|
21. |
Development of a Radioimmunoassay for the Determination of Zolpidem in Biological Samples |
|
Analyst,
Volume 122,
Issue 10,
1997,
Page 1119-1124
Isabel De Clerck,
Preview
|
|
摘要:
Development of a Radioimmunoassay for the Determination of Zolpidem in Biological Samples Isabel De Clerck and P. Daenens* Laboratory of Toxicology, Katholieke Universiteit Leuven, E. Van Evenstraat, 4, B-3000 Leuven, Belgium The development of a specific and sensitive radioimmunoassay for the detection of zolpidem and its metabolites in urine and serum samples is described. The assay can be used to pre-screen forensic and emergency samples. 6-Methyl-2-(4-methylphenyl)imidazo[1,2-a]- pyridine-3-acetic acid was prepared as a hapten and was coupled directly and indirectly to bovine serum albumin.The immunization of rabbits with the hapten–bovine serum albumin conjugates resulted in the production of highly specific antibodies, showing no significant cross-reactivities towards existing drugs. The fourparameter logistic model was used to process the calibration data into a fitting curve (r2 = 0.9764). Intraand inter-assay relative standard deviations were < 6.30 and < 12.28%, respectively.The limit of quantification was 0.1 ng ml21. Using this assay, zolpidem levels were determined in urine and serum, and could be easily detected up to 48 h after oral intake of 10 mg of zolpidem. Keywords: Radioimmunoassay; zolpidem; imidazo[1,2-a]pyridine Zolpidem [Fig. 1(a)], N,N,6-trimethyl-2-(4-methylphenyl)- imidazole[1,2-a]pyridine-3-acetamide l-(+)-hemitartrate, is a non-benzodiazepine hypnotic drug with an imidazopyridine backbone, which acts in the brain principally at receptors of the w1-receptor subtype belonging to the g-aminobutyric acid- (GABA)ergic system.Its main features are a fast onset of action and a short elimination half-life (2.4 h), unlike benzodiazepines. Stilnoct is administered orally (therapeutic dose = 10 mg). Therapeutic plasma concentrations are in the low nanogram range with average peak plasma levels of about 140 ng ml21 after 0.5–3 h.1 The drug is extensively metabolized and excreted in the urine as pharmacologically inactive metabolites.The biotransformation proceeds through three different pathways including (i) methyl oxidation on the phenyl moiety of the molecule (metabolite I, the main metabolite in urine that accounts for 52% of the administered dose), (ii) methyl oxidation on the imidazopyridine moiety leading to the corresponding carboxylic acid (metabolite II, that accounts for Å 10% of the administered dose) and (iii) hydroxylation on the imidazopyridine moiety (metabolite X, that accounts for < 10% of the administered dose).Another minor pathway, involving the hydroxylation of one of the methyl groups of the substituted amide (metabolite XI), is observed.2 Unchanged zolpidem has been observed only in trace amounts in urine ( < 10 ng ml21).3 Zolpidem has little potential for abuse because higher doses are associated with increased incidence of nausea and vomiting.4 Nevertheless, some cases of misuse have been reported.5,6 Fatal cases were always in combination with other psychotropic drugs and/or alcohol and could not be directly linked to zolpidem.7 Techniques used for the determination of zolpidem in biological samples are high-performance liquid chromatography (HPLC) combined with fluorimetric and UV detection8 and capillary gas chromatography with thermoionic detection (NPD).9 Because these methods are time consuming, they are not appropriate for the pre-screening of large numbers of samples.The aim of this work was to develop a sensitive and specific immunoassay for the detection of zolpidem and its metabolites. 6-Methyl-2-(4-methylphenyl)imidazo[1,2-a]pyridine- 3-acetic acid was chosen as the hapten and was coupled directly to bovine serum albumin (BSA) ( = immunogen 1). Alternatively, a spacer, consisting of three carbon atoms, was introduced on the carboxylic acid, prior to coupling to BSA ( = immunogen 2). Two series of four New Zealand rabbits were immunized with the two immunogens. N,N-Didemethylzolpidem- N-{2-{3-(4-hydroxy-3-[125I]iodophenyl)}methyl propionate} was synthesized and purified to serve as a radiotracer.The antisera obtained with immunogens 1 and 2 were compared and the optimum conditions for the radioimmunoassay (RIA) were determined. The test was validated and applied to urine and serum. Experimental Materials and Equipment 6-Methyl-2-(4-methylphenyl)imidazo[1,2-a]pyridine-3-acetic acid was synthesized in our laboratory by a slightly adapted procedure as described in the patent literature.10–12 Standard amounts of this product and of metabolites I, II and XI were obtained as a gift from Synth�elabo (Paris, France).Other reagents were obtained from the following sources: N,N’-carbonyldiimidazole from Aldrich (Steinheim, Germany); b-alanine methyl ester hydrochloride and l-tyrosine methyl ester from Fluka Chemika (Buchs, Switzerland); N,N-dimethylformamide (DMF) from UCB (Brussels, Belgium); Nhydroxysuccinimide (NHS), 1-(3-dimethylaminopropyl)- 3-ethylcarbodiimide hydrochloride (EDC) and N,NA-dicyclohexylcarbodiimide (DCC) from Acros Chimica Fig. 1 Structural formulae of zolpidem, (a); hapten 1, (b); hapten 2, (c); and the radiotracer, (d). Analyst, October 1997, Vol. 122 (1119–1124) 1119(Geel, Belgium); and sodium [125I]iodide IMS 30 from Amersham International (Amersham, Buckinghamshire, UK). Bovine serum albumin (BSA) (fraction 5), goat antiserum to rabbit g-globulin (GARGG), normal rabbit serum (NRS) and Freund’s complete and incomplete adjuvant were purchased from Calbiochem Biochemicals and Immunochemicals (San Diego, CA, USA).Norit Supra A was a gift from Norit (Amersfoort, The Netherlands). The Spectra/Por molecular porous cellulose membranes [relative molecular mass (Mr 12 000–14 000)] for the dialysis of the BSA conjugate were obtained from Spectrum Medical Industries (Los Angeles, CA, USA).All other reagents were obtained from Merck (Darmstadt, Germany). The acidified iodoplatinate reagent was prepared by adding 2 ml of hydrochloric acid (32%) to a solution of 0.25 g of hexachloroplatinic(iv) acid hexahydrate (40% Pt) and 5 g of potassium iodide in 100 ml of water. Thin-layer chromatography (TLC) was carried out using Polygram Sil G/UV254 plates (Machery-Nagel, D�uren, Germany). Liquid surface-assisted secondary ion mass spectrometry (L-SIMS) was performed with a Kratos Concept 1 H instrument (Kratos, Manchester, UK) using a 6 keV Cs+ beam and a thioglycerol matrix.Nuclear magnetic resonance (NMR) spectra were registered in deuteriated methanol, with addition of D2O and NaOD for hapten 1, and were taken with a Unity 500 MHz instrument (Varian, Palo Alto, CA, USA). The radioactivity of the tracer was counted on a g-counter (Berthold BF 5300, Wildbad, Germany). The degree of incorporation of the hapten was determined on a Lambda 16 UV/VIS spectrometer (Perkin- Elmer, Norwalk, CT, USA).HPLC was carried out by using a Merck–Hitachi Model L-6002 pump, equipped with a Rheodyne injector (Model 7125, Berkeley, CA, USA), supplied with a 200 ml sample loop. The analysis and purification of the radioligand were performed on an analytical LiChrospher Si-60 5 mm column (125-4) (Merck). The column eluates were monitored by using a g-counter detector (Canberra Industries, Mereden, CT, USA). Preparation of the Hapten To a suspension of 6-methyl-2-(4-methylphenyl)imidazo[1,2- a]pyridine-3-acetic acid (500 mg, 1.8 mmol) in tetrahydrofuran (THF) (12.5 ml), N,NA-carbonyldiimidazole (690 mg, 4.2 mmol) was added.The mixture was sonicated at room temperature. After 25, 40, 75 and 90 min, aliquots of the reaction mixture were taken and examined by TLC on silica plates with toluene– acetone–methanol (70 + 20 + 10, v/v/v). The reaction product (RF = 0.77) and the parent product (RF = 0.23) were localized using short-wave UV light (254 nm) and by spraying the plates with the acidified iodoplatinate reagent (brown spots).After 90 min, the reaction was complete. A suspension of b-alanine methyl ester hydrochloride (300 mg, 2.2 mmol) and triethylamine (220 mg, 2.2 mmol) in THF was added to the reaction mixture and sonicated at room temperature. The reaction was followed by TLC (RF = 0.67), using the same solvent mixture. The rean was complete after 30 min and the solvent was evaporated.Purification was carried out by adding 9 ml of water and 1 ml of saturated hydrogencarbonate solution to the residue, followed by extraction with diethyl ether (15 ml). After drying with anhydrous sodium sulfate, the solvent was evaporated under a stream of nitrogen. The ester (50 mg, 0.14 mmol) was suspended in distilled water and the pH was adjusted to 9 with sodium hydroxide (1 m). The mixture was sonicated until dissolution and concentrated hydrochloric acid (2 ml) was added.The mixture was allowed to stand at 4 °C until precipitation of the acid (RF = 0.08). The precipitate was filtered and dried. The structure of the hapten was confirmed by NMR and mass spectrometry ( = hapten 2) [Fig. 1(c)]. Alternatively, 6-methyl-2-(4-methylphenyl)imidazo[1,2- a]pyridine-3-acetic acid was used as a hapten ( = hapten 1) [Fig. 1(b)]. Preparation of the Immunogens Identical reactions were carried out for both haptens 1 and 2. The hapten (0.1 mmol) was dissolved in 7 ml of dimethyl sulfoxide (DMSO) and 11.5 mg (0.1 mmol) of NHS was added to this solution.A 20 mg (0.1 mmol) amount of EDC was dissolved in 200 ml of distilled water and slowly added to the hapten–NHS solution. The reactions were followed with the same TLC system as described above: RF hapten 1 = 0.23, RF activated ester 1 = 0.38; RF hapten 2 = 0.08 and RF activated ester 2 = 0.17. After 4 h, the reaction was almost complete and the solution was slowly added to a solution of BSA (67 mg) in 10 ml of phosphate buffer (pH 7.0, 0.05 mol l21).The mixture was stirred continuously and allowed to react for 24 h at room temperature. (60% DMSO was needed to keep the hapten in solution). The low molecular mass compounds were removed from the solution by dialysis, using a cellulose membrane with a cut-off value of 12 000–14 000 Da. The mixture was dialysed in phosphate buffer (pH 7.0, 0.05 mol l21), changing the buffer three times during the first day.After the first day, the buffer concentration was gradually decreased to 0.001 mol l21 and changed twice a day. The dialysis was stopped after 3 d and the hapten–BSA solution divided into portions of 2 ml each and stored at 220 °C. Degree of Incorporation The degree of incorporation of the hapten was estimated by UV spectrometry.13 The hapten has a maximum absorption at 308 nm. BSA does not interfere at this wavelength. Synthesis of the Tracer14 A 52 mg (0.45 mmol) amount of NHS and 93 mg (0.45 mmol) of DCC were added to a solution of 100 mg (0.36 mmol) of 6-methyl-2-(4-methylphenyl)-imidazo[1,2-a]pyridine-3-acetic acid, dissolved in 9 ml of DMF.The mixture was stirred for 1 h at room temperature and placed in a refrigerator (4 °C) overnight. The precipitate was removed by filtration and 58 mg (0.3 mmol) of l-tyrosine methyl ester were added to the filtrate. After mixing, the solution was allowed to stand at room temperature for 15 min.Purification was carried out by column chromatography. To a solution of 50 ml of methanol and 30 ml of phosphate buffer, 10 ml of l-tyrosine methyl ester conjugate solution (1 mg per 10 ml of methanol = 2.18 nmol) and 10 ml (1 mCi, 0.5 nmol) of sodium [125I]iodide solution were added. The temperature was kept at 0 °C. A 10 ml (0.75 nmol) portion of a freshly prepared chloramine-T solution was then added and the solution was vortex-mixed. Another 10 ml aliquot of the chloramine-T solution was added at 3 and 6 min.After 10 min, 20 ml of an aqueous solution of sodium metabisulfite (8.7 nmol) were added. Immunization Antisera were raised in two series of four New Zealand White Rabbits (immunogens 1 and 2). Aliquots (2 ml) of the dialysed hapten–BSA conjugate (90 nmol immunogen 1, 100 nmol immunogen 2) were emulsified with 3 ml of Freund’s complete adjuvant. The rabbits were immunized by subcutaneous injection of 1.2 ml of the water–oil emulsion.The injection volume was divided along 4–6 places on the back of the rabbit. The first two booster injections were given at a 2 week interval. The 1120 Analyst, October 1997, Vol. 122following booster solutions were made up with Freund’s incomplete adjuvant and injected at 4 week intervals. Small blood samples (10–15 ml) were collected every month from the lateral ear vein, starting 2 weeks after the second injection. Six months after the first injection, 2 weeks after the final booster, 50 ml blood samples were taken by cardiac puncture.Titration of Antisera Antisera were diluted (1 + 99 to 1 + 49 999) with phosphate buffer (pH 7.4, 0.05 mol l21) for titration experiments. The titres were determined by adding 100 ml of the antiserum dilution and 100 ml of the tracer (12 500 counts min21) to 400 ml of BSA matrix solution (2% BSA in phosphate buffer of pH 7.4). The mixtures were allowed to equilibrate at room temperature for 90 min.Bound and free radioligand were separated by a second antibody method using GARGG as described below. The serum dilution, able to bind 50% of the tracer, was calculated by constructing binder dilution curves. Optimization of the Assay Parameters that possibly influence the immunoassay such as pH, incubation time and concentration of BSA and tracer were tested. The following incubation conditions were kept constant for all experiments: to 300 ml of BSA solution (2% in phosphate buffer, pH 7.4, 0.05 mol l21) were added 100 ml of tracer, 100 ml of sample or standard solution and 100 ml of antiserum dilution.The mixture was allowed to equilibrate at room temperature for 90 min. For the optimization of the immunoassay conditions two methods for the separation of bound and free ligand were studied. First, the non-specific adsorption of the tracer onto charcoal was tested but was found to give excessive nonspecific (NSB) values (7–9%). As a second separation technique, a second antibody method using GARGG was tested.To optimize this procedure, several parameters were investigated, such as the incubation time and the optimum amounts of GARGG and NRS. After the initial incubation (90 min) at room temperature, bound and free radioligand were separated by adding 50 ml of NRS (5% in phosphate buffer) and 100 ml of GARGG (8% in phosphate buffer). After mixing, incubating for a further 20 h and centrifuging for 15 min at 3000g, both the supernatant (i.e., free fraction) and the pellet (i.e., bound fraction) were counted on a g-counter for 1 min.Calibration Graph The second antibody separation technique (GARGG) was selected for the construction of the calibration graph. Zolpidem was diluted in drug-free urine to produce a concentration range of 100–10 000 pg ml21 and 100 ml of the spiked samples were analysed with the RIA procedure. Non-specific binding was determined by replacing the antiserum by an equal volume of buffer.Calibration graphs were constructed by plotting B/B0, representing counts bound above non-specific, relative to counts bound above non-specific for zero dose of analyte, against the concentration of unlabelled ligand on a semilogarithmic scale. Human Samples To one healthy volunteer (female, 26 years), 10 mg of zolpidem (Stilnoct) were administered orally and urine and serum samples were collected over a 48 h period. Results Preparation of the Hapten, Immunogen and Tracer The main goal of our assay is to separate positive from negative urine samples and to have an idea of the concentration range of zolpidem and zolpidem-related material. 