicone rubber can be quenched by oxygen to a greater extent than in many other polymers. The successful development of optical sensors for in-contact or remote measurement of oxygen depends on an understanding of the optical characteristics of the luminescent chemicals immobilized in a suitable matrix.Demas and co-workers11~13 have studied tris(4,7-diphenyl- 1 ,lO-phenan- throline)ruthenium(Ii) perchlorate as a sensing dye for oxygen sensors. They prepared the sensing film by soaking a silicone rubber film in a dichloromethane solution of the ruthenium complex.11 In our study of the same complex as a sensing material it was found that the performance of the film is highly dependent on the method of preparation and on the concen- tration of the ruthenium complex in the polymer. This paper describes a detailed study of the optical characteristics of the tris(4,7-diphenyl-l , 10-phenanthroline)ruthenium(n) complex immobilized in a silicone rubber film, which should provide useful information for the development of a sensing material for an optical oxygen sensor.In addition, a modified form of the Stern-Volmer equation proposed previously~4 was used to fit all the oxygen quenching data obtained in this work. The results showed that this three-parameter equation could be used to correlate the oxygen quenching data obtained for a transition metal complex immobilized in a silicone rubber film as the oxygen sensing material. Experimental Materials The [ R ~ ( P h ~ p h e n ) ~ ] ( C l O ~ ) ~ (Ph2phen = 4,7-diphenyl-l,10 phenanthroline) was synthesized and purified as described by Lin et al. 18 Potassium aquapentachlororuthenate(iI1) was obtained from Johnson Matthey. The ligand Ph2phen was purchased from Aldrich. Silicon rubber (RTV 732) was purchased from VersoChem; it was a clear silicone rubber without a filler such as silica.All other chemicals were of analytical-reagent grade and were used without further purification. Oxygen and nitrogen (99.9%) were purchased from Hong Kong Oxygen. Preparation of the Oxygen Sensing Films The oxygen sensing film was made by immobilizing the ruthenium complex in the silicone rubber. Two solvents, ethanol and 1 ,2-dichloroethane7 were used to prepare a290 ANALYST, MARCH 1993, VOL. 118 solution of the ruthenium complex in order to study the effect of the solvents on the optical characteristics of the films. The ruthenium complex concentrations were 0.05 mmol dm-3 in ethanol and 0.062 mmol dm-3 in 1,2-dichloroethane. The oxygen sensing films containing various concentrations of the ruthenium complex were made by mixing various amounts of the solutions with pre-polymerized silicone rubber.The number of moles of the ruthenium complex added was calculated by using the molar concentration of the solution, the mass of the solution added and the density of the solution. The mixture of solution and silicone rubber was cast into moulds to make films with various thicknesses coated on glass slides. The preparation method was similar to that described previously.14 In order to control the film I thickness, metal shects of different thicknesses (0.1-0.5 mm) with central holes were placed on the top of the glass slides to form moulds. A small droplet of the paste-like mixture of solution and pre-polymer was placed in the central hole of the mould. The volume of the mixture added should be kept smaller than the void volume of the mould.Another piece of glass slide [with a poly( tetrafluoroethylene) (PTFE) film coating on its surface to prevent sticking] was placed on the top of the mould and pressed so that the paste became a uniform film in the mould. The final sensing film formed in the mould had the same thickness as the metal sheet without noticeable shrinking or expansion. The volume of the sensing film was calculated by dividing the mass of the film by its density, which was measured separately using a block of the same polymer. The concentration of the ruthenium complex in the sensing film was calculated by dividing the number of moles of the ruthenium complex added to make the film by the volume of the sensing film.As the solvent is removed after the silicone rubber is solidified, the final concentration of ruthenium complex in the film can be higher than that in the solution. The polymerization of the silicone rubber was retarded by prevent- ing moisture from penetrating into the polymer film, because moisture served as the catalyst for the polymerization of the RTV silicone rubber. Keeping the duration of the poly- merization process to about 48 h resulted in a satisfactory uniform distribution of the ruthenium complex in the polymer film, which was critical for obtaining consistent results. The film was evacuated under vacuum for 48 h to remove all solvent residues before use. Instrumentation The luminescence intensity measurements were conducted on a Perkin-Elmer LS-5 luminescence spectrometer coupled with a microcomputer.Two gas flow meters, individually