摘要:

Rapid screening of banknotes for the presence of controlled substances by thermal desorption atmospheric pressure chemical ionisation tandem mass spectrometry Richard Sleeman,a I. Fletcher A. Burton,a James F. Carter*b and David J. Robertsb a Mass Spec Analytical Limited, Building 20F, Golf Course Lane, PO Box 77, Bristol, UK BS99 7AR. E-mail: rs@msaltd.co.uk b Environmental and Analytical Section, School of Chemistry, University of Bristol, Cantock’s Close, Bristol, UK BS8 1TS.E-mail: jim.carter@bristol.ac.uk Received 4th November 1998, Accepted 21st December 1998 The ability to thermally desorb directly particulate matter, trapped on filter meshes, into the atmospheric pressure chemical ionisation source of a tandem mass spectrometer allowed the simultaneous detection of a range of controlled substances within complex matrices with a high degree of confidence. Dust samples were collected from bundles of banknotes using simple apparatus attached to a portable vacuum cleaner.This technique was employed without additional clean-up procedures, rendering the overall method extremely rapid. The intensities recorded for characteristic gas phase ion transitions allow the determination of the amounts of contaminants present. It has been reported that a significant proportion of bundles of banknotes are contaminated with cocaine. This study found that cocaine and heroin (and two related opiates) were present above the detection threshold on UK banknotes from general circulation.Differences in both the frequency and degree of contamination were apparent between bundles of banknotes from general circulation and those suspected of being associated with the trafficking of drugs. In addition, a significant number of bundles of banknotes, confiscated by H.M. Customs and Excise, were found to be contaminated with detectable levels of tetrahydrocannabinol and 3,4-methylenedioxymethylamphetamine. Large quantities of controlled substances are seized by authorities within the UK each year.The transportation, repackaging, sale and use of such materials will almost inevitably cause contamination of the premises, clothing and other possessions of persons involved in these activities. In addition, dealing in controlled substances frequently involves the exchange of large sums of cash, which may in turn become contaminated. The presence of trace levels of controlled substances in premises and motor vehicles and on clothing and currency is frequently used as part of the evidence to establish a link between an individual and these substances.Furthermore, English law permits the seizure1 of cash sums of £10 000 or more from persons suspected of being associated with drugs trafficking. A forfeiture order may subsequently be granted if this can be substantiated.2 The contamination of clothing, banknotes, etc., is assumed to result, at least in part, from microscopic particles of material, physically trapped between the fibres of the specimens.A strategy has therefore been devised whereby dust samples may be collected by means of vacuum sampling. Particles are collected on filter meshes using relatively simple equipment and are transferred to a laboratory for subsequent analysis. In this manner, it is not necessary for the items requiring examination to be physically transported to the laboratory. This methodology reduces the number of handling stages, which, in turn, minimises the possibility of inadvertent contamination from other sources; solvents, glassware, etc. A ‘control blank’ is also routinely taken in order to establish that the person involved in the collection of the sample and the equipment used are free from contamination with the compounds sought.The filter meshes are subsequently thermally desorbed in the laboratory, liberating volatile compounds into the atmospheric pressure chemical ionisation (APCI) source of a tandem mass spectrometer. A number of precursor–product ion transitions are monitored which are characteristic of specific compounds.In this study, filters were analysed for the presence of the nine compounds listed in Table 1. A different range of compounds may be sought on each filter mesh as required. This technique has been applied to the detection of trace levels of explosives in extremely complex matrices, with no need for additional cleanup procedures.3,4 It is evident that the removal and collection efficiency of the technique described will vary according to the chemical and physical nature of both the analyte and the substrate. The overall vacuum sampling strategy has been shown empirically to be suitable for the removal and collection of trace levels of a wide range of compounds, including the nine substances reported here, from a variety of substrates. The limits of detection of the overall approach are likely to differ from those achieved via alternative extraction procedures, such as those employed by other workers.5,6 The rapidity of this approach enables a greater number of samples to be collected and analysed from a given population in the time available, an important consideration for forensic applications.Similarly, other potentially significant evidence, present in the form of soluble components, such as fingerprints and inks, is not destroyed. Although GC-MS is perhaps the most widely used confirmatory technique in analytical chemistry, Busch et al.7 compared and contrasted it with MS-MS for the determination of targeted compounds in complex mixtures.They concluded that the analysis of targeted compounds (certain drugs) in the presence of complex mixtures (inks and dyes, finger grease and sweat, household contaminants and general dust and debris) is precisely the type of analysis for which MS-MS is suited. A limited number of studies have examined the contamination of paper currency from North America with cocaine and related substances.5,6 These two studies employed solvent Analyst, 1999, 124, 103–108 103extraction to remove controlled substances from banknotes, followed by extensive clean-up procedures and GC–MS detection.Hudson5 examined Bank of Canada notes in Regina, Saskatchewan. More recently, Oyler et al.6 examined US dollar bills from various locations within the USA. They concluded that the contamination of banknotes with cocaine was widespread.Nevertheless, Hudson also concluded that although such contamination is commonplace, there may still be grounds to associate confiscated monies with the drugs trade if it is contaminated to an abnormally high degree. The data shown illustrate the application of thermal desorption tandem mass spectrometry to the detection of nine controlled substances (Table 1) in dust collected from bundles of UK banknotes. Two groups of 97 and 96 bundles, each comprising approximately 250 banknotes, were analysed (a total of 48 250 banknotes).The first group8 had been returned to the Bank of England Printing Works after a period in general circulation, ranging from 1 year for £5 notes to several years for £50 notes.9 These were assumed to be representative of banknotes in general circulation at the time of this study, June 1995. The second group was typical of bundles of banknotes seized by H.M. Customs and Excise Officers within the UK between January 1994 and May 1996.At the time of seizure all these monies were suspected of being associated with the trafficking of controlled substances. Experimental Equipment Each dust sample was collected using a specially designed disposable brush-head cartridge assembly (DBCA), fabricated from stiff card [Fig. 1(a)] (S. H. Fiske, Bristol, UK). The opening at the front of the DBCA supports a row of bristles, designed to liberate dust from the object being sampled. The body of the DBCA holds a filter cartridge [Fig. 1(b)] which supports three quartz mesh elements designed to collect particulate matter and adsorb vapours. The mesh elements are coated with dicyanoallylsilicone gum (OV-275) to enhance collection efficiency. When used correctly, it is not possible for the operator to come into contact with these meshes. In operation, the DBCA is mounted on the nozzle of a portable domestic vacuum cleaner. Particle laden air is then drawn through the DBCA by the vacuum cleaner, which is located upstream with regard to the direction of air flow and is, therefore, unlikely to contaminate the filters.The filter cartridges are manufactured with a locating key on one face, ensuring that dust is collected on only one side of the meshes. When the meshes are desorbed, air passes through in the same direction, reducing the amount of particulate material carried to the ion source of the mass spectrometer. For routine applications, the DBCA is supplied in a heat-sealed plastic bag together with a tamper-evident bag into which it is sealed for return to the laboratory.Sampling Dust samples were collected at the Bank of England Printing Works (Loughton, Essex, UK) using an Electrolux Harmony 1300 W variable speed domestic vacuum cleaner equipped with a DBCA unit. The vacuum cleaner was adjusted to produce an air flow of 2.5 m s21, measured at the head of the DBCA unit with a cartridge installed.The flow was calibrated using an Airflow Instrumentation TA2 anemometer/thermometer (Airflow Developments, High Wycombe, Bucks., UK). Wearing a disposable paper oversuit and gloves, the operator covered a table with a layer of clean aluminium foil. A DBCA was fitted to the vacuum cleaner and passed over the entire surface of the foil and the operator’s gloves. The sample thus obtained was designated as a control blank. The bundles sampled were drawn from larger bundles which comprised 500 individual UK banknotes, which had been returned to the Bank of England in sealed, labelled, plastic wrappers, indicating the date and the location of source.A Bank of England employee separated approximately 250 notes from a bundle and handed these to the operator for sampling. The Bank of England staff were requested to handle the cash in as normal a manner as possible and took no unusual precautions to preclude crosscontamination. It was thought that this would best reflect the conditions under which cash sums are normally handled.Particulate matter was collected from the bundle of banknotes by holding it above the aluminium foil covered surface, fanning the banknotes out and passing the DBCA over the money and the foil surface so that any particles disturbed from the notes were drawn on to the filter meshes. The sampling time was of the order of 30 s. The bundle of banknotes was then handed back to the Bank of England employee and the surface of the aluminium foil again vacuumed.The cartridge was removed and sealed into a suitably labelled plastic bag. The aluminium foil, gloves and filter holder were discarded. A total of 97 bundles, comprising approximately 25 000 individual banknotes, were analysed in this manner. Dust samples from monies seized by H.M. Customs and Excise were collected by trained Customs Officers spanning the period January 1994 to May 1996. Officers responsible for collecting the samples were instructed to use the above procedures.Samples were collected using a hand held portable vacuum cleaner (Black and Decker HC150), which produced a similar air flow to the system used at the Bank of England. The sample size and collection procedure were, therefore, comparable to those employed at the Bank of England. Mass spectrometry The analytical instrument used for this study was an Aromic 9100 triple quadrupole mass spectrometer (Perkin-Elmer SCIEX, Thornhill, ON, Canada) specifically designed for sample introduction by thermal desorption and APCI.The instrument was tuned using diisopropyl methylphosphonate Fig. 1 (a) Disposable brush-head cartridge assembly (DBCA) with filter cartridge installed, shown fitted to a portable domestic vacuum cleaner, as employed during sample collection. (b) Filter cartridge assembly. 104 Analyst, 1999, 124, 103–108vapour bled into the ion source in a flow of high purity air (99.999%) (BOC, London, UK) and ionised in the positive mode.A number of reference peaks were monitored across the mass range using each of the analysing quadrupoles and the resolution was adjusted to give half-height peak widths of between 0.64 and 0.70 Da at m/z 29 and between 0.74 and 0.80 Da at m/z 361. Mass calibration was effected using these ions and the ion at m/z 181. Finally, the electron multiplier voltage was adjusted for optimum response to the ion at m/z 361.A system check was then performed which monitored the response of the instrument to a 1 ml aliquot of a standard solution containing cocaine (1 ng), diacetylmorphine (heroin) (2.5 ng) and D9-tetrahydrocannabinol (THC) (1 ng). Mass calibration, peak width adjustment and system performance checks were conducted at the beginning and end of each day’s operations and were automatically prompted following 4 h of operation. The filters were removed from their packaging and inserted into the analytical instrument by the operator without the need for any further handling steps.The relevant control blank was analysed immediately prior to each sample. Control blanks and sample cartridges were treated in an identical manner. The operator entered the identity of the sample into the data system, all subsequent operations being automated. The filter cartridge was positioned so as to align with the stream of high purity air at a flow rate of 2 l min21. A small element heated the air stream to a temperature of 330 ± 30 °C (measured immediately after the heated element) by application of a 5.0 A current for 2.0 s, followed by 6.0 A for a further 13 s.The temperature achieved varied according to the amount of dust present on the cartridges, which may have restricted the flow of air. The desorbed vapours were entrained in this stream and carried along a PTFE lined transfer tube, maintained at a temperature of 200 °C, to the ionisation source of the mass spectrometer.APCI was used with the corona discharge current maintained at 3 mA (corresponding to about 6 kV). Ions were introduced to the mass analyser region via a 250 mm orifice. Dry nitrogen curtain gas was passed across this opening to prevent the ingress of atmospheric impurities and to facilitate the break-up of adduct ions.10 Filters were screened for the presence of the nine compounds listed in Table 1. The instrument was used in the multiple reaction monitoring (MRM) mode, sequentially scanning 21 transitions for 20 ms each, for a total time of 15 s.The collision region was operated with argon at a concentration of 300 mmol cm23 and an offset voltage of 58 V for heroin and methaqualone and 41 V for all other compounds. Ion counts for each transition were recorded approximately 26 times throughout the desorption process, data being acquired and processed using the Aromic data system operating on a Digital PDP/11 computer.Data reported show the peak height of the transient signal recorded for each transition during the desorption process. Routinely, two of the filter meshes were analysed for the presence of a range of compounds. The third mesh was retained for possible subsequent analysis either by this or some other complementary technique. Although thermal desorption is considered to be a ‘destructive’ technique, it has been shown empirically that the process does not remove all material and further analyses may be performed on any of the meshes.Following each sample in which the presence of a monitored substance was detected, a purge cycle was initiated by the operator. This procedure entailed the introduction of an aluminium block with the same profile as a filter cartridge into the instrument. A sequence was then initiated whereby hot air (about 380 °C) was passed through the desorption region and heated transfer line for 15 s at each mesh position, to remove residual contamination.This was automatically followed by a normal analytical sequence with the aluminium block in situ. The purge cycle was repeated until the signals recorded returned to below established values. Two unused cartridges were then analysed and found to be blank before proceeding with any further analyses. Results and discussion The APCI mass spectra of the nine substances examined are dominated by the protonated molecule, [M + H]+. These spectra provide little information other than confirming the molecular mass of the compound.In contrast, the product ion spectra of these protonated species, shown in Fig. 2, are highly specific to the compounds of interest. In the analytical methodology adopted, a complete product ion spectrum is not recorded, and the mass spectrometer dwells for a longer period on gas phase transitions which have been empirically found to be highly specific to the compounds of interest, not necessarily the most intense transitions.In this way, an enhancement in sensitivity is obtained without a concomitant loss in selectivity. For most compounds only two precursor–product transitions are monitored, although simple molecular species require more than two transitions to be recorded for positive identification.3 Similarly, the very high detection threshold designated for the precursor– product transition m/z 136?119 (150 000 ion counts) reflects the lower specificity to amphetamine. It must also be noted that the transitions corresponding to THC are characteristic of D9- tetrahydrocannabinol, D8-tetrahydrocannabinol and certain other cannabinoids which are natural constituents of cannabis.The signal thresholds, shown in Table 1, were established empirically for each of the chosen precursor–product transitions and represent detection limits in excess of three standard deviations above the average signal observed from many thousands of analyses performed on dust samples collected from a wide range of substrates over a period of many years.For a compound to be positively identified, the response obtained from all relevant precursor–product transitions must be above the threshold levels. If the response for any of the ion transitions monitored was below the threshold, the compound was deemed not to be detected. The ratios of the ion transitions were also required to lie within a range of 0.5–2 times those recorded for an authenticated standard for positive identification. A simple criterion to confirm the presence of a compound was thus established. Results from the analysis of samples and control blanks collected by H.M. Customs and Excise and from the Bank of England were treated as four discrete data sets.For each data set the average and standard deviation of the peak height for each precursor–product transition were calculated. These results were then used to calculate the range of signal (calculated as average ± 1.96 3 standard error of measurement) expected for each transition at a 95% level of confidence,11 shown in Table 1.The signals recorded, corresponding to cocaine, THC, heroin, MDMA and amphetamine for each sample from both groups, are shown graphically in Fig. 3. For clarity, Fig. 3 shows only the intensity of the first transition listed in Table 1, data for other transitions being in close agreement. Narcotine and papaverine, which are products of the opium poppy, were found to occur with heroin and are not included in Fig. 3, also for clarity. The range for the percentage of bundles of banknotes which exceeded the detection threshold for specific compounds, at a confidence level of 95% (calculated as P ± 1.96 3 standard error of percentage),12 is given in Table 2. Only data from the analysis of the central filter meshes are shown in Fig. 3 and Table 2. Data from the analysis of the second mesh of each cartridge were in close agreement with results from the first mesh and are not presented for clarity.No controlled substances were detected on any of the control blank cartridges acquired at the Bank of England. The signals recorded were equivalent to approximately 20 pg of cocaine and 100 pg of other compounds, but these were all below the designated thresholds. There was no statistical difference Analyst, 1999, 124, 103–108 105between the range of signals recorded for the control blanks taken from the Bank of England and those collected by H.M.Customs and Excise corresponding to any of the compounds sought. This indicates that the environments in which both groups of samples were collected could be considered to be free from contamination with these compounds. Two control blank cartridges acquired by H.M. Customs and Excise showed small, but detectable, responses for cocaine (equivalent to approximately 60 and 80 pg). This was probably due to contamination from paper oversuits which are not routinely changed between the collection of different samples.In practice, data from the samples corresponding to these control blanks would be considered unreliable but are included for the purpose of this study. With these two exceptions, all control blanks yielded signals far below the designated thresholds. Cocaine, heroin, narcotine and papaverine were all deemed to be present on at least one of the bundles of banknotes sampled at the Bank of England.Of these compounds, only the signal ranges for the precursor–product transitions corresponding to cocaine were statistically different from the control blanks at a 95% confidence level, probably because the opiates were only detected on a single sample. The upper range of signal recorded for cocaine was equivalent to approximately 1.1 ng. Of the 97 bundles of banknotes sampled, 39 were deemed to be contaminated with cocaine, which was equivalent to 40.2% of the samples.Assuming that the number of banknotes in general circulation represents an infinitely large population, it may be inferred that there is at least a 31.4% chance of a bundle of 250 banknotes drawn from general circulation being detectably contaminated with cocaine. These data show a slight trend for Fig. 2 Product ion mass spectra of (a) amphetamine, (b) methamphetamine, (c) MDMA, (d) methaqualone, (e) cocaine, (f) THC, (g) papaverine, (h) heroin and (i) narcotine, showing [M + H]+ species formed by atmospheric pressure chemical ionisation and monitored product ions (marked with asterisks). 106 Analyst, 1999, 124, 103–108smaller denominations to be more frequently contaminated, £5 (57%), £10 (42%), £20 (37%) and £50 (19%), possibly reflecting a higher incidence of contact between these denominations and cocaine or the greater number of transactions which these denominations undergo. A single bundle of £5 banknotes exceeded the detection limit for heroin, equivalent to approximately 500 pg.The same bundle also contained traces of narcotine and papaverine, which are natural products of the opium poppy [Papaver somniferum L. (Papaveraceae)]. Despite this finding, ranges of both the signal level and percentage contamination, at 95% confidence, were not significantly different from the corresponding control blanks. These data, therefore, do not provide conclusive evidence that any proportion of UK banknotes in general circulation is contaminated with opiates.The detection of high frequencies or amounts of heroin contamination can, therefore, be viewed as uncommon. The extent of contamination found on bundles of banknotes confiscated by H.M. Customs and Excise was more intense and widespread than found on those sampled from general circulation, as is apparent from Fig. 3. Cocaine, THC, heroin, papaverine, narcotine, amphetamine and MDMA were all positively identified on at least one bundle of banknotes.With the exception of amphetamine and THC, the ranges of signal intensities recorded were all significantly greater than those from the corresponding control blanks. In addition, the ranges of both signal intensity and percentage contamination for cocaine, heroin, narcotine, papaverine and MDMA were significantly greater than those recorded from banknotes in general circulation. The differences in both the frequency and intensity of contamination with cocaine between confiscated banknotes and those in general circulation were less distinct than observed for other compounds and were not distinct at a 95% confidence level (Tables 1 and 2).The upper range of the signal corresponding to cocaine on confiscated banknotes was equivalent to approximately 2.9 ng. Many of the bundles of seized banknotes analysed were, however, contaminated with such large amounts of cocaine that the response of the detector was saturated. This was also obvious from the number of purge cycles required to remove all traces of the compound from the instrument following the analysis of dust from a number of bundles of confiscated banknotes. A more distinct difference between the two populations would undoubtedly have been apparent had the dynamic range of the analyser been greater.It is also unquestionable that the frequency of contamination observed is dependent on the designated detection threshold and it is essential, therefore, that comparisons are made only between samples acquired and analysed in an identical manner. Had a higher detection threshold been assigned, the percentage of bundles of confiscated banknotes deemed to be contaminated with cocaine would have been much greater than those from general circulation.The range of signals recorded corresponding to, methaqualone, methamphetamine and amphetamine from the Bank of England and H.M. Customs and Excise samples were not statistically different from each other, or from the corresponding control blanks.Methaqualone and methamphetamine were not positively detected on any of the samples analysed. Amphetamine was detected above the threshold level on a single sample collected by H.M. Customs and Excise with the outcome that the signals recorded encompassed a wide range, including zero. This is consistent with the range of the proportion of samples found to be contaminated, which also includes zero, indicating that the detection of amphetamine, methamphetamine or methaqualone on bundles of banknotes is, again, far from commonplace.Estimates of the levels of controlled substances detected have been made by comparison with the response obtained from the standard solutions analysed during the system checks. Material deposited from solution has, however, been empirically shown to exhibit a much sharper desorption profile than material collected as part of a dust sample. Also, the total amount of material collected on a filter mesh will affect the efficiency of the APCI process and may reduce the response of the instrument to any targeted substances present.Using this comparison, the Table 1 (i) Target compounds, (ii) gas phase ion transitions recorded, (iii) designated detection threshold for each transition and (iv) range of ion counts predicted for each transition, at 95% confidence (unless stated otherwise) (iv) Bank of England H.M. Customs and Excise Bank of England H.M.Customs and Excise (ii) (iii) control blanks control blanks samples samples Precursor? Detection (ion counts) (ion counts) (ion counts) (ion counts) (i) product threshold Compound (m/z) (ion counts) Lower Upper Lower Upper Lower Upper Lower Upper Cocaine 304?182 30000 4318 5627 5173 8963 123810 440411 519906 1134886 304?105 20000 2105 2616 2859 4391 43928 117091 268264 741935 Cocaine 304?182 30000 4111 5834 4574 9563 65954 491548 422639 1232154 at 99% confidence 304?105 20000 2024 2697 2617 4633 18659 228961 193347 816852 THCa 315?193 13500 703 1232 1232 1984 485 663 7478 24945 315?259 4500 651 744 798 1056 615 662 1639 4251 Heroin 370?328 3000 186 214 183 726 0 652 1746 9548 370?268 3000 340 378 884 1638 72 679 1364 6441 Papaverineb 340?202 6000 471 535 337 798 441 825 1384 5184 340?171 3000 741 890 768 1018 676 756 1001 1842 Narcotineb 414?220 12000 489 584 329 400 0 6072 13053 49544 414?179 12000 2106 2508 1705 2126 1588 2671 3323 7268 Methaqualone 251?132 30000 1524 1819 1201 1799 1703 2087 1289 1931 251?91 30000 2061 2216 3495 4196 2490 2652 3712 4591 Methamphetamine 150?119 20000 866 981 1045 1363 851 898 1055 1826 150?91 80000 4678 5212 5008 5922 4617 4833 5858 10498 150?32 30000 358 437 203 258 389 424 376 667 Amphetamine 136?119 150000 4350 5051 4744 7056 3689 3977 3125 16682 136?91 50000 7200 7724 11032 13393 7482 7768 0 69424 MDMAc 194?163 10000 557 611 478 598 636 682 2980 32794 194?135 10000 2864 3018 1746 2951 2865 2945 4037 21600 194?133 10000 1180 1273 1034 1344 1390 1496 3072 25001 194?105 10000 1231 1326 1132 1433 1423 1506 3576 30096 a Tetrahydrocannabinol, the active constituent of cannabis.b Natural products of the opium poppy. c 3,4-Methylenedioxymethylamphetamine or ‘ecstasy’. Analyst, 1999, 124, 103–108 107responses obtained from the six most heavily contaminated bundles of banknotes from the Bank of England indicate that the level of contamination collected on the filters was in excess of 5 ng of cocaine per bundle.Other contaminated bundles yielded at least 100 pg. Since each bundle contained approximately 250 banknotes, it is impossible to ascertain whether the contamination observed was distributed evenly amongst the banknotes or concentrated on a few individual specimens. It is also certain that the sampling procedure did not remove all the material present on the bundles of banknotes. In addition, it is not possible to estimate the proportion of material removed and, therefore, attempts to quantify the levels of controlled substances present on either the bundles or the individual notes are further complicated.Despite these restrictions, the technique does provide an indication of the overall concentration of controlled substances present on the bundles of banknotes and, therefore, may provide valuable evidence of contact with these compounds. The data reported in this study are similar to the findings of both Hudson5 and Oyler et al.6 as a significant number of bundles of UK paper currency examined from general circulation were found to be contaminated with cocaine.The levels of contamination detected were significantly lower than those reported by Oyler et al.,6 possibly reflecting different sampling procedures or different geographical patterns of drug use. These data also confirm the findings of Hudson5 that the levels of cocaine contamination on bundles of banknotes alleged to be associated with the trafficking of controlled substances are significantly higher than those observed on bundles of banknotes in general circulation. In addition, this study has shown that the occurrence of controlled substances, other than cocaine, on bundles of UK banknotes in general circulation was rare, in contrast to those suspected of being associated with the trafficking of such substances.The rapidity of the overall technique described allows sufficient samples to be collected and analysed in a given period for valid comparisons to be effected.Acknowledgements The authors thank the Bank of England and H.M. Customs and Excise National Investigation Service, Manchester for their cooperation in the collection of the samples for this study. Perkin-Elmer SCIEX, Canada, are thanked for their collaboration with instrument development. Pete Waller, Dave Leggett and S. H. Fiske Ltd. are acknowledged for their work in the development of the DBCA.References 1 Drug Trafficking Act 1994, Section 42, H.M. Stationery Office, London, 1994. 2 Drug Trafficking Act 1994, Section 43, H.M. Stationery Office, London, 1994. 3 J. B. French, W. R. Davidson, N. M. Reid and J. A. Buckley, in Tandem Mass Spectrometry, ed. F. W. McLafferty, Wiley, New York, 1983, ch. 18. 4 W. R. Stott, W. R. Davidson and R. Sleeman, Proc. SPIE, 1992, 1824, 68. 5 J. C. Hudson, J. Can. Soc. Forensic Sci., 1989, 22, 203. 6 J. Oyler, W. D. Darwin and J. Cone, J. Anal. Toxicol., 1996, 20, 213. 7 K. L. Busch, G. L. Glish and S. A. McLuckey, Mass Spectrometry/ Mass Spectrometry, VCH, New York, 1988. 8 I. F. A. Burton, MSc Thesis, University of Bristol, 1995. 9 The Bank of England Printing Works World Wide Web Pages, http:/ /www.bankofengland.co.uk/bndest.htm, 1998. 10 J. A. Buckley, J. B. French and N. M. Reid, Can. Aeronaut. Space J., 1974, 20, 231. 11 J. C. Miller and J. N. Miller, Statistics for Analytical Chemistry, Ellis Horwood, Chichester, 1st edn., 1984, ch. 2. 12 V. Barnett, Elements of Sampling Theory, Hodder and Stoughton, London, 1974. Paper 8/08573K Table 2 Confidence interval for proportion of bundles of banknotes found to be contaminated, at 95% confidence (unless stated otherwise) Compound Bank of England control blanks (%) H.M. Customs and Excise control blanks (%) Bank of England samples (%) H.M. Customs and Excise samples (%) Cocaine 0.0 0.0 0.0 4.9 31.4 51.0 56.1 75.1 Cocaine at 99% confidence 0.0 0.0 0.0 5.8 28.3 54.1 53.1 78.1 THC 0.0 0.0 0.0 0.0 0.0 0.0 7.5 21.6 Heroin 0.0 0.0 0.0 0.0 0.0 3.0 9.2 24.1 Papaverine 0.0 0.0 0.0 0.0 0.0 3.0 0.2 8.2 Narcotine 0.0 0.0 0.0 0.0 0.0 3.0 8.4 22.9 Methaqualone 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Methamphetamine 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Amphetamine 0.0 0.0 0.0 0.0 0.0 0.0 0.0 3.1 MDMA 0.0 0.0 0.0 0.0 0.0 0.0 3.5 15.2 Fig. 3 Bar charts showing the response recorded for each sample, corresponding to five compounds, of the two banknote groups studied. 108 Analyst, 1999, 124, 103–108