6-Methyl-2-(4-methylphenyl) imidazo[1,2-a]pyridine-3-acetic acid is the molecule of choice for realizing the synthesis of the hapten and the conjugation to BSA since the imidazo[1,2-a]pyridine part is the most specific part and the carboxylic acid can be used as the conjugation site. For the synthesis of hapten 2, 6-methyl- 2-(4-methylphenyl)imidazo[1,2-a]pyridine-3-acetic acid was first reacted with carbonyldiimidazole to form an acylimidazole derivative and the latter was subsequently reacted with b- alanine methyl ester to form an amide.Acidic hydrolysis of the ester function resulted in hapten 2. The structure was confirmed by mass spectrometry and NMR. The major fragments (m/z) of hapten 1 are 65, 92, 219, 235 and 280. For hapten 1, the assignment of the 13C peaks was based on an APT-spectrum (attached proton test) and on a heteronuclear-correlation experiment (HETCOR). 13C NMR (CD3OD/D2O + NaOD): d 177.6 (CO2H), 144.4 (C-2), 142.6 (C-1A), 138.8 (C-9), 131.8 (C-6), 130.2 (C-2A), 129.5 (C-7), 128.9 (C-3A), 123.7 (C-4A), 122.7 (C-5), 117.8 (C- 3), 115.8 (C-8), 33.6 (3-CH2), 21.2 (4A-Me), 18.3 (6-Me). 1H NMR (CD3OD/D2O + NaOD): d 2.17 (s, 3 H, 4A-Me), 2.23 (s, 3 H, 6-Me), 3.64 (s, 2 H, 3-CH2), 7.04 (dd, 1 H, J = 9.0/ 1.4 Hz, H-7), 7.15 (d, 2 H, J = 8.0 Hz, H-2A/6A), 7.31 (d, 1 H, J = 9.0 Hz, H-8), 7.41 (d, 2 H, J = 8.0 Hz, H-3A/5A), 7.71 (s, 1 H, H-5).The major fragments (m/z) of hapten 2 are : 65, 92, 219, 235 and 351. For hapten 2, the assignment of the peaks was based mainly on the assignments for hapten 1. 13C NMR (CD3OD): d 175.1 (CO2H), 169.7 (CONH), 142.4 (C-1A), 139.7 (C-9), 137.3 (C-7), 135.6 (C-2), 131.3 (C-2A), 129.6 (C-3A), 129.4 (C-6), 125.9 (C-5), 124.8 (C-4A), 118.5 (C- 3), 112.1 (C-8), 36.9 (C-b-alanyl), 34.5 (3-CH2), 30.9 (C-a- alanyl), 21.4 (4A-Me), 18.1 (6-Me). 1H NMR (CD3OD): d 2.43 (s, 3 H, 4A-Me), 2.52 (s, 3 H, 6-Me), 2.58 (t, 2 H, 3J = 6.45 Hz, CH2-a-alanyl), 3.52 (t, 2 H, J = 6.45 Hz, CH2-b-alanyl), 4.14 (s, 2 H, 3-CH2), 4.91 (br, NH, CO2H), 7.42 (d, 2 H, J = 7.8 Hz, H-2A/6A, 7.56 (d, 2 H, J = 8.1 Hz, H-3A/5A), 7.82 (d, 1 H, J = 9.2 Hz, H-8), 7.86 (dd, 1 H, J = 9.2/1.1 Hz, H-7), 8.5 (s, 1 H, H-5). The hapten–BSA conjugate was prepared by the carbodiimide technique, which was first described by Sheehan and Hess.15 By using UV spectrometry, a hapten : protein ratio (mmol) of 21.2 : 0.96 and 16.5 : 0.96 was calculated for immunogens 1 and 2, respectively, corresponding to the coupling of about 22 and 17 mol, respectively, of hapten with 1 mol of BSA.An average of 15 molecules of hapten per molecule of BSA has been recommended for an optimum immuno-response.16 The synthesis and purification of N,N-didemethylzolpidem- N-{2-{3-(4-hydroxy-3-[125I]-iodophenyl)}methyl propionate} [Fig. 1(d)] have already been described in detail.14 The freshly prepared tracer could be used for about 6 weeks without any loss of binding to the antibody. Titre of the Antibodies After the examination of the serial dilutions of all the antisera from the different bleeds, the titres were derived from binder dilution curves. The evolution of the titres of the different rabbits, immunized with immunogens 1 and 2, respectively, is shown in Fig. 2. As could be expected, great differences exist in immunogenic responses between immunogens 1 and 2, with titres ranging from 1/75 to 1/3200 for the immunogen without spacer and from 1/12 800 to 1/25 000 for the immunogen with spacer. Because of the low titres, the sera of rabbits 1.1 and 1.4 were not considered for further experiments. The antisera of the last bleeds of all the rabbits were lyophilized and stored at 4 °C until use. The optimization and specificity experiments were performed for all the antisera except for R1.1 and R1.4.All Analyst, October 1997, Vol. 122 1121other experiments were only carried out with the antiserum of R2.1. Determination of Optimum Conditions for the RIA As a separation technique, a second antibody method was evaluated.17 The optimum ratio of GARGG antiserum to NRS was investigated. Final dilutions of GARGG (1 + 124) and NRS (1 + 999) resulted in the highest concentrations of tracer. The optimum second incubation time for the precipitation of the antibody-bound tracer was 20 h at room temperature.In comparison with the adsorption technique, this method resulted in lower NSB values (1–3%) and in higher dilutions of the antiserum to reach 50% binding of the tracer (1 + 12 799). The performance characteristics of the immunoassay are demonstrated in a calibration graph obtained with the serum of R2.1 for the concentration range 100–10 000 pg ml21 (Fig. 3), using the second antibody method as the separation technique.The curve of best fit was obtained by applying the four-parametric logistic model to the experimental data.18 Limit of Quantification The limit of quantification (LOQ) of the RIA can be defined as the lowest concentration on the calibration graph that can be measured with acceptable precision and accuracy. The precision at this level has to be less than or equal to 20%. To determine the LOQ, five standard solutions of increasing concentration were used and the relative standard deviation was determined.The LOQ was 0.1 ng ml21. Precision and Recovery The precision and recovery of the method could be calculated after replicate analyses of independently prepared quality control urine samples. The concentration levels of the samples ranged from 0.25 to 1000 ng ml21. The intra- and inter-assay relative standard deviations were 4.01–6.30 and 9.13–12.28%, respectively, and the accuracy was about 100% (Table 1). Specificity The antibody specificity was assessed by measuring the crossreactivity to other hypnotics and to some widely used anxiolytics, antidepressants, analgesics and analeptics. The following compounds were found not to be detectable at a level of 10 mg ml21: chlordiazepoxide, diazepam, clobazam, nitrazepam, flunitrazepam, bromazepam, lormetazepam, triazolam, meprobamate, methaqualone, hexobarbital, fenobarbital, buspirone, hydroxyzine, chlorcyclizine, codeine, cocaine, caffein, ibuprofen, trazodone, zopiclone, chlorpromazine and acetylsalicylic acid.None of the above was found to cross-react ( < 0.005%) with the immunoassay. To rule out cross-reactivity with endogenous compounds, several urine samples of volunteers were screened and no crossreactivity could be detected. To select one antiserum out of the pool, the specificity towards the metabolites was examined. The two major metabolites of zolpidem, metabolites I and II, and a minor metabolite (metabolite XI) were tested for their cross-reactivity.The degree of cross-reactivity of the metabolites was expressed as the relative dose required for 50% displacement of the maximum tracer binding. The major metabolite (I) showed a low cross-reactivity, ranging from 2.3 3 1027 to 7.7 3 1023%. This could be expected considering that the position of the Fig. 2 Evolution of the serum titres of the two series of rabbits, immunized with the immunogens 1 and 2, respectively. (a) represents the antisera from rabbits R1.1 (/), R1.2 (~), 1.3 (-) and R1.4 (3) and (b) shows the titres of R2.1 (/), R2.2 (~), R2.3 (-) and R2.4 (3).Fig. 3 Calibration graph for zolpidem, together with 95% confidence intervals. The graph follows the equation y = d [(a 2 d)/1 + (x/c)b], where a = 0.8972 ± 0.01006; b = 9.9390 ± 0.4498; c = 6.8896 ± 0.0280; d = 0.02248 ± 0.01195; and r2 = 0.9764. Table 1 Intra- and inter-assay variance characteristics and recovery Added/ng RSD (%) n Recovery (%) Intra-assay 0.5 4.01 6 103.4 1.75 4.48 6 101.5 7.5 6.30 6 98.1 1000 5.90 6 99.3 Inter-assay 0.25 12.28 10 100.9 1 9.30 9 99.4 7.5 7.40 10 101.3 1000 9.13 9 104.0 1122 Analyst, October 1997, Vol. 122carboxylic acid function is close to the binding site with the antibodies. The cross-reactivity towards metabolite II was higher and varied from 0.9 to 6.9%. In this metabolite, the substituent is positioned further away from the binding site. Metabolite XI, which has an identical structure with zolpidem, except for the hydroxylic function on the acetamide side-chain in position 3 of the imidazopyridine skeleton, showed a very high cross-reactivity (91–116%).Because of the pre-screening purpose of this RIA, the antiserum with the highest crossreactivity for the main metabolite was selected, i.e., the antiserum of rabbit 2.1 (6th bleed). The results are shown in Table 2. Linearity To confirm the linearity of the response throughout the range of the calibration graph, samples were checked to see whether they diluted in parallel to the calibration graph.An unknown urine sample, containing approximately 4.6 ng ml21 of the drug, was analysed three times, both undiluted and diluted with blank urine (1 + 1, 1 + 3, 1 + 7 and 1 + 15). The observed (y) and the expected (x) values yielded the following linear regression equation : y = 0.0115 + 0.9703x (r = 0.9938). Human Urine and Serum Samples A set of 14 urine samples (including a blank) were collected from a volunteer during a 48 h period after the intake of 10 mg of zolpidem (Stilnoct).The samples were analysed using the described RIA procedure and the results were expressed in terms of zolpidem concentration, although there was also a contribution from the metabolites. The highest value (994 ng ml21) was reached 3 h after the ingestion of one tablet of Stilnoct. After 48 h, the hypnotic could still easily be detected (about 6.3 ng ml21).From the same volunteer and at specific time intervals, 13 serum samples (including a blank) were collected and analysed by RIA. The serum levels measured covered a concentration from 1.3 to 122 ng ml21. The highest value was recorded 2 h after administration (about 122 ng ml21). Thereafter, the level of zolpidem decreased to 1.3 ng ml21 after 48 h. Fig. 4 shows the serum concentration–time profile of zolpidem. Discussion This paper describes the development of an RIA for zolpidem and its metabolites in urine and serum. Two immunogens were synthesized to study the difference in affinity and specificity between an immunogen with and without a spacer.The antibodies from the rabbits immunized with the immunogen with a spacer showed a much higher affinity than those immunized with the other immunogen. A difference in specificity could not be seen. The synthesis and purification of N,N-didemethylzolpidem-N-{2-{3-(4-hydroxy-3-[125I]-iodophenyl)} methyl propionate} resulted in a high radiochemical purity and specific activity of the radioligand,14 thereby providing a very sensitive immunoassay (LOQ = 0.1 ng ml21).Owing to the nature of the synthesized hapten and immunogens, the detected cross-reactivity of the three metabolites is to be expected. The important fact is that zolpidem and its metabolites can be effectively detected in urine and serum up to 48 h after the intake of one tablet of Stilnoct (10 mg of zolpidem), without interference from other drugs.The proposed RIA can be used in the pre-screening of toxicological samples. Its applica- Table 2 Specificity of the antisera towards three metabolites of zolpidem in comparison with the unmetabolised drug Cross-reactivity (%) Compound R1.2 R1.3 R2.1 R2.2 R2.3 R2.4 Zolpidem 100 100 100 100 100 100 7.7 3 1023 2.3 3 1023 2.3 3 1027 4.2 3 1023 3 3 1024 1.6 3 1023 2.53 2.4 5.55 0.9 6.9 2.9 91 116 116 114 104 92 Analyst, October 1997, Vol. 122 1123bility to the analysis of blood and plasma samples is under study. Dr. R. Busson, Dr. G. Janssen and Dr. J. Rozensky are gratefully acknowledged for recording the NMR and mass spectra. References 1 Th�enot, J. P., Hermann, P., Durand, A., Burke, J. T., Allen, J., Garrigou, D., Vajta, S., Albin, H., Th�ebault, J. J., Olive, G., and Warrington, S. J., in Imidazopyridines in Sleep Disorders, ed. Sauvanet, J. P., Langer, S. Z., and Morselli, P. L., Raven Press, New York, 1988, pp. 139–153. 2 Durand, A., Th�enot, J. P., Bianchetti, G., and Morselli, P. L., Drug Metab. Rev., 1992, 24, 239. 3 Augsburger, M., Giroud, C., Lucchini, P., and Rivier, L., in Contributions to Forensic Toxicology. The Proceedings of the 31st International Meeting of the International Association of Forensic Toxicologists (TIAFT), 1993, ed. Mueller, R. K., Molinapress, Leipzig, 1994, pp. 18–22. 4 Mendelson, W. B., and Jain, B., Drug Saf., 1995, 13, 257. 5 Debailleul, G., Abi Khalil, F., and Lheureux, P., J.Anal. Toxicol., 1991, 15, 35. 6 Khodasevitch, T., and Volgram, J., Bull. Int. Assoc. Forensic Toxicol., 1996, 26(2), 37. 7 Garnier, R., Guerault, E., Muzard, D., Azoyan, P., Chaumet- Riffaud, A., and Efthymiou, M., Clin. Toxicol., 1994, 32, 391. 8 Guinebault, P., Dubruc, C., Hermann, P., and Th�enot, J. P., J. Chromatogr., 1986, 383, 206. 9 Debruyne, D., Lacotte, J., Hurault De Ligny, B., and Moulin, M., J. Pharm. Sci., 1991, 80, 71. 10 Almirante, L., and Murwann, W., Br.Pat., 1 076 089, 1967. 11 Kaplan, J. P., and George, P., Eur. Pat., 0 050 563, 1982. 12 Kaplan, J. P., and George, P., US Pat., 4 501 745, 1985. 13 Erlanger, B. F., Methods in Enzymology, Academic Press, New York, 1980, vol. 70, p. 85. 14 De Clerck, I., and Daenens, P., J. Radiolabelled Comp. Radiopharm., 1997, 39, 195. 15 Sheehan, J. C., and Hess, G. P., J. Am. Chem. Soc., 1955, 77, 1067. 16 Erlanger, B. F., Pharmacol. Rev., 1973, 25, 271. 17 Utiger, C.R., Parker, M. L., and Daughaday, W. H., J. Clin. Invest., 1962, 41, 254. 