cali- brated by using a volumetric method, were utilized to measure the relative flow rates of oxygen and nitrogen. Luminescence Intensity Measurement Oxygen and nitrogen gases were mixed in a 1 m long tube and then fed into a flow cell in which the oxygen sensing film with the immobilized ruthenium complex was exposed to the mixed gas stream, and the glass slide was facing the excitation light in the spectrometer. By varying the relative flow rates of oxygen and nitrogen, a steady environment with various concentra- tions of oxygen could be maintained in the flow cell. The oxygen concentration (YO) was calculated by dividing the oxygen flow rate by the sum of the oxygen and nitrogen flow rates.The luminescence intensities of the oxygen sensing films were measured using the spectrometer. All the samples were excited at 467 nm, which was the excited wavelength producing maximum emission intensity at 608 nm. All the emission intensities were measured at 608 nm and all the measurements were made at room temperature (25 k 2 “C) and 101.325 kPa (1 atm). Response Times of the Sensing Films Response times of the sensing films were measured by switching the gas stream alternately between oxygen and nitrogen. The luminescence intensity over the whole time course of measurement was recorded by the computer. These data were processed subsequently and the response time was defined as the time taken for the luminescence intensity to change to 95% of the whole range of the change in the luminescence intensity when the sensing film was changed from a pure oxygen to a pure nitrogen environment.Results and Discussion An ideal oxygen sensing film based on the quenching of luminescence by oxygen should have the highest luminescence intensity in an environment without oxygen and have the lowest luminescence intensity in an environment of 100% oxygen. Such a feature of the sensing film would result in a high signal-to-noise ratio, which is critical for fabricating an oxygen sensor with high sensitivity and high accuracy, and at low cost. The luminescence intensities of the sensing films (thickness 0.2 mm) in an environment without oxygen, lo, containing various concentrations of the ruthenium complex and prepared using different solvents are shown in Fig.1. The solvent effect was significant. The films using ethanol showed a much higher luminescence intensity than those using 1,2-dichloroethane. There are two possible reasons for this ‘solvent effect’. First, the solvent might affect the aggregation status of the counter ions of the ruthenium complex in the solvent because of the difference in solvent polarity. Tn a less polar environment, the complex would tend to form an ion pair with the counter ions, which might act as a quencher of the excited state of the metal complex. The aggregation status of the complex with its counter ions is fixed when the silicone rubber is solidified. The second reason for the ‘solvent effect’ might be that trace amounts of 1,2-dichloroethane are trapped in the polymer film during curing, which could cause heavy atom quenching of the fluorescence of the complex.The graph of luminescence intensity versus concentration of ruthenium complex showed a maximum at a concentration of about 0.2 mmol dm-3. As the concentration increased beyond 0.25 mmol dm-3, the luminescence intensity decreased, which suggested that self-quenching of the ruthenium complex might occur at high concentration levels. As the sensing films prepared using ethanol as solvent showed better performance than those using 1,2-dichloroethane, ethanol was selected as the solvent for the preparation of all the sensing films. All the results presented below are for those sensing films prepared using ethanol as solvent. The ratio of the luminescence intensity, Zo/Zloo, is plotted in Fig.2 as a function of the concentration of the ruthenium complex in the silicone rubber film, where IItK) denotes the 200 + + 150 -0 100 50 0 0.05 0.1 0.15 0.2 0.25 0.3 Concentration of Ru complex in film/mmol dm-3 Fig. 1 Effects of solvent and ruthenium complex concentration on the apparent luminescence intensity of oxygen sensing films (film thickness: 0.2 mm). Solvents used: A, ethanol; and B. dichloroethaneANALYST, MARCH 1993, VOL. 118 291 Table 1 Standard deviations of thc corrclation for oxygen sensing films 3 \- I 5 l l o + + + 5 0 0.05 0.1 0.15 0.2 0.25 0.3 Concentration of R u complex in film/mmol dm-3 Fig. 2 quenching ratio (film thickness, 0.2 mm) Effects of ruthcnium complex concentration on thc oxygen luminescence intensity in a 100% oxygen environment at 101.325 kPa (1 atm).This ratio represents the degree of quenching by oxygen of the luminescence of the ruthenium complex, which is directly related to the sensitivity of the sensing films to oxygen. As shown in Fig. 2, the ratio showed a moderate decrease as the concentration of the ruthenium complex increased. Considering both the intensity and the intensity ratio, the optimum