ISSN:0003-2654    

DOI:10.1039/a808573k

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Optimizing the balance between false positive and false negative error probabilities of confirmatory methods for the detection of veterinary drug residues† Waldo J. de Boer,*a Hilko van der Voet,a Wil G. de Ruig,b J. A. (Hans) van Rhijn,b Kevin M. Cooper,c D. Glenn Kennedy,c Raj K. P. Patel,d Sharon Porter,d Thea Reuvers,e Victoria Marcos,e Patricia Mu� noz,e Jaume Bosch,f Pilar Rodr�ýguezf and Josep M. Grasesf a Centre for Biometry Wageningen (CBW), P.O. Box 16, NL-6700 AA Wageningen, The Netherlands b DLO-State Institute for Quality Control of Agricultural Products, (RIKILT-DLO), P.O.Box 230, NL-6700 AE Wageningen, The Netherlands c Veterinary Sciences Division, Department of Agriculture for Northern Ireland, Stoney Road, Stormont, Belfast, UK BT4 3SD d Central Veterinary Laboratory (CVL-VLA), New Haw, Addlestone, Weybridge, Surrey, UK KT15 3NB e Centro Nacional de Alimentacion (CNAN-ISCIII), Carretesa Majadahonda a Pozuelo Km 2.2, 28220 Majadahonda, Madrid, Spain f Laboratori Agroalimentari de la Generalitat de Catalunya (LAGC), Apartat 12, 08340 Vilassar de Mar, Barcelona, Spain Received 9th September 1998, Accepted 1st December 1998 GC-MS data on veterinary drug residues in bovine urine are used for controlling the illegal practice of fattening cattle.According to current detection criteria, peak patterns of preferably four ions should agree within 10 or 20% from a corresponding standard pattern. These criteria are rigid, rather arbitrary and do not match daily practice.A new model, based on multivariate modeling of log peak abundance ratios, provides a theoretical basis for the identification of analytes and optimizes the balance between the avoidance of false positives and false negatives. The performance of the model is demonstrated on data provided by five laboratories, each supplying GC-MS measurements on the detection of clenbuterol, dienestrol and 19b-nortestosterone in urine. The proposed model shows a better performance than confirmation by using the current criteria and provides a statistical basis for inspection criteria in terms of error probabilities. European legislation prohibits the administration, for the purposes of growth-promotion, of b-agonists and other substances having a hormonal action to food-producing animals.Legislation is enforced through inspection procedures based on chemical analyses: urine samples are examined for drug residues and, depending on the outcome, classified as negative or positive.In both cases the conclusion can be true or false, yielding four categories: true positive, true negative, false positive and false negative results. This paper describes results from a project which aimed at estimating the probability of false positive (a) and false negative (b) results for given inspection procedures, as well as providing strategies for improving the criteria. Within the European Union (EU), Commission Decisions 93/256/EC1 and 93/257/EC2 establish criteria for the performance of screening and confirmatory analyses to be used in national residues control programmes.These criteria are used to establish whether a sample is positive or negative for a particular analyte. In this paper we call an analyte detected when not only the presence, but also the identity, is confirmed by the method of analysis. One of the most frequently used methods of confirmatory analysis is low-resolution gas chromatographymass spectrometry (GC-MS). For low-resolution GC-MS analysis the EU criteria are: (i) that the relative retention time should be within 0.5% of a standard obtained during the same run; (ii) that the intensity of preferably at least four diagnostic ions should be measured; and (iii) that the peak ratios of those ions or relative abundances should match those of a standard analyte preferably within a margin of ±10% for electron impact (EI) mode or ±20% for chemical ionisation (CI) mode. In addition to the European legislation, often the concentration is derived from the largest peak area by calibration using a standard, and only concentrations higher than a certain level are considered to provide sufficient evidence for confirmation.However, straightforward use of the proposed criteria is complicated for a number of reasons. The EU criteria are too rigid and do not match common laboratory practice. Even under highly controlled conditions laboratories often do not meet the test criteria for GC-MS measurements.EU criteria for mass spectrometry tend to be arbitrary, and aim to avoid false positive results.3 Generally, there is no mechanism to control the rate of false negatives. In practice, lower limits on the estimated concentration are imposed to avoid false positive results for low-level samples. In 1994 a monitoring study was started in three EU Member States to examine the analytical strategy and the quality of inspection procedures in order to offer a harmonized and costeffective system for the inspection of residues of veterinary † Presented at the Third International Symposium on Hormone and Veterinary Drug Residue Analysis, Bruges, Belgium, June 2–5, 1998.Analyst, 1999, 124, 109–114 109drugs. The research also aimed to provide strategies to quantify the risk of false positive and negative results in inspection procedures. The five laboratories provided GC-MS measurement data of growth-promoting agents in bovine urine with analyses spread out over a period of three years.The laboratories also provided screening, confirmation and inspection results in terms of numbers of negative and positive samples. The main aim was to collect a large set of data representing variability originating from different sources and different analytes that provided the basis for subsequent multivariate modeling. In this paper the emphasis is on the performance of a new method of analyzing GC-MS measurement data rather than on the theory behind it.A full description of the statistical model has been given elsewhere.4 In the present paper, a summary of the performance of the five laboratories for each of the three analytes is presented and discussed. Test materials Five participating laboratories in the UK, Spain and The Netherlands provided data over three years. Each laboratory used its own apparatus and Standard Operating Procedures (SOPs) to analyze the samples. Three analytes were studied: clenbuterol, dienestrol and 19b-nortestosterone in two types of sample: spiked and incurred urine samples.For both sample types, the test materials were prepared centrally and lyophilized prior to distribution to the participating laboratories. All laboratories used known negative urine samples to reconstitute the samples prior to analysis. Laboratories were instructed to ensure that the urine samples chosen to reconstitute the test materials were representative of the population of urine samples encountered in their local residues control programmes.For each of the three analytes the samples were stratified into ten rounds, each of which consisted of ten series, each of which consisted of six samples. The samples (600 for each analyte) were distributed to the participating laboratories over a three year period. The samples were coded to ensure that the participating laboratories were unaware of the analyte concentration in any individual sample.Six rounds (1, 2, 4, 5, 7 and 8) used spiked urine samples, and four rounds (3, 6, 9 and 10) used incurred urine samples. The incurred samples were prepared by the administration of growth-promoting doses of each of the anabolic agents to individual cattle. Clenbuterol was administered orally, whereas dienestrol and 19b-nortestosterone were administered by intramuscular injection. Urine samples were collected, and analyzed for drug concentration before being transported at 220 °C to a central point for the preparation of the test materials.The target analyte concentrations in the reconstituted test materials were chosen a priori to ensure that the samples containing the highest analyte concentration should be found positive by the participating laboratories in most cases. For clenbuterol, nominal concentrations of 0, 0.1, 0.2, 0.5, 1 anmg kg21 were chosen, while for both dienestrol and 19b- nortestosterone concentrations of 0, 0.2, 0.5, 1, 2 and 5 mg kg21 were chosen.Analysis of test materials The six samples that constituted a series were analyzed by each of the participating laboratories (usually) in a single analytical batch. Each batch of samples was accompanied by analytical standards, calibration or reference samples with known concentrations, as prescribed in the individual SOPs used by each of the laboratories. Samples were subjected to analysis exactly as if they had been official control samples.Test materials were in the majority of cases subjected to an initial screening test (see Table 1). These methods are generally inexpensive and rapid and are designed to prevent false negative results. However, unlike routine samples, all of the test materials were subjected to a confirmatory test. These analyses are expensive, more complicated and are designed to avoid false positive results.For confirmation, low-resolution GC-MS was widely used, with either EI or CI detection. One participant used HPLC for confirmatory analysis of clenbuterol (laboratory 1). Another participant used high-resolution GC-MS for the analysis of dienestrol and 19b-nortestosterone (laboratory 5), a technique not covered by current EU legislation.1 In the majority of cases abundances were recorded in EI mode. Most laboratories monitored several diagnostic ions, except for laboratory 5 for dienestrol and 19b-nortestosterone.Laboratory 4, clenbuterol, performed GC-MS using two different derivatives simultaneously and, accordingly, two internal standards (see Table 1). Table 1 Methods of analysis Screening method Confirmatory method GC-MS peaks at m/za Internal standard at m/z Laboratory 1: Clenbuterol — HPLC — — Dienestrol — GC-MS CI 381, 395, 410* — 19b-Nortestosterone — GC-MS CI 215, 256*, 331, 346 — Laboratory 2: Clenbuterol ELISA GC-MS EI 142, 331*, 333, 346 — Dienestrol RIA GC-MS EI 317, 395, 410*, 411 — 19b-Nortestosterone RIA GC-MS EI 256*, 290, 221, 346 — Laboratory 3: Clenbuterol ELISA GC-MS EI 86*, 187, 243, 262, 264 72 Dienestrol GC-MS EI GC-MS EI 244, 381, 395, 410* 435 19b-Nortestosterone GC-MS EI GC-MS EI 287, 403, 418*, 419 435 Laboratory 4: Clenbuterol EIA GC-MS EI 349*, 351; 391*, 393 355; 397 Dienestrol — GC-MS EI 179, 381, 395, 410* 419 19b-Nortestosterone — GC-MS EI 215, 256*, 331, 346 419 Laboratory 5: Clenbuterol ELISA GC-MS EI 243*, 285, 300 306 Dienestrol — HR-GC-MS EI 658 660 19b-Nortestosterone ELISA HR-GC-MS EI 666 669 a The base peak is marked with an asterisk. 110 Analyst, 1999, 124, 109–114EU and laboratory-specific criteria It was found that most participating laboratories developed their own criteria on which to base their confirmatory step. For example, one approach incorporated the standard deviation associated with the ion ratio, which was calculated from repeated analysis of standards, in the calculation of the range of ion ratios that were accepted as indicative of the presence of the analyte.Adapted criteria differed widely between laboratories. Multivariate log ratio model Some of the indicated problems related to GC-MS analysis of drug residues may be solved by using a statistical approach. Van der Voet et al.4 proposed a multivariate detection model based on the primary abundance measurements. The essentials of the multivariate model are described in the Appendix.The idea behind the model is that deviations in peak ratios increase at low analyte concentrations. At high concentrations peak ratios stabilize, except for random deviations caused by measurement error, within-run variability, the presence of interfering compounds and other sources. The relative error of the peak ratios may be expressed by means of the relative standard deviation. This is equivalent to the standard deviation after a log transformation of the peak ratios.The variability of the log peak ratios was determined and a non-linear relationship between variability and abundance level of the analyte was incorporated in the decision as to whether samples were classified as being either positive or negative. The test relies on method and/or laboratory-specific circumstances represented by a tolerance parameter e, the value of which should be specified externally. In this paper, a robust distribution-free estimate of the standard deviation, based on the median absolute deviation,5 was used to estimate the dispersion of log ratios Qk at the investigated concentrations.The estimate is: sk = 0.67 3 medk | Qk 2 medkQk | (1) where med denotes the median. In this paper an inconfidence level (false positive rate) a = 0.01 was used throughout. After fitting the models, power curves were constructed as follows: for a range of peak abundances critical values of the test statistic were calculated from a non-central chi-squared distribution [see Appendix, eqn.(a.6)]. Under H1, corresponding probabilities 1 2 b (true positive rate) were derived from a central chi-squared distribution [eqn. (a.7)]. Calibration data The dataset was split into two parts and the first half, with 50 series as a maximum, was used to calibrate the model. Log peak ratios were calculated, taking the largest analyte peak as denominator peak. For each run reference values mk were calculated, taking the mean of the log ratios for the reference samples with concentrations > 1 mg kg21. Validation data The model was validated using the second half of the data. Standardized log ratio statistics were calculated as before.For each sample a multivariate test was performed. Results depend on the tolerance parameter e. At the present state, the value of e is unspecified. Therefore, a value of e was assessed by performing the multivariate test for a set of increasing e values: e was taken as the highest value giving no false positives for the blanks. All calculations in this paper were performed by using the statistical package Genstat.6 Results The number of series analyzed for each analyte within each laboratory ranged from 63 to 100.For laboratory 2, the first three rounds (30 series) of results for dienestrol and 19b- nortestosterone were excluded from analysis because from round 4 onwards a different ion fragment was monitored. Laboratory 4 suffered from start-up problems in the determination of dienestrol and 19b-nortestosterone in the first two rounds.For all laboratories, further series were excluded for the following reasons: improbable abundances, e.g., negative or extremely high values, the occurrence of interchanged samples, missing samples, lost samples, failures, etc. Table 2 summarizes the results of the confirmatory analysis of each analyte reported by the laboratories and represents the false positive and true positive rates.Fig. 1 shows the operating characteristic (OC) curves for the confirmation results as delivered by the laboratories and, except for laboratory 1, clenbuterol, and laboratory 5, hormones, for the confirmation results after applying EU criteria. The figures in the plots are based on the number of validation samples, that is, roughly half the number of series summarized in Table 2. It is seen that performances differ widely between laboratories.Also seen is that some laboratories used their own criteria, whereas others applied strict EU criteria. It should be kept in mind that OC curves are not to be compared between laboratories because laboratories analyzed different samples (based on local urine matrix) and because of differences between laboratories in the application of EU criteria. Few false positive results were reported. In general, all laboratories reported false negative results even at the highest nominal concentrations, but results differed considerably.For clenbuterol, all laboratories performed similarly, with the exception of laboratory 2. For hormones, OC curves showed much more variability: at 1–5 mg kg21 laboratories 1 and 5 detected most positive samples, laboratories 2 and 4 failed to identify the presence of the analyte and laboratory 3 was in an intermediate position. Scatterplots per concentration level of log ratio variables Qk show the bivariate margins of the distribution for the set of samples.Ratios with a missing numerator or denominator peak in at least one of the ratios were replaced by arbitrary high or low values, respectively, and are also shown. Fig. 2 is an example, illustrating for one of the laboratory–analyte combinations that variability in the point cloud diminishes at the higher concentrations. Plots of log ratio variables Qk in all analyses showed a similar pattern. Laboratory 2, dienestrol, was the only exception where variability did not increase at lower concentration levels.Fig. 3 shows an example of the fitted relationships and empirical estimates for the standard deviation of the log ratio variables, based on median absolute deviations in each concentration group. For this example a region of constant relative variability in the ratios is reached at above 2 mg kg21. In general, the fitting of exponential curves caused no particular difficulties, except for laboratory 3, 19b-nortestosterone.Here, too many of the reported abundances were zero to allow the calculation of the distribution-free standard deviation from medians. For laboratory 2, dienestrol and 19b-nortestosterone, no exponential curve could be fitted for m/z ratios 217/410 and 290/256, respectively, and it was decided to exclude the ions at m/z 217 and 290 from the models. The application of the multivariate detection model to the validation data with specification limit e, based on calibration and confidence level 1 2 a = 0.99 for the confirmatory analyses, is presented in Fig. 4. The plots show that the Analyst, 1999, 124, 109–114 111performance of the proposed model is better than confirmation results after applying EU criteria, except for laboratory 2, dienestrol. In general, more true positives are detected at all concentrations. This is illustrated for laboratory 5, where all real positive samples were found, that is, no false negative results at all. The plots for laboratory 3 and to a lesser extent laboratory 4, hormones, show a significant improvement of the new method: EU criteria failed whereas the multivariate model performed fairly well. Application of the multivariate model failed for laboratory 3, 19b-nortestosterone, owing to the presence of more than 50% zeroes for some of the peaks.No multiple peaks were available for laboratory 1, clenbuterol, and laboratory 5, hormones (Table 1). In Fig. 5, theoretical power (1 2 b) of the multivariate model as a function of abundance level is plotted.Power is the complement of the false negative rate. It measures how many real positives are indeed identified as being positive in the procedure. At low levels, powers are already high, except for laboratory 2, hormones. Fig. 5 can be used to derive a rough indication of what the minimum detectable level would be by dropping a perpendicular line at the start of each power curve to the x-axis. Samples with levels below the minimum detectable level will not be detected in a multivariate test. Note that Fig. 5 establishes the power as a function of peak abundance level, not as a function of concentration. In Fig. 5, the median abundance levels of the non-zero nominal concentration groups are indicated by tickmarks. For instance, for laboratory 2, 19b- nortestosterone, the fourth tickmark indicates a nominal concentration of 2 mg kg21. The median abundance level of this concentration group is positioned slightly to the left of the point of minimum detection, meaning that at least 50% of the samples in this concentration group are not detectable.Discussion One of the premises at the start was that EU criteria did not match reality. As an answer to that, each laboratory developed in-house criteria to cope with the problems. This presumption was found to be true and is demonstrated in Fig. 1. For example, laboratory 3 illustrates very clearly the difference between the strict use of EU criteria and adapted criteria. The high degree of deviating peak patterns produces poor results by applying 10% criteria, but laboratory 3 manages to produce very satisfactory results by applying its own criteria.Laboratory 2 reported for all analytes abundance values with many deviating values. Obviously, at the start of the project the confirmatory step was not working satisfactorily and needed optimization. However, within the framework of the project all laboratories were requested not to change their methods as prescribed in the SOPs.New methods and improved procedures were not incorporated during the process, because comparability of confirmation results was not the aim. Results presented should be considered from this perspective. Note that in Fig. 1 the OC curves for laboratory 2, hormones, are based on four peaks, whereas in Fig. 4 results are based on three peaks. Fig. 2 and 3 show the main features of the data that are used in the multivariate model and make clear why the model is working.Standardized log ratios have a zero expectation and stabilize at high concentration levels of the analyte. Given a specified value for e, high relative variability at low levels will prevent the analyte from being detected while at high levels peak ratios deviate less and samples are found positive for the presence of the analyte. Provided that condition (a.7) (see Appendix) is fulfilled, samples can be detected as positives.Conditional on e, a value can be derived for which it is no longer possible to detect the analyte for mean levels equal to or below that value. The additional condition acts as an appropriate ‘detection limit’ determined implicitly by the proposed procedure. Ideally, blank samples always have mean levels below this minimum detectable level. This means that for fixed e, laboratories can only lower their amount of false negatives results (b errors) by improving their precision.Laboratories with better precision will detect more positives at lower levels. Thus, minimum detectable levels vary between laboratories and/or analytes depending on their precision. The multivariate test for detection is governed by conditions (a.6) and (a.7). Provided that condition (a.7) is fulfilled, samples with high deviating peak patterns have high values of D2 and, consequently, are found negative. Setting specification limits e wider increases the non-centrality parameter d until condition (a.6) is fulfilled and samples are detected as positive outcomes.Table 2 Confirmatory analysis: numbers of positive (+) and negative (2) outcomes and fraction positive results (p) for clenbuterol, dienestrol and 19b- nortestosterone Nominal Laboratory 1 Laboratory 2 Laboratory 3 Laboratory 4 Laboratory 5 concentration/ mg kg21 + 2 p + 2 p + 2 p + 2 p + 2 p Clenbuterol: 0.0 2 88 0.02 1 99 0.01 0 93 0.00 6 89 0.06 0 86 0.00 0.1 5 86 0.05 1 99 0.01 3 89 0.03 10 85 0.11 0 86 0.00 0.2 23 66 0.26 4 96 0.04 11 82 0.12 25 70 0.26 21 65 0.24 0.5 60 32 0.65 12 88 0.12 32 61 0.34 60 34 0.64 67 19 0.70 1.0 82 10 0.89 23 77 0.23 72 21 0.77 82 13 0.86 85 1 0.99 2.0 88 3 0.97 40 60 0.40 85 8 0.91 86 8 0.91 86 0 1.00 Dienestrol: 0.0 1 95 0.01 1 62 0.02 2 93 0.02 0 78 0.00 0 99 0.00 0.2 3 93 0.03 0 63 0.00 12 84 0.13 0 78 0.00 0 100 0.00 0.5 20 76 0.21 1 62 0.02 21 74 0.22 0 78 0.00 3 97 0.03 1.0 70 25 0.74 2 61 0.03 39 57 0.41 0 78 0.00 89 11 0.89 2.0 93 3 0.97 11 52 0.17 51 45 0.53 0 77 0.00 99 0 1.00 5.0 87 0 1.00 12 41 0.23 59 37 0.61 0 78 0.00 90 0 1.00 19b-Nortestosterone: 0.0 2 96 0.02 0 69 0.00 2 94 0.02 0 73 0.00 0 98 0.00 0.2 3 95 0.03 0 69 0.00 8 86 0.08 0 73 0.00 11 87 0.11 0.5 0 98 0.00 1 68 0.01 16 80 0.17 3 69 0.04 89 9 0.89 1.0 27 71 0.28 2 67 0.03 24 72 0.25 6 67 0.08 97 1 0.99 2.0 73 25 0.75 2 67 0.03 44 52 0.46 14 58 0.23 98 0 1.00 5.0 93 5 0.95 5 64 0.07 70 28 0.71 37 36 0.51 98 0 1.00 112 Analyst, 1999, 124, 109–114For values of d that are too low, condition (a.7) is never met and detection results will always be negative.This situation arises when (i) the dispersion is very high, e.g., laboratory 2 for all analytes, and/or (ii) the specification limits e are set too narrow. For laboratory 2, 19b-nortestosterone, the majority of the samples with nominal concentrations up to 2 mg kg21 are not detectable. For laboratory 5, all non-zero nominal concentrations are far above the minimum detectable level.For laboratory 2, strong deviating peak patterns prevent detection, whereas for laboratory 5 detection is governed by condition (a.6) only. Practical application of the model The statistical model presented in this paper has been applied in the context of a research project, using calibration to obtain preliminary estimates of appropriate specification limits e. A legitimate question is whether and how it would be applicable in practice.Clearly, application of the model requires many samples with known analyte concentrations over a relevant range to be analyzed. Therefore, application is restricted to laboratories with a sufficient throughput of samples. A first impression may be that analyzing (say) 20–50 concentration series to obtain calibration data for the model imposes a very large amount of additional work on the laboratory. However, modern validated methods of analysis already include a substantial number of calibration and quality control (QC) samples in their procedure.Typically, each GC-MS run contains 1–2 blank matrix samples, 1–4 spiked matrix samples, and a calibration series of about five standards. It is our impression that a cost-effective integration is possible of calibration, traditional QC, and error rate QC as proposed in this paper. Already now, some laboratories prefer to calibrate using spiked matrix samples rather than chemical standards.It seems Fig. 1 Detection of clenbuterol, dienestrol and 19b-nortestosterone in urine. Fraction positive confirmation results versus nominal concentrations after applying laboratory-specific criteria are represented by solid circles. Open circles represent confirmation results after applying EU criteria. Concentrations are 0, 0.1, 0.2, 0.5, 1 and 2 mg kg21 for clenbuterol and 0, 0.2, 0.5, 1, 2 and 5 mg kg21 for hormones. Fig. 2 Laboratory 1, 19b-nortestosterone: scatterplots per concentration level of standardized log m/z ratios 215/256 and 346/256.Open circles represent samples with a missing denominator or numerator peak in at least one of the ratios. Fig. 3 Laboratory 1, 19b-nortestosterone: standard deviation of standardized log ratios as a function of level. Level is the geometric mean of all available peak abundances internally standardized for laboratories 3, 4 and 5. Empirical estimates are based on median absolute deviations in each concentration level group (0, 0.2, 0.5, 1, 2 and 5 mg kg21).Circles represent m/z ratio 215/256, triangles m/z ratio 331/256 and squares m/z ratio 346/256. Fig. 4 Detection of clenbuterol, dienestrol and 19b-nortestosterone in urine. Fraction positive confirmation results versus nominal concentrations. Results of the multivariate model are represented by solid circles. Open circles represent confirmation results after applying EU criteria. Concentrations are 0, 0.1, 0.2, 0.5, 1 and 2 mg kg21 for clenbuterol and 0, 0.2, 0.5, 1, 2 and 5 mg kg21 for hormones. Values of e are depicted in the plots.Analyst, 1999, 124, 109–114 113useful to lower the concentrations in part of the regular QC samples in order to ascertain the performance at low levels. A further point needing practical elaboration is the inclusion of matrix variability in QC procedures. In current routine QC procedures this variability is simply ignored, with unknown consequences.The procedure employed in this paper uses local blank urine samples from individual animals. A subject for future research is how matrix variability can be practically incorporated in routine QC. Thus, we believe that, with a change of protocol with regard to standard and QC samples, data can be generated from routine practice to allow application of the model. An appropriate value for e should be found to describe the assumption that the spectral signature of interfering components will at least differ more from that of the analyte than eTS21e (see Appendix).Obviously, this requires experience with a range of interfering compounds, as well as a more practical implementation of the procedure. In addition, expert opinions will still be needed to assess appropriate values for e. Note that calibration of the model and specifications of e are specific for a protocol as operated in one laboratory. Laboratories with high precision should have different values than laboratories that invest more, e.g., in the quality of the clean-up procedures.In our view, harmonized regulations should be concerned with appropriate values of the allowable probabilities of false positive and false negative results a and b (the latter at a specified concentration). Conclusions Multivariate modeling offers statistically based criteria that may replace the 10 or 20% EU criteria for detection of veterinary drug residues.The performance of the proposed model with specification limits based on calibration is much better than results obtained after applying EU criteria. The multivariate model diminishes the number of false positives and false negatives in the majority of the investigated data. Assessing appropriate values of e remains a task for future work, requiring both chemical expertise and practical experience with the model. Appendix Analyte peaks are corrected for the internal standard, if available, giving A1…Ap.Log peak ratios yk = ln(Ak/Ap) are calculated, where Ap is the largest analyte peak and k = 1…q and q = p21. The multivariate model is based on standardized log ratios (relative peak abundances); Qk = yk 2 mk (a.1) where mk is the corresponding log ratio of a standard sample in the same analytical batch (possibly averaged over several standards and/or validation samples). The hypotheses for a test for multivariate detection are: H0 : E(Q)T S21 E(Q) ! eT S21 e H1 : E(Q)T S21 E(Q) < eT S21 e where e is a vector having q identical elements containing the tolerance parameter e, E(Q) is the expected value of Q, Q is the vector (Q1…Qq) and S is the variance–covariance matrix of Q.The diagonal and off-diagonal elements contain estimates of the variance s2 k and covariance parameters, respectively. The standard deviation is estimated by fitting simple exponential curves: sk = a + b exp (2kx) + e (a.2) where sk is the standard deviation of all samples with the same nominal concentration, a, b and k are parameters, x is the median over these samples of the geometric mean of p peaks (corrected for the internal standard) and e represents error.Covariances are estimated as sisjr, where r is the pooled within-group correlation. The geometric mean for p peak abundances is calculated as: x = (A1·A2·…·Ap)1/p (a.3) Define the test statistic: D2 = QT S21 Q (a.4) where D2, the squared Mahalanobis distance, has a non-central chi-squared distribution7 with non-centrality parameter: d = eT S21 e (a.5) The analyte is detected in a test with confidence level (1 2 a) if: D2 < c2 q,d(a) (a.6) provided that d > c2 q(1 2 a) (a.7) where c2 q,d (a) and c2 q(1 2 a) are critical values of the noncentral and central chi-squared distribution, respectively, with q degrees of freedom, non-centrality parameter d and significance level a. Acknowledgments This work was financially supported by the European Commission under contract AIR-3 CT 94 1415. References 1 Commission Decision of 14 April 1993 (93/256/EEC), Off. J. Eur. Comm., 1993, L118, 64. 2 Commission Decision of 15 April 1993 (93/257/EEC), Off. J. Eur. Comm., 1993, L118, 75. 3 W. G. De Ruig, R. W. Stephany and G. Dijkstra, J. Assoc. Off. Anal. Chem., 1989, 72, 487. 4 H. Van der Voet, W. J. De Boer, W. G. De Ruig and J. A. Van Rhijn, J. Chemometr., 1998, 12, 279. 5 F. R. Hampel, J. Am. Stat. Assoc., 1974, 69, 383. 6 Genstat 5 Committee, Genstat 5 Release 3 Reference Manual, Clarendon Press, Oxford, 1993. 7 E. S. Pearson, Biometrika, 1959, 46, 364. Paper 8/07051B Fig. 5 Detection of clenbuterol, dienestrol and 19b-nortestosterone in urine. Theoretical power of the multivariate model as a function of level. Level is the geometric mean of all available peak abundances internally standardized for laboratories 3, 4 and 5. Ticks indicate median levels for nominal concentrations 0.1, 0.2, 0.5, 1 and 2 mg kg21 for clenbuterol and 0.2, 0.5, 1, 2 and 5 mg kg21 for hormones. Levels of laboratories 1 and 2 are multiplied by 1025. The horizontal line shown represents a = 0.01. 114 Analyst, 1999, 124, 109–114