18 Dudley, R., Edwards, P., Ekins, R. P., Finney, D. J., McKenzie, I. G. M., Raab, G. M., Rodbard, D., and Rodgers, R. P. C., Clin. Chem. (Winston-Salem, N.C.), 1985, 31, 1264. Paper 7/02869E Received April 28, 1997 Accepted June 6, 1997 Fig. 4 Serum levels, expressed in terms of zolpidem concentration, in one healthy volunteer after a single oral intake of one tablet (10 mg) of Stilnoct. 1124 Analyst, October 1997, Vol. 122 Development of a Radioimmunoassay for the Determination of Zolpidem in Biological Samples Isabel De Clerck and P. Daenens* Laboratory of Toxicology, Katholieke Universiteit Leuven, E. Van Evenstraat, 4, B-3000 Leuven, Belgium The development of a specific and sensitive radioimmunoassay for the detection of zolpidem and its metabolites in urine and serum samples is described. The assay can be used to pre-screen forensic and emergency samples. 6-Methyl-2-(4-methylphenyl)imidazo[1,2-a]- pyridine-3-acetic acid was prepared as a hapten and was coupled directly and indirectly to bovine serum albumin. The immunization of rabbits with the hapten–bovine serum albumin conjugates resulted in the production of highly specific antibodies, showing no significant cross-reactivities towards existing drugs.The fourparameter logistic model was used to process the calibration data into a fitting curve (r2 = 0.9764). Intraand inter-assay relative standard deviations were < 6.30 and < 12.28%, respectively.The limit of quantification was 0.1 ng ml21. Using this assay, zolpidem levels were determined in urine and serum, and could be easily detected up to 48 h after oral intake of 10 mg of zolpidem. Keywords: Radioimmunoassay; zolpidem; imidazo[1,2-a]pyridine Zolpidem [Fig. 1(a)], N,N,6-trimethyl-2-(4-methylphenyl)- imidazole[1,2-a]pyridine-3-acetamide l-(+)-hemitartrate, is a non-benzodiazepine hypnotic drug with an imidazopyridine backbone, which acts in the brain principally at receptors of the w1-receptor subtype belonging to the g-aminobutyric acid- (GABA)ergic system.Its main features are a fast onset of action and a short elimination half-life (2.4 h), unlike benzodiazepines. Stilnoct is administered orally (therapeutic dose = 10 mg). Therapeutic plasma concentrations are in the low nanogram range with average peak plasma levels of about 140 ng ml21 after 0.5–3 h.1 The drug is extensively metabolized and excreted in the urine as pharmacologically inactive metabolites.The biotransformation proceeds through three different pathways including (i) methyl oxidation on the phenyl moiety of the molecule (metabolite I, the main metabolite in urine that accounts for 52% of the administered dose), (ii) methyl oxidation on the imidazopyridine moiety leading to the corresponding carboxylic acid (metabolite II, that accounts for Å 10% of the administered dose) and (iii) hydroxylation on the imidazopyridine moiety (metabolite X, that accounts for < 10% of the administered dose).Another minor pathway, involving the hydroxylation of one of the methyl groups of the substituted amide (metabolite XI), is observed.2 Unchanged zolpidem has been observed only in trace amounts in urine ( < 10 ng ml21).3 Zolpidem has little potential for abuse because higher doses are associated with increased incidence of nausea and vomiting.4 Nevertheless, some cases of misuse have been reported.5,6 Fatal cases were always in combination with other psychotropic drugs and/or alcohol and could not be directly linked to zolpidem.7 Techniques used for the determination of zolpidem in biological samples are high-performance liquid chromatography (HPLC) combineduorimetric and UV detection8 and capillary gas chromatography with thermoionic detection (NPD).9 Because these methods are time consuming, they are not appropriate for the pre-screening of large numbers of samples.The aim of this work was to develop a sensitive and specific immunoassay for the detection of zolpidem and its metabolites. 6-Methyl-2-(4-methylphenyl)imidazo[1,2-a]pyridine- 3-acetic acid was chosen as the hapten and was coupled directly to bovine serum albumin (BSA) ( = immunogen 1). Alternatively, a spacer, consisting of three carbon atoms, was introduced on the carboxylic acid, prior to coupling to BSA ( = immunogen 2).Two series of four New Zealand rabbits were immunized with the two immunogens. N,N-Didemethylzolpidem- N-{2-{3-(4-hydroxy-3-[125I]iodophenyl)}methyl propionate} was synthesized and purified to serve as a radiotracer. The antisera obtained with immunogens 1 and 2 were compared and the optimum conditions for the radioimmunoassay (RIA) were determined. The test was validated and applied to urine and serum. Experimental Materials and Equipment 6-Methyl-2-(4-methylphenyl)imidazo[1,2-a]pyridine-3-acetic acid was synthesized in our laboratory by a slightly adapted procedure as described in the patent literature.10–12 Standard amounts of this product and of metabolites I, II and XI were obtained as a gift from Synth�elabo (Paris, France).Other reagents were obtained from the following sources: N,N’-carbonyldiimidazole from Aldrich (Steinheim, Germany); b-alanine methyl ester hydrochloride and l-tyrosine methyl ester from Fluka Chemika (Buchs, Switzerland); N,N-dimethylformamide (DMF) from UCB (Brussels, Belgium); Nhydroxysuccinimide (NHS), 1-(3-dimethylaminopropyl)- 3-ethylcarbodiimide hydrochloride (EDC) and N,NA-dicyclohexylcarbodiimide (DCC) from Acros Chimica Fig. 1 Structural formulae of zolpidem, (a); hapten 1, (b); hapten 2, (c); and the radiotracer, (d). Analyst, October 1997, Vol. 122 (1119–1124) 1119(Geel, Belgium); and sodium [125I]iodide IMS 30 from Amersham International (Amersham, Buckinghamshire, UK).Bovine serum albumin (BSA) (fraction 5), goat antiserum to rabbit g-globulin (GARGG), normal rabbit serum (NRS) and Freund’s complete and incomplete adjuvant were purchased from Calbiochem Biochemicals and Immunochemicals (San Diego, CA, USA). Norit Supra A was a gift from Norit (Amersfoort, The Netherlands). The Spectra/Por molecular porous cellulose membranes [relative molecular mass (Mr 12 000–14 000)] for the dialysis of the BSA conjugate were obtained from Spectrum Medical Industries (Los Angeles, CA, USA).All other reagents were obtained from Merck (Darmstadt, Germany). The acidified iodoplatinate reagent was prepared by adding 2 ml of hydrochloric acid (32%) to a solution of 0.25 g of hexachloroplatinic(iv) acid hexahydrate (40% Pt) and 5 g of potassium iodide in 100 ml of water. Thin-layer chromatography (TLC) was carried out using Polygram Sil G/UV254 plates (Machery-Nagel, D�uren, Germany). Liquid surface-assisted secondary ion mass spectrometry (L-SIMS) was performed with a Kratos Concept 1 H instrument (Kratos, Manchester, UK) using a 6 keV Cs+ beam and a thioglycerol matrix.Nuclear magnetic resonance (NMR) spectra were registered in deuteriated methanol, with addition of D2O and NaOD for hapten 1, and were taken with a Unity 500 MHz instrument (Varian, Palo Alto, CA, USA). The radioactivity of the tracer was counted on a g-counter (Berthold BF 5300, Wildbad, Germany). The degree of incorporation of the hapten was determined on a Lambda 16 UV/VIS spectrometer (Perkin- Elmer, Norwalk, CT, USA).HPLC was carried out by using a Merck–Hitachi Model L-6002 pump, equipped with a Rheodyne injector (Model 7125, Berkeley, CA, USA), supplied with a 200 ml sample loop. The analysis and purification of the radioligand were performed on an analytical LiChrospher Si-60 5 mm column (125-4) (Merck). The column eluates were monitored by using a g-counter detector (Canberra Industries, Mereden, CT, USA).Preparation of the Hapten To a suspension of 6-methyl-2-(4-methylphenyl)imidazo[1,2- a]pyridine-3-acetic acid (500 mg, 1.8 mmol) in tetrahydrofuran (THF) (12.5 ml), N,NA-carbonyldiimidazole (690 mg, 4.2 mmol) was added. The mixture was sonicated at room temperature. After 25, 40, 75 and 90 min, aliquots of the reaction mixture were taken and examined by TLC on silica plates with toluene– acetone–methanol (70 + 20 + 10, v/v/v). The reaction product (RF = 0.77) and the parent product (RF = 0.23) were localized using short-wave UV light (254 nm) and by spraying the plates with the acidified iodoplatinate reagent (brown spots).After 90 min, the reaction was complete. A suspension of b-alanine methyl ester hydrochloride (300 mg, 2.2 mmol) and triethylamine (220 mg, 2.2 mmol) in THF was added to the reaction mixture and sonicated at room temperature. The reaction was followed by TLC (RF = 0.67), using the same solvent mixture. The reaction was complete after 30 min and the solvent was evaporated.Purification was carried out by adding 9 ml of water and 1 ml of saturated hydrogencarbonate solution to the residue, followed by extraction with diethyl ether (15 ml). After drying with anhydrous sodium sulfate, the solvent was evaporated under a stream of nitrogen. The ester (50 mg, 0.14 mmol) was suspended in distilled water and the pH was adjusted to 9 with sodium hydroxide (1 m). The mixture was sonicated until dissolution and concentrated hydrochloric acid (2 ml) was added.The mixture was allowed to stand at 4 °C until precipitation of the acid (RF = 0.08). The precipitate was filtered and dried. The structure of the hapten was confirmed by NMR and mass spectrometry ( = hapten 2) [Fig. 1(c)]. Alternatively, 6-methyl-2-(4-methylphenyl)imidazo[1,2- a]pyridine-3-acetic acid was used as a hapten ( = hapten 1) [Fig. 1(b)]. Preparation of the Immunogens Identical reactions were carried out for both haptens 1 and 2.The hapten (0.1 mmol) was dissolved in 7 ml of dimethyl sulfoxide (DMSO) and 11.5 mg (0.1 mmol) of NHS was added to this solution. A 20 mg (0.1 mmol) amount of EDC was dissolved in 200 ml of distilled water and slowly added to the hapten–NHS solution. The reactions were followed with the same TLC system as described above: RF hapten 1 = 0.23, RF activated ester 1 = 0.38; RF hapten 2 = 0.08 and RF activated ester 2 = 0.17. After 4 h, the reaction was almost complete and the solution was slowly added to a solution of BSA (67 mg) in 10 ml of phosphate buffer (pH 7.0, 0.05 mol l21).The mixture was stirred continuously and allowed to react for 24 h at room temperature. (60% DMSO was needed to keep the hapten in solution). The low molecular mass compounds were removed from the solution by dialysis, using a cellulose membrane with a cut-off value of 12 000–14 000 Da. The mixture was dialysed in phosphate buffer (pH 7.0, 0.05 mol l21), changing the buffer three times during the first day.After the first day, the buffer concentration was gradually decreased to 0.001 mol l21 and changed twice a day. The dialysis was stopped after 3 d and the hapten–BSA solution divided into portions of 2 ml each and stored at 220 °C. Degree of Incorporation The degree of incorporation of the hapten was estimated by UV spectrometry.13 The hapten has a maximum absorption at 308 nm. BSA does not interfere at this wavelength.Synthesis of the Tracer14 A 52 mg (0.45 mmol) amount of NHS and 93 mg (0.45 mmol) of DCC were added to a solution of 100 mg (0.36 mmol) of 6-methyl-2-(4-methylphenyl)-imidazo[1,2-a]pyridine-3-acetic acid, dissolved in 9 ml of DMF. The mixture was stirred for 1 h at room temperature and placed in a refrigerator (4 °C) overnight. The precipitate was removed by filtration and 58 mg (0.3 mmol) of l-tyrosine methyl ester were added to the filtrate. After mixing, the solution was allowed to stand at room temperature for 15 min.Purification was carried out by column chromatography. To a solution of 50 ml of methanol and 30 ml of phosphate buffer, 10 ml of l-tyrosine methyl ester conjugate solution (1 mg per 10 ml of methanol = 2.18 nmol) and 10 ml (1 mCi5 nmol) of sodium [125I]iodide solution were added. The temperature was kept at 0 °C. A 10 ml (0.75 nmol) portion of a freshly prepared chloramine-T solution was then added and the solution was vortex-mixed. Another 10 ml aliquot of the chloramine-T solution was added at 3 and 6 min.After 10 min, 20 ml of an aqueous solution of sodium metabisulfite (8.7 nmol) were added. Immunization Antisera were raised in two series of four New Zealand White Rabbits (immunogens 1 and 2). Aliquots (2 ml) of the dialysed hapten–BSA conjugate (90 nmol immunogen 1, 100 nmol immunogen 2) were emulsified with 3 ml of Freund’s complete adjuvant. The rabbits were immunized by subcutaneous injection of 1.2 ml of the water–oil emulsion.The injection volume was divided along 4–6 places on the back of the rabbit. The first two booster injections were given at a 2 week interval. The 1120 Analyst, October 1997, Vol. 122following booster solutions were made up with Freund’s incomplete adjuvant and injected at 4 week intervals. Small blood samples (10–15 ml) were collected every month from the lateral ear vein, starting 2 weeks after the second injection.Six months after the first injection, 2 weeks after the final booster, 50 ml blood samples were taken by cardiac puncture. Titration of Antisera Antisera were diluted (1 + 99 to 1 + 49 999) with phosphate buffer (pH 7.4, 0.05 mol l21) for titration experiments. The titres were determined by adding 100 ml of the antiserum dilution and 100 ml of the tracer (12 500 counts min21) to 400 ml of BSA matrix solution (2% BSA in phosphate buffer of pH 7.4). The mixtures were allowed to equilibrate at room temperature for 90 min.Bound and free radioligand were separated by a second antibody method using GARGG as described below. The serum dilution, able to bind 50% of the tracer, was calculated by constructing binder dilution curves. Optimization of the Assay Parameters that possibly influence the immunoassay such as pH, incubation time and concentration of BSA and tracer were tested. The following incubation conditions were kept constant for all experiments: to 300 ml of BSA solution (2% in phosphate buffer, pH 7.4, 0.05 mol l21) were added 100 ml of tracer, 100 ml of sample or standard solution and 100 ml of antiserum dilution.The mixture was allowed to equilibrate at room temperature for 90 min. For the optimization of the immunoassay conditions two methods for the separation of bound and free ligand were studied. First, the non-specific adsorption of the tracer onto charcoal was tested but was found to give excessive nonspecific (NSB) values (7–9%). As a second separation technique, a second antibody method using GARGG was tested. To optimize this procedure, several parameters were investigated, such as the incubation time and the optimum amounts of GARGG and NRS.After the initial incubation (90 min) at room temperature, bound and free radioligand were separated by adding 50 ml of NRS (5% in phosphate buffer) and 100 ml of GARGG (8% in phosphate buffer). After mixing, incubating for a further 20 h and centrifuging for 15 min at 3000g, both the supernatant (i.e., free fraction) and the pellet (i.e., bound fraction) were counted on a g-counter for 1 min.Calibration Graph The second antibody separation technique (GARGG) was selected for the construction of the calibration graph. Zolpidem was diluted in drug-free urine to produce a concentration range of 100–10 000 pg ml21 and 100 ml of the spiked samples were analysed with the RIA procedure. Non-specific binding was determined by replacing the antiserum by an equal volume of buffer.Calibration graphs were constructed by plotting B/B0, representing counts bound above non-specific, relative to counts bound above non-specific for zero dose of analyte, against the concentration of unlabelled ligand on a semilogarithmic scale. Human Samples To one healthy volunteer (female, 26 years), 10 mg of zolpidem (Stilnoct) were administered orally and urine and serum samples were collected over a 48 h period.Results Preparation of the Hapten, Immunogen and Tracer The main goal of our assay is to separate positive from negative urine samples and to have an idea of the concentration range of zolpidem and zolpidem-related material. 