concentration of the ruthenium complex in the silicone rubber film for oxygen sensing is about 0.2 mmol dm-3. It is well known that the Stern-Volmer plots of the oxygen quenching data for luminescent transition metal complexes in silicone rubber films are not linear.11,'4,17 In order to use the sensing films for the measurement of oxygen, a model should be utilized to correlate the luminescence intensity data with the oxygen concentration. A two-parameter model requires a minimum of two data points, which can only be used to correlate data having a linear relationship.A three-parameter model requires a minimum of three data points, which is the minimum requirement for a non-linear data set. In theory, any data curve can be fitted by using an equation with an infinite number of parameters. However, a good correlation using the minimum number of parameters could reduce the effort required to obtain the calibration data. A modified form of the Stern-Volmer equation, based on the kinetics of oxygen quenching and the solubility equation of oxygen in the polymer, has been proposed by Li and WongI4 for correlating the luminescence intensity of a transition metal complex in a polymer with the oxygen concentration where I,, is the luminescence intensity of the sensing film in an environment without oxygen, I is the luminescence intensity of the sensing film exposed to an environment with an oxygen partial pressure of po2 (Pa) outside the polymer film, A and B are parameters combined with the kinetic constants of oxygen quenching and the parameters of the solubility equation, and b is a parameter in the solubility equation.Unlike the two-site model, it considers all the photoexcited molecules in the silicone-rubber film can be quenched by oxygen. Details of the theoretical analysis can be found elsewhere. 14 Twenty sets of oxygen quenching data for oxygen sensing films with various concentrations of ruthenium complex in silicone rubber were fitted by using the modified form of the Stern-Volmer equation, i.e., eqn.(1). The standard deviations of the correlation are listed in Table 1. The maximum relative error, estimated by dividing the standard deviation by Idlloo, was 0.85%. Many of the relative errors were below 0.5%. Two of the quenching curves for the oxygen sensing films containing high and low ruthenium complex concentrations are shown in Fig. 3. The solid lines in Fig. 3 connecting the data points are the fitting curves obtained by using eqn. (1). The results demonstrate that eqn. (1) can be used for correlating the oxygen quenching data for a luminescent transition metal complex immobilized in a silicone rubber film and that this equation can serve as a calibration equation for oxygen Concentration of ruthenium complcx in films/ No.mmol dm-3 IdIiw Standard deviation 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0.038 0.058 0.067 0.085 0.100 0.109 0.110 0.131 0.139 0.158 0.167 0.177 0.189 0.210 0.21s 0.234 0.249 0.264 0.264 0.283 13.9 12.2 12.1 14.3 13.4 12.4 14.2 11.8 12.8 14.0 12.9 14.4 11.0 12.4 11.9 12.6 11.4 11 .5 10.6 11.4 0.1 182 0.0448 0.0493 0.0642 0.0357 0.0433 0.0537 0.0521 0.0391 0.0454 O.ob11 0.0446 0.0376 0.0387 0.0364 0.0579 0.0362 0.0618 0.0263 0.0484 16 14 - I l l l l l l l l l l 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 Oxygen partial pressure/MPa Fig. 3 Fitting of thc oxygen quenching curves (film thickness: 0.2 mm).Concentration of complex in film: A, 0.086; and B, 0.264 mmol I + I + I I I 1 1 0 0.05 0.1 0.15 0.2 0.25 0.3 Concentration of R u complex in film/mrnol dm-3 Fig. 4 Fitting parameters of the modified Stern-Volmcr equation [eqn. (l)]. Mean values of h = 4.17 sensing devices using this type of sensing film. The parameters obtained by using a least-squares fitting algorithm for the fitting equations of all the samples are plotted in Fig. 4. The value for the parameter b in eqn. (1) is almost constant at 4.17, which is consistent with the theoretical analysis14 that b is only affected by the solubility of oxygen in the silicone rubber film. The parameters A and B show a higher concentration dependence than parameter b. The concentration dependence of A and B is not explicitly included in the analysis, which292 ANALYST, MARCH 1993.VOL. 118 0 100 200 300 400 500 600 700 800 900 Ti me/s Fig. 5 Dynamic response of the luminescence intensity to a change in the oxygen concentration in the environment of an oxygen sensing film (film thickness: 0.2 mm; ruthenium complex concentration: 0.234 mmol dm-3) 250 9 I + 200 v) 150 . E F 100 50 +- B I n v Y I I 0 0.1 0.2 0.3 0.4 0.5 Fi I m t h ic kness/m m Fig. 6 Effect of thc film thickness on thc rcsponsc time (ruthenium complex Concentration: 0.149 mmol dm-3). A. 95% recovery; and B. 95% quenching 200 I 0 0.1 0.2 0.3 0.4 0.5 Film thicknesdmm Effect of the film thickness on the apparent luminescence Fig. 7 intensity (ruthenium complex concentration: 0.149 mmol dm-3) allows for further improvement of the