ISSN:0003-2654    

DOI:10.1039/a807051b

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

On-line dialysis–SPE–CE of acidic drugs in biological samples J. R. Veraart,* M. C. E. Groot, C. Gooijer, H. Lingeman, N. H. Velthorst and U. A. Th. Brinkman Vrije Universiteit, Department of Analytical Chemistry, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands Received 1st October 1998, Accepted 27th November 1998 A fully automated method is presented for the determination of acidic drugs in urine and serum using on-line dialysis–solid-phase extraction (SPE)–capillary electrophoresis (CE) with UV detection.With non-steroidal anti-inflammatory drugs (NSAIDs) as test compounds, detection limits in the biological samples were 0.05–1.0 mg ml21. Calibration plots were linear over two orders of magnitude and the within-day and between-day repeatability were better than 10%. The CE capillary and SPE column were used for over 500 analyses; the dialysis membrane was replaced after 250 analyses. A general protocol for dialysis–SPE–CE which can be used for amphoteric and acidic drugs was devised.The present results show that this protocol has general validity and can be recommended for future work on other classes of drugs. Introduction Capillary electrophoresis (CE) has a high separation power and can be used for the separation of charged analytes present in biological samples. Unfortunately, such matrices contain a number of compounds which can interfere with the analysis, such as salts, proteins and particulate matter. These compounds must be removed prior to injection by applying a suitable sample preparation procedure.Off-line methods such as liquid– liquid extraction (LLE), precipitation and off-line solid-phase extraction (SPE) are most frequently used. These are, however, time consuming and degradation of the analytes can occur because of (nearly) irreversible adsorption effects and if solvent evaporation steps cannot be avoided. However, in bioanalysis, large numbers of samples often have to be processed, which means that automated procedures are preferred.Papers describing CE procedures for the determination of NSAIDs are mainly directed at the separation of chiral isomers,1,2 aqueous standards3 and pharmaceutical formulations. 4 The determination of NSAIDs using CE in serum has been reported using off-line sample preparation methods such as LLE5,6 and protein precipitation.7 Detection limits varied from 1 to 10 mg ml,6,7 which are not sufficient for therapeutic drug monitoring for some of the compounds (see Table 1).The principle of using an automated dialysis system connected to an HPLC system has been presented.8 In a recent study, on-line dialysis–SPE–CE was applied for the first time to the determination of sulfonamides in urine and serum.9 Dialysis is applied to remove proteins and particulate matter. The analytes are diluted during this procedure; therefore, in a second step, they are trapped on an SPE column while salts are simultaneously removed.During the subsequent washing step, compounds that will interfere during the CE run can be removed. Finally, the analytes are desorbed and injected into the CE capillary via a laboratory-made interface. In this paper, it will be shown that this set-up and the optimization strategy in general can be used also to determine other compound classes such as strongly acidic drugs, with non-steroidal anti-inflammatory drugs (NSAIDs) as test compounds. The result is a fully automated method for the direct determination of NSAIDs in biological samples, presented here for the first time.Experimental Chemicals and samples Ibuprofen, naproxen, fenoprofen, ketoprofen and flurbiprofen were obtained from Sigma (St. Louis, MO, USA), acetonitrile, boric acid, methanol and phosphoric acid from J. T. Baker (Deventer, The Netherlands), decanoic acid, sodium dihydrogenphosphate monohydrate and disodium hydrogenphosphate dodecahydrate from Merck (Darmstadt, Germany) and acetic acid and sodium acetate from Riedel-de Haën (Seelze, Germany).Water was de-mineralized and distilled before use. In all cases, chemicals of the best available quality were used. Urine was collected from five healthy volunteers on three consequent days. The samples were pooled and divided into 100 ml portions and frozen at 218 °C. Bovine serum from untreated animals was purchased from Sigma; it was divided into 10 ml portions and frozen at 218 °C.The biological samples were stored for a maximum period of 3 months. Methods The set-up of the dialysis–SPE–CE system and the instruments used have been described in detail,9 and here just a schematic diagram of the set-up is presented in Fig. 1. The experimental parameters are discussed in the next section and are summarized in Table 2. Results and discussion The main aim of this project was to devise a generally valid optimization strategy for dialysis–SPE–CE. An overview of all Table 1 Therapeutic levels of the NSAIDs in serum and their absorbance maxima Compound Concentration/ mg ml21 Absorbance maximum/nm Ibuprofen 25–50 264 Naproxen 25–70 273 Fenoprofen 20–50 273 Ketoprofen 0.5–6 261 Flurbiprofen 2–12 247 Analyst, 1999, 124, 115–118 115relevant parameters is given in Table 2; data relating to an earlier first attempt are included for convenience.Because of practical aspects, several parameters were kept the same. These included the sample injection into the dialysis block (which can certainly be improved if there were insufficient analyte detectability), the dimensions of the CE capillary and the CE capillary rinsing procedure.The method development steps are briefly introduced below and the results found for the NSAIDs are presented and, if relevant, compared with those of the sulfonamides. (i) In order to optimize the CE analysis, the first choice of buffer composition is a relatively high pH to ensure complete ionization of the analytes.If co-migration is observed, optimization is carried out using pH values closer to the pKa values of the analytes involved. With the NSAIDs, complete resolution required pH adjustment to 4.6, a value very close to the pKa of all test analytes; furthermore, 10% v/v of methanol had to be added to reduce the electroosmotic flow and to improve the separation. NSAIDs have absorption maxima in the 247–273 nm range (cf., Table 1), but the absorption wavelengths of the maxima differ considerably.Therefore, to detect all analytes, ‘end absorption’, i.e., detection at 200 nm, had to be used, a wavelength choice which, of course, adversely affects the selectivity of the detection. (ii) The composition of the dialysis acceptor solution and the type of SPE cartridge largely determine the selection range of most of the other parameters. For example, the acceptor solution should not contain compounds which can interfere in the CE analysis (such as cationic ion-pairing reagents).To facilitate dialysis and increase analyte retention on hydrophobic SPE sorbents, it should also have a pH at which the analytes are neutral. As regards analyte trapping, in many instances the use of a small SPE cartridge (20 3 2.1 mm id) packed with a 5 mm sorbent is a good choice. To be on the safe side, the SPE cartridge should have a breakthrough volume for the analytes that is sufficient to retain them from a volume twice the volume of the acceptor phase generated during the dialysis step (typically 5–20 ml).For obvious reasons, an acceptor solution with low pH was selected for the NSAIDs. No breakthrough then occurred for any analyte after loading up to 50 ml of a pH 2 phosphate buffer when using C18-bonded silica as SPE sorbent. This is in marked contrast with the sulfonamides, which required a hydrophobic styrene–divinylbenzene copolymer, PLRP-S, to achieve sufficient retention.(iii) Interfacing of the SPE and CE systems involves three steps, desorption of the analytes from the SPE column, their transport to the interface and their injection into the CE capillary. A critical aspect in SPE–CE interfacing is the selection of an organic desorption buffer which, preferably, should have a low ionic strength to ensure a good stacking effect in the CE system. In order to desorb the analytes in a relatively small volume, at least 70% v/v of acetonitrile should be added to the desorption buffer.In addition, the pH should be increased to 7 to ensure that all analytes are fully deprotonated during the desorption step. Because of the relatively small volume of the SPE cartridge, the time between starting the elution from the SPE cartridge and the analytes passing the tip of the CE capillary, Dt, and the total time in which the analytes pass the tip of the CE capillary, tinj, are critical. Since a decrease in tinj requires a significantly higher pressure for a short period during Fig. 1 Set-up of the on-line dialysis–SPE–CE–UV system. 116 Analyst, 1999, 124, 115–118injection, the repeatability of the procedure will be adversely affected. As a result, the maximum desorption flow rate is 0.2 ml min21. (iv) To improve sensitivity, more sample can be injected. Because the time during which injection is performed has already been set (tinj; see above), the only way to inject more analyte is to increase the pressure during injection.Optimization can be performed by injecting increasing amounts of sample and plotting the peak height versus the injection pressure. With our set-up, analyte detectability can be improved by injecting at pressures of up to 80 mbar without a dramatic decrease in efficiency due to band broadening. Increasing the pressure eightfold resulted in an eightfold increase in the peak heights, with a concomitant decrease in plate number of only 20%. (v) Optimization of the dialysis step usually involves the selection of a suitable (often pulsed-type) dialysis mode and the sample volume for analysis.Regarding the mode, pulsed dialysis is often preferred since it improves the concentration detection limits. To facilitate the comparison of the protocols for NSAIDs and sulfonamides, the above dialysis parameters were kept the same as in previous work.9 However, because of the good retention of the NSAIDs on the SPE cartridge, a higher acceptor solution flow rate was used to decrease the dialysis time, as the risk of early analyte breakthrough was negligible (see above).(vi) The dialysis–SPE–CE–UV system is used for the analysis of urine. If too many interferences show up in the electropherogram, the SPE cartridge has to be subjected to a wash step prior to desorption of the analytes. The volume of the wash solution, its pH (often around the pKa values of the analytes) and the fraction of organic modifier (more rapid removal of interferences versus increasing loss of analytes) are the main variables. Fig. 2(A) and (B) show that a number of interfering peaks occur near those of the analyte if no wash step is included. The beneficial effect of a 1.5 min clean-up procedure in the presence of 10% v/v of organic modifier and with a phosphate buffer of the expected pH value becomes obvious from Fig. 2(C) and (D). (vii) The analysis of serum generally requires a modified strategy because of the more complex sample configuration and its higher viscosity (which tend to affect analyte detectability) and drug–protein bonding.If total drug concentrations in serum have to be determined, a significant amount of an organic solvent and/or a properly selected displacer have to be added to the sample to disrupt the drug–protein bonding. For the acidic drugs concerned, the addition of acetonitrile and decanoic acid (final concentrations 20% v/v and 8 mm, respectively) was found to give the best results.If such a step has to be included in the final protocol, it is recommended that the duration of the dialysis procedure is re-optimized. For the NSAIDs, the optimum was found to be 19 min. The increase in the dialysis time can probably be attributed to an unfavourable drug–protein equilibrium which is shifted during the dialysis procedure. With serum, a wash step was also necessary, even though the situation is less critical than with urine [Figs. 3(A) and (B)]. Table 2 Experimental conditions used for NSAIDs (this study) and sulfonamides9 Determination of NSAIDs Sulfonamides Sample treatment— Urine None Serum, free concentration None Serum, total concentration Addition of 20% ACN + 8 mm decanoic acid Dialysis— Dialysis block Donor phase 100 ml, acceptor phase 170 ml Membrane Cellulose acetate, MMCOa 15 kDa Sample injection 0.4 ml prior to starting dialysis and 0.6 ml at 0.18 ml min21 after starting dialysis Acceptor solution 10 mm phosphate buffer (pH 2) 20 mm phosphate buffer (pH 3) Acceptor solution flow rate/ml min21 1.0 0.5 Dialysis time for urine/min 10 9 Dialysis time for serum/min 19 9 SPE— Column 20 3 2.1 mm id, 5 mm C18 50 3 2 mm id, 5 mm PLRP-S Desorption buffer 27 mm phosphate buffer (pH 7)–ACN (30 + 70 v/v) water–THF (25 + 75 v/v) Desorption buffer flow rate/ml min21 0.2 0.4 Wash buffer 12.5 mm phosphate buffer (pH 4.5)–ACN (80 + 20 v/v) 10 mm phosphate buffer (pH 6)–ACN (90 + 10 v/v) Wash buffer flow rate/ml min21 1.0 Wash time/min 1.5 2.0 SPE–CE interfacing— Interface Laboratory-made interface CE buffer flow rate/ ml min21 0.2 Dtb 18 20 tinj c 18 15 CE— Capillary Bare silica, 50 mm id, 375 mm od, total length 106 cm, effective length 40 cm Capillary rinsing 2000 mbar for 2 min Injection pressure 280 mbar during tinj CE buffer 50 mm acetate buffer (pH 4.6)–MeOH (90 + 10 v/v) 20 mm phosphate buffer (pH 7.0) UV detection/nm 200 260 Analysis time/min 15 10 a Molecular mass cut-off.b Time between start of elution and analytes passing the tip of the CE capillary. c Total length of time in which the analytes pass the tip of the CE capillary. Fig. 2 On-line dialysis–SPE–CE of (B and D) blank and (A and C) spiked (10 mg ml21 level of each NSAID) urine. (A and B) no washing of the SPE cartridge after loading or (C and D) washing with 1.5 ml of 12.5 mm phosphate buffer (pH 4.5)–acetonitrile (80 + 20 v/v). Fig. 3 On-line dialysis–SPE–CE of (B and D) blank and (A and C) spiked (10 mg ml21 level of each NSAID) serum.(A and B) no washing of the SPE cartridge after loading or (C and D) washing with 1.5 ml of 12.5 mm phosphate buffer (pH 4.5)–acetonitrile (80 + 20, v/v). Analyst, 1999, 124, 115–118 117When the same washing procedure as for urine was applied, satisfactory results were obtained [Fig. 3(C) and (D)]. Finally, the optimized procedure for the determination of the NSAIDs was validated with serum and urine.The detection limits were 0.05–0.1 mg ml21 for urine and 0.1–1.0 mg ml21 for serum (Table 3). As indicated above, the higher detection limits in serum can be attributed to small losses caused by drug– protein binding and to a higher background. Calibration curves were constructed from the detection limits up to 10 mg ml21 (urine) or 100 mg ml21 (serum). They were linear (Table 3) and covered the therapeutic levels in serum (0.5–70 mg ml21) listed in Table 1.The within-day and between-day precisions (RSDs) were < 8% for urine and < 9% for serum (Table 3). This is significantly better than the value of 15% which is an often used acceptance level for quantification results for biological samples.10 Obviously, the present protocol can be used to optimize dialysis–SPE–CE procedures for the low-mg ml21 determination of different classes of drugs. Since the set-up is fully automated, analyses can be, and indeed were, run unattended.The CE capillary and SPE cartridge were used without any problems for over 500 analyses. The dialysis membrane was replaced every 250 analyses. Conclusions This study has shown that the optimization procedure used earlier for sulfonamides can be applied also for other classes of drugs; there are, of course, differences in the numerical values for several parameters, but no major changes occur. However, as we shall show in a subsequent paper, complicating effects do occur if unusual CE conditions are encountered.This was observed when cationic drugs were studied and a non-standard CE buffer had to be used.11 On-line dialysis–SPE–CE of NSAIDs with UV detection was performed in a fully automated way, being controlled by the SPE unit and dialysis module. The procedure is robust, and 250–500 analyses can be performed without any need for maintenance or exchange of parts. The analytical characteristics are satisfactory for both urine and serum and allow the use of dialysis–SPE–CE–UV for metabolic studies and clinical analyses. 10 References 1 A. Guttman and N. Cooke, J. Chromatogr. A, 1994, 685, 155. 2 S. Fanali and Z. Aturki, J. Chromatogr. A, 1995, 694, 297. 3 M. G. Danato, E. van den Eeckhout, W. van den Bossche and P. Sandra, J. Pharm. Biomed. Anal., 1993, 11, 197. 4 M. G. Danato, W. Baeyens, W. van den Bossche and P. Sandra, J. Pharm. Biomed. Anal., 1994, 12, 21. 5 H. Soini, M. V. Novotny and M.-L. Riekkola, J. Microcol. Sep., 1992, 4, 313. 6 C. W. Maboundou, G. Paintaud, M. Berard and P. R. Bechtel, J. Chromatogr. B, 1994, 657, 173. 7 Z. K. Shibabi and M. E. Hinsdale, J. Chromatogr. B, 1996, 683, 115. 8 D. C. Turnell and J. D. H. Cooper, J. Autom. Chem., 1985, 7, 177. 9 J. R. Veraart, J. van Hekezen, M. C. E. Groot, C. Gooijer, H. Lingeman, N. H. Velthorst and U. A. Th Brinkman, Electrophoresis, 1998, 19, 2944. 10 F. Lapique, P. Netter, B. Bannwarth, P. Trechot, P. Gillet, H. Lambert and R. J. Royer, J. Chromatogr., 1989, 496, 301. 11 J. R. Veraart, P. J. Lagas, C. Gooijer, H. Lingeman, N. H. Velthorst and U. A. Th. Brinkman, in preparation. Paper 8/07636G Table 3 Calibration data for the determination of NSAIDs using dialysis–SPE–CE. Concentration range: from detection limit to 10 mg ml21 for urine and from detection limit to 100 mg ml21 for serum Analyte Detection limit/ mg ml21 Within-day precision (%)a Between-day precision (%)a Interceptb Slopeb Correlation coefficient (r2) Urine (n = 7–8)— Ibuprofen 0.10 5.9 6.5 20.03 (0.01) 0.46 (0.002) 0.9997 Naproxen 0.05 1.5 4.0 0.06 (0.05) 1.47 (0.01) 0.9992 Fenoprofen 0.05 2.5 3.6 0.01 (0.02) 0.84 (0.004) 0.9997 Ketoprofen 0.10 2.2 5.1 0.00 (0.05) 0.69 (0.01) 0.9973 0.10 8.0 4.7 20.01 (0.02) 0.73 (0.004) 0.9997 Serum (n = 7–10)— Ibuprofen 1.00 3.5 6.3 20.08 (0.04) 0.10 (0.001) 0.9997 Naproxen 0.10 3.0 5.6 20.04 (0.11) 0.52 (0.003) 0.9997 Fenoprofen 0.50 2.8 8.5 20.13 (0.10) 0.20 (0.002) 0.9991 Ketoprofen 0.20 4.6 7.7 0.03 (0.11) 0.29 (0.003) 0.9992 Flurbiprofen 1.00 6.1 8.8 20.10 (0.08) 0.12 (0.002) 0.9988 a Determined using a concentration of five times the LOD. b Standard deviation in parentheses. 118 Analyst, 1999, 124, 115–118

ISSN:0003-2654    

DOI:10.1039/a807636g

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Determination of hexamethylene diisocyanate-based isocyanates in spray-painting operations Part 2.† Comparison of high performance liquid chromatography with capillary zone electrophoresis Walter E. Rudzinski,*a Jian Yin,a Ellen Englandb and Gary Carltonb a Department of Chemistry, Southwest Texas State University, San Marcos, TX 78666, USA b DET1 HSE/OEMI, 2402 E. Drive, Brooks AFB, San Antonio, TX 78235-5501, USA Received 11th August 1998, Accepted 25th November 1998 A capillary zone electrophoresis (CZE) approach was developed for the determination of hexamethylene diisocyanate (HDI) monomer and HDI-based oligomers. A comparison of CZE with high performance liquid chromatography (HPLC) indicates that the CZE separation completely isolates isocyanates from excess solvent, derivatizing reagent and pigment while offering a fivefold increase in sensitivity.The CZE approach allows for the quantification of HDI monomer and oligomer within a 1 min time window under the run conditions selected.For the determination of HDI-based oligomer, provided that the relative response with respect to HDI monomer is calculated, there is no significant difference (p < 0.05, n = 10) in the isocyanate air concentration when using either HPLC or CZE. The results are significant because they indicate that CZE has advantages for the determination of both HDI-based oligomer and HDI monomer generated during spray-painting operations. Introduction Isocyanates are a significant hazard in the workplace.1-7 Workers who are involved in the production of polyurethane foams, elastomers and fibers and in the application of polyurethane paints and coatings are all exposed to diisocyanate and polyisocyanates.These compounds irritate the nose, throat and lungs and may eventually lead to the development of bronchial asthma.2–6 A number of methods have been developed for the sampling and determination of atmospheric organic isocyanates.8–10 The Occupational Safety and Health Administration (OSHA) prescribes a protocol for sample collection on a filter followed by high performance liquid chromatography (HPLC).11 The National Institute for Occupational Safety and Health (NIOSH) promulgates Method 552112 for the determination of isocyanates in the USA. This method is an adaptation of MDHS 25, Method for Total Isocyanate in Air, developed for use in the UK.13 The last two methods use an impinger filled with 1-(2-methoxyphenyl)piperazine (MOP) in toluene for sample collection followed by HPLC analysis.In this work, we developed and validated a capillary zone electrophoresis (CZE) approach to the determination of total HDI-based isocyanates, since CZE has several advantages relative to HPLC: low solvent consumption, a relatively short analysis time and higher resolution.14 This work extends our previous efforts to find analytical approaches to the determination of isocyanates when collected on polyurethane foam (PUF)-based samplers.15 Experimental Reagents 1-(2-Methoxyphenyl)piperazine (MOP) was obtained from Fluka (Milwaukee, WI, USA), HDI monomer (HDI) from Eastman Kodak (Rochester, NY, USA), Desmodur N-100 (N- 100), which contains 99.3% polyisocyanate (predominately in the form of the biuret trimer of HDI), from Bayer Chemical (Pittsburgh, PA, USA),16 dimethyl sulfoxide from Aldrich (Milwaukee, WI, USA), acetonitrile (HPLC grade) and methanol from EM Science (Gibbstown, NJ, USA), acetic anhydride, glacial acetic acid, toluene and anhydrous sodium acetate from Fisher (Fairlawn, NJ, USA) and phosphoric acid and sodium phosphate from Baker (Phillipsburg, NJ, USA).Preparation of HDI–MOP and N-100–MOP standard solutions Isocyanate–MOP standards were prepared by directly reacting HDI-based isocyanates (which contain the NCO moiety) with MOP (which contains an accessible amine group) to form ureas (containing the –NCON– moiety), as described in NIOSH Method 5521.12 A stock standard solution containing 500 mm HDI–MOP was prepared by dissolving 27.6 mg of HDI–MOP in 100.0 ml of methanol.A stock standard solution of N- 100–MOP containing 1200 mmol of isocyanate groups per liter was prepared by dissolving 158.1 mg of N-100–MOP (equivalent mass of Desmodur N-100 = 191) in 100.0 ml of methanol. A 1000 mg l21 MOP stock standard solution was prepared by dissolving 0.100 g of MOP in 100 ml of methanol.In order to compare CZE with HPLC, five working standard solutions, containing final concentrations of 2.5–40 mm of HDI– MOP (5–80 mm NCO), 2.0–32 mmol of isocyanate groups per liter of N-100–MOP (6.0–96.0 mm NCO) and 10 mg l21 MOP, were prepared by mixing the appropriate amount of a 500 mm HDI–MOP stock standard solution and 1200 mmol of isocyanate group per liter N-100–MOP stock standard solution with 0.1 ml of a 1000 mg l21 MOP stock standard solution and then diluting to 10 ml with methanol. To half of each standard solution 10 ml of acetic anhydride were added.For the analysis of samples obtained during spray-painting operations, at least six working standard solutions were prepared. These HDI–MOP standard solutions contained 20 ml † For Part 1 see ref. 15. Analyst, 1999, 124, 119–123 119of acetic anhydride, 100 mg l21 of MOP and 0.25–16 mg l21 of HDI–MOP in methanol. In order to identify polyisocyanates generated during spraypainting operations, a 50 ml portion of the isocyanate hardener (bulk catalyst) was diluted in 10 ml of toluene and 50 ml of this solution were then diluted to 10 ml with 1000 mg l21 MOP in toluene.The solution was allowed to react for 12 h before heating to dryness under a gentle stream of nitrogen. A 30 ml aliquot of acetic anhydride was added to the dry residue, which was reconstituted in 5 ml of methanol while agitating the solution in an ultrasonic water-bath for 15 min. The preparation of methoxyphenylpiperazine derivatives was performed in a fume-hood in order to avoid exposure to isocyanate and solvent vapors.Isocyanates are known respiratory irritants. Preparation of sponge samplers The polyether-based polyurethane foam (PUF) sponge (No. 2405; Supelco, Bellefonte, PA, USA) was cut to the correct thickness using a razor blade. Sponges for use in the IOM sampler (SKC, Eighty Four, PA, USA) were 15 mm thick and sponges for use in the cassette sampler (Nuclepore, Pleasanton, CA, USA) were 17.5 mm thick. The sponges for both types of samplers were then punched from the slice using a No. 15 cork borer (23 mm id). The sponges were soaked in dimethyl sulfoxide for at least 2 h before extracting three times with 10 ml of acetonitrile, then air drying. Glass-fiber filters (25 mm) (Omega, Chelmsford, MA, USA) were impregnated with 50 mg l21 MOP in acetonitrile, then air dried. The cassette samplers were assembled by placing an aluminum mesh screen (2.5 cm in diameter) in the bottom of the cassette sampler, followed by the impregnated glass-fiber filter, then the clean sponge on top.The IOM samplers were assembled by placing the impregnated glass-fiber filter in the bottom, then positioning the sponge on top before screwing the cap back on. To each type of sampler, 2 ml of 1000 mg l21 MOP in dimethyl sulfoxide were added and the excess solution was removed by pumping on the sampler at a flow rate of 1.0 l min21.The DuPont (Wilmington, DE, USA) Alpha 1 constant-flow, air-sampling pump used to remove the excess solvent is capable of drawing up to 2.0 l min21 of air and was also used to obtain air samples during spray-painting (field) operations. Description of spray-painting operations Six different spray-painting operations (12 sampling events) were evaluated at Hill Air Force Base. Operations 41 and 42 involved painting the lower fuselage of an F-16 fighter aircraft. Operations 43 and 44 involved painting aircraft parts and under the wing of a C-130 cargo aircraft, respectively.Operation 45 involved fully painting an F-16 fighter aircraft. Operation 46 involved painting the landing gear and door area of the airplane. All spray guns were high volume, low pressure (HVLP). Airverter, Devilbiss, Binks and Sata spray guns were used for operations 41 and 42, 43, 44, 45 and 46, respectively. In all operations, the painters used cloth coveralls and gloves and a supplied air hood.In operations 41, 42, 44 and 45, a gray polyurethane paint formulation (Deft, Irvine, CA, USA) was used containing a 3:1 (polyenamel-to-hardener) ratio. In operations 43 and 46, sea foam green and white gloss polyurethane paint formulations (Deft) were used containing a 1:1 (polyenamel-to-hardener) ratio, respectively. Sampling strategy For all operations, sampling was performed by positioning a cassette sampler and an IOM sampler about 2.5 cm apart on a cart at approximately breathing zone height.For operations 41, 42 and 45, the cart was positioned about 1.8 m behind the fuselage. For operation 43, the cart was 0.6 m behind the parts. For operation 44, the cart was positioned under the aircraft wing. For operation 46, the cart was positioned 1.2 m to the right of the parts being painted. All cart positions were in the maximum overspray area. The sampling protocols described in NIOSH Method 5521 were used throughout.12 The isocyanate samples were collected at an air flow rate of 1.0 l min21.Sampling times ranged from 13 to 62 min. After collection, the sponges in the samplers were immediately removed from the cassette and IOM samplers and placed in a sampling bottle containing 10 ml of 50 mg l21 MOP in acetonitrile. The sampling bottles were shipped within 2 d of sampling. Preparation of samples When the samples arrived in the analytical laboratory, they were prepared for analysis within 1 d.The sponges were extracted three times with 10 ml of acetonitrile and the extracts were combined and then evaporated under a gentle stream of nitrogen to a known fixed volume (2–5 ml). Each sample was separated into two equivalent fractions; one was used unchanged (nonacetylated sample) whereas to the second an additional 30 ml of acetic anhydride were added (acetylated sample). Samplers which were not used for the collection of isocyanate were also treated in the same way as the actual samplers and served as method blanks.Instrumentation for analysis The extract from the sponges was analyzed using an HPLC system consisting of a Beckman (Fullerton, CA, USA) Model 110 B solvent delivery system, a Beckman Model 270A sample injector and a Beckman Model 160 ultraviolet–visible (UV) detector set at 254 nm and connected to a Hitachi (San Jose, CA, USA) Model D-2500 integrator. A BAS (Lafayette, IN, USA) electrochemical (EC) detector operated in the oxidative mode (+0.8 V versus Ag/AgCl) followed the UV detector.The EC cell was controlled by a BAS Model LC4B amperometric detector and the output was delivered to a Varian (Palo Alto, CA, USA) Model 4400 integrator. The column used was a Phenomenex (Torrance, CA, USA) Prodigy (100 3 4.60 mm id) C8 PEEK column (5 mm with 150 Å pores). The mobile phase varied between 30 + 70 and 40 + 60 acetonitrile–methanolic acetate buffer (0.6% sodium acetate in 50 + 50 methanol–water adjusted to pH 6.0 with glacial acetic acid).The mobile phase flow rate was set at 1.0 ml min21. The injection volume was 20 ml. A Waters (Milford, MA, USA) Quanta 4000 system equipped with a Hewlett-Packard (Avondale, PA, USA) 3390A integrator was used for the CZE analysis of samples. All samples were injected using hydrostatic injection for 10 s. The capillary column had an effective separation length of 40–50 cm and an id of 75 mm. The total column length was 48–58 cm.The operating voltage was set at 20 kV. The detector wavelength was set at 185 nm. The capillary was purged with 0.5 m KOH for 3 min, then water for 3 min, then running buffer for 5 min between each run. The running buffer varied between 0 + 100 and 50 + 50 acetonitrile–phosphate buffer [0.010 m anhydrous sodium phosphate (Na3PO4) in distilled water]. The pH was adjusted to 120 Analyst, 1999, 124, 119–1233.0 with concentrated phosphoric acid after the addition of acetonitrile.Identification and quantification For HPLC analysis, the peak produced by derivatized HDI monomer was identified by matching its retention time with that of HDI–MOP standards. The N-100 oligomer peaks were also identified by matching their chromatographic retention times with those of the N-100–MOP standards. The HDI-based oligomer peaks in field samples were identified by comparing them with those produced from the derivatized bulk catalyst used in the spray-painting operation.In addition, HDI-based oligomer peaks were confirmed by comparing their EC/UV response ratio with that obtained using HDI–MOP standards. For CZE analysis, the peak produced by derivatized HDI monomer was identified by matching its migration time with those of HDI–MOP standards. The N-100 oligomer peaks were also identified by matching their migration times with those of the N-100–MOP standards. HDI monomer in field samples was identified by the method of standard additions, a sample being run with and without added HDI–MOP standard.The HDIbased oligomer peaks in field samples were identified by comparing them with those produced from the derivatized bulk catalyst used in the spray-painting operation. For comparison of the HPLC and CZE methods, quantification of HDI monomer and HDI-based oligomers was based on calibration curves prepared with HDI–MOP and N-100–MOP standard solutions. For field studies, quantification of HDI monomer and HDI-based oligomers was based on calibration curves prepared from HDI–MOP standard solutions.All quantifications were based on the average of at least two runs. Results and discussion CZE method development Effect of added organic modifier. A mixture of 40 mm HDI– MOP, 32 mmol of isocyanate groups per liter N-100–MOP and excess MOP can easily be separated using CZE, but N- 100–MOP gives a very poor peak shape in an aqueous phosphate buffer (pH 3.0). By adding acetonitrile to the running buffer, the N-100–MOP peak becomes narrower with an improved peak profile. This is attributed to the fact that the acetonitrile helps to dissolve the N-100–MOP in the running buffer.The electropherograms which illustrate the effect of added acetonitrile on the CZE separation are shown in Fig. 1. The peak for N-100–MOP becomes sharper upon adding acetonitrile up to a concentration of 40 + 60 acetonitrile–phosphate buffer. At a concentration of 50 + 50, the HDI–MOP peak begins to broaden.Concurrently, as the acetonitrile concentration in the mobile phase varies from 0 to 50%, the UV response factor for HDI–MOP and N-100–MOP double and triple, respectively. The peak area ratio of N-100–MOP to HDI–MOP increases from 56 to 105% as the mobile phase composition increases from 0 + 100 to 40 + 60 acetonitrile–phosphate buffer (see Fig. 2). The peak area ratio for a solution containing an equimolar amount of HDI and N-100 should be 1.50 if N-100 is presumed to have 3 mol NCO per mole of N-100 and HDI has 2 mol NCO per mole of HDI, and all the absorbance can be attributed to the MOP derivatizing reagent.If the N(CNO)N moiety between the piperazine of the MOP reagent and the hexamethylene of HDI and N-100 also absorbs, then the total absorbance will increase but the absorbance ratio will remain 1.5. The migration time (tm) for both HDI–MOP and N-100–MOP also increases with increasing concentration of acetonitrile.This can be attributed to the addition of organic solvent, which probably decreases the degree of protonation of the MOP–isocyanate derivatives while not affecting the degree of protonation of the more basic free MOP. Effect of added acetic anhydride. In HPLC analysis, acetic anhydride is used to improve the analytical efficiency and to prevent excess derivatizing reagent (amine) from attaching to the silica substrate. The same problems as observed in HPLC also have been observed in CZE.If the excess of derivatizing reagent (MOP) is not converted into the amide (acetylated), then the excess MOP sticks to the capillary wall. In a standard solution containing acetic anhydride, the migration times of HDI–MOP and N-100–MOP (5.60 and 5.85 min, respectively) differ from those in the solution without acetic anhydride (6.03 and 6.33 min, respectively), but the excess MOP peak shifts from before to after the isocyanate– MOP peaks upon addition of acetic anhydride (tm = 4.13 min without added acetic anhydride and 6.47 min with added acetic anhydride).Fig. 1 Effect of acetonitrile on CZE analysis. Capillary, fused silica (50 cm 3 75 mm id); applied voltage, 20 kV; UV detector, 185 nm; running buffer, 0 + 100 to 50 + 50 acetonitrile–aqueous phosphate buffer: (a) 0 + 100; (b) 10 + 90; (c) 20 + 80; (d) 30 + 70; (e) 40 + 60; (f) 50 + 50. *, HDI monomer; **, HDI oligomer; ***, excess MOP. Fig. 2 Effect of acetonitrile on the CZE response of HDI and N-100.Capillary, fused silica (50 cm 3 75 mm id); applied voltage, 20 kV; UV detector, 185 nm; running buffer, 0 + 100 to 50 + 50 acetonitrile–aqueous phosphate buffer. Analyst, 1999, 124, 119–123 121The average relative standard deviation (RSD) was based on five standards run in triplicate. If the migration times are compared, the average RSD for the migration time for acetylated samples is 1.8% for HDI–MOP and 1.9% for N- 100–MOP. The average RSD for non-acetylated samples is 4.0% for HDI–MOP and 4.5% for N-100–MOP.The average RSD for the integrated peak area for acetylated samples is 8.1% for HDI–MOP and 14.2% for N-100–MOP. The average RSD for non-acetylated samples is 9.5% for HDI–MOP and 13.3% for N-100–MOP. Both acetylated and non-acetylated samples have a good linear relationship between the UV response and the concentration of NCO groups. The Pearson correlation coefficient (r) values for acetylated samples (r > 0.999 for HDI–MOP, r > 0.996 for N-100–MOP) are better than those obtained for nonacetylated samples (r > 0.993 for HDI–MOP, r > 0.991 for N- 100–MOP).The sensitivity (slope of the calibration curve) for acetylated samples (3051 per mm of NCO in HDI and 2746 per mm of NCO in N-100) also is better than that for non-acetylated samples (2526 per mm of NCO in HDI and 2411 per mM of NCO in N-100). Because of the more reproducible migration times and better sensitivity after treating standards with acetic anhydride, all further CZE analyses involved samples that had been acetylated.Comparison of HPLC with CZE Both calibration curves obtained for HDI–MOP when using HPLC and CZE are linear (r > 0.990). The HDI sensitivity for CZE (3051 per mm of NCO) is about five times larger than that for HPLC (584 per mm of NCO). The HDI detection limit for CZE (0.08 mg l21; 1.0 mm NCO) is about the same as that for HPLC (0.08 mg l21; 1.0 mm NCO). A comparison of calibration curves obtained for N-100–MOP when using HPLC and CZE shows a more linear relationship for CZE analysis (r > 0.996) than for HPLC analysis (r > 0.987).The N-100 sensitivity for CZE analysis (2746 per mm of NCO) is about nine times larger than that for HPLC (291 per mm of NCO). The N-100 detection limit for CZE (5.0 mm NCO) is not as low as that for HPLC (2.5 mm NCO). If the slope of the N-100–MOP calibration curve is compared with that of the HDI–MOP calibration curve from HPLC analysis, the slope ratio (N-100–MOP/HDI–MOP) is 50%, which is lower than the value of 90% obtained from CZE analysis.This has important implications when trying to compare polyisocyanate results obtained from HPLC and CZE when using HDI–MOP standards. The higher response of HDI oligomer when using HDI–MOP as a standard for CZE partially explains why CZE analysis always gives higher HDI oligomer concentrations than HPLC analysis for field samples (see below). Analysis of samples obtained during spray-painting operations Fig. 3 illustrates (a) HPLC and (b) CZE results for a bulk catalyst used in spray-painting. In the HPLC trace, three peaks (**) are obtained from the HDI oligomer, but no HDI monomer peak (*) because of a huge excess of MOP (***). In the CZ electropherogram, both HDI monomer (*) and two oligomer (**) peaks are separated from the excess MOP peak (***). There is usually a third peak in CZE which follows the two HDI oligomer peaks, but because of its small size, it is difficult to confirm and quantify as an isocyanate.A number of samples were acquired during spray-painting operations at Hill Air Force Base and subjected to both HPLC and CZE analyses. The samples from operation 42 were not included in Table 1 because of a short sampling time (13 min). Table 1 illustrates the results for HPLC and CZE analysis of HDI and HDI oligomer during spray-painting operations. In nine out of ten samples HDI oligomer data obtained from CZE analysis (column 3; 0.53–3.19 mg m23) are significantly higher than those obtained from HPLC analysis (column 2; 0.0–2.30 mg m23).The reason lies in a different N-100–MOP/HDI– MOP response ratio for CZE and HPLC. If the CZE data are adjusted by multiplying by a relative response factor of 50/90 (the response ratio of N-100–MOP/HDI–MOP from HPLC divided by the response ratio of N-100–MOP/HDI–MOP from CZE), the CZE results can be normalized for a direct comparison with the HPLC results.If a paired t-test is then used to compare HDI-based oligomer concentrations obtained using CZE and HPLC (see data pairs in columns 2 and 4), the difference between means is not significant at the 0.05 level. The results indicate that either HPLC or CZE may be used for the determination of HDI oligomer in the spray-painting environment. HDI monomer peaks were not detected using HPLC (Table 1, column 5), but HDI monomer was detected using CZE at 185 nm with a concentration in the range 0.003–0.033 mg m23.Comparing HPLC and CZE analysis, the CZE method offers better sensitivity for both HDI monomer and oligomer, shorter Fig. 3 Comparison of HPLC and CZE separations of a bulk catalyst. (a) HPLC. UV detector, 254 nm; column, Phenomenex (100 3 4.60 mm id) C8 PEEK (5 mm); mobile phase, 30 + 70 acetonitrile–methanolic acetate buffer. (b) CZE. UV detector, 185 nm; capillary, fused silica (50 cm 3 75 mm id); applied voltage, 20 kV; running buffer, 40 + 60 acetonitrile–phosphate buffer.*, HDI monomer; **, HDI oligomer; ***, excess MOP. Table 1 Comparison of HPLC and CZE determination of isocyanates during spray-painting operations at Hill AFB HDI oligomer/mg m23 HDI monomer/ mg m23 Sampler (volume)a HPLC CZE CZE (adj.)b HPLC CZE 41021-CAS (32 l) 1.55 1.86 1.03 < LODc 0.026 41101-IOM (30 l) 1.07 1.50 0.84 < LOD 0.013 43023-CAS (23 l) 0.46 1.87 1.04 < LOD 0.031 43103-IOM (24 l) 2.30 3.19 1.79 < LOD 0.012 44024-CASd (62 l) 0.41 0.67 0.37 < LOD 0.021 44104-IOMd (62 l) 0.82 1.80 1.00 < LOD 0.007 45025-CAS (62 l) 0.74 1.26 0.71 < LOD 0.033 45105-IOM (62 l) 0.83 1.28 0.72 < LOD 0.003 46026-CAS (31 l) 0.04 0.53 0.30 < LOD 0.023 46106-IOM (31 l) 1.49 0.69 0.39 < LOD 0.005 a Sampler (volume) refer to the laboratory sample number followed by the acronym for the sampler (CAS = cassette, IOM = personal sampler) followed by the volume of air sampled.b Normalized CZE results. c < LOD = below limit of detection. LOD = 0.08 mg of HDI per sample for HPLC analysis and 0.08 mg of HDI per sample for CZE analysis. d All samples used in CZE analysis contained acetic anhydride except for 44024-CAS and 44104-IOM. 122 Analyst, 1999, 124, 119–123analysis time, lower buffer consumption and smaller sample requirements. However, there are a few problems in the CZE method. First, the capillary used in CZE analysis is very easily ruined by some compounds (e.g., MOP) in samples so that the capillary has to be replaced every few weeks.Second, the migration time and the resolution of isocyanate peaks are very sensitive to the pH of the buffer so that it is very difficult to obtain reproducible migration times during routine analysis. This necessitates the use of the method of standard additions in order to be certain of the identity of HDI–MOP. These two disadvantages, however, are mitigated by the ability of CZE to determine both HDI monomer and HDI oligomer under the same optimum run conditions even in the presence of a huge excess of derivatizing reagent.Acknowledgements W.E.R. thanks the National Institute for Occupational Safety and Health, USA (Grant No. R01OH03295-01), for support of this work and also thanks Irene D. DeGraff of Supelco for providing samples of the PUF sponge (No. 2405). References 1 S. Silk and H. Hardy, Ann. Occup. Hyg., 1996, 27, 333. 2 L. Belin, U. Hjortsberg and U. Wass, Scand. J. Work Environ. Health, 1981, 7, 310. 3 J. L. Malo, G. Ouimet, A. Cartier, D. Levitz and R. Zeiss, J. Allergy Clin. Immunol., 1983, 72, 413. 4 J. Nielsen, C. Sango, G. Winroth, T. Hallberg and S. Skerfving, Scand. J. Work Environ. Health, 1985, 11, 51. 5 O. Vandenplas, A. Cartier, J. LeSage, Y. Cloutier, G. Perreault, L. C. Grammer, M. A. Shaughnessy and J.-L. Malo, J. Allergy Clin. Immunol., 1993, 91, 850. 6 G. Tornling, R. Alexandersson, G. Hedenstierna and N. Plato, Am. J. Ind. Med., 1990, 17, 299. 7 M. Janko, K. McCarthy, M. Fajer and J. van Raalte, Am. Ind. Hyg. Assoc. J., 1992, 53, 331. 8 C. J. Purnell and R. Walker, Analyst, 1985, 110, 893. 9 V. Dharmarajan, R. D. Lingg, K. S. Booth and D. Hackathorn, in Sampling and Calibration for Atmospheric Measurements, ed. J. K. Taylor, ASTM Special Technical Publication 957, American Society for Testing and Materials, Philadelphia, PA, 1988, pp. 190. 10 S. P. Levine, K. J. D. Hillig, V. Dharmarajan, M. W. Spence and M. D. Baker, Am. Ind. Hyg. Assoc. J., 1995, 56, 581. 11 Method 42: Diisocyanates, in OSHA Methods Manual, Occupational Safety and Health Administration, Salt Lake City, UT, USA, 1994, pp. 42-1–42-39. 12 Method No. 5521: Isocyanates, Monomeric, Issue 2, in NIOSH Manual of Analytical Methods, ed. P. M. Eller, DHHS (NIOSH) Publication No. 94-113, vol. 2, National Institute for Occupational Safety and Health, Cincinnati, OH, 4th edn., 1994. 13 MDHS 25: Organic Isocyanates in Air, in Methods for the Determination of Hazardous Substances, Health and Safety Executive, Occupational Safety and Hygiene Laboratory, London, 1987. 14 W. E. Rudzinski, L. Pin, R. Sutcliffe, A. Richardson and T. Thomas, Anal. Chem., 1994, 66, 1664. 15 W. E. Rudzinski, J. Yin, S. H. Norman and D. A. Glaska, Analyst, 1998, 123, 2079 16 Desmodur N-100 Material Safety Data Sheet, Bayer, Pittsburgh, PA, 1996. Paper 8/06351F Analyst, 1999, 124, 119–123 123