6-Methyl-2-(4-methylphenyl) imidazo[1,2-a]pyridine-3-acetic acid is the molecule of choice for realizing the synthesis of the hapten and the conjugation to BSA since the imidazo[1,2-a]pyridine part is the most specific part and the carboxylic acid can be used as the conjugation site.For the synthesis of hapten 2, 6-methyl- 2-(4-methylphenyl)imidazo[1,2-a]pyridine-3-acetic acid was first reacted with carbonyldiimidazole to form an acylimidazole derivative and the latter was subsequently reacted with b- alanine methyl ester to form an amide. Acidic hydrolysis of the ester function resulted in hapten 2. The structure was confirmed by mass spectrometry and NMR. The major fragments (m/z) of hapten 1 are 65, 92, 219, 235 and 280.For hapten 1, the assignment of the 13C peaks was based on an APT-spectrum (attached proton test) and on a heteronuclear-correlation experiment (HETCOR). 13C NMR (CD3OD/D2O + NaOD): d 177.6 (CO2H), 144.4 (C-2), 142.6 (C-1A), 138.8 (C-9), 131.8 (C-6), 130.2 (C-2A), 129.5 (C-7), 128.9 (C-3A), 123.7 (C-4A), 122.7 (C-5), 117.8 (C- 3), 115.8 (C-8), 33.6 (3-CH2), 21.2 (4A-Me), 18.3 (6-Me). 1H NMR (CD3OD/D2O + NaOD): d 2.17 (s, 3 H, 4A-Me), 2.23 (s, 3 H, 6-Me), 3.64 (s, 2 H, 3-CH2), 7.04 (dd, 1 H, J = 9.0/ 1.4 Hz, H-7), 7.15 (d, 2 H, J = 8.0 Hz, H-2A/6A), 7.31 (d, 1 H, J = 9.0 Hz, H-8), 7.41 (d, 2 H, J = 8.0 Hz, H-3A/5A), 7.71 (s, 1 H, H-5).The major fragments (m/z) of hapten 2 are : 65, 92, 219, 235 and 351. For hapten 2, the assignment of the peaks was based mainly on the assignments for hapten 1. 13C NMR (CD3OD): d 175.1 (CO2H), 169.7 (CONH), 142.4 (C-1A), 139.7 (C-9), 137.3 (C-7), 135.6 (C-2), 131.3 (C-2A), 129.6 (C-3A), 129.4 (C-6), 125.9 (C-5), 124.8 (C-4A), 118.5 (C- 3), 112.1 (C-8), 36.9 (C-b-alanyl), 34.5 (3-CH2), 30.9 (C-a- alanyl), 21.4 (4A-Me), 18.1 (6-Me). 1H NMR (CD3OD): d 2.43 (s, 3 H, 4A-Me), 2.52 (s, 3 H, 6-Me), 2.58 (t, 2 H, 3J = 6.45 Hz, CH2-a-alanyl), 3.52 (t, 2 H, J = 6.45 Hz, CH2-b-alanyl), 4.14 (s, 2 H, 3-CH2), 4.91 (br, NH, CO2H), 7.42 (d, 2 H, J = 7.8 Hz, H-2A/6A, 7.56 (d, 2 H, J = 8.1 Hz, H-3A/5A), 7.82 (d, 1 H, J = 9.2 Hz, H-8), 7.86 (dd, 1 H, J = 9.2/1.1 Hz, H-7), 8.5 (s, 1 H, H-5). The hapten–BSA conjugate was prepared by the carbodiimide technique, which was first described by Sheehan and Hess.15 By using UV spectrometry, a hapten : protein ratio (mmol) of 21.2 : 0.96 and 16.5 : 0.96 was calculated for immunogens 1 and 2, respectively, corresponding to the coupling of about 22 and 17 mol, respectively, of hapten with 1 mol of BSA.An average of 15 molecules of hapten per molecule of BSA has been recommended for an optimum immuno-response.16 The synthesis and purification of N,N-didemethylzolpidem- N-{2-{3-(4-hydroxy-3-[125I]-iodophenyl)}methyl propionate} [Fig. 1(d)] have already been described in detail.14 The freshly prepared tracer could be used for about 6 weeks without any loss of binding to the antibody. Titre of the Antibodies After the examination of the serial dilutions of all the antisera from the different bleeds, the titres were derived from binder dilution curves. The evolution of the titres of the different rabbits, immunized with immunogens 1 and 2, respectively, is shown in Fig. 2. As could be expected, great differences exist in immunogenic responses between immunogens 1 and 2, with titres ranging from 1/75 to 1/3200 for the immunogen without spacer and from 1/12 800 to 1/25 000 for the immunogen with spacer. Because of the low titres, the sera of rabbits 1.1 and 1.4 were not considered for further experiments. The antisera of the last bleeds of all the rabbits were lyophilized and stored at 4 °C until use.The optimization and specificity experiments were performed for all the antisera except for R1.1 and R1.4. All Analyst, October 1997, Vol. 122 1121other experiments were only carried out with the antiserum of R2.1. Determination of Optimum Conditions for the RIA As a separation technique, a second antibody method was evaluated.17 The optimum ratio of GARGG antiserum to NRS was investigated. Final dilutions of GARGG (1 + 124) and NRS (1 + 999) resulted in the highest concentrations of tracer.The optimum second incubation time for the precipitation of the antibody-bound tracer was 20 h at room temperature. In comparison with the adsorption technique, this method resulted in lower NSB values (1–3%) and in higher dilutions of the antiserum to reach 50% binding of the tracer (1 + 12 799). The performance characteristics of the immunoassay are demonstrated in a calibration graph obtained with the serum of R2.1 for the concentration range 100–10 000 pg ml21 (Fig. 3), using the second antibody method as the separation technique. The curve of best fit was obtained by applying the four-parametric logistic model to the experimental data.18 Limit of Quantification The limit of quantification (LOQ) of the RIA can be defined as the lowest concentration on the calibration graph that can be measured with acceptable precision and accuracy. The precision at this level has to be less than or equal to 20%. To determine the LOQ, five standard solutions of increasing concentration were used and the relative standard deviation was determined.The LOQ was 0.1 ng ml21. Precision and Recovery The precision and recovery of the method could be calculated after replicate analyses of independently prepared quality control urine samples. The concentration levels of the samples ranged from 0.25 to 1000 ng ml21. The intra- and inter-assay relative standard deviations were 4.01–6.30 and 9.13–12.28%, respectively, and the accuracy was about 100% (Table 1).Specificity The antibody specificity was assessed by measuring the crossreactivity to other hypnotics and to some widely used anxiolytics, antidepressants, analgesics and analeptics. The following compounds were found not to be detectable at a level of 10 mg ml21: chlordiazepoxide, diazepam, clobazam, nitrazepam, flunitrazepam, bromazepam, lormetazepam, triazolam, meprobamate, methaqualone, hexobarbital, fenobarbital, buspirone, hydroxyzine, chlorcyclizine, codeine, cocaine, caffein, ibuprofen, trazodone, zopiclone, chlorpromazine and acetylsalicylic acid.None of the above was found to cross-react ( < 0.005%) with the immunoassay. To rule out cross-reactivity with endogenous compounds, several urine samples of volunteers were screened and no crossreactivity could be detected. To select one antiserum out of the pool, the specificity towards the metabolites was examined. The two major metabolites of zolpidem, metabolites I and II, and a minor metabolite (metabolite XI) were tested for their cross-reactivity. The degree of cross-reactivity of the metabolites was expressed as the relative dose required for 50% displacement of the maximum tracer binding.The major metabolite (I) showed a low cross-reactivity, ranging from 2.3 3 1027 to 7.7 3 1023%. This could be expected considering that the position of the Fig. 2 Evolution of the serum titres of the two series of rabbits, immunized with the immunogens 1 and 2, respectively.(a) represents the antisera from rabbits R1.1 (/), R1.2 (~), 1.3 (-) and R1.4 (3) and (b) shows the titres of R2.1 (/), R2.2 (~), R2.3 (-) and R2.4 (3). Fig. 3 Calibration graph for zolpidem, together with 95% confidence intervals. The graph follows the equation y = d [(a 2 d)/1 + (x/c)b], where a = 0.8972 ± 0.01006; b = 9.9390 ± 0.4498; c = 6.8896 ± 0.0280; d = 0.02248 ± 0.01195; and r2 = 0.9764. Table 1 Intra- and inter-assay variance characteristics and recovery Added/ng RSD (%) n Recovery (%) Intra-assay 0.5 4.01 6 103.4 1.75 4.48 6 101.5 7.5 6.30 6 98.1 1000 5.90 6 99.3 Inter-assay 0.25 12.28 10 100.9 1 9.30 9 99.4 7.5 7.40 10 101.3 1000 9.13 9 104.0 1122 Analyst, October 1997, Vol. 122carboxylic acid function is close to the binding site with the antibodies. The cross-reactivity towards metabolite II was higher and varied from 0.9 to 6.9%. In this metabolite, the substituent is positioned further away from the binding site.Metabolite XI, which has an identical structure with zolpidem, except for the hydroxylic function on the acetamide side-chain in position 3 of the imidazopyridine skeleton, showed a very high cross-reactivity (91–116%). Because of the pre-screening purpose of this RIA, the antiserum with the highest crossreactivity for the main metabolite was selected, i.e., the antiserum of rabbit 2.1 (6th bleed). The results are shown in Table 2. Linearity To confirm the linearity of the response throughout the range of the calibration graph, samples were checked to see whether they diluted in parallel to the calibration graph.An unknown urine sample, containing approximately 4.6 ng ml21 of the drug, was analysed three times, both undiluted and diluted with blank urine (1 + 1, 1 + 3, 1 + 7 and 1 + 15). The observed (y) and the expected (x) values yielded the following linear regression equation : y = 0.0115 + 0.9703x (r = 0.9938). Human Urine and Serum Samples A set of 14 urine samples (including a blank) were collected from a volunteer during a 48 h period after the intake of 10 mg of zolpidem (Stilnoct).The samples were analysed using the described RIA procedure and the results were expressed in terms of zolpidem concentration, although there was also a contribution from the metabolites. The highest value (994 ng ml21) was reached 3 h after the ingestion of one tablet of Stilnoct. After 48 h, the hypnotic could still easily be detected (about 6.3 ng ml21).From the same volunteer and at specific time intervals, 13 serum samples (including a blank) were collected and analysed by RIA. The serum levels measured covered a concentration from 1.3 to 122 ng ml21. The highest value was recorded 2 h after administration (about 122 ng ml21). Thereafter, the level of zolpidem decreased to 1.3 ng ml21 after 48 h. Fig. 4 shows the serum concentration–time profile of zolpidem. Discussion This paper describes the development of an RIA for zolpidem and its metabolites in urine and serum.Two immunogens were synthesized to study the difference in affinity and specificity between an immunogen with and without a spacer. The antibodies from the rabbits immunized with the immunogen with a spacer showed a much higher affinity than those immunized with the other immunogen. A difference in specificity could not be seen. The synthesis and purification of N,N-didemethylzolpidem-N-{2-{3-(4-hydroxy-3-[125I]-iodophenyl)} methyl propionate} resulted in a high radiochemical purity and specific activity of the radioligand,14 thereby providing a very sensitive immunoassay (LOQ = 0.1 ng ml21).Owing to the nature of the synthesized hapten and immunogens, the detected cross-reactivity of the three metabolites is to be expected. The important fact is that zolpidem and its metabolites can be effectively detected in urine and serum up to 48 h after the intake of one tablet of Stilnoct (10 mg of zolpidem), without interference from other drugs.The proposed RIA can be used in the pre-screening of toxicological samples. Its applica- Table 2 Specificity of the antisera towards three metabolites of zolpidem in comparison with the unmetabolised drug Cross-reactivity (%) Compound R1.2 R1.3 R2.1 R2.2 R2.3 R2.4 Zolpidem 100 100 100 100 100 100 7.7 3 1023 2.3 3 1023 2.3 3 1027 4.2 3 1023 3 3 1024 1.6 3 1023 2.53 2.4 5.55 0.9 6.9 2.9 91 116 116 114 104 92 Analyst, October 1997, Vol. 122 1123bility to the analysis of blood and plasma samples is under study. Dr. R. Busson, Dr. G. Janssen and Dr. J. Rozensky are gratefully acknowledged for recording the NMR and mass spectra. References 1 Th�enot, J. P., Hermann, P., Durand, A., Burke, J. T., Allen, J., Garrigou, D., Vajta, S., Albin, H., Th�ebault, J. J., Olive, G., and Warrington, S. J., in Imidazopyridines in Sleep Disorders, ed. Sauvanet, J. P., Langer, S. Z., and Morselli, P. L., Raven Press, New York, 1988, pp. 139–153. 2 Durand, A., Th�enot, J. P., Bianchetti, G., and Morselli, P. L., Drug Metab. Rev., 1992, 24, 239. 3 Augsburger, M., Giroud, C., Lucchini, P Rivier, L., in Contributions to Forensic Toxicology. The Proceedings of the 31st International Meeting of the International Association of Forensic Toxicologists (TIAFT), 1993, ed. Mueller, R. K., Molinapress, Leipzig, 1994, pp. 18–22. 4 Mendelson, W. B., and Jain, B., Drug Saf., 1995, 13, 257. 5 Debailleul, G., Abi Khalil, F., and Lheureux, P., J. Anal. Toxicol., 1991, 15, 35. 6 Khodasevitch, T., and Volgram, J., Bull. Int. Assoc. Forensic Toxicol., 1996, 26(2), 37. 7 Garnier, R., Guerault, E., Muzard, D., Azoyan, P., Chaumet- Riffaud, A., and Efthymiou, M., Clin. Toxicol., 1994, 32, 391. 8 Guinebault, P., Dubruc, C., Hermann, P., and Th�enot, J. P., J. Chromatogr., 1986, 383, 206. 9 Debruyne, D., Lacotte, J., Hurault De Ligny, B., and Moulin, M., J. Pharm. Sci., 1991, 80, 71. 10 Almirante, L., and Murwann, W., Br. Pat., 1 076 089, 1967. 11 Kaplan, J. P., and George, P., Eur. Pat., 0 050 563, 1982. 12 Kaplan, J. P., and George, P., US Pat., 4 501 745, 1985. 13 Erlanger, B. F., Methods in Enzymology, Academic Press, New York, 1980, vol. 70, p. 85. 14 De Clerck, I., and Daenens, P., J. Radiolabelled Comp. Radiopharm., 1997, 39, 195. 15 Sheehan, J. C., and Hess, G. P., J. Am. Chem. Soc., 1955, 77, 1067. 16 Erlanger, B. F., Pharmacol. Rev., 1973, 25, 271. 17 Utiger, C. R., Parker, M. L., and Daughaday, W. H., J. Clin. Invest., 1962, 41, 254. 18 Dudley, R., Edwards, P., Ekins, R. P., Finney, D. J., McKenzie, I. G. M., Raab, G. M., Rodbard, D., and Rodgers, R. P. C., Clin. Chem. (Winston-Salem, N.C.), 1985, 31, 1264. Paper 7/02869E Received April 28, 1997 Accepted June 6, 1997 Fig. 4 Serum levels, expressed in terms of zolpidem concentration, in one healthy volunteer after a single oral intake of one tablet (10 mg) of Stilnoct. 1124 Analyst, October 1997, Vol. 1