theoretical model.However, such a shortcoming of the model does not hinder its application as an accurate calibration equation. In addition to luminescence intensity and sensitivity, the dynamic response time is another important parameter for oxygen sensing. Different applications have different require- ments for the response time. For example, the response time should be less than 1 s for pressure mapping. In contrast, a response time of the order of 1 min is satisfactory for the monitoring of aerobic cell cultures. In theory, a thinner film requires less time for the oxygen concentration inside the film to reach equilibrium with the environment outside the film. However, a thinner film would be expected to emit a lower luminescence intensity when making comparative measure- ments at the same concentration level.The dynamic response of the luminescence intensity of an oxygen sensing film (thickness 0.2 mm) to a change in the oxygen concentration of the environment is shown in Fig. 5 . The recovery time, &,is significantly longer than the quenching time, t,, as shown in Fig. 5. This might be because the oxygen molecules are adsorbed in the silicone rubber matrix, and hence the adsorption process is faster than the desorption process. The effect of film thickness on the response time is shown in Fig. 6 and its effect on the apparent luminescence intensity is shown in Fig. 7. As the film thickness is increased, the apparent luminescence intensity is also increased; this can be explained by the fact that more of the luminescent complex is presented to the path of the light beam in a thicker film.However, the luminescence intensity is not directly proportional to the film thickness. As the film thickness is increased to 0.5 mm, the magnitude of the increase in the luminescence intensity decreases, as the silicone rubber is able to reflect and absorb part of the excitation and emission light. The selection of film thickness in a specific application should be a compromise between the luminescence intensity and the response time. The authors thank the Hong Kong Polytechnic and the University and Polytechnic Grant Committee of Hong Kong for financial support. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Skoog, D. A., West, D. M., and Hollcr, F.J., Fundamentals of Analytical Chemistry, Saunders, Philadelphia, 5th edn., 1988, Clark, L. C., US Pat., 2913386, 1959. Li, X. M., and Wang, H. Y., in Transient D.O. Measurement Using a Computerized Membrane Electrode. Horizons of Biochemical Engineering, ed. Aiba, S . , University of Tokyo Press, Tokyo, 1987, p. 213. Li, X. M., and Wang. H. Y., US Pat., 4921 582. 1990. Gehrich, J. L., Lubbers, D. W., Opitz, N., Hammann, D. R.. Miller, W. W., Tusa, J. K., and Yafuso, M . , IEEE Trans. Biomed. Eng., 1986, BME-33 (2), 117. Grishaeva, T. I . , and Zakharov, A. I . , Zh. Anal. Khim., 1990, 45, 1333. Peterson, J. I., Fitzgerald, R. V., and Buckhold, D. K., Anal. Chem., 1984, 56, 62. Kroneis, H. W., and Marsoncr, H. J., Sens. Actuators, 1983, 4, 587. Optiz, N., Graf, H. J., and Lubbers, D. W., Sens. Acruutors, 1988. 13, 159. Wolfbeis, 0. S . . Leiner, M. J. P., and Posch, H. E., Mikrochim. Acta, Part //Z, 1986, 359. Bacon, J. R . , and Demas, J. N., Anal. Chem., 1987, 59. 2780. Wolfbeis, 0. S . , Weis, L., Leiner, M. J . P., and Ziegler, W. E., A n d . Chem., 1988, 60,2028. Carraway, E. R., Demas, J. N., DeGraff, B. A., and Bacon, J. R., Anal. Chem., 1991, 63, 337. Li, X. M., and Wong, K. Y., Anal. Chim. Acta, 1992, 262,27. Surgi, M. R., in Applied Biosensors, ed. Wise, D. L., Buttcrworth, Guildford, 1989, ch. 9. Kavandi. J., Callis, J., Gouterman, M., Khalil, G., Wright, D., Green, E., Burns, D., and McLachlan, B . , Rev. Sci. Instrum., 1990, 61, 3340. Carraway, E. R., Ph.D. Dissertation, Univesity of Virgina, 1989. Lin, C. T., Bottcher, W., Chou, M., Creutz, C., and Sutin, N., J. Am. Chem. Soc., 1976, 98, 6536. p. 344. Paper 2104126J Received July 31, 1992 Accepted November 6, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800289

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYST, MARCH 1993, VOL. 118 293 Adsorptive Differential-pulse Voltammetric Determination of Trace Amounts of Ruthenium R. Palaniappan and T. Ashok Kumar Applications Development Laboratory, Electrochemical and Spectral Division, Elico Pvt. Ltd., B-90, Sanath Nagar Industrial Estate, Hyderabad-500 018, India An electrochemical technique for the convenient determination of trace amounts of ruthenium was developed, based on the adsorptive accumulation of Ru"-salicylaldehyde thiosemicarbazone on the surface of a hanging mercury drop electrode, followed by the reduction of the adsorbed complex during the cathodic scan. The adsorptive differential-pulse voltammetric curve exhibited a well-defined cathodic peak a t -0.750 V versus a saturated calomel electrode. Cyclic voltammetry was used to characterize the interfacial and redox behaviour.The optimum analytical conditions