ISSN:0003-2654    

DOI:10.1039/a806351f

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Determination of thyreostatics in animal feed by micellar electrokinetic chromatography Josep Esteve-Romero,a Ima Escrig-Tena,a Ernesto F. Simó-Alfonso,b and Guillermo Ramis-Ramos*b a Area de Química Analítica, ESTCE, Universitat Jaume I, 12060 Castelló, Spain b Departament de Química Analítica, Facultat de Química, Universitat de València, 46100 Burjassot, Spain Received 20th October 1998, Accepted 17th December 1998 The determination of the thyreostatics 2-thiouracil, its derivatives (4-methyl-2-thiouracil, 4-propyl-2-thiouracil and 4-phenyl-2-thiouracil) and methimazole in manufactured dried animal feed by micellar electrokinetic chromatography (MEKC) is described.A 99 ± 5% extraction yield at the 20 mg g21 level (n = 8) was achieved by shaking the milled fodder with methanol–1 m NaOH (80 + 20). Aliquots of the supernatant were injected in a 75 mm 3 33.5 cm uncoated silica capillary using pressure; separation was performed at 23 °C with 15 kV (positive polarity) in a background electrolyte (BGE) containing 40 mm sodium dihydrogenphosphate, 50 mm sodium dodecyl sulfate and 15 mm Tween 20 at pH 9.When the surfactants were added to the BGE, all the thyreostatics were well resolved and the fodder extracts showed lower backgrounds. The peaks appeared within the 2.25–5.2 min range with efficiencies in the 2.5 3 104–8 3 104 range; methimazole appeared in the vicinity of the electroosmotic migration time. Calibration curves were linear within the studied range (20–200 mg ml21, r2 > 0.998).Limits of detection in the extracts of spiked fodder samples ranged from 0.25 to 0.4 mg ml21, which corresponded to 0.6–1.0 mg of drug per gram of fodder. Peak area repeatabilities were about 4% at the 20 mg ml21 level. 2-Thiouracil, its derivatives and methimazole are thyreostatic drugs used in the treatment of hyperthyroidism1, and also to increase the growth rate in animals for human consumption with harmful consequences for human health.These drugs have been determined in pharmaceuticals by titrimetry,2,3 spectrophotometry, 4–7 nuclear magnetic resonance,8 electroanalytical methods9 –11 and high-performance liquid chromatography (HPLC),12 and in biological samples including flour, fluids and tissues by a spectrophotometric kinetic procedure,7 thin-layer chromatography,13,14 gas chromatography,15,16 HPLC12,17 and radioimmunoassay.18 Capillary zone electrophoresis has been used to determine thyreostatics in human, bovine and horse urine.19 Using 100 mm id capillaries, a background electrolyte (BGE) containing 20 mm N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid adjusted with NaOH at pH 7.5 and spectrophotometric detection at 278 nm, the limits of detection (LODs) were 0.5 and 0.1 mg ml21 for the direct injection of urine and for the injection of ethyl acetate extracts, respectively.19 These drugs are customarily screened by the inspection services of the sanitary authorities in the thyroid gland of killed animals. Thyreostatics are illegally added to fodders in the final weeks of animal growth to increase meat production.For this reason, there is a need to extend the inspection to animal feed. Dried fodders which are manufactured as small bricks or pellets contain wheat, barley, maize, soybean, herring, phosphate, limestone and other ingredients.20 As far as we know, suitable analytical methods to determine the drugs in these matrices have not been described.The direct injection of methanol extracts into a liquid chromatograph gives rise to an excessively noisy background, and clean-up procedures addressing this particular problem were not found in the literature. The aim of this work was to develop a simple and rapid procedure for the screening, identification and determination of thyreostatics in manufactured dried fodder. This was accomplished by extracting the analytes with a methanol–aqueous NaOH mixture followed by micellar electrokinetic chromatography (MEKC).Experimental Apparatus Spectrophotometric studies were carried out with a Lambda 19 spectrophotometer (Perkin-Elmer, Norwalk, CT, USA). A shaker (Vibromatic 384, Selecta, Barcelona, Spain) was used to prepare fodder extracts. Separations were performed with a Hewlett-Packard 3D capillary electrophoresis system, including a built-in diode-array detector and a ChemStation (DOS Series) with HPCHEM 3D software; 75 mm id 3 33.5 cm total length (25 cm effective length) uncoated fused-silica capillaries were used.All buffer and sample solutions were filtered by pressing with a syringe through 0.45 mm nylon filters (Micron Separations, Westboro, MA, USA). Reagents, solutions and samples Methanol (Scharlau, Barcelona, Spain), dimethylformamide (DMF), sodium hydroxide, sodium dihydrogenphosphate and potassium hydrogenphthalate (Panreac, Barcelona, Spain) were used. 2-Thiouracil (4-hydroxy-2-thiopyrimidine), its derivatives 4-methyl-2-thiouracil (4-hydroxy-6-methyl-2-thiopyrimidine), 4-propyl-2-thiouracil (2-thio-4-hydroxy-6-n-propylpyrimidine) and 4-phenyl-2-thiouracil (4-phenoxy-2- thiopyrimidine), and methimazole (1-methylimidazole-2-thiol) were obtained from Sigma (St.Louis, MO, USA). Sodium dodecyl sulfate (SDS) (Merck, Darmstadt, Germany) and Analyst, 1999, 124, 125–128 125Tween 20 (Sigma) were used. De-ionized water (Barnstead deionizer, Sybron, Boston, MA, USA) was used throughout.Samples of manufactured dried fodder of several types were used. Extractions from spiked samples were performed with methanol–1 m NaOH (80 + 20 v/v); 1 mm stock standard solutions of the thyreostatics were prepared in this medium and refreshed weekly. The absorption spectra of these solutions showed no changes at least in 1 week. A 20 mg ml21 stock standard solution of potassium hydrogenphthalate (used as an internal standard) was prepared in water.Procedures New capillaries were flushed with 1 m NaOH at 60 °C for 3 min, followed by the BGE for a further 10 min; the BGE contained 40 mm sodium dihydrogenphosphate adjusted up to pH 9.0 with 1 m NaOH, 50 mm SDS and 15 mm Tween 20. Between runs, the capillary was flushed only with the BGE for 10 min. About 50 g of fodder were milled, and about 10 g were weighed, shaken for 10 min with 25 ml of methanol–1 m NaOH (80 + 20) and centrifuged; a 1 ml aliquot of supernatant was taken (0.1 ml of internal standard solution can be added) and introduced in the injection vial with a syringe by pressing through a filter; 50 mbar for 2 s was used for the injection in the capillary, separation was performed at 23 °C with 15 kV positive polarity and the signal was monitored at 254 nm (10 nm bandwidth) with 400 nm as reference wavelength (20 nm bandwidth).To identify a peak, an absorption spectrum scan within the 200–350 nm region in the vicinity of the maximum was obtained with the diode-array detector, and compared with the spectrum of the standard.Also, another electropherogram of the same extract with standard added was obtained. Results and discussion Preliminary studies The solubility of the thyreostatics was low in water, methanol, ethanol and acetone and high in 0.2 and 1 m NaOH and in methanol–1 m NaOH mixtures. The spectra of the drugs were obtained in 40 mm phosphate buffer at pH 3, 6 and 9. 2-Thiouracil and its derivatives showed a spectral change, indicating that the respective anionic forms predominate at pH 9. This indicated the possibility of using a BGE of pH 9 and positive polarity to separate 2-thiouracil and its derivatives on the basis of different electrophoretic mobilities as anions. A high pH also has the advantages of a higher solubility of the drugs and a shorter separation time owing to the larger electroosmotic flow. On the other hand, spectral changes were not observed between pH 3 and 11 with methimazole, suggesting that the uncharged form was maintained.At pH 9, 2-thiouracil and its derivatives showed an absorption maximum at about 266 nm, but the maximum was located at 254 nm for methimazole. At the wavelength used for detection, 254 nm, the sensitivity for 2-thiouracil and its derivatives was about 80% of their respective maximum values. Optimization of separation conditions Using a BGE containing 40 mm sodium dihydrogenphosphate adjusted to pH 9 with 0.1 m NaOH, and injecting a mixture of standards, methimazole appeared in the vicinity of the electroosmotic migration time (2.3 min, 2 3 104 plates), and 2-thiouracil and its derivatives gave peaks between 2.5 and 6 min, with efficiencies in the 2 3 104–6 3 104 plate range.The migration order was methimazole > 4-methyl-2-thiouracil > 4-propyl-2-thiouracil > 2-thiouracil Å 4-phenyl-2-thiouracil. This order indicated that separation was not exclusively due to the differences in charge/mass ratios.Since the peaks of 2-thiouracil and 4-phenyl-2-thiouracil overlapped, surfactants were added to the BGE to modify the selectivity. In the presence of 50 mm SDS, the positions of the peaks were modified, but 4-methyl-2-thiouracil and 4-propyl-2-thiouracil overlapped. As shown in Fig. 1, in the presence of 50 mm SDS and 15 mm Tween 20, all the drugs gave well resolved peaks within the 2.5–5.2 min range, with efficiencies in the 2.5 3 104–8 3 104 plate range.Methimazole gave a peak at 2.25 min, with an efficiency of 4.2 3 104 plates, whereas the electroosmotic migration time was 2.1 min (established with DMF). To increase the difference between the electroosmotic and methimazole migration times, several modifications of the BGE composition (e.g., addition of 60 mm cholic acid) were tried without success. An advantage of the use of surfactants in the BGE was the reduction of noise in the electropherograms of the fodder extracts, i.e., a much lower background immediately after the methimazole peak was obtained. Application to real samples and figures of merit To evaluate the extraction recovery, 1 ml of standard solution containing 1 mg of drug was shaken with 1 g of fodder.The mixture was extracted with 4 ml of the methanol–1 m NaOH solution, centrifuged and the supernatant was injected; on the other hand, 1 ml of the same standard solution was diluted to 5 ml, injected and the peak area was used as reference.With this procedure, the extraction yield at the 1 mg g21 level was about 100% for all the drugs and 12 fodder samples. Similarly, the extraction yield was established at the 20 mg g21 level. For this purpose, a series of 200 ml aliquots of a 100 mg ml21 solution of 2-thiouracil was added to eight 1 g fodder samples. After mixing, the samples were extracted. Another series of 200 ml aliquots of the same 2-thiouracil solution was added to eight extracts of the same fodder.By dividing the average values of the peak areas, an extraction yield of 99 ± 5% was achieved. Fig. 1 Electropherogram of a mixture of standards obtained using the recommended BGE (40 mm phosphate, 50 mm SDS, 15 mm Tween 20, pH 9): A, methimazole (7 mg ml21); B, 2-thiouracil (10 mg ml21); C, 4-methyl- 2-thiouracil (7 mg ml21); D, 4-propyl-2-thiouracil (7 mg ml21); E, 4-phenyl- 2-thiouracil (7 mg ml21); and IS, potassium hydrogenphthalate (20 mg ml21). 126 Analyst, 1999, 124, 125–128For each drug, the same migration time was obtained when either a standard solution or an extract of a spiked fodder was analysed using the recommended BGE. The electropherograms of the same fodder before and after being spiked with the drugs are shown in Fig. 2, parts 1 and 2. The baseline noise increased slightly when fodder extracts instead of standard solutions were injected (the fodder used to obtain electropherograms 1 and 2 in Fig. 2 was the noisiest case). The standards used in calibration studies were prepared as follows: standard solutions with increasing concentrations of the drugs were prepared in methanol–1 m NaOH (80 + 20) and 1 ml aliquots of these solutions were mixed with 4 ml of a fodder extract (also prepared with the same methanol–NaOH mixture). All the thyreostatics gave linear calibration curves within the studied range (0, 20, 40, 60, 100 and 200 mg ml21, r2 > 0.998 in all cases).The intercepts were not significantly different from zero. Relative sensitivities (with respect to the 2-thiouracil peak) and repeatabilities are given in Table 1. The possible use of potassium hydrogenphthalate as internal standard was studied; however, the same repeatabilities were found with and without applying the internal standard correction to the peak areas. Finally, the same sensitivities were obtained by the external calibration and standard addition methods, indicating that the procedure had no matrix effect.Two procedures were used to estimate the LODs. First, a series of 200 ml aliquots of a solution containing 100 mg ml21 of each thyreostatic was added to eight 1 g fodder samples. After mixing, the samples were extracted and injected (the extracts contained 20 mg ml21 of each drug). The standard deviations of the peak areas, sa, were calculated; for each drug, the LOD was obtained as 3sa divided by the slope of the respective calibration curve obtained from peak areas.The LODs obtained in this way ranged from 0.1 to 0.2 mg ml21 in the fodder extracts. The LODs were also calculated as three times the standard deviation of the baseline, sh, divided by the slope of the calibration curves obtained from peak heights; sh was calculated from the baseline peak-to-peak width, which was taken as 5sh.21 This procedure gave LODs which ranged from 0.25 to 0.4 mg ml21, which corresponded to 0.6 to 1.0 mg of drug per gram of dry sample.Twelve samples collected by the inspection services of the Conselleria de Agricultura of the Generalitat Valenciana (Comunidad Valenciana, Spain) in various cattle and pig growing farms were analysed using the recommended procedure. Methimazole (20–120 mg g21) was found in four samples, and another sample contained 2-thiouracil (1 mg g21). Conclusions A simple procedure for the determination of thyreostatics in manufactured dried animal feed by MEKC has been developed.Using methanol–1 m NaOH (80 + 20), about a 100% extraction recovery is achieved at concentration levels as low as 20 mg g21. The background noise is largely reduced by adding SDS to the BGE, and the use of non-ionic surfactants to modify the selectivity allows different thyreostatic pairs to be separated. Sample preparation and manual extraction of a single sample can be performed in less than 10 min, separation takes about 6 min and re-equilibration of the capillary is completed in a further 10 min.Less than 1 mg of drug per gram of fodder can be detected. This work was supported by the DGICYT of Spain, Project PB97-1384. I. Escrig-Tena thanks the Universitat Jaume I for a grant. Thanks are also due to Dr. Adolfo Sáez of the Laboratorio Agrario de la Conselleria d’Agricultura de la Generalitat Valenciana (Spain) for the drugs, samples and useful comments. References 1 A. Burger, in Burger’ Medicinal Chemistry, 3rd edn., Wiley, New York, 1970, part II, p. 853. 2 M. G. El-Bardicy, Y. S. El-Saharty and M. S. Tawakkol, Talanta, 1993, 40, 577. 3 I. López García, P. Viñas and J. A. Martínez Gil, Fresenius’ J. Anal. Chem., 1993, 345, 723. 4 M. G. El-Bardicy, Y. S. El-Saharty and M. S. Tawakkol, Spectrosc. Lett., 1991, 24, 1079. 5 C. S. P. Sastry, P. Satyanarayana, N. R. P. Singh and A. R. M. Rao, J. Inst. Chem., 1988, 60, 162. 6 S. M. Sultan, J. Pharm. Biomed. Anal., 1992, 10, 1059. 7 M. S. García, M. I.Albero, C. Sánchez Pedreño and L. Tobal, Analyst, 1995, 120, 129. 8 H. Y. Aboul-Enein, J. Pharm. Pharmacol, 1979, 31, 196. 9 A. Berka, K. Velasevic and K. Nikolic, Pharmazie, 1989, 44, 499. Fig. 2 Part 1, electropherogram of an extract of a fodder spiked with 7 mg ml21 of each drug: A, methimazole; B, 2-thiouracil; C, 4-methyl- 2-thiouracil; D, 4-propyl-2-thiouracil; and E, 4-phenyl-2-thiouracil. Parts 2 and 3, electropherograms of extracts of the same fodder as in Part 1 and a fodder of a different type, without drugs added.The recommended BGE was used. Table 1 Figures of merit for spiked animal feed samples using the recommended BGE Analyte Sensitivitya RSD (%)b tm (min)c 2-Thiouracil 1.0 3; 7 5.12 ± 0.08 4-Methyl-2-thiouracil 1.4 4; 9 3.10 ± 0.02 4-Propyl-2-thiouracil 2.0 4; 6 2.96 ± 0.04 4-Phenyl-2-thiouracil 2.8 4; 8 3.90 ± 0.06 Methimazole 1.8 4; 6 2.25 ± 0.03 a In mg 2-thiouracil mg21 analyte. b Relative standard deviation at two concentration levels (n = 8): 20 and 200 mg ml21.c Migration times and standard deviations calculated with eight different animal feed samples. Analyst, 1999, 124, 125–128 12710 K. Nikolic and K. Velasevic, Pharmazie, 1987, 42, 689. 11 M. B. Thomas and T. A. Last, Anal. Chem., 1988, 60, 2158. 12 G. R. Cannell, J. P. Williams, A. S. Yap and R. H. Mortimer, J. Chromatogr. A, 1991, 564, 310. 13 H. F. De Brabander, P. Batjoens and J. Van Hoff, J. Planar Chromatogr. Mod. TLC, 1992, 5, 124. 14 Th. Reuvers De Lange, Institute of Health Carlos III, Centro Nacional de Alimentación, Ministerio de Sanidad y Consumo, Madrid (Spain), Method 7.1. 15 D. G. Watson, C. D. Bates, G. G. Skellern, R. Mairs and S. Martin, Rapid Commun. Mass Spectrom., 1991, 5, 141. 16 R. Schilt, J. M. Weseman, H. Hooijerink, H. J. Korbee and W. A. Traag, J. Chromatogr., Biomed. Appl., 1989, 81, 127. 17 Th. Reuvers De Lange, R. Díaz Díaz, J. Gómez Herrero, Institute of Health Carlos III, Centro Nacional de Alimentación, Ministerio de Sanidad y Consumo, Madrid (Spain), Method 7.2. 18 R. Halpern, D. S. Cooper, J. D. Kieffer, V. Saxe, H. Mover, F. Mallof and E. C. Ridgway, Endocrinology, 1983, 113, 915. 19 L. Krivánková, S. Krásensky and P. Boöcek, Electrophoresis, 1996, 17, 1959. 20 Veevoeders en Dieren Specialiteiten, http://users.skynet.be/vds/gene. htm. 21 M. J. Adams, Chemometrics in Analytical Spectroscopy, Royal Society of Chemistry, Cambridge, 1995, pp. 32–34. Paper 8/08152B 128 Analyst, 1999, 124, 125–128