ISSN:0003-2654
DOI:10.1039/a702869e
出版商:RSC
年代:1997
数据来源: RSC
|
22. |
Microbial Detection by a Glucose Biosensor Coupled to a Microdialysis Fibre |
|
Analyst,
Volume 122,
Issue 10,
1997,
Page 1125-1128
F. Palmisano,
Preview
|
|
摘要:
Microbial Detection by a Glucose Biosensor Coupled to a Microdialysis Fibre F. Palmisano*a, A. De Santisa, G. Tantillob, T. Volpicellaab and P. G. Zambonina a Dipartimento di Chimica, Universit`a degli Studi, Via Orabona, 4-70126 Bari, Italy b Istituto di Ispezione degli Alimenti, Universit`a degli Studi, Via per Casamassima Km, 3-70010 Valenzano, Bari, Italy The use of a glucose biosensor coupled to microdialysis sampling in a flow injection analysis system is described to follow the growth of Escherichia coli in a glucose-containing liquid culture medium.The experimental set-up permitted a throughput rate of 25 samples h21. Growth curves were modelled by a modified Gompertz equation, which permitted the determination of lag time and maximum specific growth rate. The time required to produce an appreciable variation in the biosensor response (minimum detection time, MDT) was determined. A plot of MDT versus microbial concentration was found to be linear in the range 106–1010 colony forming units (cfu) ml21.A microbial concentration of 106 cfu ml-1 can be detected after about 5 h. Keywords: Biosensor; microbial detection; Escherichia coli; glucose; microdialysis The quantification of micro-organisms plays a vital role in fermentation processes, food industry, medical practice and industrial waste water monitoring. Thus, an accurate method for the rapid (possibly real-time) determination of biomass is an important goal to be achieved.A large number of detection methods have been developed utilising the optical, electrochemical, biochemical and physical properties of microorganisms (see ref. 1 for an excellent review). An ideal microbial sensor should fulfil several requirements; it must be accurate, sensitive, easy to calibrate and robust. In addition, there should be no need for sample pre-treatment, no interference from the culture conditions, no added reagents and on-line capabilities. The analysis time is, obviously, a characteristic of paramount importance, particularly when microbial detection/quantification is required on a production line, e.g., food processing/ packaging.The term ‘rapid method’ is usually applied to any method presenting an analysis time significantly shorter (e.g., less than 24 h) than that of conventional detection procedures. Among these, viable cell counting methods2–4 (e.g., plate count) are widely used to estimate microbial populations; the main disadvantage of the plate count method is the long incubation period (24–72 h) and the high degree of operator skill required.The microbial content of a sample can be determined by monitoring the microbial metabolism instead of the biomass. Several electrochemical detection systems have been proposed: impedimetry,5 conductivity,6 potentiometry,7 voltammetry8 and amperometry.9The use of biosensors for biomass detection has been scarcely investigated. A glucose biosensor, based on the amperometric mediated enzyme electrode principle, has been adapted for the development of a biosensor ‘knife probe’ and applied to the ultra-rapid in situ assessment of meat freshness.10 Depletion of glucose at the surface relative to the bulk of the meat is indicative of microbial activity.In this paper, a different approach is explored. A glucose biosensor,11,12 based on glucose oxidase (GOD) immobilised in an electrochemically synthesised poly(pyrrole) (PPY) film, is used to monitor in near-real-time the glucose consumption in a liquid culture medium which has been inoculated with a given microbial mass.A microdialysis fibre12,13 is used for sampling purposes; hence, there is no need for biosensor sterilisation (which can denature the enzyme). At the same time on-line dilution of the sample is provided in order to obtain glucose signals in the linear range of response. The calibration status of the sensor can be conveniently checked by injecting, at regular time intervals, a glucose standard.Microbial growth causes glucose depletion in the culture medium which can be followed by the biosensor the response of which changes accordingly. The minimum detection time (MDT), i.e., the time required to detect a significant change in the sensor response, can be related (see below) to the initial microbial concentration through a suitable calibration plot. Hence, a microbial concentration of 106 colony forming units (cfu) ml21 can be detected after about 5 h.Experimental Chemicals Escherichia coli (isolated in our laboratory) was grown in two different liquid culture media: the first contained peptone, NaCl and glucose (2.0 g l21) while the second (Koser-modified medium) contained Na2NH4PO4, MgSO4·7H2O, bovine serum and glucose (2.0 g l21). Plate counts were performed using ‘Agar nutritive’ and ‘Agar nutritive with brilliant green’ (Oxoid, Basingstoke, Hampshire, UK).Glucose oxidase (EC 1.1.3.4 from Aspergillus niger, Type VII S) was obtained from Sigma (St. Louis, MO, USA). A glucose stock solution was prepared from b-D-(+)-glucose (Sigma) and allowed to mutarotate overnight; glucose standard solutions were prepared just before use by dilution of the stock with phosphate buffer. Pyrrole (Aldrich, Milwaukee, WI, USA) was purified by vacuum distillation at 62 °C. All other chemicals were of analytical-reagent grade. Biosensor Preparation Glucose enzyme electrodes (Pt–GOD–PPY) were prepared, as previously described,11 by electrochemical polymerisation at +0.7 V versus Ag/AgCl from a 10 mmol l21 KCl supporting electrolyte containing 0.4 mmol l21 pyrrole and 250 U ml21 of GOD.Pt–GOD–PPY electrodes were overoxidised at +0.7 V overnight in phosphate buffer. The detection potential was +0.7 V versus Ag/AgCl. Apparatus A PAR 273 (EG&G Princeton Applied Research, Princeton, NJ, USA) potentiostat–galvanostat was used for the elec- Analyst, October 1997, Vol. 122 (1125–1128) 1125trosynthesis of the PPY film containing immobilised GOD. A PAR Model 400 electrochemical detector coupled to a Kipp & Zonen (Delft, The Netherlands) BD112 Y–t recorder was used to monitor the response of the glucose biosensor. A Gilson (Villiers le Bel, France) Minipuls 3 peristaltic pump was used in flow experiments. Spectra/por hollow fibres (regenerated cellulose, 150 mm id, 9 mm wall thickness) having a molecular weight cut-off of 9000 Da were obtained from Spectrum Medical Industries (Los Angeles, CA, USA).A Heto (Allerød, Denmark) Type 21 DT-2 thermostatically controlled bath was used for temperature control. The set-up used to follow the microbial metabolism is shown in Fig. 1. A two-channel peristaltic pump (4 in Fig. 1) was used to pump the carrier solution (1) through the microdialysis fibrebased sampler (6), previously described in ref. 13. Unless otherwise stated, the flow rate outside/inside the microdialysis fibre was 300 ml min21.A six-way low pressure injection valve (5) equipped with a 110 ml injection loop (5A) permitted discontinuous sampling of the culture medium (3) or of glucose standards (8) required for the initial calibration of the sensor and checking of the calibration status during the experiment. The maximum throughput allowed by the above-described apparatus was 25 samples h21. Results and Discussion Microdialysis is a dynamic sampling method based on analyte diffusion across a semi-permeable membrane in the presence of a concentration gradient.The concentration ratio at the two sides of the microdialysis membrane is dependent on a number of factors, the most important being (for a given probe) the perfusion rate, temperature, analyte species and physicochemical characteristics of the external medium. The microdialysis probe can, of course, be inserted directly, once sterilised, into the growth medium and glucose in the perfusate monitored continuously.A requirement that must be fulfilled is that the fibre recovery should remain essentially constant during the experiment. This requirement might not be met since fouling of the fibre surface by the growing bacteria can occur, particularly over long periods of time. For this reason a different approach was followed based on the sampler previously described and fully characterised in ref. 13. In this approach, the fibre is continuously washed by the carrier buffer and intermittently contacted by the growth medium only during injections.Control of the calibration status of the sensor is more easily achieved owing to the relatively high sample throughput of the system depicted in Fig. 1. Fig. 2 shows a typical example of the sensor responses for different glucose standards and the relevant calibration plot. As can be seen, the system response is repeatable, linear and sufficiently fast to permit a throughput of about 25 samples h21.The sampling frequency can of course be increased, by increasing the flow rate inside/outside the fibre,13 but at the expense of fibre recovery (i.e., of sensitivity), which decreases on increasing the flow rate. When a glucose-enriched sterile liquid culture medium is inoculated with a given microbial mass, the microbial growth will cause a depletion of the glucose in the culture medium which can be followed in near-real-time by a glucose biosensor the response of which S(t) will change with time accordingly.If S(0) is the sensor response before inoculation (i.e., at t = 0) then the variable y y S t S = - 1 0 ( ) ( ) (1) which is related to the microbial growth, could be described in terms of existing models such as the Gompertz equation modified by Zwietering et al.14 y A A t = × - × - + é ë ê ù û ú ìí ï îï üý ï �ï exp ( ) exp e m m l 1 (2) where t is the time, mm is the maximum specific growth rate, l is the lag time and A is the asymptotic value reached by y for t approaching infinity.Fig. 3 shows some typical sensor outputs at different times in an experiment where the growth medium is inoculated with a known microbial mass (107 cfu ml21). The peaks marked Std1 and Std2 refer to a 10 mmol l21 glucose standard injected before and 30 h after inoculation, respectively. As can be seen, the peak height remained virtually unchanged, demonstrating that the biosensor response remained stable over the typical time frame of such an experiment, and that the fibre recovery is also constant (absence of fouling).Fig. 4 shows typical growth data obtained at different microbial concentrations; the solid lines in Fig. 4 represent the best fit of the experimental data obtained using eqn. (2). Maximum specific growth rate, mm, values of Fig. 1 Schematic diagram of the experimental set-up. 1: Carrier solution reservoir; 2: thermostatically controlled bath; 3: culture medium; 4: twochannel peristaltic pump; 5: six-way low pressure injection valve; 5A injection loop; 6: microdialysis fibre sampler; 7: three-way valve; 8: glucose solution (calibration standard); 9: syringe for manual filling of the injection loop; 10: flow cell with glucose amperometric biosensor; 11: potentiostat; and 12: recorder.Fig. 2 Glucose responses obtained at a Pt–PPYox–GOD biosensor with the experimental configuration shown in Fig. 1. Flow rate inside/outside the microdialysis fibre, 300 ml min21.Inset: calibration plot. 1126 Analyst, October 1997, Vol. 1220.093, 0.094 and 0.095 h21 were estimated at microbial concentrations of 109, 108 and 106 cfu ml21, respectively. Linear interpolation of the experimental points after the lag phase gives a straight line the intercept of which on the time axis can be assumed to be the MDT; note that the MDT values so obtained are essentially the same as the lag time, l, values obtained through eqn.(2). Modelling of the growth curves indicates that the MDT values correspond to about a (7 ± 1)% decrease in the biosensor response measured before microbial inoculation in the growth medium (hence, in practical applications the actual growth curve does not have to be followed entirely). A plot of the MDT values versus the logarithm of the initial microbial concentration (see Fig. 5) was found to be linear, giving a ‘working curve’ from which an unknown microbial concentration can be determined from the measured MDT.As can be seen, under the present experimental conditions, the MDT varies between 25 min and about 5 h for microbial concentrations varying between 1011 and 106 cfu ml21; of course, smaller microbial concentrations require longer times. In any case, it is evident that the total analysis time is significantly shorter than that required by plate count methods, and is essentially dictated by the kinetics of the growth process since the biosensor responds in near-real-time. Furthermore, the flow system described here can monitor several different microbial cultures simultaneously, potentially can be automated and does not require a skilled operator.The usefulness of such an approach in practical applications was demonstrated by preliminary experiments on a contaminated meat sample the microbial mass of which was evaluated by the proposed method and the result compared with that of the conventional plate count method.For this purpose, a naturally contaminated meat sample was washed three times with 10 ml aliquots of a sterile physiological solution to remove the surface microbial population. A 20 ml aliquot was then transferred into 100 ml of ‘modified Koser’ growth medium, and glucose depletion versus time followed. The microbial concentration was also determined, in a parallel experiment, by the plate count method. From the experimental MDT value and the working curve, a microbial concentration of (2.0 ± 1.4) 3 109 cfu g21 was calculated, which was found not to be significantly different (according to a t-test at the 95% confidence level) from the value obtained by the plate count method.It should be noted that since the growth medium may not be selective for E. coli, data obtained on the meat sample might have been influenced by the presence of micro-organisms (e.g., mesophilic bacteria) other than coliforms. The above results on real samples must therefore be considered as preliminary; further work in this direction is underway in our laboratory.The authors thank Professor G. Tiecco for helpful suggestions. Financial support from MURST and National Research Council (CNR, Rome) is gratefully acknowledged. References 1 Hobson, N. S., Tothill, I., and Turner A. P. F., Biosens. Bioelectron., 1996, 11, 455. 