for the determination of ruthenium, for the working range 5-80 ng cm-3, were etablished. A statistical evaluation of the experimental results is reported. The method was applied to the determination of ruthenium in synthetic solutions of various compositions as well as in ca t a I ysts. Keywords: Adsorptive differential-pulse voltammetry; ruthenium determination; adsorptive wave; salic ylalde h yde th iosemica rbazo ne Adsorptive voltammetry involves the preliminary adsorptive concentration of an electroactive complex on the surface of a hanging mercury drop electrode (HMDE) at a fixed potential more positive than the reduction potential of the complex, and subsequent measurement of the reduction peak of the adsorbed complex during the cathodic scan.The principles as well as some analytical applications of this method have been described by various workers. 1-8 Several polarographic methods have been reported for the determination of trace amounts of ruthenium,9-14 but neither their sensitivity or selectivity is very satisfactory. Because of the importance of ruthenium, a sensitive method is required for its reliable determination. In particular, the quantification of ruthenium at trace levels is desired for geological surveys, catalytic applications and material sciences. In previous electrochemical studiesls-19 the polarographic properties of the ruthenium complex with salicylaldehyde thiosemicarbazone (SAT) were examimed and it was found that the complex is adsorbed on mercury electrodes, after undergoing single-electron reduction.This adsorptive property of the ruthenium-SAT complex can be successfully utilized for the highly sensitive determination of ruthenium by adsorptive differential-pulse voltammetry (ADPV). The aim of the present study was to establish the optimum conditions for this determination and its applicability to reakynthetic sample analysis. Experimental Apparatus Pulse polarographic and absorptive voltammetric measure- ments were performed with an Elico (Hyderabad, India) Model CL-90 instrument equipped with a Polarecord x-y recorder. The three-electrode system consisted of a dropping mercury electrode (DME) or HMDE, a platinum auxiliary electrode and a saturated calomel reference electrode.Cyclic voltammetric measurements were carried out by using an EG&G Princeton Applied Research (Princeton, NJ, USA) Model 174A polarographic analyser in conjunction with a Metrohm (Herisau, Switzerland) 663 VA stand, operated in the HMDE mode. The pH measurements were carried out using an Elico digital pH meter (Model LI-120). Materials and Reagents Analytical-reagent grade chemicals and doubly distilled water were used throughout. Mercury metal (Merck, Darmstadt, Germany) of 99.8% purity was washed successively with 5% nitric acid, conductivity water and doubly distilled water under vacuum. Sodium perchlorate solution (1.0 rnol dm-3) was used as the supporting electrolyte. Walpole acetate buffers with pH values from 2.0 to 6.0 were prepared by mixing 0.1 to 1.0 rnol dm-3 hydrochloric acid and 0.1 to 1.0 rnol dm-3 sodium acetate, depending on the requirement.A stock solution of RuL" was prepared by dissolving RuC13.3H20 (Arora Metthey, Calcutta, India) in the mini- mum volume of 1 mol dm-3 hydrochloric acid (approximately 0.01 rnol dm-3) and standardized by spectrophotometry.20 An aliquot of this solution was diluted as required. For the interference studies, solutions of other noble metals such as Pd", Rhlll, Ir"' and Pt" in 1 rnol dm-3 hydrochloric were prepared fresh as stock solutions from the analytical- reagent grade salts PdClz, RhC13.3H20, IrCI3-3H20 and PtCl, (Arora-Metthey), respectively. The stock solution of Pd" was standardized gravimetrically by precipitation with dimethyl- glyoxime .21 The Rh"' solution was also standardized gravi- metrically by precipitation as the sulfide, followed by ignition to the oxide and then reduction to the metal in the presence of hydrogen.21 The Ir"' solution was standardized with 2-mercap- tobenzothiazole, and the Pt" solution as described by Ayres and McCrory.20721 Osmium tetroxide (Aldrich, Milwaukee, W1, USA) was weighed carefully in an ampoule and dissolved in 0.5 rnol dm-3 sodium hydroxide, and the solution was neutralized with sulfuric acid and diluted to a known volume before being standardized by titrimetry.20 Salicylaldehyde thiosemicarbazone (m .p.235 "C) was pre- pared by the method of Sah and Daniels.22 A stock solution of SAT (0.01 rnol dm-3) was prepared in 96% ethanol. General Procedure To a 50 cm3 calibrated flask, an appropriate volume of ruthenium solution to yield a final concentration of ruthenium between 5.0 x 10-8 and 8.0 X 10-7 rnol dm-3 (5-80 ng cm-3) was placed. The pH of the solution was adjusted to 4.5 with 10 cm3 of 2.5 rnol dm-3 acetate buffer, and 5 cm3 of 0.01 mol dm-3 ethanolic SAT solution and 10 cm3 of 1 .0 rnol dm-3 sodium perchlorate were added.The solution was made up to 50 cm3 with water and