ISSN:0003-2654    

DOI:10.1039/a808152b

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Optical biosensing of nitric oxide using the metalloprotein cytochrome cA David J. Blyth,a Jonathan W. Aylott,a James W. B. Moir,†b David J. Richardsonb and David A. Russell*a a School of Chemical Sciences, University of East Anglia, Norwich, UK NR4 7JT. E-mail: d.russell@uea.ac.uk b School of Biological Sciences, University of East Anglia, Norwich, UK NR4 7JT Received 4th September 1998, Accepted 15th December 1998 The metalloprotein cytochrome cA was extracted and purified from the bacterium Paracoccus denitrificans in order to develop a specific biosensing system for nitric oxide (NO).The metalloprotein was encapsulated in a porous silicate sol–gel glass to enable spectroscopic changes in the haem centre as a function of NO ligation to be quantified using absorption measurements. Spectroscopic evidence suggested that, between 2 and 4 d after encapsulation, the cytochrome cA protein changed conformation in the locality of the haem moiety, possibly from a five to a six coordinate haem centre. Such conformational changes were also observed when the cytochrome cA was stored in solution, although over a 2–3 month period.The conformational changes occurring in the protein altered the spectral characteristics of the reduced, oxidised and nitrosyl complex of the cytochrome cA and appear to change the binding affinity of the protein towards NO. However, the encapsulated (reconformed) cytochrome cA was shown to retain its selectivity towards NO with good reproducibility (seven consecutive measurements of NO produced an intensity value with a relative standard deviation of 0.28%).An NO calibration curve, using the in situ release of NO from the donor diethylamine NONOate, was obtained for the encapsulated cytochrome cA with an approximate working range of 10–400 mmol l21. Introduction Nitric oxide (NO), described as the ‘molecule of the year 1992’,1 has attracted a great deal of scientific attention over the past decade since it has been established that NO has a number of significant roles in physiology, microbiology and atmospheric chemistry.For example, NO has been shown to behave as an intracellular messenger molecule acting as the transduction mechanism for physiological processes such as vasodilation and neurotransmission in the brain.2 Elevated physiological NO levels have been shown to be indicative of human organ transplant failure.3 NO is an intermediary in the nitrogen cycle, although its role in the bacterial denitrification process is not yet fully understood.4,5 Additionally, the presence of NO and NO2, known collectively as NOx, in photochemical smogs6,7 poses a hazard to health.With such diverse interest in NO, there is a consequent demand for a specific technique for the quantitative measurement of NO in both aqueous and gaseous media. Several methods for nitric oxide detection exist, including GC–MS,8 spectrophotometric detection using haemoglobin9 and via a nitrite azo-coupling reaction.10 However, more recently developed techniques such as chemiluminescence and electrochemically based sensors show more promise in the field of NO detection.Chemiluminescent methods for the detection of nitric oxide utilise the reaction of NO with ozone11 and the NO–luminol–H2O2 12 chemiluminescence mechanism. The reported limit of detection of the ozone based system is 10213 mol l21. However, the method suffers the disadvantage of requiring a large sample size (about 1 l).The luminol system has been incorporated into an optical fibre-based sensor13 with a reported response time of 8–17 s and limit of detection of 1.3 mmol l21; however, the system is subject to systematic errors induced by compounds such as dopamine, uric acid, ascorbic acid and cysteine under ambient oxygen conditions. A number of NO sensing systems have been developed using electrochemistry. For example, Malinksi and Taha14 described the deposition of a polymeric nickel porphyrin on carbon fibres.The metalloporphyrin acts as a catalyst for the electrochemical oxidation of small molecules, including NO, allowing amperometric detection of NO at nanomolar levels. Friedemann et al.15 described an o-phenylenediamine-modified carbon fibre electrode for the detection of NO to a limit of 35 ± 7 nmol l21. Additionally, there exists a commercially available NO microsensor (World Precision Instruments) exhibiting nanomolar detection and fast response using 30 mm diameter disposable sensing tips.In order to achieve optimum selectivity for the detection of NO, we and other researchers have investigated the use of metalloproteins.16–19 The aim of the research reported here was to investigate the potential of using one such metalloprotein, cytochrome cA, as the recognition element in an optical absorption based biosensor for NO. Cytochrome cA is a dimeric metalloprotein found in a large number of photosynthetic and denitrifying bacteria.Each monomer contains, as the active site, a five-coordinate protoporphyrin IX (haem) prosthetic group. The haem iron, postulated to exist in a ‘quantum mechanically admixed’ (S = 5/2, 3/2) spin-state,20 possesses a nearby histidine residue as an axial ligand yet the sixth coordination site remains empty.21 The haem group is situated in a hydrophobic pocket in the protein structure such that only nitric oxide and carbon monoxide may bind to the reduced (Fe2+) haem iron, and only nitric oxide may bind to the oxidised (Fe3+) haem iron22,23.(Suzuki et al.24 reported the binding of CN2 to reduced and oxidised cytochrome cA from some bacteria, although with low affinity). Consequently, cytochrome cA may represent ideal biosensing † Present address: Department of Molecular Biology and Biotechnology, University of Sheffield, Sheffield, UK SN10 2TT. Analyst, 1999, 124, 129–134 129selectivity, with potentially no interferents, for NO when in the native oxidised state.Barker et al.19 recently demonstrated the use of cytochrome cA in a fluorescence based biosensor for NO, using both entrapment within a polyacrylamide matrix and the direct adsorption of the protein to a colloidal gold substrate as immobilisation strategies. However, nitric oxide binding to polyacrylamide entrapped cytochrome cA was found to be nonreversible, with the authors speculating that the protein was modified in the polymer matrix.Additionally, the direct attachment of a protein to a gold surface may not be the ideal method for the immobilisation of the recognition element of the biosensing system, as it has been shown that proteins denature when adsorbed directly on gold surfaces,25,26 an effect which we have also recently observed with the protein concanavalin A.27 Barker et al.19 also found that the cytochrome cA protein was bacterially destroyed when stored in solution for up to 2 weeks, which meant that their sensors were used within 1 week of fabrication.It is clear that for the development of biosensors, using proteins or enzymes as the key recognition element, it is imperative that the bioactive component is immobilised in a manner such that its primary function is retained and that the biomolecule is not denatured. The immobilisation process often takes the form of attaching the molecule to, or entrapping the molecule within, a porous matrix such that analyte molecules may permeate and interact with the recognition element. Methods of entrapment may include, for example, the use of polyacrylamide gels,28 Teflon membranes29 or porous silica30 –32 (from the sol–gel process). Our research efforts in optical biosensing have recently been focused on the sol–gel process.We have used methods similar to the two-stage sol–gel method developed by Ellerby et al.,33 who were able to show that the sol–gel encapsulation of metalloproteins can be achieved with no loss of the characteristic spectral changes associated with ligand binding.Using a variety of proteins, we subsequently developed sol–gel biosensing systems for dissolved NO and CO,17 nitrate34 and calcium35 ions and gaseous NO.18 The success of bioencapsulation via the sol–gel technique stems from the stability of the biological recognition molecule in the silicate matrix. Indeed, reports suggest that sol– gel immobilisation may even stabilise biological materials against denaturation.36,37 In this paper, we report the application of the sol–gel process to the immobilisation of cytochrome cA in order to couple high analyte selectivity with the optical transparency and relatively inert chemistry of the silica sol–gel for the development of an optical biosensing system for NO.Experimental Reagents and instrumentation Diethylamine NONOate was obtained from Cayman Chemical (Ann Arbor, MI, USA).All other reagents were obtained from Aldrich (Poole, Dorset, UK) and used as received. Aqueous solutions were prepared using doubly distilled, de-ionised water. UV/VIS absorption spectra were recorded on a Hewlett- Packard (Palo Alto, CA, USA) Model 8452A diode array spectrometer. Extraction and purification of cytochrome cA Cytochrome cA was isolated from the bacterium Paracoccus denitrificans (formerly known as Thiosphaera pantotropha38) strain LMD 82.5.Cells were grown in anaerobic batch culture (180 l) at 37 °C in a growth medium containing acetate as the carbon source and nitrate as electron acceptor. When the optical absorbance at 650 nm had reached a value of 1, the cells were harvested by cross-flow ultrafiltration (Sartorius, Epsom, Surrey, UK) to produce a concentrated cell suspension (6 l), which was then centrifuged at 5500 rpm for 20 min to produce a pellet. The cell periplasm was obtained by addition of lysozyme.Cytochrome cA was isolated from the lysate using ion-exchange and size-exclusion column chromatography. More detailed descriptions of the extraction of proteins, and their subsequent purification, from this bacterium have been reported by Aylott et al.34 and Moir et al.39 Using these methods, approximately 50 mg of cytochrome cA were extracted and purified from the Paracoccus denitrificans bacterium. The purified extract was shown to possess one haem protein using a haem stained sodium dodecyl sulfate polyacrylamide gel electrophoresis gel.The spectroscopic purity of the cytochrome cA, using an A410/A280 nm purity ratio, was determined to be 1.15. Sol–gel manufacture Bulk sol–gels were prepared using a method similar to that reported previously17 by the acid-catalysed sonication of a water–tetramethyl orthosilicate (TMOS) precursor solution followed by the addition of a buffered protein solution. The final water to TMOS mole ratio (r value) used was 33 : 1.Sol–gels were formed on one optical face of a 1 cm pathlength polystyrene cuvette to a thickness of approximately 1.5 mm and a height of 2 cm. Gelation took approximately 15 min and the sol–gels were subsequently stored in 50 mm phosphate buffer (pH 7) at 4 °C. After 24 h, the gels were washed with buffer solution to remove any alcohol by-product from the sol–gel process. Oxidation state UV/VIS absorption spectra were obtained for cytochrome cA in its native oxidation (Fe3+) state, both in buffered solution and in sol–gel.The addition of an excess (about 1mg) of sodium dithionite permitted the acquisition of the spectrum of the protein in the reduced (Fe2+) oxidation state. Binding of NO to cytochrome cA The nitrosyl complex of cytochrome cA was formed by bubbling the gas through a septum-sealed cuvette containing a solution of the oxidised metalloprotein that had been previously deoxygenated by bubbling with argon to prevent the oxidation of NO to NO22.The formation of the nitrosyl derivative was determined by monitoring changes in the UV/VIS absorption spectrum. Nitric oxide titration of sol–gel immobilised cytochrome cA using controlled release from diethylamine NONOate In order to obtain quantitative NO titration data for the cytochrome cA in both solution and encapsulated forms, a novel calibration method involving in situ release of NO from a donor molecule was developed. Diethylamine NONOate is known40 to release 1.5 equiv.of nitric oxide upon dissociation of the molecule. At pH 5 the NONOate has a half-life of approximately 9 s; therefore, 98% of potential NO release is achieved within 1 min. The half-life reduces with pH but a pH lower than 5 was avoided to reduce the risk of denaturation of the protein samples. Solutions of diethylamine NONOate were prepared in an oxygen free ( < 5 ppmv) glove-box at room temperature. Approximately 1 mg of NONOate was dissolved in 3 ml of sodium hydroxide solution (10 mmol l21) previously deoxygen- 130 Analyst, 1999, 124, 129–134ated by prolonged bubbling with argon.The exact concentration of NONOate present in the solution was determined by measuring the intensity of absorption of the chromophore at 250 nm (e250 = 6.5 l mmol21 cm21). Solutions were kept frozen and thawed prior to use. NONOate solution (equivalent to an NO concentration of 4.4 mmol l21) was introduced by a microlitre syringe into a cuvette containing a cytochrome cA encapsulated sol–gel [containing 7 nmol of protein, calculated assuming eSoret Å 90 l mmol21 cm21 (ref. 23)] surrounded by 2 ml of potassium phosphate solution (pH 5, 50 mmol l21). The intensity of absorption at 418 nm (Soret band of nitrosyl-cytochrome cA) was monitored with time (typically 10–15 min) until a maximum intensity was obtained. Calibration points were then plotted as a function of NONOate added. Reproducibility of NO–cytochrome cA titrations Consecutive titrations of 20 ml NONOate aliquots were performed, as described above, on the same cytochrome cA immobilised sol–gel.After each titration the immobilised protein was allowed to revert fully to the native Fe3+ cytochrome cA by bubbling argon through the surrounding solution (taking typically 1 h), before the next NONOate addition. The titration process was repeated seven times in order to determine the reproducibility of the measurement signal. Interferences Although in solution cytochrome cA is documented to bind selectively with NO and CO (and CN2 with cytochrome cA from some bacterial sources24), conformational changes of the protein which occurred in solution and which were accelerated during immobilisation (see Results and discussion) potentially may have affected the reported specificity.Therefore, experiments were performed to investigate the binding of other species to the metalloprotein. The selectivity of the ‘reconfigured’ cytochrome cA was investigated by recording the response of a solution of the protein (approximately 15 mmol l21, in phosphate buffer, 50 mmol l21, pH 7) to three potential interferents, oxygen, nitrite and cyanide ions, chosen to reflect those molecules and ions which bind to other haemoproteins.41 Results and discussion UV/VIS absorption properties of cytochrome cA in solution and sol–gel The oxidised and reduced absorption spectra of a freshly prepared solution of cytochrome cA (approximately 10 mmol l21) and cytochrome cA immobilised in a sol–gel are shown in Fig. 1(a) and 1(b), respectively. The intense absorption band below 360 nm in the reduced spectra is due to the dithionite ion which was required for protein reduction. The solution spectrum [Fig. 1(a)] of oxidised cytochrome cA exhibits the Soret (g) absorption band at 408 nm and a low intensity absorption band at 640 nm, indicative of the high-spin (S = 5/2) character of the haem iron.Upon reduction with dithionite, the Soret absorption band is red shifted to 428 nm with the formation of a characteristic shoulder at 438 nm. Additionally, absorption bands at 568 nm (sh) and 550 nm (termed a and b bands, respectively) are apparent in the Fe2+ cytochrome cA absorption spectrum. The oxidised and reduced spectra of the cytochrome cA encapsulated in the sol–gel matrix (Fig. 1(b)] appear to be different from those of the native solution spectrum, which suggests that the metalloprotein has undergone a degree of reconfiguration during the entrapment process.The oxidised protein spectrum displays a blue shifted Soret absorption band upon immobilisation (solution 408 nm ? sol– gel 404 nm). The reduced cytochrome cA spectrum exhibits clear resolution of the absorption band in the a/b region (solution 550 and 568 nm (sh) ? sol–gel 520 and 550 nm), accompanied by a significant blue shift of the Soret band (solution 428 nm ? sol–gel 418 nm) with the disappearance of the shoulder at 438 nm.These results suggest a significant change in the environment of the haem moiety, possibly a transformation from a high spin five-coordinate to a low spin six-coordinate haem iron species comparable to that found in the protein cytochrome c. While the structural change of the protein in the sol–gel occurred over 2–4 d it should be noted that the same spectral changes were observed for solutions of cytochrome cA stored at 4 °C, but over a period of 2–3 months.Such an ‘aged’ solution spectrum is shown in Fig. 1(c). By comparison of the spectra in Fig. 1 Absorption spectra of oxidised (dashed line) and reduced (solid line) cytochrome cA: (a) freshly prepared solution, (b) encapsulated in a sol– gel and (c) an ‘aged’ solution. Analyst, 1999, 124, 129–134 131Fig. 1(b) and 1(c), it is apparent that the same reconfiguration process occurs in solution and in the sol–gel matrix although the reconfiguration is clearly accelerated during the encapsulation of the cytochrome cA.The bacterial destruction of the cytochrome cA protein in solution observed by Barker et al.19 would not account for the spectroscopic changes observed in our work. Whereas previously we have not experienced any adverse effects with the sol–gel encapsulation of haemoproteins for the development of biosensing systems,17,18,34 other researchers have observed restriction of protein mobility42,43 and conformational changes44 upon sol–gel immobilisation. It is therefore possible that a similar mechanism could be contributing to advance the protein reconfiguration and consequent spectroscopic changes observed in this present work.Nitrosyl complex of cytochrome cA: absorption spectra Fig. 2(a) shows the formation of the nitrosyl complex of a ‘nonaged’ sample of cytochrome cA freshly prepared in solution. The NO–Fe3+ cytochrome cA spectrum agrees with literature observations; 23 the Soret band, with greatly enhanced intensity, is located at 418 nm and the pronounced a and b absorption bands are located at 562 and 528 nm, respectively.Fig. 2(b) shows the equivalent nitrosyl spectrum of sol–gel immobilised (and reconformed) Fe3+ cytochrome cA, formed by the titration of 50 ml of NONOate solution (containing an equivalent NO concentration of 4.6 mmol l21). While both the Soret absorption bands of the Fe3+ cytochrome cA and the NO complex are blue shifted in comparison with their solution values (sol–gel 398 and 408 nm; solution 406 and 417 nm, respectively), the clearly resolved a and b absorption bands, indicative of NO ligation, are identical for the cytochrome cA in both solution and sol–gel forms (562 and 528 nm, respectively).The Soret absorption band of the NO–cytochrome cA complex does not show an enhanced absorbance value as seen with the solution spectrum. This suggests that the haem centre of the encapsulated metalloprotein was not saturated with regard to the NO ligand, possibly as a result of a change in the binding affinity upon encapsulation.Nitric oxide titration of sol–gel immobilised cytochrome cA using controlled release from diethylamine NONOate While the binding affinity of the encapsulated cytochrome cA towards NO appeared to be lower as a result of conformational changes in the haem environment, it is clear that the protein was still able to bind NO. It was therefore possible to obtain a calibration curve using the controlled release NONOate calibration technique.Fig. 3 shows the calibration curve for the addition of aliquots of NONOate solution, suggesting an approximate working range of 10–400 mmol l21 (estimated from calibration points between 5 and 200 ml additions of NONOate) dissolved NO. The 200 ml NONOate addition calibration point corresponds to a total release of 876 nmol of NO. Such a quantity of NO is at least 100 times greater than the total amount of protein present (approximately 7 nmol). A similar calibration experiment performed on a fresh solution of cytochrome cA required 55 nmol of NO released from NONOate to saturate 20 nmol of the metalloprotein. While some loss of NO to the headspace of the sealed cuvette in which the experiment was performed may have occurred, the large excess of NONOate required to saturate the sol–gel cytochrome cA with NO, as compared with a freshly prepared cytochrome cA solution, does suggest that the binding affinity of the immobilised cytochrome cA has decreased upon encapsulation.It is possible that the microenvironment within the sol–gel structure has a different pH to that of the surrounding solution. Such a variation in pH may account for the change in binding affinity of NO with the solution versus encapsulated cytochrome cA. However, hydrogels which are constantly surrounded with buffer solution, as in this work, are less likely to exhibit localised pH variations.Perhaps more likely is that the binding affinity of cytochrome cA towards NO is reduced as a consequence of a structural change in the vicinity of the haem binding moiety. It should be noted that decreases in the apparent binding affinity of other biological molecules upon entrapment within sol–gel media have been reported previously.45,46 Fig. 2 Absorption spectra of oxidised (dashed line) and nitrosyl complex (solid line) cytochrome cA: (a) freshly prepared solution and (b) encapsulated in a sol–gel.Fig. 3 NO calibration curve for the encapsulated cytochrome cA generated using the in situ release of NO from diethylamine NONOate at pH 5. 132 Analyst, 1999, 124, 129–134Reproducibility of NO-cytochrome cA titrations in sol–gel The addition of a 20 ml aliquot of NONOate to the sol–gel immobilised metalloprotein followed by the dissociation of the NO ligand from the cytochrome cA was repeated seven times using the same sol–gel with measurement of the absorption intensity at 418 nm.To obtain meaningful reproducibility data it was imperative to remove all dissolved oxygen, by flushing with argon, prior to the addition of the NONOate solution to avoid the formation of nitrite ions. It was evident that the soformed nitrite ions did not bind to the haem centre of the protein as there was an apparent decrease in the absorbance intensity at 418 nm. This does suggests that the selectivity of cytochrome cA towards NO was retained upon immobilisation and subsequent reconfiguration.However, upon complete removal of all dissolved oxygen the repeated addition (seven measurements) of the NONOate solution to the sol–gel encapsulated cytochrome cA gave a reproducible signal with a relative standard deviation (RSD) of 0.28%. Interferences No spectral changes were observed on bubbling oxygen gas through solutions of both oxidised and reduced cytochrome cA. The addition of 0.1 mol l21 sodium nitrite was also observed to have no effect upon the metalloprotein when in either the oxidised or reduced state.Conversely, cyanide ions were shown to interact with the reconfigured cytochrome cA. In the oxidised form, the Soret band was observed to shift from 404 to 410 nm accompanied by a reduction in the absorption intensity at 640 nm. In the reduced form, cytochrome cA bound to cyanide ions resulted in a 2 nm shift in each of the Soret (from 418 to 420 nm), b (from 524 to 526 nm) and a (from 552 to 554 nm) absorption bands.These absorption band maxima for CN2– cytochrome cA are not in full agreement with documented values24 (424, 529 and 559 nm). Indeed the absorption band maxima more closely resemble those of CN2-cytochrome c (421, 525 and 554 nm).24 This supports the suggestion that a structural change in the vicinity of the haem group has occurred. While it is apparent that cyanide ions do bind to both reduced and oxidised cytochrome cA from Paracoccus denitrificans, which is in agreement for cytochromes cA from other bacterial sources,24 it is clear that the encapsulated cytochrome cA retains its inherent selectivity.Conclusions The NO specific metalloprotein cytochrome cA can be readily extracted and purified from the bacterium Paracoccus denitrificans. Upon encapsulation of the metalloprotein in a silica sol–gel, it is apparent that conformational changes in the vicinity of the haem moiety alter the spectral characteristics of the reduced, oxidised and nitrosyl complex of cytochrome cA and, perhaps more significantly, appear to change the binding affinity of the protein towards NO.Such spectroscopically observed conformational changes are apparent when cytochrome cA is stored in solution, although the changes are accelerated when the protein is encapsulated in the sol–gel. While the observed conformational changes appear to lower the NO binding affinity of the cytochrome cA, the protein remains selective to the NO analyte.An NO calibration curve, using controlled release of NO from diethylamine NONOate, was obtained from the encapsulated (reconfigured) cytochrome cA, which suggested an approximate working range of 10–400 mmol l21. Additionally, excellent reproducibility data could be obtained for the sol–gel encapsulated cytochrome cA. It is therefore apparent that cytochrome cA could prove to be extremely useful as a molecular recognition molecule for the biosensing of NO.The exact nature of the conformational change which appears to occur in the locality of the active haem binding site should be investigated using techniques such as magnetic circular dichroism spectroscopy to elucidate the haem–ligand coordination. Such information and further modification of the sol–gel process, e.g., the use of substituted precursor molecules, may aid the stabilisation of the entrapped protein, delaying or preventing the onset of the reconfiguration process.Acknowledgements The authors thank the Trustees of the Analytical Chemistry Trust Fund of the Royal Society of Chemistry for the award of a SAC Research Studentship to J.W.A. and the University of East Anglia for the award of a studentship to D.J.B. References 1 E. Culotta and D. E. Koshland, Science, 1992, 258, 1862. 2 S. Moncada and R. M. J. Palmer, Pharmacol. Rev., 1991, 43, 109. 3 C. Benvenuti, P. N. Bories and D. Loisance, Transplantation, 1996, 61, 745. 4 W. G. Zumft, Arch. Microbiol., 1993, 160, 253. 5 R. T. St. John and T. C. Hollocher, J. Biol. Chem., 1977, 252, 212. 6 B. J. Finlayson and J. N. Pitts, Science, 1976, 192, 111. 7 P. Brimblecombe, Air Composition and Chemistry, Cambridge University Press, Cambridge, 2nd edn., 1996. 8 R. M. J. Palmer, D. S. Ashton and S. Moncada, Nature (London), 1988, 333, 664. 9 M. Kelm and J. Schrader, Circ. Res., 1990, 66, 1561. 10 D. S. Bredt and S. H. Snyder, Proc.Natl. Acad. Sci. USA, 1989, 86, 9030. 11 O. C. Zafiriou and M. McFarland, Anal. Chem., 1980, 52, 1662. 12 K. Kikuchi, T. Nagano, H. Hayakawa, Y. Hirata and M. Hirobe, Anal. Chem., 1993, 65, 1794. 13 X. Zhou and M. A. Arnold, Anal. Chem., 1996, 68, 1748. 14 T. Malinski and Z. Taha, Nature (London), 1992, 358, 676. 15 M. N. Friedemann, S. W. Robinson and G. A. Gerhardt, Anal. Chem., 1996, 68, 2621. 16 B. C. Dave, B. Dunn, J. S. Valentine and J. I. Zink, Anal. Chem., 1994, 66, 1120A. 17 D. J. Blyth, J. W. Aylott, D. J. Richardson and D. A. Russell, Analyst, 1995, 120, 2725. 18 J. W. Aylott, D. J. Richardson and D. A. Russell, Chem. Mater., 1997, 9, 2261. 19 S. L. R.Barker, R. Kopelman, T. E. Meyer and M. A. Cusanovich, Anal. Chem., 1998, 70, 971. 20 M. M. Maltempo, J. Chem. Phys., 1974, 61, 2540. 21 T. Yoshimura, S. Suzuki, A. Nakahara, H. Iwasaki, M. Masuko and T. Matsubara, Biochim. Biophys. Acta, 1985, 831, 267. 22 T. Yoshimura, S. Suzuki, A.Nakahara, H. Iwasaki, M. Masuko and T. Matsubara, Biochemistry, 1986, 25, 2436. 23 T. Yoshimura, S. Fujii, H. Kamada, K. Yamaguchi, S. Suzuki, S. Shidara and S. Takakuwa, Biochim. Biophys. Acta, 1996, 1292, 39. 24 S. Suzuki, A. Nakahara, T. Yoshimura, H. Iwasaki, S. Shidara and T. Matsubara, Inorg. Chim. Acta, 1988, 153, 227. 25 R. F. DeBono, U. J. Krull and Gh. Rounaghi, ACS Symp. Ser., 1992, 511, 121. 26 K. Umemura, M. Hara, H. Sasabe and W. Knoll, Jpn. J. Appl. Phys., 1997, 36, L945. 27 D. J. Revell, J. R. Knight, D. J. Blyth, A. H. Haines and D. A. Russell, Langmuir, 1998, 14, 4517. 28 G. G. Guilbault and J. G., Montalvo, J. Am. Chem. Soc., 1970, 92, 2533. 29 Z. Zhujun and W. R. Seitz, Anal. Chem., 1986, 58, 220. 30 U. Narang, P. N. Prasad, F. V. Bright, K. Ramanathan, D. Kumar, B. D. Malhotra, M. N. Kamalasanan and S. Chandra, Anal. Chem., 1994, 66, 3139. 31 K. E. Chung, E. H. Lan, M. S. Davidson, B. S. Dunn, J. S. Valentine and J. I. Zink, Anal. Chem., 1995, 67, 1505. Analyst, 1999, 124, 129–134 13332 S. Shtelzer, S. Rappoport, D. Avnir, M. Ottolenghi and S. Braun, Biotechnol. Appl. Biochem., 1992, 15, 227. 33 L. M. Ellerby, C. R. Nishida, F. Nishida, S. A.Yamanaka, B. Dunn, J. S. Valentine and J. I. Zink, Science, 1992, 255, 1113. 34 J. W. Aylott, D. J. Richardson and D. A. Russell, Analyst, 1997, 122, 77. 35 D. J. Blyth, S. J. Poynter and D. A. Russell, Analyst, 1996, 121, 1975. 36 S. Braun, S. Rappoport, R. Zusman, D. Avnir and M. Ottolenghi, Mater. Lett., 1990, 10, 1. 37 O. Heichal-Segal, S. Rappoport and S. Braun, Biotechnology, 1995, 13, 798. 38 L. A. Robertson and J. G. Kuenen, J. Gen. Microbiol., 1983, 129, 2847. 39 J. W. B. Moir, D. Barrata, D. J. Richardson and S. J. Ferguson, Eur. J. Biochem., 1993, 212, 377. 40 C. M. Maragos, D. Morley, D. A. Wink, T. M. Dunams, J. E. Saavedra, A. Hoffmann, A. A. Bove, L. Isaac, J. A. Hrabie and L. K. Keefer, J. Med. Chem., 1991, 34, 3242. 41 E. Antonini and M. Brunori, Haemoglobin and Myoglobin in Their Reactions With Ligands, North-Holland, Amsterdam, 1971. 42 B. C. Dave, H. Soyez, J. M. Miller, B. Dunn, J. S. Valentine and J. I. Zink, Chem. Mater., 1995, 7, 1431. 43 N. Shibayama and S. Saigo, J. Mol. Biol., 1995, 251, 203. 44 C. Zhongping, D. L. Kaplan, K. Yang, J. Kumar, K. A. Marx and S. K. Tripathy, J. Sol–Gel Sci. Technol., 1996, 7, 99. 45 R. Wang, U. Narang, P. N. Prasad and F. V. Bright, Anal. Chem., 1993, 65, 2671. 46 K. Flora and J. D. Brennan, Anal. Chem., 1998, 70, 4505. Paper 8/06921B 134 Analyst, 1999, 124, 129–134