2 Hope, C. F. A., and Tubb, R. S., J. Inst. Brewing, 1985, 91, 12. 3 Ding, T., and Schmidt, R.D., Anal. Chim. Acta, 1990, 234, 247. 4 Ashley, N., Dairy Ind. Int., 1991, 56, 39. 5 Zafari, Y., and Martin, W. J., J. Clin. Microbiol., 1977, 5, 545. 6 Richards, J. C. S., Jason, A. C., Hobbs, G., Gibson, D. M., and Christie, R. H., J. Phys., 1978, 11, 560. 7 Wilkins, J. R., Young, R., and Boykin, E., J. Appl. Environ. Microbiol., 1978, 35, 214. 8 Matsunaga, T., and Namba, Y., Anal. Chim. Acta, 1984, 159, 87. 9 Turner, A. P. F., Ramsey, G., and Higgins, I. J. H., Biochem.Soc. Trans., 1983, 11, 445. 10 Kress-Rogers, E., D’Costa, E. J., Sollars, J. E., Gibbs, P. A., and Turner, A. P. F., J. Food Control, 1993, 4, 149. Fig. 3 Glucose responses obtained with a Pt–PPYox–GOD biosensor at different times (a = 6.5; b = 9.5; c = 13 h), during an experiment in which the microbial growth was followed with the experimental configuration shown in Fig. 1. The growth medium was inoculated with E. coli at a known concentration of 107 cfu ml21. Std1 and Std2 refer to a 10 mmol l21 glucose standard injected before inoculation and 30 h after.Fig. 4 Microbial growth curves obtained at three different E. coli concentrations: 109 (5), 108 (-) and 106 (~) cfu ml21. Solid lines are calculated as the best fit of eqn. (2). Fig. 5 MDT versus log of microbial concentration for E. coli. MDT values are obtained by extrapolation on the time axis of the linear portion of the growth curve. Analyst, October 1997, Vol. 122 112711 Centonze, D., Guerrieri, A., Malitesta, C., Palmisano, F., and Zambonin, P.G., Fresenius’ J. Anal. Chem., 1992, 342, 729. 12 Palmisano, F., Centonze, D., Guerrieri, A., and Zambonin, P. G., Biosens. Bioelectron., 1993, 8, 393. 13 Palmisano, F., Centonze, D., Quinto, M.,nd Zambonin, P. G., Biosens. Bioelectron., 1996, 11, 419. 14 Zwietering, M. H., Jongenburger, I., Rombouts, F. M., and van’t Riet, K., Appl. Environ. Microbiol., 1990, 56, 1857. Paper 7/03594B Received May 23, 1997 Accepted August 4, 1997 1128 Analyst, October 1997, Vol. 122 Microbial Detection by a Glucose Biosensor Coupled to a Microdialysis Fibre F. Palmisano*a, A. De Santisa, G. Tantillob, T. Volpicellaab and P. G. Zambonina a Dipartimento di Chimica, Universit`a degli Studi, Via Orabona, 4-70126 Bari, Italy b Istituto di Ispezione degli Alimenti, Universit`a degli Studi, Via per Casamassima Km, 3-70010 Valenzano, Bari, Italy The use of a glucose biosensor coupled to microdialysis sampling in a flow injection analysis system is described to follow the growth of Escherichia coli in a glucose-containing liquid culture medium.The experimental set-up permitted a throughput rate of 25 samples h21. Growth curves were modelled by a modified Gompertz equation, which permitted the determination of lag time and maximum specific growth rate. The time required to produce an appreciable variation in the biosensor response (minimum detection time, MDT) was determined. A plot of MDT versus microbial concentration was found to be linear in the range 106–1010 colony forming units (cfu) ml21.A microbial concentration of 106 cfu ml-1 can be detected after about 5 h. Keywords: Biosensor; microbial detection; Escherichia coli; glucose; microdialysis The quantification of micro-organisms plays a vital role in fermentation processes, food industry, medical practice and industrial waste water monitoring. Thus, an accurate method for the rapid (possibly real-time) determination of biomass is an important goal to be achieved.A large number of detection methods have been developed utilising the optical, electrochemical, biochemical and physical properties of microorganisms (see ref. 1 for an excellent review). An ideal microbial sensor should fulfil several requirements; it must be accurate, sensitive, easy to calibrate and robust. In addition, there should be no need for sample pre-treatment, no interference from the culture conditions, no added reagents and on-line capabilities.The analysis time is, obviously, a characteristic of paramount importance, particularly when microbial detection/quantification is required on a production line, e.g., food processing/ packaging. The term ‘rapid method’ is usually applied to any method presenting an analysis time significantly shorter (e.g., less than 24 h) than that of conventional detection procedures. Among these, viable cell counting methods2–4 (e.g., plate count) are widely used to estimate microbial populations; the main disadvantage of the plate count method is the long incubation period (24–72 h) and the high degree of operator skill required.The microbial content of a sample can be determined by monitoring the microbial metabolism instead of the biomass. Several electrochemical detection systems have been proposed: impedimetry,5 conductivity,6 potentiometry,7 voltammetry8 and amperometry.9The use of biosensors for biomass detection has been scarcely investigated.A glucose biosensor, based on the amperometric mediated enzyme electrode principle, has been adapted for the development of a biosensor ‘knife probe’ and applied to the ultra-rapid in situ assessment of meat freshness.10 Depletion of glucose at the surface relative to the bulk of the meat is indicative of microbial activity. In this paper, a different approach is explored. A glucose biosensor,11,12 based on glucose oxidase (GOD) immobilised in an electrochemically synthesised poly(pyrrole) (PPY) film, is used to monitor in near-real-time the glucose consumption in a liquid culture medium which has been inoculated with a given microbial mass.A microdialysis fibre12,13 is used for sampling purposes; hence, there is no need for biosensor sterilisation (which can denature the enzyme). At the same time on-line dilution of the sample is provided in order to obtain glucose signals in the linear range of response.The calibration status of the sensor can be conveniently checked by injecting, at regular time intervals, a glucose standard. Microbial growth causes glucose depletion in the culture medium which can be followed by the biosensor the response of which changes accordingly. The minimum detection time (MDT), i.e., the time required to detect a significant change in the sensor response, can be related (see below) to the initial microbial concentration through a suitable calibration plot.Hence, a microbial concentration of 106 colony forming units (cfu) ml21 can be detected after about 5 h. Experimental Chemicals Escherichia coli (isolated in our laboratory) was grown in two different liquid culture media: the first contained peptone, NaCl and glucose (2.0 g l21) while the second (Koser-modified medium) contained Na2NH4PO4, MgSO4·7H2O, bovine serum and glucose (2.0 g l21). Plate counts were performed using ‘Agar nutritive’ and ‘Agar nutritive with brilliant green’ (Oxoid, Basingstoke, Hampshire, UK).Glucose oxidase (EC 1.1.3.4 from Aspergillus niger, Type VII S) was obtained from Sigma (St. Louis, MO, USA). A glucose stock solution was prepared from b-D-(+)-glucose (Sigma) and allowed to mutarotate overnight; glucose standard solutions were prepared just before use by dilution of the stock with phosphate buffer. Pyrrole (Aldrich, Milwaukee, WI, USA) was purified by vacuum distillation at 62 °C. All other chemicals were of analytical-reagent grade.Biosensor Preparation Glucose enzyme electrodes (Pt–GOD–PPY) were prepared, as previously described,11 by electrochemical polymerisation at +0.7 V versus Ag/AgCl from a 10 mmol l21 KCl supporting electrolyte containing 0.4 mmol l21 pyrrole and 250 U ml21 of GOD. Pt–GOD–PPY electrodes were overoxidised at +0.7 V overnight in phosphate buffer. The detection potential was +0.7 V versus Ag/AgCl. Apparatus A PAR 273 (EG&G Princeton Applied Research, Princeton, NJ, USA) potentiostat–galvanostat was used for the elec- Analyst, October 1997, Vol. 122 (1125–1128) 1125trosynthesis of the PPY film containing immobilised GOD. A PAR Model 400 electrochemical detector coupled to a Kipp & Zonen (Delft, The Netherlands) BD112 Y–t recorder was used to monitor the response of the glucose biosensor. A Gilson (Villiers le Bel, France) Minipuls 3 peristaltic pump was used in flow experiments. Spectra/por hollow fibres (regenerated cellulose, 150 mm id, 9 mm wall thickness) having a molecular weight cut-off of 9000 Da were obtained from Spectrum Medical Industries (Los Angeles, CA, USA).A Heto (Allerød, Denmark) Type 21 DT-2 thermostatically controlled bath was used for temperature control. The set-up used to follow the microbial metabolism is shown in Fig. 1. A two-channel peristaltic pump (4 in Fig. 1) was used to pump the carrier solution (1) through the microdialysis fibrebased sampler (6), previously described in ref. 13. Unless otherwise stated, the flow rate outside/inside the microdialysis fibre was 300 ml min21. A six-way low pressure injection valve (5) equipped with a 110 ml injection loop (5A) permitted discontinuous sampling of the culture medium (3) or of glucose standards (8) required for the initial calibration of the sensor and checking of the calibration status during the experiment. The maximum throughput allowed by the above-described apparatus was 25 samples h21.Results and Discussion Microdialysis is a dynamic sampling method based on analyte diffusion across a semi-permeable membrane in the presence of a concentration gradient. The concentration ratio at the two sides of the microdialysis membrane is dependent on a number of factors, the most important being (for a given probe) the perfusion rate, temperature, analyte species and physicochemical characteristics of the external medium. The microdialysis probe can, of course, be inserted directly, once sterilised, into the growth medium and glucose in the perfusate monitored continuously. A requirement that must be fulfilled is that the fibre recovery should remain essentially constant during the experiment.This requirement might not be met since fouling of the fibre surface by the growing bacteria can occur, particularly over long periods of time. For this reason a different approach was followed based on the sampler previously described and fully characterised in ref. 13. In this approach, the fibre is continuously washed by the carrier buffer and intermittently contacted by the growth medium only during injections. Control of the calibration status of the sensor is more easily achieved owing to the relatively high sample throughput of the system depicted in Fig. 1. Fig. 2 shows a typical example of the sensor responses for different glucose standards and the relevant calibration plot. As can be seen, the system response is repeatable, linear and sufficiently fast to permit a throughput of about 25 samples h21.The sampling frequency can of course be increased, by increasing the flow rate inside/outside the fibre,13 but at the expense of fibre recovery (i.e., of sensitivity), which decreases on increasing the flow rate. When a glucose-enriched sterile liquid culture medium is inoculated with a given microbial mass, the microbial growth will cause a depletion of the glucose in the culture medium which can be followed in near-real-time by a glucose biosensor the response of which S(t) will change with time accordingly.If S(0) is the sensor response before inoculation (i.e., at t = 0) then the variable y y S t S = - 1 0 ( ) ( ) (1) which is related to the microbial growth, could be described in terms of existing models such as the Gompertz equation modified by Zwietering et al.14 y A A t = × - × - + é ë ê ù û ú ìí ï îï üý ï �ï exp ( ) exp e m m l 1 (2) where t is the time, mm is the maximum specific growth rate, l is the lag time and A is the asymptotic value reached by y for t approaching infinity.Fig. 3 shows some typical sensor outputs at different times in an experiment where the growth medium is inoculated with a known microbial mass (107 cfu ml21). The peaks marked Std1 and Std2 refer to a 10 mmol l21 glucose standard injected before and 30 h after inoculation, respectively. As can be seen, the peak height remained virtually unchanged, demonstrating that the biosensor response remained stable over the typical time frame of such an experiment, and that the fibre recovery is also constant (absence of fouling).Fig. 4 shows typical growth data obtained at different microbial concentrations; the solid lines in Fig. 4 represent the best fit of the experimental data obtained using eqn. (2). Maximum specific growth rate, mm, values of Fig. 1 Schematic diagram of the experimental set-up. 1: Carrier solution reservoir; 2: thermostatically controlled bath; 3: culture medium; 4: twochannel peristaltic pump; 5: six-way low pressure injection valve; 5A injection loop; 6: microdialysis fibre sampler; 7: three-way valve; 8: glucose solution (calibration standard); 9: syringe for manual filling of the injection loop; 10: flow cell with glucose amperometric biosensor; 11: potentiostat; and 12: recorder.Fig. 2 Glucose responses obtained at a Pt–PPYox–GOD biosensor with the experimental configuration shown in Fig. 1. Flow rate inside/outside the microdialysis fibre, 300 ml min21. Inset: calibration plot. 1126 Analyst, October 1997, Vol. 1220.093, 0.094 and 0.095 h21 were estimated at microbial concentrations of 109, 108 and 106 cfu ml21, respectively. Linear interpolation of the experimental points after the lag phase gives a straight line the intercept of which on the time axis can be assumed to be the MDT; note that the MDT values so obtained are essentially the same as the lag time, l, values obtained through eqn.