placed in a voltammetric cell. After 10294 ANALYST. MARCH 1993. VOL. 118 t Y 2 0 4- 22 3 0 Q 0 0 0.5 pA L A 1 -0.4 -0.6 -0.8 -1.0 Applied potentialN versus SCE Fig. 1 Differential-pulse polarograms of the Ru"'-SAT system. A, 0.5 mol dm--3 acetate buffer and 0.2 mol dm--3 NaC104 at pH 4.5; B, A plus 5.0 x 10-5 mol dm-3 Ru"'; C, A plus 1.0 x 10-3 rnol dm-3 SAT; and D; C plus 5.0 x mol dm-3 Ru"' under the following conditions: drop time, 1 s; potential sweep rate, 6 mV s-1; and pulse modulation aniplitudc, -50 mV.W1 and W,, are the first and second reduction waves, whereas Wcw is a catalytic wave min de-aeration with no voltage applied, ruthenium was accumulated on the HMDE at -0.540 V for a period of 90 s from a stirred solution. The stirring was then stopped and after 15 s the ADPV curve was recorded over the potential range from -0.540 to -0.850 V under the following instrumental conditions: pulse amplitude, 50 mV; scan rate, 12 mV s-1; and sensitivity, 0.1 pA V-1. The peak for ruthenium was measured at about -0.750 V. Procedure for the Analysis of Ruthenium Catalysts Catalyst (0.1 Ifi- 0.0001 g) was accurately weighed into a poly(tetrafluoroethy1ene) (PTFE) dish.Approximately 10 cm3 of water, 2 cm3 of conc. sulfuric acid and an excess of hydrofluoric acid (@YO), to remove silicon if present, were added. The contents of the dish were evaporated to dryness on a hot-plate. On cooling, water and conc. hydrochloric acid were added to the residue in the dish so that the final solution would be 10% in hydrochloric acid. The contents of the dish were then heated until dissolution was complete; the solution was cooled, transferred into a 50 cm3 calibrated flask and brought to volume with water. Diluted sample solutions were treated as described under the General Procedure. Results and Discussion Polarographic and Cyclic Voltammetric Study of the Complex The DP polarogramB for 5.0 x 10-5 rnol dm-3 Ru"' with 1.0 x 10-3 mol dm-3 SAT in 0.5 rnol dm-3 acetate buffer (pH 4.5) containing 0.2 rnol dm-3 sodium perchlorate, under DME conditions, are shown in Fig. 1 (D).The polarogram of the solution, including either ruthenium (B) or SAT (C), was almost the same as that of the acetate buffer solution (A). Neither the reduction wave of Ru"' nor that of SAT was observed in the potential window studied here. Two waves for the reduction of Ru"' appear at -0.450 V (W,) and -0.750 V (W,,) apart from the catalytic wave (W,-+,) around - 1.050 V.23 Under these conditions, it was found that the first wave (W,) corresponds to the quasi-reversible one-electron reduction of Ru"' to Ru''17 and that the second wave (W,,) is caused by the two-electron reduction of the adsorbed Ru"-SAT complex.23 The second wave was chosen for this work because of its adsorptive character and its well-defined shape and sensitivity.The repetitive cyclic voltammetry of the system was investigated with an HMDE to evaluate the interfacial and redox behaviour. Typical cyclic voltammetric curves are shown in Fig. 2 (A,C) preceded by accumulation at -0.540 V t 4- C 22 3 0 3 0.550 -0.750 -0.950 -0.550 -0.750 -0.950 PotentialN Fig. 2 Repetitive cyclic voltammograms for A, 5.0 x lo-' mol dm-3 and C, 5.0 X 10-8 rnol dm-3 Ru"' with 1.0 x 10-3 mol dm-3 SAT solution, A and C, with; and B, without a 90 s accumulation time at -0.540 V, with stirring at 400 rev min-1 and a scan rate of 100 mV s-1. Electrolyte as in Fig.1 for 90 s. In the absence of prior accumulation, only a very minute peak was observed at around -0.750 V [Fig. 2(B)]. The cathodic peak was found at almost the same potential as the DP polarographic peak, and the reduction peak in the first scan, after an accumulation time of 90 s, was much higher than that in the second scan. The peak current (ip) was a linear function of the scan rate (v) in the range 10-100 mV s-1 for both 5.0 x 10-7 and 2.0 X 10-8 rnol dm-3 Ru"', and the peak potential became more negative with increasing scan rate, v , indicating an adsorption process.24.2S When the electrode was exposed to the solution at -0.540 V, the peak current gradually increased with exposure time, which indicates that the longer accumulation time leads to the adsorption of more of the metal complex on the electrode surface, thereby yielding a larger peak current.At very low concentrations (approximately 2.5 x 10-8 rnol dm-3 of Ru"'), only the second wave (attributed to the reduction of t h e adsorbed complex) is observed, with an accumulation time of 90 s. The redox processes associated with diffusion are not sensitive to surfactants, whereas the adsorption is known to be inhibited by the addition of surfactants. The effect of surfactants on the peak current of the DP polarographic curve was examined. The peak height was considerably reduced byANALYST, MARCH 1993, VOL. 118 295 Table 1 Effect of pH on the peak current obtained for the ADPV of the Ru"l-SAT complex after accumulation at -0.540 V for 90 s. Ru"' = 5.0 X mol dm-3; [SAT] = 1.0 x 10-3 