ISSN:0003-2654    

DOI:10.1039/a806921b

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Determination of live clay in foundry greensand by sodium-selective electrode Michael C. Cooper,a Michael J. Hey,*a Nicholas J. Milesb and Brian P. Atkinb a School of Chemistry, University of Nottingham, University Park, Nottingham, UK NG7 2RD. E-mail: michael.hey@nottingham.ac.uk b School of Chemical, Environmental and Mining Engineering, University of Nottingham, University Park, Nottingham, UK NG7 2RD Received 27th October 1998, Accepted 2nd December 1998 The rapid determination of ‘live’ clay in greensand is an essential procedure for the efficient running of a foundry production line, since progressive calcination of the clay binder during casting affects mould competence and quality of finish.A new method for assessing the active fraction of clay in foundry sand is described that is based on the response of an ion-selective electrode (ISE) to sodium ions liberated from the clay by ammonium ions. The method possesses advantages of speed and objectivity over the Methylene Blue titration method presently employed in foundries.Results appear to be unaffected by the presence of coal dust in greensand. 1 Introduction Natural smectite clays, especially sodium montmorillonite or sodium-exchanged Ca–Mg smectites, are used extensively in the production of the greensand moulds employed in many foundries. Fresh, moist clay (‘live’ clay) allows greensand to flow around a pattern while imparting both the compressive and tensile strengths needed to maintain the mould shape.1 After casting, clay in the proximity of the hot metal is degraded, causing these properties to be lost.This is termed ‘dead’ clay. Consequentially, there exists the need to constantly bleed off part of the greensand process stream and to replace it by fresh material, in order to maintain mould performance.2 To date, this waste has been consigned to landfill, or to alternative use. These options may, however, attract a financial penalty and there is an increasing interest in the possibility of recycling clay binders, in order to achieve significant reductions in operating costs.Rapid analysis of the amount of live clay in greensand wastes would be an essential component of such a process. The current method utilised for the analysis of live clay is the Methylene Blue ‘halo’ test.3 Methylene Blue (MB), a cationic dye, is sorbed strongly by clay particles until the cation exchange capacity (CEC) of the clay is reached, after which the excess remains in solution.4 In practice, MB is titrated against a known weight of greensand, dispersed in aqueous suspension.Aliquots of 1 cm3 of dye solution are usually added, after each addition of which a drop of suspension is removed and ‘spotted’ on to filter paper. MB present in excess of the CEC results in the formation of a halo of colour around a central area of dyed clay particles. The first appearance of the halo may be subtle, and some skill is required in order to consistently detect it.In addition, this procedure is slow, prone to subjective error, and accuracy is furthermore limited by the strength of MB solution used and the volume of each aliquot added. The present study suggests that use of an electrode sensitive to sodium ions in solution could provide an alternative, objective method for the quantification of live clay in greensand. The sodium-selective electrode has found use in many applications, for example in the analysis of foodstuffs,5 the determination of feldspars,6 and in the study of the rate of cation exchange on clay minerals.7 The electrode incorporates a sodium-containing glass membrane, whose electrical potential (E), when placed in contact with an aqueous solution containing sodium ions, is described by the Nicolsky–Eisenman form of the Nernst equation: E = E0 + (2.303RT)/F log10 (aNa+ + SkNa+,B aB 1/nB) (1) where E0 is the standard potential, R the gas constant, T the absolute temperature, F the Faraday constant, aNa+ the sodium ion activity, and aB the activity of an interfering ion B, with charge nB.The potentiometric selectivity coefficient kNa+,B expresses the degree of selectivity of the sodium electrode with respect to an interfering ion B. When selectivity coefficients are much less than unity for all other ions present in solution, the electrode response is Nernstian and a plot of potential versus sodium ion activity has a slope equal to 59 mV per decade at 298 K.Addition of an electrolyte of non-interfering ions at high ionic strength ensures that the activity coefficient of sodium ions in solution is effectively constant, so that the observed potential is given by: E = E0 + S log10 [Na+] (2) where S is the response of the electrode, normally reported as mV per molar decade. In the present study, the required ionic strength was provided by the addition of a solution of ammonium chloride and ammonium hydroxide (referred to as the Ionic Strength Adjuster or ISA), made up according to the recommendations of the electrode manufacturer.The ISA solution additionally fulfilled the purpose of adjusting the analyte to ca. pH 9 (hydrogen is an interferent ion), whilst simultaneously supplying ammonium ions which displace sodium ions from the clay matrix.8 As ammonium ions show a low level of interference for the sodium electrode (selectivity coefficient ca 1 3 1024)9 they have a negligible effect on the observed potential even in solutions where their concentration exceeds that of sodium.The response of the sodium-selective electrode is therefore governed only by the concentration of the displaced sodium ions and hence the amount of live clay present in a sample of foundry sand. 2 Experimental The electrode assembly consisted of a sodium-selective electrode in combination with a calomel reference electrode Analyst, 1999, 124, 135–138 135(Russell type 97-7129; Russell pH Ltd., Auchtermuchty, Scotland). A high precision digital pH/mV meter (Radiometer PHM84, Radiometer A/S, Copenhagen, Denmark) was used for the potentiometric measurements.A stock solution of ISA was prepared by dissolving reagent-grade ammonium chloride (20 g) in purified water (ca. 50 cm3; Elgastat UHQ II, specific resistivity 18 MW cm21), to which 27 cm3 of concentrated ammonia solution (specific gravity 0.880) was added, before making up to a final volume of 100 cm3 with purified water. The electrode was calibrated using solutions containing 0.001, 0.01, 0.1, 1 and 2 mol dm23 NaCl.A plot of E against log10[Na+] was linear (R2 = 0.995) with a gradient of 62 mV per molar decade, compared to the theoretical Nernstian slope of 59 mV. When not in use, the electrode was kept immersed in a storage solution of 5 mol dm23 NaCl. After the electrode had been allowed to equilibrate for three days in the storage solution, repeated checks on the calibration showed the response to be sufficiently reproducible for recalibration to be unnecessary before each set of determinations.Analyses were performed on a Mediterranean soda-activated montmorillonite, as used by a local foundry. The CEC for this product was specified by the original supplier as ca. 90 meq 100 g21, of which 80–84 meq 100 g21 were due to Na+ ions. ‘Dead’ clay was produced by heating live clay, in a platinum crucible, placed in a muffle furnace at 850 °C for five hours.Measurements of CEC for dead clay revealed very low values (< 0.001 meq 100 g21), indicating that the product was comparable to that obtained from a foundry. Samples of simulated foundry greensand, produced by the dry mixing of sand (ca. 70% w/w), coal dust (ca. 20% w/w) and clay (ca. 10% w/w), were representative of those used in practice. 2.1 Determination of CEC A series of suspensions of clay (ca. 0.1 g, accurately weighed) in acidified water (20 cm3) was prepared, containing increasing volumes of 1% (w/v) MB dye solution. Each sample was shaken continuously for ten minutes and then allowed to settle.An aliquot (1 cm3) of the clear supernatant was removed and the dye concentration determined spectrophotometrically (lmax = 663 nm), from a previously acquired calibration curve. Dilution of the supernatant was necessary at higher concentrations of MB to ensure that the measured absorbance was within the calibration range (<2 absorbance units).Initial addition of dye, at levels below CEC, resulted in irreversible physical adsorption, 4 with no dye detectable in the supernatant. Larger volumes of dye solution resulted in an excess of dye remaining in the supernatant, the concentration of which was then quantified. The sodium exchange capacity of another sample of the same clay was then determined by ISE. Aliquots (1 cm3) of ISA solution (stock ISA diluted 1 + 99 v/v; 0.077 mol dm23, pH 12) were titrated into a suspension of an accurately weighed (ca. 1 g in 50 cm3 water) sample of clay. The potential developed by the electrode was measured after the addition of each aliquot and the amount of sodium ions in solution calculated. 2.2 Determination of clay in greensand A sample of greensand (ca. 5 g, accurately weighed) was suspended in water (50 cm3) and simmered for ca. 20 min, in order to ensure complete disaggregation of particles and to allow the clay to dilate.The solution was then acidified (H2SO4, 2 mol dm23, 1 cm3), and allowed to cool to room temperature, before titration with MB solution in aliquots of 1 cm3. Samples for ISE analysis (ca. 5 g) were accurately weighed into a 100 cm3 beaker, a solution of ISA (1 cm3) in deionised water (50 cm3) added, and a suspension formed by stirring magnetically for two minutes. The sodium-selective electrode was then immersed in the suspension and the equilibrium voltage recorded.In order to determine whether sufficient ammonium ions had been added to displace all sodium ions from the clay, a further aliquot of 1 cm3 of ISA was added and the potential of the electrode remeasured. No significant change was observed, indicating that complete exchange had been achieved. In order to compare the two methods, mixtures (ca. 10 g, accurately weighed) of simulated foundry greensand (sand, coal dust and live clay) were prepared with different clay contents: 0, 1, 3, 5, and 10% (w/w), comparable to those found in actual foundry greensand.Each sample was then divided into two ca. 5 g sub-samples. Sodium determination by ion-selective electrode was carried out on one sub-sample, while the other was subjected to the MB halo test. The determination of live clay in simulated greensand was then extended to sands containing both live and dead clays. Two series of samples (ca. 10 g) were prepared, one by mixing 20% coal dust, 10% total clay, and 70% sand, and the other 20% coal dust, 5% total clay and 75% sand.The clay fractions contained 0, 20, 40, 60, 80, and 100% (w/w) live clay, the balance being made up by dead clay. Again, an attempt was made to prepare mixtures representative of those which might be encountered in foundry operations. Samples were thoroughly dry mixed, by intermittent shaking over a period of 24 h, and then separated into two ca. 5 g sub-samples. After dispersion in ISA solution, a sodium ion determination by ion-selective electrode was carried out for both sub-samples. 3 Results and discussion 3.1 Cation exchange capacity by MB sorption Fig. 1 shows the results of the determination of CEC for live clay by MB sorption. For low levels of added MB, the amount of dye sorbed by the clay was always equal to the amount added. This equality was maintained until the clay became saturated at the CEC, after which the amount sorbed remained constant with the excess MB appearing in the supernatant.4,10 The CEC determined by this method was calculated to be 80 ± 2.5 milliequivalents (meq) 100 g21 of clay, in reasonable agreement with the manufacturer’s specification of ca. 90 meq 100 g21. 3.2 ISE determination of sodium exchange capacity The sodium exchange capacity (meq 100 g21 of clay) of the live clay was obtained from the plot shown in Fig. 2. Titration of the clay suspension by diluted ISA solution resulted in sodium ions Fig. 1 Spectrophotometric determination of CEC of a Mediterranean soda-activated clay, utilising Methylene Blue (MB) sorption.Error bars denote the standard deviations obtained from four independent determinations. 136 Analyst, 1999, 124, 135–138being displaced from the clay until a constant concentration of sodium ions in solution was reached as indicated by a constant electrode potential, indicating that no more were available for exchange. The sodium exchange capacity was calculated to be 80 ± 3 meq 100 g21, in agreement with the CEC value resulting from the MB determination.The plot also indicates that about 45 meq 100 g21 of sodium ions were released from the clay into the suspension, prior to the addition of ISA, probably as a result of exchange with protons. 3.3 Determination of live clay in greensand: comparison of MB and ISE methods A comparison of the results for live clay in simulated greensand as determined by the MB halo test and by ISE is presented in Fig. 3.Both methods give linear correlations between estimated clay content and known clay content, but the ISE method was superior in both accuracy and precision. For the MB results the slope of the plot was 1.2845 (R2 = 0.9494) compared with a slope of 1.0967 (R2 = 0.9924) for the ISE results. A perfect correlation would obviously result in a slope of 1.0000 (R2 = 1.0000). The results from ISE analyses of simulated greensand mixtures containing both live and dead clay are displayed in Fig. 4, where percentage live clay is plotted against concentration of sodium ions in solution. A linear relationship is observed, with R2 = 0.997, indicating a consistent efficiency of exchange of sodium with ammonium ions. A sodium ion concentration of ca. 0.5 ppm was detected in samples containing no live clay, and therefore would appear to represent a lower limit to the detection of sodium in greensand. It is felt most unlikely, however, that these levels of sodium ions would be encountered during normal foundry operations. 4 Conclusion The use of a sodium-selective electrode for the determination of live clay in foundry greensand offers several advantages over the traditional method based on uptake of dye.Users of sodium montmorillonite, or sodium-modified clays, could benefit from the objective nature of the analysis which leads to increased accuracy and precision over the MB halo test. The ISE method, which possesses a detection limit adequate for all foundry situations, allows an analysis to be performed in a few minutes, in contrast to the much slower MB halo test.The sodiumcombination electrode is robust, simple to use, and requires only an inexpensive voltmeter for its operation. The technique could be modified for use outside the laboratory to give rapid determinations of live clay in the foundry. In foundries employing clays for which exchangeable cations are only partially replaced by sodium, frequent calibration would have to be carried out in order to compensate for variations in sodium content. Alternatively, dual analysis using both sodium- and calcium-selective electrodes (for sodiumbeneficiated calcium montmorillonites) could be used to monitor variations thus providing a method for in-house quality control of the raw material. Acknowledgements The authors gratefully acknowledge the award of a research grant from the Engineering and Physical Sciences Research Council under its Waste Minimisation (WMR3) programme.Mr Graham Cooper (William Lee Ltd) and Mr David Goring (Stanton Plc) are thanked for their support, expertise and encouragement. References 1 I. E. Odom, Phil. Trans. R. Soc. Lond., 1984, A311, 391. 2 W. B. Parkes, Clay Bonded Foundry Sand, Applied Science Publishers, London, 1971, pp. 117–124. 3 D. F. Knight, W. Amos, M. H. Lavington and R. Wootton, Meehanite Sand Control Manual. Green Sand Systems, Meehanite Institute of Great Britain, International Meehanite Metal Co. Ltd., Reigate, UK, 1992, pp. 3.32–3.34. 4 P. T. Hang and G. W. Brindley, Clays Clay Minerals, 1970, 18, 203. 5 F. G. McNerney, J. A. O. A. C. Int., 1976, 59, 1131. 6 J. H. Puffer and R. S. Cohen, Chem. Geol., 1975, 15, 217. 7 V. C. Kennedy and T. E. Brown, Clays Clay Minerals, 1964, 13, 351. Fig. 2 Determination of sodium exchange capacity of a Mediterranean soda-activated clay, utilising a sodium-selective electrode. Error bars indicate standard deviations obtained from four independent determinations. Fig. 3 Comparison of results obtained from determination of live clay in simulated greensand mixtures, by ISE and MB halo tests. Fig. 4 Sodium ion concentrations in suspensions of simulated foundry greensand mixtures, determined by ISE. Analyst, 1999, 124, 135–138 1378 E. Grim, Clay Mineralogy, McGraw-Hill Book Company, New York, 2nd edn., 1968, pp. 184–233. 9 P. L. Bailey, Analysis with Ion-selective Electrodes, Heyden, London, 2nd edn., 1980, pp. 70–71. 10 G. W. Brindley and T. D. Thompson, Israel J. Chem., 1970, 8, 409. Paper 8/08344D 138 Analyst, 1999, 124, 135–138

ISSN:0003-2654    

DOI:10.1039/a808344d

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Ion-selective electrode for the determination of trazodone in tablets Sabry Khalil Department of Chemistry, Faculty of Science, Cairo University, Fayoum Branch, 63514 - Fayoum, Egypt Received 10th November 1998, Accepted 2nd December 1998 A coated wire trazodone-selective electrode based on incorporation of trazodone–tetraphenylborate ion pair in a poly(vinylchloride) coating membrane was constructed. The influences of membrane composition, temperature, pH of the test solution, and foreign ions on the electrode performance were investigated. The electrode showed a Nernstian response over a trazodone concentration range from 1.41 3 1025 to 0.89 3 1022 M, at 25 °C, and was found to be very selective, precise, and usable within the pH range 2.4–9.0.The standard electrode potentials, E°, were determined at 20, 25, 30, 35, 40 and 45 °C and used to calculate the isothermal temperature coefficient (dE°/dT) of the electrode. Temperatures higher than 45 °C seriously affected the electrode performance.The electrode was successfully used for potentiometric determination of trazodone hydrochloride both in pure solutions and in pharmaceutical preparations. Trazodone or 2-[3-(4-m-chlorophenylpiperazin-1-yl)propyl]- 1,2,4-triazolo[4,3-a]pyridin-3(2H)-one is a triazolo pyridine derivative. The distinguishing property of trazodone is its capacity to act selectively on the system of emotional integration, correcting the two main mechanisms responsible for depression: (a) an excessive input of unpleasant information as for example in secondary depression and (b) an intrinsic defect in integration, as in endogenous depression. Trazodone also acts selectively on the serotoninergic system both at the central level, where it inhibits the uptake phenomenon, and at the vascular level, where its antiserotonin effect prevails since serotonin is involved in cerebral ischemia; this latter effect may prove to be useful in pathological conditions accompanied by a diminished cerebral blood flow.Several methods have been reported for the determination of this important compound.1–10 However, most of these methods involve several manipulation steps before the final result of the analysis is obtained. Although potentiometric methods of analysis using ion-selective electrodes are simple, cheap and applicable to samples, no selective electrode is, so far, available for the determination of trazodone.The present work, thus, describes a new selective membrane electrode of the coated wire type, for determination of trazodone in pure solutions and in pharmaceutical preparations. This electrode is based on incorporation of an ion-pair complex of tetraphenylborate anion (TPB2) with trazodone cation (TZH+) in a poly(vinylchloride) matrix. It is noteworthy that all previously reported investigations using poly(vinylchloride) (PVC) membrane selective electrodes for determination of species of pharmaceutical and/or medical importance have been carried out at only one temperature, mostly 20 or 25 °C.No attention has been paid to the higher temperature range, 25–45 °C, although many potentiometric measurements concerning biological media and fluids are made at such temperatures.11 In this paper, the effect of the temperature of the test solution on the performance characteristics of the proposed coated wire electrode (CWE) is reported. Experimental Reagents and materials All chemicals used were of analytical or pharmacopeial grade (can be used for manufacturing pharmaceutical preparations). Bi-distilled water was used throughout all experiments.The pharmaceutical preparations containing trazodone (Deprax, Trazolan and Trittico tablets) were obtained from local drug stores. The TZH–TPB ion pair was prepared by a method similar to that described previously.12 The base component of the produced ion pair has been determined by the non-aqueous titration method.13 The agreement between calculated and found values was very good confirming the postulated stoichiometry; the 1 : 1 (TZH : TPB) molar ratio stoichiometry was also confirmed by elemental analysis.Construction of electrode Spectroscopic pure copper wires of 2.0 mm diameter and 12 cm length were tightly insulated by polyethylene tubes leaving 1.0 cm at one end of the coating and 0.5 cm at the other end for connection. The coating solutions were prepared by dissolving varying amounts of powdered PVC, dioctylphthalate, DOP (plasticizer), and the TZH–TPB in the least amount of tetrahydrofuran possible (3–4 ml), Table 1.Prior to coating, the polished copper surface was washed with a detergent and water, thoroughly rinsed with de-ionized water, and dried with acetone. Then the wire was rinsed with chloroform and allowed to dry. Afterwards, the copper wire was coated by quickly dipping it into the coating solutions, (a), (b), (c), or (d), several times and allowing the film left on the wire to dry for about 2 min. The process was repeated several times until a plastic membrane of approximately 1.0 mm thickness was formed as measured by an electronic linear measuring gauge head, Tesa- GT41.The prepared electrodes were preconditioned by soaking them for 1.5 h in 1023 M TZHCl solution daily. Analyst, 1999, 124, 139–142 139Potentiometric studies and electrochemical system Potentiometric measurements were carried out with an Orion (Cambridge, MA, USA) Model 701 A digital pH/mV-meter.A Techne circulator thermostat, Model C-100, was used to control the temperature of the test solution. The electrochemical system was as follows: Cu|membrane|test solution||KCl salt bridge|| KCl(sat.)|Hg2Cl2–Hg. Construction of the calibration graphs Suitable increments of standard TZHCl solution were added to 50 ml of 1 3 1026 M TZHCl solution so as to cover the concentration range from 1 3 1026 to 3.2 3 1022 M.In this solution the sensor and the reference electrode were immersed and the emf was recorded after 10 s, at 25 °C, for each addition. The electrode potentials, Eelec, were calculated from the emf values and plotted versus pTZH (2log[TZH]). The process was repeated at 25, 30, 35, 40 and 45 °C. To determine the linearity range in the case of the pharmaceutical samples, calibration graphs were constructed by using standardized drug solutions, at 25 °C, by measuring the electrode potential in solutions containing varying amounts of the respective drug.The electrode was repeatedly calibrated over a period of four months. Selectivity of the electrode The selectivity coefficients, Kpot TZH, Jz+, were evaluated by the separate solution method described by Badawy et al.14 Potentiometric determination of trazodone The standard addition method was applied in which small increments of standard trazodone hydrochloride solution (1 3 1022 M) were added to 50 ml aliquot samples of various concentrations (3.0 31024 to 1.5 31023 M).The change in the potential reading (at constant temperature of 25 °C) was recorded for each increment and used to calculate the concentration of TZHCl sample solution. For analysis of trazodone formulations 8.50–32.65, 9.65–28.16 or 7.35–34.25 mg of Deprax (16 tablets), Trazolan (12 tablets) or Trittico (20 tablets), respectively, were dissolved in 50 ml of distilled water and the standard addition technique was applied as described above.Results and discussion Composition of the coating membrane Four coating membrane compositions were investigated as given in Table 1. CWE made by using coating solution (d) exhibited a calibration plot of very good Nernstian slope (59.0 mV per concentration decade, at 25 °C, Table 1) over a relatively wide range of TZH+ concentration (1.41 3 1025 to 0.89 3 1022 M) with a response time < 10 s. Consequently, the electrode made by using coating solution (d) was selected for carrying out all the following studies.Effect of soaking The performance characteristics of the TZH+ CWE were studied as a function of soaking time. For this purpose the electrode was soaked in a 1 3 1023 M solution of TZHCl and the calibration graphs (pTZH vs. Eelec, mV) were plotted after 5 min and 0.5, 1.0, 1.5, 2, 3, 4, 8, 24, and 48 h. The optimum soaking time was found to be 1.5–2.0 h, at which the slopes of the calibration curves were 57.0–59.0 mV per pTZH decade, at 25 °C.Soaking for longer than 24 h is not recommended to avoid leaching, though very little, of the electroactive species into the bathing solution. The electrode should be kept dry in an opaque closed vessel and stored in a refrigerator while not in use. The reproducibility of repeated measurements on the same solutions was ±1 mV. Effect of temperature of the test solution Calibration graphs constructed, as previously described, at test solution temperatures of 20, 25, 30, 35, 40, 45, and 50 °C are represented in Fig. 1 (a–g, respectively). The slope, usable concentration range, and response time of the electrode corresponding to each temperature are reported in Table 2. From the table it is clear that the electrode gave a good Nernstian response in the temperature range 20–45 °C. From Fig. 1, the standard electrode potentials (E°) were determined, as the intercepts of the calibration graphs at pTZH = 0, and used to obtain the isothermal temperature coefficient (dE°/dT) of the electrode by aid of the following equation:15 E° = E°25 + (dE°/dT) (t 2 25) A plot of Eo vs.(t 2 25) gave a straight line, the slope of which was taken as the isothermal temperature coefficient. It amounts to 20.0009 V per °C, revealing a fairly good thermal stability of the electrode. Effect of pH The effect of pH of the TZHCl test solution on the electrode potential is graphically represented in Fig. 2. The pH of the initial solution is altered by the addition of very small volumes of HCl and/or NaOH (0.1–1.0 M each). Fig. 2 indicates that the pH has a negligible effect within the pH range of 2.3–9.0. In this range the electrode can be safely used for trazodone determination. During the operative life of the electrode (four months), no significant change in the potential–pH behaviour was observed. Table 1 Composition of the coating membranes and slopes of the corresponding calibration graphs at 25 °C Coating solution/mga Membrane composition (%, m/m) Slope/ mV decade21 RSD Membrane PVC DOP Ion pair PVC DOP Ion pair 1.5 h Presoak (%)b (a) 120.0 112.5 17.5 48 45 7 48.0 1.3 (b) 115.0 112.5 22.5 46 45 9 51.5 1.1 (c) 112.5 100.0 37.5 45 40 15 54.0 1.2 (d) 120.0 100.0 30.0 48 40 12 59.0 1.1 a Dissolved in the least amount of tetrahydrofuran possible (3–4 ml).b Relative standard deviation values of slopes (six determinations). 140 Analyst, 1999, 124, 139–142The decrease in potential readings at pH < 2.3 and pH > 9.0 until pH ~ 10.3 may be attributed to penetration of Cl2 and OH2 ions, respectively. At pH 10.3, a turbidity due to precipitation of trazodone base was first detected and associated with a concurrent increase in the electrode potential up to pH 11.0.This increase is most probably due to a corresponding decrease in the penetration of the OH2 ions as a result of their reaction with the protonated trazodone species.Beyond pH 11.0, the sharp decrease in potential may be attributed to two reasons. The first is the disappearance of the TZH+ species from the medium as a result of precipitation. The second reason is the penetration of the OH2 ions into the gel layer of the membrane replacing, partially, the TPB2 anions of the ion pair. Thus the electrode works as a sensor for the OH2 ions in highly alkaline media, exhibiting a decrease in potential as the pH value increases.Selectivity of the electrode The selectivity coefficients Kpot TZH, Jz+ presented in Table 3 clearly showed that the proposed CWE is very selective toward TZH+ with respect to many common inorganic and organic cations, sugars, and amino acids which are frequently present in biological fluids and pharmaceutical preparations. Analytical applications The present CWE has been successfully used for the determination of trazodone in aqueous solution and in the pharmaceutical preparations Deprax, Trazolan, and Trittico (tablets) by using the standard addition method described above.The recovery and standard deviation values given in Table 4 were calculated from ten determinations in the case of pure TZHCl solution and from six determinations in the case of pharmaceutical preparations. The present method is not applicable to cream products since the presence of greasy material poisons the membrane surface. In pharmaceutical analysis it is important to test the selectivity toward excipients and fillers added to the pharmaceutical preparations.Fortunately, such materials mostly do not interfere. It is clear from the results obtained for the pharmaceutical preparations (Table 4) that these excipients do not interfere. Fig. 1 Calibration graphs at 20 (a), 25 (b), 30 (c), 35 (d), 40 (e), 45 (f), and 50 °C (g) using a trazodone-coated wire electrode [membrane (d)] soaked for 1.5 h. Table 2 Performance characteristics of trazodone CWEa at different temperatures as determined in aqueous solutions Slope (expt)/ Response Intercept at Temp/°C mV decade21 Usable range/M times/s pTZH = 0 E°elec 20 56.0 1.12 3 1025–2.81 3 1022 @10 368.0 25 59.0 1.41 3 1025–0.89 3 1022 @10 362.5 30 62.5 1.58 3 1025–0.61 3 1022 @10 356.6 35 66.0 1.51 3 1025–0.61 3 1022 @10 352.0 40 68.0 1.62 3 1025–0.60 3 1022 @10 347.0 45 71.0 1.54 3 1025–0.49 3 1022 @10 343.5 a Preconditioned by soaking for 1.5 h, approximate film thickness is 1.0 mm.Fig. 2 Effect of pH of the test solution on the potential reading: (5) 4.5 3 1023 M TZHCl, (2) 2.8 3 1023 M TZHCl solution at 25 °C, using electrode (d). Table 3 Selectivity coefficients of the TZH+ CWE calculated by the separate solution method (1 3 1023 M of both TZH+ and the interferent) at 25 °C Interferent Kpot TZH, Jz+ Interferent Kpot TZH, Jz+ Na+ 1.35 3 1023 Lactose 2.54 3 1023 K+ 1.31 3 1023 Sucrose 1.58 3 1023 NH4 + 1.41 3 1023 Glycine 1.10 3 1023 Mg2+ 3.65 3 1024 Alanine 9.44 3 1024 Ca2+ 1.24 3 1024 Phenylalanine 1.08 3 1023 Fe2+ 1.44 3 1024 (Me)2NH+ 1.55 3 1023 Fe3+ 1.01 3 1024 (Et)2NH2 + 2.17 3 1023 Glucose 2.42 3 1023 (Et)3NH+ 1.11 3 1023 Maltose 1.99 3 1023 (Et)4N+ 2.58 3 1023 Analyst, 1999, 124, 139–142 141References 1 N.Rifai, C. B. Levtzow, C. M. Howlett, C. M. Phillips, N. C. Parker and R. E. Cross, J. Anal. Toxicol., 1988, 12, 150. 2 J. M. Kauffmann, J. C. Vire, G. J. Patriarche, L. J. Nunez-Vergara and J.A. Squella, Electrochim. Acta, 1987, 32, 1159. 3 T. J. Siek, J. Anal. Toxicol., 1987, 11, 225. 4 R. T. Sane, V. R. Nerurkar, R. V. Tendolkar, D. P. Ganagal, P. S. Mainkar and S. N. Dhumal, Indian Drugs, 1990, 27, 251. 5 Z. Liangyua, Zhongguo Yaoka Daxue Xuebao, 1989, 20, 208. 6 L. J. Lovett. G. A. Nygard and S. K. W. Khalid, J. Liq. Chromatogr., 1987, 10, 909. 7 I. M. Roy and T. M. Jefferies, Pharm. Biomed. Anal., 1990, 8, 831. 8 N. Beaulieu, R. W. Sears and E. G. Lovering, J. AOAC Int., 1994, 77, 857. 9 G. Esposito, J. Anal. Toxicol., 1996, 20, 59. 10 W. Lambert, J. Van Bocxlaer, M. Piette and A. P. Deheenheer, J. Anal. Toxicol., 1996, 20, 60. 11 G. Nagy, J. Tarcall, K. Toth, R. N. Adams and E. Pungor, Fourth Symposium on Ion-selective Electrodes, Matrafured, Hungary, 1984, p. 567. 12 A. F. Shoukry, S. S. Badawy and Y. M. Issa, Anal. Chem., 1987, 59, 1078. 13 L. G. Chatten, M. Pernarowski and L. Levi, J. Am. Pharm. Assoc., Sci. Ed., 1959, 48, 276. 14 S. S. Badawy, A. F. Shoukry and Y. M. Issa, Analyst, 1986, 111, 1363. 15 L. I. Antropov, Theoretical Electrochemistry, Mir Publishers, Moscow, 1972, p. 378. Paper 8/08800D Table 4 Potentiometric determination of trazodone in aqueous solution and in pharmaceutical preparations with a TZH electrode by the standard addition method, at 25 °C Sample Amount taken/mg Recovery (%) RSD (%) Pure TZH+ solution 5.75–36.30 100.15 0.85 Deprax tabletsa 8.50–32.65 99.70 1.13 Trazolan tabletsb 9.65–28.16 99.10 0.80 Trittico tabletsc 7.35–34.25 98.70 1.20 a Farma Lepori, Spain. b Searle, Netherlands. c Egyptian International Pharmaceutical Industries Co., Tenth of Ramadan City A.R.E. 142 Analyst, 1999, 124, 139–142