(2). Modelling of the growth curves indicates that the MDT values correspond to about a (7 ± 1)% decrease in the biosensor response measured before microbial inoculation in the growth medium (hence, in practical applications the actual growth curve does not have to be followed entirely). A plot of the MDT values versus the logarithm of the initial microbial concentration (see Fig. 5) was found to be linear, giving a ‘working curve’ from which an unknown microbial concentration can be determined from the measured MDT.As can be seen, under the present experimental conditions, the MDT varies between 25 min and about 5 h for microbial concentrations varying between 1011 and 106 cfu ml21; of course, smaller microbial concentrations require longer times. In any case, it is evident that the total analysis time is significantly shorter than that required by plate count methods, and is essentially dictated by the kinetics of the growth process since the biosensor responds in near-real-time.Furthermore, the flow system described here can monitor several different microbial cultures simultaneously, potentially can be automated and does not require a skilled operator. The usefulness of such an approach in practical applications was demonstrated by preliminary experiments on a contaminated meat sample the microbial mass of which was evaluated by the proposed method and the result compared with that of the conventional plate count method.For this purpose, a naturally contaminated meat sample was washed three times with 10 ml aliquots of a sterile physiological solution to remove the surface microbial population. A 20 ml aliquot was then transferred into 100 ml of ‘modified Koser’ growth medium, and glucose depletion versus time followed. The microbial concentration was also determined, in a parallel experiment, by the plate count method.From the experimental MDT value and the working curve, a microbial concentration of (2.0 ± 1.4) 3 109 cfu g21 was calculated, which was found not to be significantly different (according to a t-test at the 95% confidence level) from the value obtained by the plate count method. It should be noted that since the growth medium may not be selective for E. coli, data obtained on the meat sample might have been influenced by the presence of micro-organisms (e.g., mesophilic bacteria) other than coliforms.The above results on real samples must therefore be considered as preliminary; further work in this direction is underway in our laboratory. The authors thank Professor G. Tiecco for helpful suggestions. Financial support from MURST and National Research Council (CNR, Rome) is gratefully acknowledged. References 1 Hobson, N. S., Tothill, I., and Turner A. P. F., Biosens. Bioelectron., 1996, 11, 455. 2 Hope, C. F. A., and Tubb, R. S., J. Inst. Brewing, 1985, 91, 12. 3 Ding, T., and Schmidt, R. D., Anal. Chim. Acta, 1990, 234, 247. 4 Ashley, N., Dairy Ind. Int., 1991, 56, 39. 5 Zafari, Y., and Martin, W. J., J. Clin. Microbiol., 1977, 5, 545. 6 Richards, J. C. S., Jason, A. C., Hobbs, G., Gibson, D. M., and Christie, R. H., J. Phys., 1978, 11, 560. 7 Wilkins, J. R., Young, R., and Boykin, E., J. Appl. Environ. Microbiol., 1978, 35, 214. 8 Matsunaga, T., and Namba, Y., Anal. Chim. Acta, 1984, 159, 87. 9 Turner, A. P. F., Ramsey, G., and Higgins, I. J. H., Biochem. Soc. Trans., 1983, 11, 445. 10 Kress-Rogers, E., D’Costa, E. J., Sollars, J. E., Gibbs, P. A., and Turner, A. P. F., J. Food Control, 1993, 4, 149. Fig. 3 Glucose responses obtained with a Pt–PPYox–GOD biosensor at different times (a = 6.5; b = 9.5; c = 13 h), during an experiment in which the microbial growth was followed with the experimental configuration shown in Fig. 1. The growth medium was inoculated with E. coli at a known concentration of 107 cfu ml21. Std1 and Std2 refer to a 10 mmol l21 glucose standard injected before inoculation and 30 h after. Fig. 4 Microbial growth curves obtained at three different E. coli concentrations: 109 (5), 108 (-) and 106 (~) cfu ml21. Solid lines are calculated as the best fit of eqn. (2). Fig. 5 MDT versus log of microbial concentration for E. coli. MDT values are obtained by extrapolation on the time axis of the linear portion of the growth curve. Analyst, October 1997, l. 122 112711 Centonze, D., Guerrieri, A., Malitesta, C., Palmisano, F., and Zambonin, P. G., Fresenius’ J. Anal. Chem., 1992, 342, 729. 12 Palmisano, F., Centonze, D., Guerrieri, A., and Zambonin, P. G., Biosens. Bioelectron., 1993, 8, 393. 13 Palmisano, F., Centonze, D., Quinto, M., and Zambonin, P. G., Biosens. Bioelectron., 1996, 11, 419. 14 Zwietering, M. H., Jongenburger, I., Rombouts, F. M., and van’t Riet, K., Appl. Environ. Microbiol., 1990, 56, 1857. Paper 7/03594B Received May 23, 1997 Accepted August 4, 1997 1128 Analyst, October 1997, Vol. 122
ISSN:0003-2654
DOI:10.1039/a703594b
出版商:RSC
年代:1997
数据来源: RSC
|
23. |
Characterization of Polymer Films of Pyrrole Derivatives for Chemical Sensing by Cyclic Voltammetry, X-ray Photoelectron Spectroscopy and Vapour Sorption Studies |
|
Analyst,
Volume 122,
Issue 10,
1997,
Page 1129-1138
Zhiping Deng,
Preview
|
|
摘要:
Characterization of Polymer Films of Pyrrole Derivatives for Chemical Sensing by Cyclic Voltammetry, X-ray Photoelectron Spectroscopy and Vapour Sorption Studies Zhiping Deng, David C. Stone and Michael Thompson Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario, Canada M5S 3H6 Eight different conducting polymer films formed from pyrrole and N-substituted pyrrole derivatives were characterized by cyclic voltammetry, X-ray photoelectron spectroscopy (XPS) and scanning electron microscopy.In particular, the XPS of poly[N-butylpyrrole], poly[N-(2-carboxyethyl)pyrrole], poly[N–(6–hydroxyhexyl)pyrrole] and poly[N-(6-tetrahydropyranylhexyl)pyrrole] is reported for the first time. The vapour sorption properties of these films were also examined by forming the films onto the electrodes of thickness-shear mode acoustic wave sensors. The influence of the pendant side chain is apparent in both the electrochemical behaviour, composition, doping level, morphology and the nature and extent of polymer–vapour interactions.The latter can be rationalized by consideration of vapour physical properties and solvatochromic parameters. Keywords: Cyclic voltammetry; X-ray photoelectron spectroscopy; poly(pyrrole) and derivatives; vapour sorption; thickness-shear mode; quartz crystal; acoustic wave chemical sensor Organic conducting polymers such as poly(pyrrole) have found extensive application in highly diverse fields such as biomaterials, chemistry, electronics, microfabrication, non-linear optics, sensors and textiles.1–5 In the area of analytical chemistry, for example, such materials have been used to form chemically modified electrodes,6 permeation membranes,7,8 liquid chromatographic stationary phases9,10 and chemical and biological sensors.11–16 These chemical applications are all based on the physico-chemical interactions that occur between the conducting polymer and the respective chemical species.Our own interest in these materials is their use as selective (or partially selective) coatings for thickness-shear mode (TSM) acoustic wave sensors. One reason for using conducting polymers in this way is the ease of film formation by direct electropolymerization onto the TSM electrode surface. Other advantages lie in the ease with which redox state and incorporated counter ions can be changed electrochemically, and the wide range of functionalities that can be introduced through chemical reaction with the pyrrole ring nitrogen. This in turn provides various mechanisms for modifying the extent of different reversible interactions between the polymer films and different analytes, thus affecting both the sensitivity and selectivity of the sensor system.In a series of papers,17–20 we have described the application of various N-substituted poly(pyrroles) to organic vapour sensing. These include the parent compound, poly(pyrrole) (PPY), and poly(N-methylpyrrole) (PMPY), poly(N-butylpyrrole) (PBPY), poly[N-(2-cyanoethyl)pyrrole] (PCPY), poly[N- (2-carboxyethyl)pyrrole] (PCbPY), poly(N-phenylpyrrole) (PPPY), poly[N–(6–hydroxyhexyl)pyrrole] (PHPY) and poly[N-(6-tetrahydropyranylhexyl)pyrrole] (PTHPY).In the preceding papers, we have described the preparation, characterization and effect of redox state of PCPY18,20 as well as the application of all eight polymers to the selective detection of aroma components.19 In the present paper, we provide a more detailed characterization and comparison of the eight conducting polymers, with an emphasis on their formation and application as coatings for organic vapour sensors.Experimental Reagents and Materials The solvents hexane, toluene, methanol, butanal, acetonitrile, triethylamine, ethanol, butan-1-ol, hexan-1-ol and nonan-1-ol (analytical-reagent grade, Aldrich, Milwaukee, WI, USA) were used as received. Pyrrole (98%), N-methypyrrole (99%) and N- (2-cyanoethyl)pyrrole (99%) (Aldrich) were vacuum-distilled before use.Both N-phenylpyrrole (99%) (Aldrich) and tetrabutylammonium perchlorate (TBAP) (electrochemical grade, Fluka, Buchs, Switzerland) were used as received. N-(2-Carboxyethyl) pyrrole, N-butylpyrrole, N-(6-tetrahydropyranylhexyl) pyrrole and N-(6-hydroxyhexyl)pyrrole were synthesized as described previously.19 All other chemicals were obtained from Aldrich and used as received. The piezoelectric crystals were rough or optically polished 9 MHz AT-cut quartz crystals with gold electrodes (International Crystal Manufacturing, Oklahoma City, OK, USA).A combined Pt–Ag/AgCl (3 mol l21 KCl) electrode (Metrohm, Herisau, Switzerland) was utilized as the counter and reference electrode, respectively. Polymerization Procedures One electrode of the TSM device was used as a working electrode for the electropolymerization of pyrrole and its derivatives using a cell as described previously.17 Unpolished crystals were used for vapour sorption experiments while optically polished crystals were used for the cyclic voltammetry studies.All devices were washed with acetone and dried under nitrogen before use. Polymerization was achieved using a solution containing monomer (0.1 mol l21) and TBAP (0.1 mol l21) in de-oxygenated acetonitrile using either constant potential or cyclic voltammetric deposition with a potentiostat/ galvanostat (Model 273, EG&G Princeton Applied Research, Princeton, NJ, USA).The polymerization potentials for the different monomers have been listed previously.19 The coated TSM sensors were then rinsed with acetonitrile, dried with nitrogen, washed with acetone, dried with nitrogen and finally placed in an oven for 1 h at 100 °C. The coating mass for each TSM device was determined from the in-air frequency difference before and after polymer deposition. Analyst, October 1997, Vol. 122 (1129–1138) 1129Cyclic Voltammetry Cyclic voltammograms for film deposition and growth on the electrodes of polished TSM devices were obtained for ten cycles at a sweep rate of 100 mV s21 for potentials varying between 20.8 and +1.2 V versus Ag/AgCl, depending on the monomer used.The resulting films were then studied using sweep rates of 20–100 mV s21 and potentials varying between 20.8 and +1.2 V versus Ag/AgCl using TBAP (0.1 mol l21) in acetonitrile as the supporting electrolyte.X-ray Photoelectron Spectroscopy (XPS) XPS was performed using a Leybold Max 200 XPS (Leybold, Cologne, Germany) instrument using an unmonochromatized Mg Ka source and an analysis area of 2 3 4 mm2. Survey and low resolution spectra were obtained using a pass energy of 192 eV; high resolution spectra were acquired with a pass energy of 48 eV. All spectra were satellite-subtracted and normalized using software and elemental sensitivity factors provided by the manufacturer.The binding energy scale was further calibrated to 285.0 eV for the main C (1s) feature in order to compensate for sample charging effects. Scanning Electron Microscopy Scanning electron micrographs were obtained using a Hitachi Model S-570 microscope and recorded using the Quartz PCI image capture system. In order to reduce charging effects, the coated TSM devices were sputter-coated with a thin gold film using an argon plasma vapour deposition system (Polaron Model E5100, Polaron, Watford, Hertfordshire, UK).All micrographs were acquired using an accelerating voltage of 18 kV. Vapour Sorption Studies Vapour generation was achieved by bubbling helium gas through the corresponding liquid at room temperature and pressure using the flow system described previously.18,19 The flow rates of the sample and purge streams were held at 30 ml min21 for all experiments. The frequency shift of the polymer-coated TSM sensors was measured using a universal frequency counter (Model HP5334B, Hewlett-Packard, Avondale, PA, USA).Data were collected, displayed and stored using a Macintosh II computer equipped with an IEEE 488 interface bus using software written in-house. The coated TSM sensors were initially purged using pure helium carrier gas for 1–2 h to allow the device to stabilize. Once a stable baseline had been obtained, the stream was switched to the sample vapour and returned to the purge gas once the steady-state frequency shift had been obtained.Measurements were performed in triplicate for each coating– vapour combination. Vapour concentrations were obtained using a liquid nitrogen trap for a fixed time interval and weighing the resulting condensate. The measured vapour concentrations are given in Table 1, together with the relevant physical properties for the compounds studied. Results and Discussion Cyclic Voltammetry The cyclic voltammograms for the electropolymerization of all eight monomers are fairly similar, the main differences being in the position and magnitude of the anodic and cathodic peaks.That for N-(2-carboxyethyl)pyrrole is shown in Fig. 1 as a representative example. As can be seen, the first oxidation peak occurs at a higher oxidation potential (+1.06 V) than that of pyrrole (+0.85 V). In fact, the oxidation potentials of all seven N-substituted pyrroles are higher than that of the parent compound (Table 2). This is attributable to the steric and electronic effects of the pendant side chain, the inductive effect rendering the oxidation process more difficult for the mono- Table 1 Physical data and concentration values for the test solvents and corresponding vapours Cv/mg l21 Boiling- Density/ Molecular (at 21 °C and Solvent point/°C g cm23 mass/g mol21 30 ml min21) Hexane 69 0.659 86.16 31 Toluene 110 0.865 110.6 10 Water 100 1.000 18.00 15 Acetonitrile 82 0.786 41.05 34 Triethylamine 88.8 0.726 101.19 30 Butanal 75 0.800 72.11 22 Methanol 64.7 0.791 32.04 40 Ethanol 78 0.785 46.07 27 Butan-1-ol 117.7 0.810 74.12 8.3 Hexan-1-ol 156.5 0.814 102.18 1.2 Nonan-1-ol 215 0.827 144.26 0.3 Fig. 1 Cyclic voltammetric polymerization of N-(2-carboxyethyl)pyrrole (0.1 mol l21 in 0.1 mol l21 TBAP–acetonitrile) at scan rate of 100 mV s21 versus Ag–AgCl (3 mol l21). Table 2 Anodic and cathodic peak potentials (versus Ag/AgCl) for the cyclic voltammetric polymerization of N-substituted pyrrole monomers (0.1 mol l21 monomer in 0.1 mol l21 TBAP–acetonitrile, 100 mV s21 scan rate) Epa Epa Epc Polymer (monomer)/V* (polymer)/V† (polymer)/V† PPY 0.85 20.08 20.23 PMPY 0.90 0.46 0.42 PBPY 1.08 0.65 0.47 PCPY 1.10 0.70 0.64 PCbPY 1.06 0.68 0.56 PPPY 1.19 0.67 0.58 PHPY 1.05 0.68 0.52 PTHPY 1.08 0.69 0.58 * From first scan.