rnol dm-3 Peak current PH at -0.750 V/nA 2.0 3.0 4.0 4.5 5.0 6.0 14 126 278 280 243 150 the addition of 0.05 or 0.1 cm3 of solution (at the concentra- tions given in parentheses) of surfactants, such as sodium dodecyl sulfate (0.5 mg cm3), gelatin (0.5 mg cm3) and Triton X-100 (0.1 mg cm3), to the medium used.The peak disappeared at gelatin concentrations greater than 0.02% . Ordinary diffusion waves were not affected by the addition of 0.02% of gelatin. Therefore, adsorption is responsible for the appearance of the peak. Electrocapillary curves can yield some information about the adsorption of a particular species. An electrocapillary curve was obtained by measuring the drop times of the DME in acetate buffer solution containing 1.0 x 10-7 rnol dm--7 Rul" and 1.0 x 10-3 rnol dm-3 SAT.A substantial change in drop time was observed in the potential region more positive than -0.750 V. This indicates that the adsorption of the RuII-SAT complex on the surface of the DME changes the surface tension of the mercury drop. Optimum Analytical Conditions To determine the optimum conditions for the adsorption preconcentration of the Ru"'-SAT complex at the HMDE, the effects of varying accumulation time, accumulation potential, SAT concentration and voltage scan rate, as well as the effects of other elements on the peak current, were investigated by ADPV. The influence of the pH on the ADPV peak current of a 5.0 x 10-7 rnol dm-3 solution of Ru"' in the presence of 1.0 x 10-3 rnol dm-3 SAT is summarized in Table 1.At pH 2.0 (0.3 rnol dm-3 acetate buffer) no significant adsorption of the complex at the electrode surface was observed. The current was at a maximum in the pH range 4.0-4.5 in acetate buffer solutions containing 0.1 rnol dm-3 sodium perchlorate, 1 .O x 10-3 rnol dm-3 SAT and 5.0 X 10-7 rnol dm-3 Ru"'. As the concentration of the acetate buffer solution was increased from 0.1 to 0.3 mol dm-3 at the optimum ph (4.5) the peak current also increased. However, the peak current was independent of concentration in the buffer concentration range 0.34.8 rnol dm-3. The effect of convection mass transport was also evaluated. Accumulation of the Ru'I'-SAT complex from a stirred solution gave rise to an ADPV peak current 2.4 times as large as that obtained with the quiescent solution in the same accumulation period (90 s).In the example of a stirred solution the voltammetric curves were recorded after a 15 s rest period. The effect of accumulation time at an HMDE under 400 rev min-1 stirring was studied by ADPV in 0.5 rnol dm-3 acetate buffer solution4.1 rnol dm-3 sodium perchlorate-1 .0 X 10-7 rnol dm-3 Ru1If--1 .0 X 10-3 rnol dm-3 SAT. It was found that the peak height increased with accumulation time and that the limiting current occurred within 60 s, corresponding to complete coverage of the electrode surface by the adsorbed species. If the process is diffusion controlled, with no adsorptive accumulation, the peak height will be independent of the accumulation time before scanning.26.27 This also proves that the DP curve is an adsorption curve.All the ADPV curves were recorded with an accumulation time of 90 s. Table 2 Results for the ADPV determination of Ru in synthetic mixtures Composition of Ruthenium Standard synthetic mixture/ found* deviation/ Sample ng cm -3 ng cm-3 ng cm-3 1 Ru"l 10.5, Mo"' 65.0 10.9 0.6 2 Ru"' 15.0, Rh"' 10.5 15.6 0.8 3 Run" 20.5, Ir'll 350 21.3 1.1 4 Ru"'25.5, Osvfi' 200 26.1 0.7 Pd" 50.0, Rh"' 5.5 OsV1l1 110, Fell1 1000 Agl300, Rh"' 12.5 0 ~ ~ " ~ 150, Pd" 50.5 IrlII 500 Au"' 150 * Avcragc of five determinations. Table 3 Results for the ADPV determination of Ru in various catalysts Standard Ru present" Ru found? deviation/ Catalyst samplc ( Y o ) (Yo ng cm-3 Ruthenium-alumina 1.08 1.12 0.05 Ruthenium-silica- 1.98 2.03 0.08 Ruthenium-platinum 0.46 0.49 0.03 alumina (1.08%)-alumina * Ccrtified value.? Avcrage of six determinations. The SAT concentration had a pronounced effect on the DP adsorptive peak current. The ADPV curve peak current for 5.0 x 10-7 rnol dm-3 ruthenium increased almost linearly with increasing SAT concentrations between 1.0 x 10-4 and 5.0 x 10-4 mol dm-3, then remained constant in acetate buffer solution, with an accumulation potential, E,,,, at 0.540 V for 90 s in stirred solution. The influence of the accumulation potential on the peak current for 1.0 x 10-7 mol dm-3 Ru"' in the presence of 1 .0 x 10-3 rnol dm-3 SAT was studied. At potentials more negative than -0.700 V, reduction of the complex began. The best accumulation potential was about -0.540 V, which was slightly negative relative to the first single-electron reduction peak.This implies that the adsorbed species is a SAT complex of Ru". The ADPV current was almost unchanged with scan rates from 1 to 12 mV s-1, so a scan rate of 12 mV s-1 was chosen. The peak current increased with the pulse amplitude, and the potential shifted to more positive values. The recommended pulse amplitude was 