ISSN:0003-2654    

DOI:10.1039/a808800d

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Amperometric determination of the water content in acetone, butter and margarine using a wall-tube platinum microelectrode in a flow injection system Fredrik Björefors,a Leif Nyholm,*a Mikolaj Dontenb and Zbigniew Stojekb a Department of Analytical Chemistry, Uppsala University, P.O. Box 531, SE-751 21 Uppsala, Sweden. E-mail: Leif.Nyholm@kemi.uu.se b Department of Chemistry, University of Warsaw, ul. Pasteura 1, PL 02-093 Warsaw, Poland Received 2nd September 1998, Accepted 8th December 1998 Amperometric determinations of the water content in acetone, butter and margarine were performed using a wall-tube platinum microelectrode in combination with a flow injection system.For the butter and margarine samples, acetone containing small amounts of supporting electrolyte was used both as solvent and mobile phase. Linear calibration curves for oxidation of water in acetone were obtained in the range 0.005–0.3% v/v and the limit of detection was determined as 0.003% v/v.The sizes of the currents and the linear range were found to be independent of the supporting electrolyte concentration in the range 1.0 3 1024–2.0 3 1023 m, provided that the detection potential was altered to ensure the attainment of steady state currents. The results also indicate that the methanol content in acetone can be determined with the present flow injection system. Introduction Electrochemical measurements in flowing solutions have become increasingly popular,1–4 e.g., owing to the possibilities of performing fast on-line analyses and of combining electrochemistry with other analytical techniques.An especially important area is the utilisation of microelectrodes as sensitive and low dead volume detectors in miniaturised systems such as capillary flow injection analysis, capillary liquid chromatography and capillary electrophoresis.3–6 An interesting advantage of using microelectrodes is the possibility of carrying out measurements also in media with very low conductivities such as solutions containing no or very small amounts of added supporting electrolyte, pure organic solvents or supercritical media.2,7–15 Studies performed with microelectrodes in low conductivity media have shown that the migrational mass transport of ionic species can be significant under such experimental conditions. 13,16–24 Migrational effects can, however, also be observed for uncharged species in solutions containing small amounts of added supporting electrolyte since the generation of a charged product requires the build-up of an ionic layer in the vicinity of the electrode.13,21–23,25 Recently, it has been shown26 that the positioning of a wall tube microelectrode with respect to the capillary end and also the flow rate, potential and ionic strength influence this build-up of the ionic layer significantly. The results of this study also indicated that such a wall-tube based flow injection system can be used for, e.g., determinations of water in various organic solvents and samples dissolved in organic solvents.The aim of this work was to demonstrate the analytical applicability of a wall-tube based flow amperometric system for the determination of the water content in acetone, butter and margarine. For this purpose, the effects of the supporting electrolyte concentration and detection potential on the magnitude of the currents and linear range were also investigated.In addition, it is shown that the flow injection system can be used for determinations of the methanol content in acetone. Experimental The experimental set-up consisted of a PU-980 HPLC pump (Jasco, Tokyo, Japan), a Rheodyne (Cotati, CA, USA) injection valve with an 80 ml injection loop and a previously described26 laboratory made impinging jet flow cell made of Kel-F. In the cell, the flow was directed upwards, perpendicular to the surface of the working microdisk electrode.The distance between the wall-tube working electrode and the end of the capillary tubing was set to 1.0 mm and the electrode was positioned in the centre of the flow outlet using a laboratory made 3D precision positioner. The capillary tubing used was made of PEEK, and had an id of 0.76 mm and an od of 1.6 mm. The working electrode was a 10 mm diameter platinum disk manufactured by Project (Warsaw, Poland) and the reference electrode consisted of a platinum coil to eliminate leakage of water from a conventional reference electrode.The half-wave potentials measured versus the quasi reference platinum Fig. 1 Amperometric signals due to injections of increasing amounts of water in acetone containing 2.0 31023 m LiClO4. The water concentrations were (a) 0.025, (b) 0.05, (c) 0.07 and (d) 0.15% v/v, and the detection potentials were (a) 2.8, (b) 2.8, (c) 2.8 and (d) 3.0 V. The flow rate was 0.1 ml min21 and the distance between the tubing outlet and the electrode was 1.0 mm.Analyst, 1999, 124, 143–146 143electrode have been found previously to be constant within 10 mV.27 The stainless steel nut used for fixing the outlet capillary tubing in the flow cell served as the counter electrode. All potentials reported in the paper are given versus the quasi reference platinum electrode. The working electrode, embedded in a 6 mm diameter glass cylinder, was polished on a wet pad with 0.3 mm Al2O3 particles (BDH, Poole, Dorset, UK) before use.A PAR Model 273 potentiostat (EG&G Princeton Applied Research, Princeton, NJ, USA) together with the ECHEM software was used in the experiments. Extra dry SeccoSolv quality acetone (Merck, Darmstadt, Germany) was employed. The water and methanol contents in the acetone did not exceed 0.008 and 0.05% v/v, respectively, according to the manufacturer. The methanol used was of LiChrosolv quality (Merck) and water free purum p. a. LiClO4 (Fluka, Buchs, Switzerland) was used as supporting electrolyte.Distilled water was further purified using a Milli-Q system (Millipore, Milford, MA, USA). Results and discussion Oxidation of water in acetone In a previous study,26 the key parameters influencing the oxidation of neutral compounds in methanol or acetone containing small amounts of supporting electrolyte in a flow injection system were investigated. One of the objectives with the present study was to determine amperometrically the water content in samples such as butter and margarine that are easily dissolved in acetone, which has been shown27 to be a suitable solvent for this type of experiment.To determine the linear range and limit of detection for the determinations based on the oxidation of water in acetone, increasing amounts of water were added to acetone in the range 0.005–0.5% v/v. In these experiments the acetone contained 2.0 3 1023 m LiClO4 as supporting electrolyte. The flow rate of the mobile phase, i.e., acetone containing the same amount of LiClO4, was 0.1 ml min21.Some of the peaks obtained upon injection are shown in Fig. 1. The equation for the linear part of the calibration curve, i.e., from 0.005 to 0.3% v/v, was i = 13.1c 2 0.05, where i is the current in mA and c is the concentration in % v/v. The number of additions was seven and the correlation coefficient (r) was 0.9984. The standard deviations for the slope and intercept were 0.33 mA (% v/v)21 and 0.05 mA, respectively. The limit of detection was determined as about 0.003% v/v, based on a signal-to-noise ratio of three.As can be seen in Fig. 1, the detection potential was altered when more water was added to the acetone. Using a constant detection potential, i.e., 2.8 V, a linear calibration curve was obtained from 0.005 to about 0.07% v/v. At higher concentrations of water the response became non-linear, most likely as a result of the increased iR drop associated with increasing water concentration.As indicated above, the linear range could, however, be extended to about 0.3% v/v if the detection potential was made more positive. The detection potential range used in the linear range study for the oxidation of water in acetone containing 2.0 3 1023 m LiClO4 was 2.8–3.0 V. As can be seen in Fig. 1, the background current for a detection potential of 2.8 V was about 0.3 mA. This current stems from oxidation of trace compounds, mostly water and methanol, initially present in the acetone.The peak current after subtraction of this background current, i.e., 0.3 mA, was used in the evaluations of the data as the electrode potential was chosen according to the water concentration in the injected sample and was not based on the concentration of water present in the mobile phase. This was done since the more positive potential used resulted in a higher background current prior to and after the peak due to increased oxidation of acetone itself.As will be shown below, it may not be necessary to alter the detection potential if iR compensation is used and/or the range of water concentration studied is sufficiently small, e.g., when using determinations based on standard additions. The necessity to alter the detection potential to reach the steady state current was further studied by cyclic voltammetry. A shift in the half-wave potential was seen upon increasing the water concentration, as illustrated in Fig. 2, showing cyclic voltammograms for 0.05 and 0.25% v/v water in acetone. These experiments were performed in a stationary solution of acetone containing two different supporting electrolyte concentrations, i.e., 2.0 3 1023 and 1.0 3 1024 m LiClO4. As indicated above, the shift in the half-wave potential was a result of the larger oxidation current, and hence iR drop, associated with increasing water concentration. The magnitude of this shift therefore also depended on the concentration of ions in the solution.For higher water concentrations, the half-wave potential was shifted about 180 mV to more positive potential when the acetone contained 2.0 3 1023 m LiClO4 [Fig. 2(a)], and for the lower salt concentration [Fig. 2(b)] the corresponding shift in halfwave potential was about 370 mV. The larger shift in the latter case reflects the higher resistance in this solution but also the fact that fewer ions were available to form the ionic layer.For a lower salt concentration, the half-wave potential will hence be more positive regardless of the water concentration. As is also exemplified in Fig. 2, the steady state current for water oxidation was virtually independent of the supporting electrolyte concentration. This indicates that diffusion controlled currents were recorded and that the influence of migrational effects on the currents were negligible provided that an appropriate detection potential was chosen.This is also supported by the fact that the equations for the linear calibration curve for water concentrations in the range 0.05–0.3% v/v obtained in the flow system with acetone containing 2.0 31023 and 1.0 3 1024 m LiClO4 were in good agreement. In the latter case, the equation was found to be i = 13.4c + 0.11 (n = 6, r = 0.9996). The standard deviation for the slope and intercept were 0.20 mA (% v/v)21 and 0.04 mA, respectively. For an LiClO4 concentration of 1.0 3 1024 m, the range of detection potentials used for water additions in the range 0.05–0.3% v/v was 3.3–4.0 V.The linear range for oxidation of water in acetone was the same for the two LiClO4 concentrations employed. In both cases, the upper limit of the linear range was found to be set by the necessity to increase the detection potential when the water concentration was increased. When performing cyclic voltammetry in acetone in a stationary solution, it was seen that a water concentration higher than about 0.3% (v/v) required such a positive potential that also the solvent, i.e., acetone, started to be oxidised. After this point, the steady state plateau could never be reached for water oxidation, which resulted in deviations from the linear dependence between the current and the water Fig. 2 (a) Cyclic voltammograms for the oxidation of 0.05 and 0.25% v/v water in a stationary solution of acetone containing either 2.0 3 1023 or 1.0 3 1024 m LiClO4.The scan rate was 20 mV s21. 144 Analyst, 1999, 124, 143–146concentration. The lower limit of linearity and the detection limit could, most likely, be improved if purer acetone (with respect to water, methanol and other electroactive trace substances) was available, as small peaks would be easier to detect when the background signal is decreased. However, very dry acetone will be very sensitive to water contamination, predominantly from humid air, which may require analysis in a closed system.In order to compare the results obtained in the flow injection system with the corresponding results obtained in the batch mode, increasing amounts of water were added to a stationary solution of acetone. Although the linear range was the same as in the flow system for both 1.0 3 1024 and 2.0 3 1023 m LiClO4, the water oxidation current in the stationary solution of acetone was found to be about 25% smaller than that in the flow injection system. The equation for the calibration curve for water concentrations in the range 0.05–0.3% v/v obtained in a stationary solution of acetone containing 1.0 3 1024 m LiClO4 was i = 10.4c 2 0.003, n = 6 and r = 0.9999.The standard deviations for the slope and intercept were 0.07 mA (% v/v)21 and 0.01 mA, respectively. The higher currents obtained in the flow system were most likely a result of the convective mass transport contribution to the current. As described previously,26 the detection potential must be set higher in the flow system than in a stationary solution. Consequently, the detection potential range used for the water calibration curve in a stationary solution of acetone containing 1.0 3 1024 m LiClO4 was 3.1–3.9 V compared with 3.3–4.0 V in the flow system at a flow rate of 0.1 ml min21. The necessity to employ a higher detection potential was also influenced by the fact that higher currents, and hence iR drop, were obtained in the flow system. Acetone also contains traces of methanol which can be oxidised at the platinum microelectrode.As in the case of water, it was possible to obtain a linear curve for additions of methanol to a stationary solution of acetone also containing 2.0 3 1023 m LiClO4. A lower sensitivity was found for the oxidation current for methanol compared with that for water, most likely due to a lower diffusion coefficient for methanol. These results indicate that the present flow system also can be used for determinations of the methanol content in samples. Determination of the water content in butter and margarine To determine the water content in butter, a sample was first dissolved in acetone containing 2.0 3 1023 m LiClO4 to give a water concentration of approximately 0.015% v/v. A sample of margarine was treated in the same way as the butter sample.Standard addition was chosen for the evaluation of the water content in the samples to minimise the influence of matrix effects.Butter and margarine, for example, contain salt, the content of which varies from sample to sample. After analysing a sample from which the water had been removed by heat evaporation prior to dissolution in the acetone, it was concluded that no significant oxidation of fat from the butter and margarine samples was present. The influence of any oxidation of other compounds, e.g., vitamins, could also be ruled out as their concentrations in butter and margarine were much lower than that of water and as their diffusion coefficients most likely are smaller.A constant detection potential was used for the butter and margarine standard addition curves. In Fig. 3, the current peaks due to three injections of the margarine sample solution are displayed. In this case, it was possible to employ iR compensation by positive feedback to ensure a linear standard addition curve. The value of the resistance R used for the iRcompensation for the sample solution was 325 k½.The addition of water to the sample solutions sometimes resulted in current oscillations and, to avoid overcompensation, the value of the resistance had to be decreased. The range used for the iRcompensation for the standard addition curves was 325–225 k½. As the detection potential setting was constant, the peak height was used to evaluate the data. The relative standard deviation (RSD) of the peak heights was approximately 0.8% (n = 3). The equation for the standard addition curve resulting from four additions of water to the butter sample solution was i = 12.6c + 0.21, with standard deviations for the slope and intercept of 0.16 mA (% v/v)21 and 0.006 mA, respectively.The correlation coefficient was 0.9991 and the standard deviation for the extrapolation to the concentration axis was 0.66%. This gives a water content in the butter of 16.1 ± 2.1% (the uncertainty is given as a 95% confidence interval) which should be compared with the value of 17.7% (average based on two samples: X1 = 17.9% and X2 = 17.5%) obtained with the Karl Fischer method used for reference.The equation resulting from three additions of water to the margarine sample solution was i = 11.8c + 0.24. The standard deviations for the slope and intercept were in this case 0.24 mA (% v/v)21 and 0.007 mA, respectively, the correlation coefficient was 0.9996 and the standard deviation for the extrapolation to the concentration axis was 0.98%.This gives a water content in the margarine of 20.7 ± 4.2% which should be compared with the value of 17.3% (average based on two samples: X1 = 18.0% and X2 = 16.5%) obtained with the Karl Fisher method. The large uncertainties obtained for the butter and margarine determinations presented here stem from the less precise extrapolation procedure inherent in the standard addition technique.28 Nevertheless, the method presented in this work may be seen as an alternative approach for water determinations as it offers the possibility of obtaining a high sample throughput owing to injections in a flowing carrier and possibilities for incorporation in miniaturised systems, such as on-chip devices.There is also no need for an initial blank experiment as in the Karl Fischer method. In conclusion, the results indicate that amperometric determinations using a wall-tube microelectrode based flow injection system can be used for determinations of the water content of samples dissolved in acetone.Our findings also indicate that methanol can be determined in the same way. The magnitude of the currents and the linear ranges for the oxidation of water in acetone do not depend on the supporting electrolyte concentration provided that the detection potential can be set sufficiently positive to ensure mass transport controlled currents. If matrix effects from the sample can be made negligible, e.g., by adding more salt to the mobile phase and sample solvent, evaluation of the water content by a calibration curve would be possible.This would most likely decrease the uncertainties in the determina- Fig. 3 Peak currents due to three injections of the margarine sample dissolved in acetone containing 2.0 3 1023 m LiClO4. The detection potential was 2.5 V, the flow rate was 0.1 ml min21 and the distance between the tubing outlet and the electrode was 1.0 mm. iR compensation using R = 300 k½ was utilised.Analyst, 1999, 124, 143–146 145tions. Increasing the salt concentration and/or using a narrow water range for the calibration curve would also result in the possibility of employing a constant detection potential without the need for iR compensation. Acknowledgements This work was supported in part by a University of Warsaw grant (BST-562/5/97) and by grant No. K-AA/KU 09368-320 from the Swedish Natural Science Research Council. References 1 H. Gunasingham and B.Fleet, in Electroanalytical Chemistry, ed. A. J. Bard, Marcel Dekker, New York, 1989, vol. 16, pp. 89–180. 2 J. A. Cooper and R. G. Compton, Electroanalysis, 1998, 10, 141. 3 A. M. Bond, Analyst, 1994, 119, R1. 4 A. G. Ewing, J. M. Mesaros and P. F. Gavin, Anal. Chem., 1994, 66, 527A. 5 F.-M. Matysik, F. Björefors and L. Nyholm, Anal. Chim. Acta, in the press. 6 M. Zhong, J. Zhou, S. M. Lunte, G. Zhao, D. M. Giolando and J. R. Kirchhoff, Anal. Chem., 1996, 68, 203. 7 R. M.Wightman and D. O. Wipf, in Electroanalytical Chemistry, ed. A. J. Bard, Marcel Dekker, New York, 1988, vol. 15, pp. 267–353. 8 A. M. Bond, M. Fleischmann and J. Robinson, J. Electroanal. Chem., 1984, 168, 299. 9 M. Ciszkowska and Z. Stojek, J. Electroanal. Chem., 1986, 213, 189. 10 D. O. Wipf and R. M. Wightman, Anal. Chem., 1990, 62, 98. 11 S. A. Olsen and D. E. Tallman, Anal. Chem., 1996, 68, 2054. 12 C. J. Amatore, M. R. Deakin and R. M. Wightman, J. Electroanal. Chem., 1987, 220, 49. 13 K. B. Oldham, J. Electroanal. Chem., 1988, 250, 1. 14 K. Aoki, Electroanalysis, 1993, 5, 627. 15 S. R. Wallenborg, K. E. Markides and L. Nyholm, Anal. Chem., 1997, 69, 439. 16 D. R. Baker, M. W. Verbrugge and J. Newman, J. Electroanal. Chem., 1991, 314, 23. 17 A. M. Bond, M. Fleischmann and J. Robinson, J. Electroanal. Chem., 1984, 172, 11. 18 J. D. Norton, H. S. White and S. W. Feldberg, J. Phys. Chem., 1990, 94, 6772. 19 C. Amatore, B. Fosset, J. Bartelt, M. R. Deakin and R. M. Wightman, J. Electroanal. Chem., 1988, 256, 255. 20 J. B. Cooper and A. M. Bond, J. Electroanal. Chem., 1991, 315, 143. 21 M. Palys, Z. Stojek, M. Bos and W. van der Linden, Anal. Chim. Acta, 1997, 337, 5. 22 W. Hyk and Z. Stojek, J. Electroanal. Chem., 1997, 422, 179. 23 J. C. Myland and K. B. Oldham, J. Electroanal. Chem., 1993, 347, 49. 24 J. B. Cooper, A. M. Bond and K. B. Oldham, J. Electroanal. Chem., 1992, 331, 877. 25 J. Gadomska, M. Donten, Z. Stojek and L. Nyholm, Analyst, 1996, 121, 1869. 26 F. Björefors, J. Gadomska, M. Donten, L. Nyholm and Z. Stojek, in preparation. 27 M. Ciszkowska and Z. Stojek, Analyst, 1994, 119, 239. 28 J. C. Miller and J. N. Miller, Statistics for Analytical Chemistry, Ellis Horwood, Chichester, 2nd edn., 1992. Paper 8/06844E 146 Analyst, 1999, 124, 143–146