50 mV, which afforded reasonable sensitivity and peak shape. The ADPV peak current (i,)-concentration relationship was found to be linear for 5.0 X 10-8 to 8.0 X 10-7 rnol dm-3 Ru"' under the optimized analytical conditions. Linear calib- ration graphs were obtained over two concentration ranges, viz., from 5.0 x 10-8 to 5.0 x 10-7 and from 5.0 x 10-7 to 8.0 X rnol dm-3. The calibration plot started to deviate from linearity when the Ru"' concentration was further increased and a plateau was finally reached at above 8.0 x 10-7 rnol dm-3 Ru"', probably owing to saturation of the electrode surface by the adsorbed complex.The average sensitivities were 562.5 and 507.7 nA pmol-1 for 5.0 X 10-8-5.0 x 10-7 and 5.0 X 10-7-8.0 X 10-7 rnol dm-3 Ru"', respectively. The limit of determination was 5.0 x 10--8 mol dm-3. The precision of the method was calculated from nine successive measurements on 5.0 X 10-7 and 1.0 X 10-7 rnol dm-3 Ru"l (90 s accumulation time); the mean peak currents were 283.0 and 57.4 nA, respectively (the relative standard deviations were 2.6 and 3.1%). Correlation coefficients of 0.9954 and 0.9928 were obtained for 5.0 X 10-8-5.0 x 10-7 and 5.0 x 10-7-8.0 x 10-7 mol dm-3 Ru"', respectively.296 ' In adsorptive voltammetry, interference can arise from competitive adsorption of ions or their complexes on the surface of the HMDE or from reduction peaks in the vicinity of the analyte peak.Of the various cations tested individually in the determination of 1.0 X 10-7 mol dm-3 Ru"', no interference was observed in the presence of the following ions at the amount stated: Zn2+, Ni2+, Fe*+, Fe3+, Mn2+ and Mo6+ (100-fold excess); Cd'+, Ag+, Pb2+, Ir3+, Cr3+ and Sn2+ (50-fold excess); Au3+, Co*+, Os8+, Cu2+ and As3+ (20-fold excess); Pd2+, Ptz+ and V5+ (10-fold excess); and Rh3+ (equimolar amount). An additional peak was observed in the presence of 1.0 x 10-5 mol dm-3 cobalt, at 120 mV negative to the peak of interest, but this did not affect the determination of Ru"'.Tervalent Rh seriously interfered by increasing the peak current above the equimolar amount. Analytical Applications of the Method As the proposed method proved to be sensitive and free from most of the associated metal ions, it was applied to the determination of ruthenium in synthetic mixtures (Table 2) and a few catalysts (Table 3). The results obtained for the determination of ruthenium (Table 2 and 3) further confirmed the analytical usefulness of the proposed method. The method described provides a simple approach to the determination of trace levels of ruthenium. The interfacial accumulation results in a substantial enhancement of the voltammetric response, permitting convenient determination at the ng cm-3 level.References Kalvoda, R . , Ann. Chim. (Rome), 1983, 73, 239. Kalvoda, R., Anal. Chim. Acta, 1982, 138, 11. Vydra, F., Stulik, K., and Julakova, E., Electrochemical Stripping Analysis, Ellis Horwood, Chichester, 1979. Nurnberg, H. W., Instrumentelle Multielementanalyse, VCH, Weinheim, 1985. 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 ANALYST, MARCH 1993, VOL. 118 Pihlar, B., Valenta, P., and Nurnberg, H. W., 2. Anal. Chem., 1981,307, 337. Wang, J., Am. Lab., 1985, 17, 41. van den Berg, C. M. G., J. Electroanal. Chem., lnterfacial Electrochem., 1986, 215. 111. Wang, J., Lin, M. S . , and Villa, V., Analyst, 1987, 112, 1303. Love, D. L., and Greendale, A. E., Anal. Chem., 1960,32,780. Wagnerova, M., Collect. Czech. Chem. Commun., 1962, 27, 1130. Buckley, J. P., Anal. Chim. Acta, 1970, 52, 379. Hojman, J. Stefanovic, A., Stankovic, B., and Zuman, P., 1. Electroanal. Chem. Interfacial Electrochem., 1971, 30, 469. Medyantseva, E. P., Ulaklovich, N. A., Romanova. 0. N., and Budnikov, G. K., Zh. Anal. Khim., 1989, 44, 695. Medyantseva, E. P., Budnikov, G. K., Romanova, 0. N., and Zhivolup, I . V., Zh. Anal. Khim., 1987, 42, 1846. Palaniappan, R., Analyst, 1989, 114, 1043. Palaniappan, R., and Revathy, V., Analyst, 1989, 114, 517. Palaniappan, R., J. Electrochem. Soc. India, 1990, 39, 21. Palaniappan, R., and Paul, A., Proc. Indian Acad. Sci. (Chem. Sci.), 1989, 101, 115. Palaniappan, R., Bull. Electrochem., 1991, 7, 367. Beamish, F. E., and van Loon, J. C., Recent Advances in Analytical Chemistry of Noble Metals, Pergamon Press, London 1972, pp. 357, 358, 466. Beamish, F. E., and van Loon, J . C., Analysis of Noble Metals, Academic Press, New York, 1977, pp. 131-135, 141, 142, 145, 146. Sah, P. T., and Daniels, T. C., Recl. Trav. Chim. Pays-Bas, 1950, 69, 1545. Palaniappan, R., Ph.D. Thesis, Madras University, India, 1989. Osteryoung, R. A., Lauer, G., and Anson, F. C.. Anal. Chem., 1962, 34, 1833. Li, Q., and Li, S., Dianfenxi Huaxue, 1989, 1, 65. Koryta, J.. Collect. Czech. Chem. Commun., 1953, 18, 206. Li, Q., and Li, S . , Fenxi Ceshi Tongbao, 1989, 3, 45. Paper 2104739J Received September 3, I992 Accepted October 22, 1992

ISSN:0003-2654    

DOI:10.1039/AN9931800293

出版商:RSC    

年代:1993

数据来源: RSC