ISSN:0003-2654    

DOI:10.1039/a806844e

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Determination of Cl2, Br2, I2, Mn2+, malonic acid and quercetin by perturbation of a non-equilibrium stationary state in the Bray–Liebhafsky reaction Vladana B. Vukojevi�c,* Nataösa D. Peji�c, Dragomir R. Stanisavljev, Slobodan R. Ani�c and Ljiljana Z. Kolar-Ani�c Faculty of Physical Chemistry, University of Belgrade, P.O. Box 137, YU-11001 Belgrade, Yugoslavia Received 30th September 1998, Accepted 2nd December 1998 A new method applying a non-linear chemical system under conditions far from thermodynamic equilibrium in microvolume/microconcentration quantitative analysis is described.The chemical system used as a matrix is the Bray–Liebhafsky reaction in a non-equilibrium stationary state close to a bifurcation point. The method is based on monitoring the response of this system to perturbations by Cl2, Br2, I2, Mn2+, malonic acid and quercetin analyte solutions, which are followed potentiometrically either by an Ag+/S22 ion-sensitive or by a Pt electrode.A linear response of the potential shift versus the logarithm of the analyte concentrations is found in the following ranges: 1.3 3 1026 mol dm23 � [Cl2] � 1.6 3 1024 mol dm23, 1.0 3 1026 mol dm23 � [Br2] � 8.3 3 1025 mol dm23, 2.0 3 1026 mol dm23 � [I2] � 1.0 3 1024 mol dm23, 8.4 3 1027 mol dm23 � [Mn2+] � 8.3 3 1025 mol dm23, 3.8 3 1027 mol dm23 � [malonic acid] � 2.1 3 1025 mol dm23 and 1.5 3 1028 mol dm23 � [quercetin] � 3.7 3 1025 mol dm23. Under the investigated conditions an improved detection limit for all halides tested is obtained.Unknown concentrations of the analytes can be determined from a standard series of calibration curves to an accuracy within ±5%. In addition, the application of potentiometric measurements in microvolume/microconcentration quantitative analysis is diversified. Introduction A novel method in quantitative analysis that builds on specific features of non-linear chemical systems far from thermodynamic equilibrium1 is described.It relies on disturbing a delicate balance between the reaction’s intermediates, established in a non-equilibrium stationary state in the vicinity of a bifurcation point. Although advantages of the application of non-linear chemical systems under far from equilibrium conditions in quantitative analysis have been reported, e.g., the methodological approach termed the analyte pulse perturbation technique was described by Jiménez-Prieto et al.,2 so far only one nonequilibrium dynamic structure, viz., simple periodic oscillations, has been considered for such a purpose.2 We have pursued this approach further and, in so doing, we enhanced its ease of application.In particular, it is not necessary to test the oscillation phases for the best response to perturbations and to measure their frequencies. This simplifies the procedure and shortens considerably the time required for a full analysis. The aim of this paper is to introduce this method and to demonstrate its advantages.As the non-linear system in a state far from thermodynamic equilibrium, the Bray–Liebhafsky oscillatory reaction3,4 is used. We show that in this system the possibility of the quantitative determination of halides with the Ag+/S22 ion-sensitive electrode is improved. Moreover, the application of potentiometric measurements for quantitative analysis is extended to include species that cannot be determined directly by the applied sensor.Survey of the Bray–Liebhafsky reaction The Bray–Liebhafsky (BL) reaction involves the catalytic decomposition of hydrogen peroxide in the presence of H+ and IO32:3,4 H+, IO32 2H2O2 ––––––? O2 + 2H2O This deceptively simple reaction proceeds through a complex mechanism involving a number of intermediates, such as I2, I2, HOI and HIO2.5 During the course of the reaction they undergo oxido-reductions and their concentrations, and also those of all other intermediates, are kept several orders of magnitude lower than the concentrations of the reactants by the delicately balanced reaction mechanism.It is one of the longest known reactions to exhibit non-linear properties, self-organisation temporal dynamic structures such as non-equilibrium stationary states, simple and complex oscillations, when driven under conditions far from thermodynamic equilibrium.1b,4,6 The dynamic structures achieved in the BL reaction can be sustained over an extended period of time by continuously supplying the reactants and removing the surplus volume of the reaction mixture, as is done in the CSTR (continuously fed well stirred tank reactor).1,6 Hence, the system remains in the same stable state as long as all parameters are kept constant.1 A dynamic structure established in the CSTR may be intentionally changed by varying a control parameter such as temperature, Analyst, 1999, 124, 147–152 147specific flow rate† or mixed inflow concentrations of the reactants.‡ Consequently, transition of one dynamic structure to another, i.e., a bifurcation,§ can occur.1,6,7 In the BL system, the bifurcation, like a transition from sustained oscillations to a stable stationary state, for example, occurs at a certain critical value of temperature as the control parameter (see Fig. 1). In the vicinity of a bifurcation point the system is found in the vicinity of two nearby attractors.¶ This creates an extremely fragile balance, which is susceptible to perturbations.Even very small changes in the concentrations of the intermediates, caused by addition of the analytes, may disturb this balance and induce detectable changes in the dynamic pattern. Experimental We utilized a version of the Bray–Liebhafsky (BL) reaction, in which potassium iodate is the source of IO32 and sulfuric acid is the source of H+. The reaction was carried out in a CSTR. The reactants, i.e., aqueous solutions of KIO3, H2SO4 and H2O2, were introduced into the reaction vessel made of glass as three separate inflows.The volume of the reaction mixture was kept constant at V = 24.0 ± 0.2 mL, by suction through a U-shaped glass tube. The reaction mixture and the gaseous phase above it were removed at a volume flow rate of 5.8 mL min21. The flows were generated by peristaltic pumps (Ole Dich, Copenhagen, Denmark). The temperature was controlled within ±0.2 K by a circulating water thermostat (Serie U, MLW Freital, Germany).The experimental conditions were as follows: mixed inflow concentrations of the reactants, [KIO3]0 = 5.9 3 1022 mol dm23, [H2SO4]0 = 5.5 3 1022 mol dm23 and [H2O2]0 = 2.0 3 1021 mol dm23; specific flow rate, j0 = 2.96 3 1022 min21 (equivalent to a residence time of t = 33.8 min); and stirring rate, 900 rpm. Temporal evolution of the system was monitored potentiometrically (MA 5730 potentiometer, Iskra, Hojrul, Slovenia) by a solid-state membrane Ag+/S22 ion-sensitive electrode (Model 6.0502.180, Metrohm, Herisau, Switzerland) versus a double junction Ag/AgCl electrode (Metrohm Model 6.0726.100) as a reference.The inner solution was 3 mol dm23 KCl, and the outer solution was saturated K2SO4. Occasionally a Pt electrode (Metrohm Model 6.0301.100) was used as a monitoring sensor. The potential output was recorded on a chart recorder (Servogor 220, BBC Goerz, Nuenen, Austria). The start-up procedure was performed in the following way.First, the thermostated reaction vessel was filled by the three separate inflows of the reactants at the maximum speed (16 mL min21). It was allowed that double the reaction mixture volume (48 mL) flowed into the vessel (3 min at the maximum flow rate). The inflows were then stopped and the excess of the reaction mixture was sucked out through the U-shaped glass tube. Hence, the reaction commenced under batch conditions. After three batch oscillations (about 40 min), the inflows were turned on at the required flow rate.Under the investigated conditions it is sufficient to wait for three residence times (about 1.7 h) to obtain sustained oscillations at 333.2 K. To find a bifurcation point from the sustained oscillations to a stationary state, the temperature, which was the control parameteries showing dynamic structures observed under the above conditions at different temperatures are presented in Fig. 1(a) and a bifurcation diagram showing the transition from sustained oscillations to a stable stationary state with decreasing temperature is presented in Fig. 1(b). The bifurcation point was found at 321.8 K by linear extrapolation of a plot of the square of the amplitude of the limit cycle oscillations versus temperature [Fig. 1(c)]. It is not necessary to determine the bifurcation point freshly each time. Once a bifurcation point has already been determined, in the subsequent routine analysis it is sufficient to reduce the temperature to the working value immediately after achieving the sustained oscillations at 333.2 K.In this way, the preparatory procedure takes about 3 h. Since the bifurcation point was found at 321.8 K, we have chosen to work as close as possible to it, that is, at 319.6 K. At this temperature the system is still sufficiently close to the bifurcation point that all exceptional advantages are preserved. On the other hand, it is sufficiently far from it that small spontaneous perturbations will not induce transitions to the oscillatory side of the bifurcation point.Perturbations were performed by additions of microvolumes of the analyte solutions by micropipettes (Transferpette, Brand, Wertheim, Germany). Volumes from 5 to 500 ml were tested. Perturbations with volumes larger than 500 ml were not reliable, under the current set-up, because the dynamic pattern may be additionally changed due to dilution.Not only the amount of the injected analyte but also the rate and the duration of that injection may have an effect.4k Therefore, fast and reproducible injection is required. We applied manual injections of approximate duration 0.5 s. The time required after each perturbation to reach the initial stationary state depends on the concentration of the analyte and its chemical characteristics (Fig. 2). Too frequent perturbations, i.e., perturbation before the system has returned to the initial † The specific flow rate (j0) is a parameter in the CSTR which is equal to the overall volume flow rate (expressed in units of volume per unit time) divided by the volume of the reaction mixture.It has units of (time)21 and so is like a first order rate constant. By introducing this parameter, the flow rates of all species can be described as first order reactions. Its reciprocal value is called the residence time (t) and equals the time required for a total replacement of the reaction mixture by the flow.1 ‡ Mixed inflow concentrations of the reactants are their concentrations obtained after mixing the separate inflows, before any chemical reaction has taken place.1.§ A qualitative change of one dynamic structure into another caused by variation of the control parameter is called a bifurcation.1,7 ¶ An intuitive definition of a stable attractor is that it is essentially the limit structure to which a system is ‘attracted’ after a long time.There are three basic types of attractors: point attractors, corresponding to systems reaching a stable equilibrium (stationary state), periodic attractors, corresponding to periodic oscillations, and strange attractors, corresponding to chaotic systems.1,7 Fig. 1 (a) Time series showing stable dynamic structures observed in the BL reaction at different temperatures, under the following experimental conditions: [KIO3]0 = 5.9 3 1022 mol dm23, [H2SO4]0 = 5.5 3 1022 mol dm23 and [H2O2]0 = 2.0 3 1021 mol dm23, specific flow rate j0 = 2.96 3 1022 min21 and stirring rate = 900 rpm.The temperature was controlled within ±0.2 K. (b) Bifurcation diagram showing transition from sustained periodic oscillations (open circles) to a stable stationary state (solid circles) for decreasing temperature. (c) Plot of the square of the oscillation amplitudes (top to bottom) as a function of temperature. The bifurcation point is determined from the intercept on the abscissa. 148 Analyst, 1999, 124, 147–152stationary state, should be avoided to avert any possible cumulative effect. For example, in the case of perturbations by Cl2, Br2, I2 and malonic acid, establishment of the individual calibration curves requires about 2 h. Cl2, Br2 and I2 were introduced into the system as aqueous solutions of the corresponding potassium salts; Mn2+ as an aqueous solution of MnSO4·H2O and malonic acid (C3H4O4) as its aqueous solution. Quercetin (3,3A,4A,5,7-pentahydroxyflavone) was introduced as a solution of quercetin dihydrate (C15H10O7·2H2O) in ethanol.Addition of ethanol alone does not perturb the BL system. Points on the calibration curves were obtained from two independent additions, for all series of standard solutions. The best linear fit of the experimental points was determined by the least-squares method. For determination of unknown concentrations from the standard series calibration curves, an accuracy within ±5% was obtained for all species in the concentration ranges investigated.Reagents All reagents were of analytical-reagent grade, and were used without further purification. KIO3, H2SO4, H2O2, KCl, KBr, KI, MnSO4·H2O, C3H4O4 and 95% C2H5OH were obtained from Merck (Darmstadt, Germany) and C15H10O7·2H2O from Fluka (Deisenhofen, Germany). De-ionized water (Milli-Q, Millipore, Bedford, MA, USA), with a specific resistance of 18 MWcm21, was used as a solvent throughout. Results All results obtained by perturbing a non-equilibrium stationary state established in the BL reaction are discussed in turn in this section.Other important results regarding the electrode calibrations and response of the Ag+/S22 ion-sensitive electrode are gathered in the Appendix. By perturbations of non-equilibrium stationary states in the BL reaction it is possible to determine quantitatively only the species that interfere with the reaction system and hence influence its mechanism.Therefore, in addition to I2, which is a constituent intermediate of the BL system, other analytes, viz., Cl2, Br2, Mn2+, malonic acid and quercetin, were also tested. Since, for a long time, we have had positive results in monitoring the BL reaction with the Ag+/S22 ion-sensitive and Pt electrodes,4h–n,8 for the determination of microvolumes/ microconcentrations of the examined species by perturbing a non-equilibrium stationary state in the same reaction only these electrodes were used.Further, it is known that the Ag+/S22 ionsensitive electrode has good sensitivity to all investigated halides9 and an extended working temperature range compared with other Ag+/X2 electrodes.9a The system’s response to additions of microvolumes of all the analytes examined is given on the left side of Fig. 2. As can be seen, the system’s response to the perturbations results in a readily detectable change in the potential. To evaluate the experimental results obtained we analysed the potential response curves presented on the left side of Fig. 2, i.e., we separated the response of the monitoring sensor introduced by the perturbation from the following system’s response to that perturbation. As can be seen, an initial overshoot–decay response is observed only for perturbations with Br2 and larger amounts of I2 (Fig. 2). This initial response, due to the monitoring sensor, is irreproducible and cannot be considered for quantitative determination (see Appendix).Therefore, Ep and Es, where Ep is the potential after the perturbation and Es is the potential of the stationary state, are used for analytical purposes as representative potential values. The change in potential, defined as the difference DE = Ep 2 Es, is proportional to the analyte concentration (left side of Fig. 2). The linear dependence obtained between the potential shift DE and the logarithm of the analyte concentration is presented on the right side of Fig. 2. Detection limits obtained in this way are given in Table 1. By comparison of the data given in Table 1, one can see that for all halides considered an improved detection limit is achieved. We underline that these remarkable results are not due to the electrode, but rather to the BL system used as the matrix. Moreover, it is shown that the same system used as a matrix is very convenient for quantitative determination of Mn2+, malonic acid and quercetin by the same monitoring sensor, although the monitoring sensor does not respond to any of them directly.The obtained detection limits are given in Table 1. Discussion The obvious dependence between the added amount of the analytes and the provoked potential shift (Fig. 2) enables using the proposed method, under the defined conditions, for quantitative determination. Although, knowing the reaction’s mechanism is not necessary for practical application, we would like to give some general features of this subject together with a review of the relevant literature.Fig. 2 Typical response curves obtained after perturbing a stationary state in the BL reaction by addition of microvolumes of the analytes, together with the thus obtained standard series of calibration curves. The analyte concentrations (from left to right) are as follows: [Cl2] = 1.2 31026; [Cl2] = 1.5 3 1025; [Cl2] = 1.0 3 1024; [Br2] = 1.0 3 1026; [Br2] = 4.2 3 1026; [Br2] = 8.4 3 1026; [I2] = 2.0 3 1026; [I2] = 2.1 3 1025; [I2] = 6.2 3 1025; [Mn2+] = 8.3 3 1027; [Mn2+] = 4.2 3 1026; [Mn2+] = 1.7 3 1025; [MA] = 3.8 3 1027; [MA] = 2.1 3 1026; [MA] = 3.1 3 1025; [Q] = 1.5 3 1028; [Q] = 1.5 3 1027; and [Q] = 3.1 3 1025 mol dm23.MA = malonic acid and Q = quercetin. In experiments with quercetin, a Pt electrode was used as the monitoring sensor. Analyst, 1999, 124, 147–152 149The most important fact is that the chosen potential shift DE is due to the system’s response to the applied perturbations but not the monitoring sensor.This is obvious for Mn2+, malonic acid and quercetin, since the monitoring sensor does not respond to any of them directly. In the case of Cl2, Br2 and I2 the applied monitoring sensor can respond directly to their concentration changes under equilibrium conditions (see Appendix), but this is not the case under non-equilibrium conditions of the BL reaction. Under such conditions the slopes of the DE = f(log[X2]) curves differ significantly from the corresponding one obtained under equilibrium conditions. In particular, in the BL system the slope in respect to the logarithm of concentration in the case of Cl2 is 7.7 mV, Br2 is 2.2 mV and I2 is 8.4 mV, compared to 263.8 mV obtained in all cases under the reference equilibrium conditions. It is evident that changes in the chemical mechanism occur owing to the addition of the analytes to the system.Chemical interactions with analytes change the ratio between [HIO]ss and [I2]ss established in the stationary state and this change is followed by the monitoring sensor (see Appendix).This is valid in all cases although the explanation would differ from one case to the other. Perturbations of the BL reaction by I2 are different from perturbations with any other species discussed later, since I2 is an intermediate of the investigated system. Because of this, no new chemical species are formed after injecting I2 and the system evolves along its inherent trajectories.The subsequent dynamic response is entirely governed by the chemical mechanism underlying the BL reaction. Addition of analytes which are not originally present in the matrix BL reaction will inevitably introduce new chemical reactions and species. In fact, we shall have a new chemical dynamic system. The relevant explanations, different in all cases considered, are very difficult since the mechanism of the BL system as the matrix to which we add the examined species is extremely complex.5 The numerous explanations related to the mechanism of the BL reaction5 and combinations of the BL system with the examined species,3a,4a,4b,10 except quercetin, illustrate the complexity of the problem.To our knowledge, the influence of quercetin or any other hydroxyflavone on the BL reaction has not been investigated. Some hydroxyflavones have been used as organic substrates in the BL reaction, but their reaction mechanisms were not discussed.11 The low detection limit for quercetin obtained by this method may perhaps be interesting for the determination of hydroxyflavones in general.Anyhow, the detailed mechanism of the reaction considered is outside the scope of this paper. Our aim, namely the determination of unknown concentrations of the considered analytes under non-equilibrium conditions in the vicinity of a bifurcation point in the BL system, was successfully achieved.The advantages of perturbing a stationary state in the vicinity of a bifurcation point, rather than any other non-equilibrium dynamic structure, will be well appreciated by investigators with little experience with non-linear systems. Here, any change in the dynamic pattern following the perturbation is easily observable and unmistakably recognized (Fig. 2). Perturbations of more complex dynamic structures may be less pronounced; for example, a phase shift may occur,2,12 the response may be phase specific,2,13 the period of the oscillations may change2,13b or the dynamic structure may change from subharmonic to periodic.14 Such subtle changes require a confident and experienced experimentalist to recognize them.In addition, longer recordings of the unperturbed and perturbed system are required to relate quantitatively the observed response to the perturbation. The estimated time for a full analysis is dependent on the system’s relaxation rate to the initial stationary state, after perturbation with the analyte, and its distance from the bifurcation point.In general, the time required may be shortened by working slightly further from the bifurcation point, but a lower sensitivity should be expected as a consequence. Conclusion We have described a new way of applying non-linear chemical systems under conditions far from thermodynamic equilibrium in quantitative analysis. Specific features of such systems that have been exploited are the existence of non-equilibrium stationary states and their extreme sensitivity to perturbations in the vicinity of a bifurcation point.The reaction used as a matrix is the well known Bray– Liebhafsky (BL) oscillating reaction. A stationary state in the BL reaction in the vicinity of a bifurcation point is perturbed by additions of microvolumes of the analytes. The change induced in the potential is proportional to the analyte concentration. Advantages of the described method have been demonstrated for the determination of Cl2, Br2, I2, Mn2+, malonic acid and quercetin.Under the non-equilibrium conditions established in the BL reaction, an improved detection limit for all tested halides with the Ag+/S22 ion-sensitive electrode is obtained. The practical detection limits for these species by the described method are: 1.3 3 1026 mol dm23 for [Cl2], 1.0 3 1026 mol dm23 for [Br2] and 2.0 3 1026 mol dm23 for [I2]. The detection limit for Cl2 is about 30 times lower than the theoretical detection limit of the applied monitoring sensor! The practical detection limit for Br2 approaches the theoretical limit and an improved sensitivity for I2 is achieved.The applicability of potentiometric measurements in microvolume/ microconcentration quantitative analysis is extended to species which cannot be determined directly by the applied sensor. This is demonstrated for the following concentration ranges: 8.4 3 1027 mol dm23 � [Mn2+] � 8.3 3 1025 Table 1 Concentration detection limits for the investigated analytes [Cl2] [Br2] [I2] [Mn2+] [Malonic acid] [Quercetin]b Detection limitsa obtained by perturbing a non-equilibrium stationary state established in the BL reaction at 46.4 °C Practical detection limit/mol dm23 1.3 3 1026 1.0 3 1026 2.0 3 1026 8.4 3 1027 3.8 3 1027 1.5 3 1028 Detection limits under equilibrium conditions Theoretical detection limit of Ag+/X2 electrodes at 50 °C 9a/mol dm23 4.5 3 1025 2.8 3 1026 7.9 3 1028 — — — Practical detection limit of Ag+/X2 electrodes at 46.4 °C in sulfuric acid/mol dm23 (see Appendix) 3.2 3 1023 7.1 3 1025 5.0 3 1025 — — — Practical detection limit of Ag+ electrodes at 46.4 °C in sulfuric acid/mol dm23 (see Appendix) 1.3 3 1022 6.3 3 1024 9.0 3 1026 — — — a The detection limit is defined as the concentration of the tested species which will produce a signal-to-noise ratio of 3 (see the response curves in Fig. 2, obtained for the lowest concentration of each species).b A Pt electrode was used as a monitoring sensor. 150 Analyst, 1999, 124, 147–152mol dm23, 3.8 3 1027 mol dm23 � [malonic acid] � 2.1 3 1025 mol dm23 and 1.5 3 1028 mol dm23 � [quercetin] � 3.7 3 1025 mol dm23. Concentrations of all tested analytes in a sample may be obtained from a standard series of calibration curves, within an accuracy of ±5%. A microvolume of the sample, as small as 20 ml, may suffice for a complete analysis.Considering the widespread applications of potentiometric measurements in general chemistry laboratories and ionsensitive electrodes for the determination of halides in different scientific and applied branches, we believe that the proposed method with a significantly improved performance will be suitable in this area. Appendix Theoretically, the response of Ag+/X2 ion-sensitive electrodes to halides under equilibrium conditions is described by the equilibrium Nernst equation: E E RT F e X = - � � - ° ln log[ ] 10 (1) where Ee is the zero-current cell potential, E° is the standard cell potential, R is the gas constant, F is the Faraday constant, T is absolute temperature and [X2] is the concentration of the corresponding halide (note that activities are replaced by concentrations).The detection limit is restricted by the solubility product of the corresponding silver halides AgX of which the membranes are made.9 However, since it is difficult to perform measurements close to the detection limit, in practice the practical detection limit is usually at higher concentrations. 9 We tested the sensitivity of Ag+/X2 and Ag+/S22 ionsensitive electrodes to halides. Calibration of the electrodes was performed in an acidic medium (H2SO4) at a fixed ionic strength I = 0.45 m. The ionic strength was adjusted with K2SO4. The results of a standard series of measurements are presented in Fig. 3. As can be seen, when an increase in concentration of halides is introduced by sudden addition, a non-monotonic transient response, i.e., an overshoot–decay response, is usually observed.A typical response curve, magnified in the inset in Fig. 3, shows that this short excursion is followed by a steady potential value (Ee) corresponding to the concentration of halides in the solution. This feature of precipitate-based ionsensitive electrodes is well known, and it is determined by both adsorption–desorption and surface reactions of the membrane. 15 The subsequently attained potential shows a Nernstian response down to the concentration values given in Table 1, which determine the practical detection limit. For concentrations lower than the practical detection limit, the sensitivity of the electrodes is no longer in accordance with the theoretical sensitivity, i.e., it is not possible to perform direct measurements below the practical detection limit. Although it is not necessary, for practical purposes, to understand how it is possible to determine concentrations lower than the practical, or even theoretical, detection limit of the applied monitoring sensor, it is important to try to analyse the prerequisites for this extraordinary feature under non-equilibrium conditions. It is known that in certain oxy-halogen oscillatory reactions the established potential of the applied ion-sensitive electrodes is up to several hundred millivolts higher than attainable under corresponding equilibrium conditions, and would correspond to concentrations of halides lower than allowed by the solubility product limit.8,16,17c,17d To explain this experimental fact, two main explanations evolved: the kinetic–buffer theory (KBT)16 and the corrosion–potential theory (CPT).17 Therefore, we also tested the response of the Ag+/S22 ion-sensitive electrode to hypoiodous acid and some other relevant species under equilibrium conditions.These calibrations were performed in an acidic medium (4.0 3 1022 mol dm23 H2SO4) at 325.2 K with standard solutions of I2, Ag+, IO32 and KMnO4, whereas the calibration with HIO was performed by the procedure described by Noszticzius et al.17c The results obtained are presented in Fig. 4.As can be seen from Fig. 4, the Ag+/S22 ion-selective electrode responds to all tested species, and a Nernstian response was obtained for all of them except IO32 in a certain concentration range. Of particular importance are the results obtained with KMnO4.Calibration with KMnO4 was performed both on a new Ag+/S22 electrode and on an electrode used for several years for Fig. 3 Standard series of calibration curves for halides, obtained under equilibrium conditions. Calibration of the electrodes was performed in an acidic medium, [H2SO4]0 = 5.5 3 1022 mol dm23, with fixed ionic strength I = 0.45 m, at 321.8 K. The ionic strength was adjusted with K2SO4. Solid symbols present the response of the Ag+/S22 ion-sensitive electrode to chlorides (È, Ó), bromides (½, :) and iodides (2, 5).Open symbols present the response of Ag+/X2 ion-sensitive electrodes to the corresponding halides (same symbol representation is used). The slopes of the linear parts agree to within ±1 mV with the theoretical value 2(RT/ F)ln10 = 263.8 mV. Magnified in the inset is a typical response curve of the ion-sensitive electrodes obtained when a sudden increase in the concentration of halides is introduced by addition.It shows a short nonmonotonic transient response, i.e., an overshoot–decay response, followed by a steady potential value (Ee) corresponding to the actual concentration of halides in the solution. Fig. 4 Standard series of calibration curves of the Ag+/S22 ion-sensitive electrode for I2, Ag+, IO32, HIO and KMnO4, obtained under equilibrium conditions. Calibration of the electrode was performed in an acidic medium, [H2SO4]0 = 4.0 3 1022 mol dm23, with fixed ionic strength I = 0.24 m, at 325.2 K.Solid circles present the response to I2, crosses to Ag+, diamonds to IO32, open circles to HIO and squares to KMnO4. The slopes of the linear parts agree to within ±0.5 mV with the corresponding theoretical values of 2(RT/nF)ln10, except for IO32. Analyst, 1999, 124, 147–152 151monitoring of the BL reaction. These parallel experiments were performed for comparison since the membrane surface can be altered after being exposed to conditions prevailing in the BL system.8 There is no difference between their response to KMnO4.The results obtained unambiguously confirm that the electrode surface can be corroded by strong oxidizing agents. Furthermore, the number of exchanged electrons in the corrosion reaction: Mn7+ + 5e ––? Mn2+ is correctly reflected by the slope of the calibration curve of 12.8 mV compared with the theoretical slope RT/5F = 12.9 mV. We therefore conclude that excess of Ag+, remaining on the membrane surface after corrosion, determines the electrode potential.It is also evident from Fig. 4 that the calibration curve for HIO overlaps that determined for Ag+, indicating that corrosion with HIO is indeed possible. Also in the corrosion reaction one electron is exchanged as assumed from the corrosion reaction: H+ + X2 + HXO ––? X2 + H2O Although IO32 is a strong oxidizing agent, it does not corrode the electrode surface. Iodous acid (HIO2) also influences the electrode potential through corrosion, but at a slower rate compared with corrosion with HIO.17c,17d Hence we conclude that the Ag+/S22 ion-sensitive electrode is sensitive to both I2 and HIO under equilibrium conditions.When both species are present, which is the case in the BL system, the potential of the electrode attained depends on the actual amounts of HIO and I2 present. When the investigated analytes are introduced into the BL system, they take part in the underlying mechanism, changing the previously established ratio of [HIO]ss and [I2]ss. Asare important intermediates in the investigated system, the potential shift due to changes in their concentrations reflects the system’s response to the applied perturbation.References 1 (a) G. Nicolis and I. Prigogine, Self-Organization in Nonequilibrium Systems, Wiley, New York, 1977; (b) Oscillations and Traveling Waves in Chemical Systems, ed. R. J. Field and M. Burger, Wiley, New York, 1985; (c) G.Nicolis and I. Prigogine, Exploring Complexity, Freeman, San Francisco, 1989; (d) P. Gray and S. Scott, Chemical Oscillations and Instabilities: Nonlinear Chemical Kinetics, Oxford University Press, Oxford, 1990; (e) G. Nicolis, Introduction to Nonlinear Science, Cambridge University Press, Cambridge, 1995. 2 (a) R. Jiménez-Prieto, M. Silva and D. Pérez-Bendito, Anal. Chem., 1995, 67, 729; (b) R. Jiménez-Prieto, M. Silva and D. Pérez-Bendito, Anal.Chim. Acta, 1996, 321, 53, and references cited therein. 3 (a) W. C. Bray, J. Am. Chem. Soc., 1921, 43, 1262; (b) W. C. Bray and H. A. Liebhafsky, J. Am. Chem. Soc., 1931, 53, 38. 4 (a) M. G. Peard and C. F. Cullis, Trans. Faraday Soc., 1951, 47, 616; (b) H. Degn, Acta. Chem. Scand., 1967, 21(4), 26; (c) I. Matsuzaki, J. H. Woodson and H. A. Liebhafsky, Bull. Chem. Soc. Jpn., 1970, 43, 3317; (d) H. A. Liebhafsky and L. S. Wu, J. Am. Chem. Soc., 1974, 96, 7180; (e) K. R.Sharma and R. M. Noyes, J. Am. Chem. Soc., 1975, 97, 202; (f) K. R. Sharma and R. M. Noyes, J. Am. Chem. Soc., 1976, 98, 4345; (g) J. A. Odutola, C. A. Bohlander and R. M. Noyes, J. Phys. Chem., 1982, 86, 818; (h) S. Ani�c and Lj. Kolar-Ani�c, Ber. Bunsenges. Phys. Chem., 1986, 90, 539; (i) S. Ani�c and Lj. Kolar-Ani�c, J. Chem. Soc., Faraday Trans. 1, 1988, 84, 3413; (j) S. Ani�c, D. Stanisavljev, G. Krnajski Belovljev and Lj. Kolar-Ani�c, Ber. Bunsenges. Phys. Chem., 1989, 93, 488; (k) G.Schmitz, in Spatial Inhomogeneities and Transient Behavior in Chemical Kinetics, ed. P. Gray, G. Nicolis, P. Borckmans and S. K. Scott, Manchester University Press, Manchester, 1990, p. PA16; (l) S. Ani�c, ö Z. ö Cupi�c and J. � Ciri�c, in Physical Chemistry ’94, ed. S. Ribnikar, The Society of Physical Chemists of Serbia, Belgrade, 1994, pp. 141–142; (m) D. Stanisavljev and V. Vukojevi�c, J. Serb. Chem. Soc., 1995, 60, 1125; (n) S. Ani�c, Lj. Kolar-Ani�c and E.Ko"rös, React. Kinet. Catal. Lett., 1996, 57, 37; (o) S. Ani�c, D. Stanisavljev, ö Z. ö Cupi�c, M. Radenkovi�c, V. Vukojevi�c and Lj. Kolar-Ani�c, Sci. Sintering, 1998, 30, 49, and references cited therein. 5 (a) H. A. Liebhafsky, W. C. McGavock, R. J. Reyes, G. M. Roe and S. L. Wu, J. Am. Chem. Soc., 1978, 100, 87; (b) D. Edelson and R. M. Noyes, J. Phys. Chem., 1979, 83, 212; (c) G. Schmitz, J. Chim. Phys. 1987, 84, 957; (d) Lj. Kolar-Ani�c and G. Schmitz, J. Chem. Soc., Faraday Trans., 1992, 88, 2343; (e) L.Treindl and R. M. Noyes, J. Phys. Chem., 1993, 97, 11354; (f) Lj. Kolar-Ani�c, Dj. Miösljenovi�c, S. Ani�c and G. Nicolis, React. Kinet. Catal. Lett., 1995, 54, 35; (g) Lj. Kolar-Ani�c, ö Z. ö Cupi�c, S. Ani�c and G. Schmitz, J. Chem. Soc., Faraday Trans., 1997, 93, 2147; (h) Lj. Kolar-Ani�c, ö Z. ö Cupi�c and S. Ani�c, Chem. Ind., 1998, 52, 337, and references cited therein. 6 (a) J. Chopin-Dumas, C. R. Acad. Sci., Ser. C, 1987, 287, 553; (b) J.Chopin-Dumas and M. N. Papel, in Synergetics: Non-Equilibrium Dynamics in Chemical Systems, ed. C. Vidal and A. Pacault, Springer, New York, 1984, pp. 69–73. 7 P. G. Drazin, Nonlinear Systems, Cambridge University Press, Cambridge, 1994. 8 (a) S. Ani�c and Lj. Kolar-Ani�c, Ber. Bunsenges. Phys. Chem., 1986, 90, 1084; (b) S. Ani�c and D. Miti�c, J. Serb. Chem. Soc., 1988, 53, 371, and references cited therein. 9 (a) ISE Instructions for Use, Metrohm, Herisau, 1988, p. 41; (b) J. Vesle�y, O. J. Jensen and B. Nicolaisen, Anal. Chim. Acta, 1972, 62, 1; (c) R. A. Durst and B. T. Duhart, Anal. Chem., 1970, 42, 1002. 10 (a) S. D. Furrow and R. M. Noyes, J. Am. Chem. Soc., 1982, 104, 38; (b) H. D. Försterling and Z. Noszticzius, J. Phys. Chem., 1989, 93, 2740; (c) S. Ani�c and D. Miti�c, Glasnik na Hemicarite i Tehnolozite na Makedonija, 1989, 7, 203; (d) D. Edelson, R. M. Noyes and R. J. Field, Int. J. Chem. Kinet., 1979, 11, 155; (e) A. N. Kappanna and E. R. Talaty, J. Indian Chem. Soc., 1951, 28, 675; (f) K. R. Leopold and A. Haim, Int. J. Chem. Kinet., 1977, 9, 83; (g) G. Davies, J. Kirchenbaum and K. Kustin, Nature (London), 1968, 7, 146. 11 V. Sridevi and R. Ramaswamy, in Polyphenols Communications 96, eds. J. Vercauteran, C. Cheze, M. C. Dumon and J. F. Weber, Universit�e Bordeaux, 1996, pp. 139–140. 12 J. Kosek, P. Graae Sørensen, M. Marek and F. Hynne, J. Phys. Chem., 1994, 98, 6128. 13 (a) V. Vukojevi�c, P. Graae Sørensen and F. Hynne, J. Phys. Chem., 1993, 97, 4091; (b) J. Stemwedel, J. Ross and I. Schreiber, Adv. Chem. Phys., 1995, 89, 327. 14 P. Graae Sørensen, F. Hynne and K. Nielsen, React. Kinet. Catal. Lett., 1990, 42, 309. 15 T. R. Berube, R. P. Buck, E. Lindner, M. Gratzl and E. Pungor, Anal. Chem., 1989, 61, 453. 16 (a) J. H. Woodson and H. A. Liebhafsky, Anal. Chem., 1969, 41, 1894; (b) R. J. Field, E. K�orös and R. M. Noyes, J. Am. Chem. Soc., 1972, 94, 8649; (c) N. Ganapathisubramanian and R. M. Noyes, J. Phys. Chem., 1982, 86, 3217, and references cited therein. 17 (a) R. P. Buck, Anal. Chem., 1968, 40, 1432; (b) R. P. Buck, Anal. Chem., 1968, 40, 1439; (c) Z. Noszticzius, E. Noszticzius and Z. A. Schelly, J. Am. Chem. Soc., 1982, 104, 6194; (d) Z. Noszticzius, E. Noszticzius and Z. A. Schelly, J. Phys

ISSN:0003-2654    

DOI:10.1039/a807608a

出版商:RSC    

年代:1999

数据来源: RSC