† From second scan. 1130 Analyst, October 1997, Vol. 122mers.21 When the potential is reversed on the first scan, the anodic current continues to increase for a short time before decreasing.The following cathodic current is also higher than that in the forward scan until the potential is +0.76 V, resulting in a loop in the voltammogram. This is indicative of a nucleation mechanism, and is commonly observed for the electropolymerization of conducting polymers.22 The effect is attributed to the fact that polymerization occurs faster at a polymeric nucleus than at the uncoated electrode surface. The anodic peak for the second scan occurs at a much lower potential (+0.68 V), and corresponds to polymer growth with incorporation of perchlorate counter ions (Table 2).The peak is broad because there will be a distribution of oligomer sizes, while diffusion of counter ions in and out of the film will be slow. Adams23 has concluded that whenever the second and subsequent sweeps of a cyclic voltammogram differ markedly from the first, a followup chemical reaction has occurred. These results, therefore, show the characteristics of an ECE reaction, in which electron transfer is followed by a chemical reaction and subsequent electron transfer reaction.24 Therefore, the overall polymerization mechanism for the pyrrole derivatives is the same as that for the parent compound.25 Another feature of the voltammograms is that, on successive scans, the anodic peak shifts to increasingly positive potential with a corresponding increase in current until a limiting value is reached.A similar effect is observed for the cathodic peak, except here the shift is towards a more negative potential.The shift is, at least in part, due to the polymer film resistance giving rise to a larger overpotential. The ratio of the cathodic to anodic peak currents is also less than one, confirming that the anodic peak consists of both reversible and irreversible contributions. The anodic peak potentials for the different films are very similar with the exception of PPY and PMPY. This may be explained by considering the effect of side-chain length on the porosity of the polymer.It is anticipated that as the length of the substituent chain increases, the polymer chains will be forced farther apart making it easier for counter ions to diffuse in and out of the film. This is confirmed by the cathodic peak potentials for the series PPY, PMPY, PBPY and PTHPY, which show an increasing shift towards positive potentials. Following film formation, the cyclic voltammogram for each polymer film was obtained at various scan rates using TBAP in acetonitrile without monomer as the supporting electrolyte.Well-defined, broad anodic and cathodic peaks were obtained in all cases (Fig. 2), which reflects the slow rate of counter ion transfer in and out of the film as well as interactions between electroactive sites and the electrochemical non-equivalence of these sites.26 The magnitude of the anodic and cathodic peak currents increases linearly with scan rate, while peak separation also increases.The polymer redox processes are, therefore, not diffusion-controlled. Table 3 summarizes the positions and separation of the anodic and cathodic peaks. Both the anodic (Epa) and cathodic (Epc) peak potentials become increasingly positive relative to the parent PPY as the chain length of the Nsubstituent increases. This trend arises from increasing distortion of inter-ring coplanarity, which would destabilize the cationic (oxidized) form of these polymers.One potential benefit of the positive shift in Epa is that films formed from Nsubstituted pyrrole derivatives will be less sensitive to atmospheric oxidation. This implies improved handling, storage and lifetime characteristics for chemical sensors fabricated using such films. Another observation is that the peak separation (DE) is considerably lower for PPY and PMPY compared with the other polymer films. In fact, the oxidation and reduction of both PPY and PMPY films are essentially reversible, while other irreversible processes must be involved for the remaining films.As an example, PHPY shows two anodic waves, the second appearing at Å 1.0 V for scan rates of 80 and 100 mV s21. This is likely associated with oxidation of the hydroxyl group, although further work is needed to confirm this. X-ray Photoelectron Spectroscopy Although the X-ray photoelectron spectra of PPY, PMPY, PPPY and PCPY films have been previously reported, 17,18,20,26–31 we have included the relevant data with those for PBPY, PCbPY, PHPY and PTHPY in Table 4 for ease of comparison. This lists the results of peak deconvolution on the C(1s), N(1s), O(1s) and Cl(2p) regions for all eight as-prepared polymer films.In addition, Fig. 3 shows the C(1s) spectrum obtained for PCbPY. All eight films show the expected principal C(1s) component at 285 eV arising from the pyrrole ring a and b carbons26 and methylene groups within the pendant side chain.These species are unresolvable under the experimental conditions used here. A second C(1s) component occurring between 286.1 and 286.4 eV can arise from a number of different sources. For PCPY, for example, it is attributed to the nitrile group as discussed previously.18 For PHPY, there is also a contribution from the side chain –CH2OH group, while PTHPY contains two ether groups. It is, therefore, not surprising that these three films show significant enhancement of the 286 eV component relative to the remaining polymers.The origin of this peak for PPY, PMPY, PBPY, PCbPY and PPPY is amenable to several interpretations. Pfluger and Street26 attribute this band in PPY to ‘disorder type’ carbons, which they define as being crosslinked, chain-terminating and non-a,aA bonded carbons as well as partially saturated rings. There will also be a small contribution from unavoidable hydrocarbon contaminants, while Atanasoska et al.28 have suggested that electrostatic interaction of ring a carbons with counter ions will also have an effect.A fourth possibility is the presence of covalently bound chlorine species, which we will consider shortly. The final C(1s) component between 288 and 289 eV is generally attributed to carbonyl or carboxyl species resulting from chain termination.21,28 This has been partly confirmed by FTIR studies of chemically polymerized PPY samples,29 which clearly show the presence of a small band at 1705 cm21 that scales in intensity with the 288 eV C(1s) component. For PCbPY, there is also the pendant carboxyl group (289.1 eV), which clearly increases the relative contribution of this component.Turning now to the N(1s) region, all the films show two peaks located at approximately 400 and 402 eV, with the major component being the lower binding energy signal. This is in general agreement with other XPS studies of different PPY species,26,28 a number of which also show a small shoulder at Å 397 eV.17,20,29,30 The main peak at 400 eV arises from the neutral pyrrole ring nitrogen, while the higher binding energy component is generally attributed to partially charged nitrogens within bipolaron sub-units.The presence of the shoulder at 397 eV depends, to some extent, on the experimental signal-to-noise ratio but also on the film preparation conditions since it is more pronounced for neutral than as-prepared or oxidized PPY and PCPY.17,20 Lei et al.29 have attributed it to –CNN– defects in the pyrrole backbone, while Vigmond et al.17 ascribed it to interchain hydrogen bonding effects.In this latter interpretation, equal intensity peaks on either side of the principal N(1s) line are expected due to electron donation from one nitrogen to another. This being the case, one would expect enhancement of the higher binding energy component in the as-prepared films for polymers in which electron donation from one ring nitrogen to an adjacent ring side chain is possible.Such an effect is, in fact, observed for all the polymers relative to the phenyl- and alkyl-substituted pyrroles (PMPY, PBPY and PPPY). The exception here is PCPY, although the situation in this case may Analyst, October 1997, Vol. 122 1131be masked by the additional nitrogen per pyrrole unit contributed by the nitrile group. The Cl(2p) spectra show peaks with the 2p3/2 component occurring at 207.6 and 200.7 eV.The higher binding energy peak is due to incorporated perchlorate ion, while the lower peak is assigned to covalently incorporated chlorine as discussed previously.20 Support for this assignment comes from the results of Kang et al.30 and Toshima and Tayanagi,31 who observed similar Cl(2p) spectra for chemically polymerized PPY samples. In this respect, it should be noted that the covalently bound chlorine will also result in an increase in the C(1s) component at 286 eV.It is possible to estimate the doping ratio of the as-prepared polymer films from the XPS data using the Cl/N ratio, although Fig. 2 Cyclic voltammograms for the polymer films in 0.1 mol l21 TBAP–acetonitrile without monomer at different scan rates (mV s21). 1132 Analyst, October 1997, Vol. 122this will not be as accurate as that calculated from the composition obtained by bulk elemental analysis since only a thin layer of an irregular surface is being analysed. The estimated doping ratio for each film is given in Table 5, together with the corresponding number of electrons per monomer unit (n) involved in polymerization and subsequent oxidation.It is also possible to estimate a value of n from the cyclic voltammetry data using Nicholson’s model for irreversible charge transfer32 Ip = nFAC*n1/2(pDanF/RT)1/2c(bt) (1) where F is Faraday’s constant, A is the area, C* is the surface concentration of electroactive species, v is the scan rate, D is the diffusion coefficient, a is the transfer coefficient and c(bt) is the current function for irreversible charge transfer.A difficulty here, however, is obtaining a reasonable estimate of the surface concentration for an amorphous, porous polymer film. One crude approach is to use the generally accepted value of n = 2.3–2.5 for PPY and use this to calculate the surface concentration, then assume an identical value for the Nsubstituted derivatives. Agreement between the XPS data and the values of n derived in this manner is surprisingly good, although it breaks down for PCPY and PCbPY.The O(1s) region can generally be deconvoluted into two components at 533.5 and 532.0 eV, representing singly and doubly bonded oxygen species, respectively. For the nonoxygen containing polymers, the only sources of oxygen in the final film will be incorporated counter ion and carboxy groups introduced by chain termination reactions.28 Note that perchlorates also show oxygen binding energies of Å 533 eV, which is unresolvable from the other species for the instrumental conditions used in this study.For these reasons, the O(1s) region has rarely been discussed in previous XPS studies of poly- (pyrroles). Simple composition calculations assuming that the perchlorate and carbonyl oxygen species overlap yield RNO: R– O ratios close to the theoretical 1 : 1 for PPY, PCbPY and PPPY, but break down for PMPY, PBPY, PCPY and PTHPY. This is anticipated since the relative amount of chlorine species other than perchlorate was not taken into account. What the results do show, however, is the expected progressive increase in R–O species for PHPY and PTHPY relative to the parent compound.One curious feature of the O(1s) deconvolution is the apparent presence of an additional oxygen species with a high binding energy of 535 eV. Such species can be observed on plasma-cleaned gold surfaces. This can be discounted, however, since no gold peaks are observed in any of the survey scans.Table 3 Anodic and cathodic peak potentials (versus Ag/AgCl) of Nsubstituted pyrrole polymers in 0.1 mol l21 TBAP–acetonitrile for a 20 mV s21 scan rate Polymer Epa/V Epc/V DE/V PPY 0.28 0.25 0.027 PMPY 0.51 0.48 0.030 PBPY 0.72 0.56 0.15 PCPY 0.82 0.72 0.10 PCbPY 0.88 0.76 0.12 PPPY 0.84 0.70 0.14 PHPY 0.81 0.64 0.16 PTHPY 0.86 0.76 0.10 Table 4 Polymer elemental composition (at.-%) and peak analysis from the XPS data C(1s) N(1s) O(1s) Cl(2p) Polymer BE*/eV at.-% BE*/eV at.-% BE*/eV at.-% BE*/eV at.-% PPY 66.9 11.6 18.2 3.3 285.0 76.9 400.2 71.3 532.4 87.5 207.6 100.0 286.3 18.4 402.1 28.7 533.7 12.5 288.1 4.7 PMPY 74.5 10.4 12.7 2.4 285.1 69.1 400.1 87.3 531.6 64.7 200.9 100.0 286.4 18.7 401.8 12.7 533.0 35.3 288.4 12.2 PBPY 71.9 13.4 12.8 1.9 285.0 75.7 398.5 87.9 531.8 52.9 201.1 78.6 286.4 16.6 400.3 12.1 533.0 47.1 208.1 21.4 288.2 7.7 PCPY 73.7 7.8 15.4 2.3 284.8 51.3 399.9 93.8 532.0 70.7 200.7 40.1 286.4 43.8 401.7 6.2 533.6 29.3 207.5 59.7 288.6 5.0 PCbPY 63.9 9.1 25.0 2.0 284.9 63.7 399.9 59.8 532.3 60.8 200.6 68.2 286.2 18.9 400.4 34.9 533.5 33.6 207.5 31.8 289.1 17.3 401.6 9.4 534.6 5.6 PPPY 73.3 8.1 16.7 1.3 284.9 75.0 400.7 92.0 532.6 63.0 200.4 100.0 286.1 21.0 402.4 8.0 533.5 37.0 287.9 4.0 PHPY 74.6 7.3 16.6 1.5 284.9 58.5 400.1 69.0 532.5 56.8 207.4 100.0 286.1 36.1 401.4 31.0 533.8 36.8 288.5 5.4 535.6 6.4 PTHPY 64.6 5.4 27.5 2.6 285.0 61.0 400.3 62.8 532.4 22.4 200.8 18.8 286.3 29.5 402.3 37.3 533.4 67.5 207.9 81.2 288.3 9.5 534.6 10.2 * BE = Binding energy.Analyst, October 1997, Vol. 122 1133There are several other possible explanations, including differential charging affecting the incorporated counter ion and the presence of an oxidized nitrogen species within the polymer. It is difficult to envisage the latter, however, since the peak is only observed for N-substituted polymers. A final possibility is that it is a fitting artifact arising from the unresolved perchlorate oxygen, since there are an infinite number of mathematical solutions that would equally well fit the O(1s) envelope.Further work is, therefore, needed to clarify this matter. Scanning Electron Microscopy In previous studies, we have found that surface morphology can have an effect on the response of the polymer-coated sensors to different vapours.17–20 Polymer morphology is, in turn, influenced to various degrees by substrate morphology, film thickness and level of oxidation, and the nature of the counter ion. PPY films are frequently described as having a ‘cauliflower- like’ appearance, which arises from the nucleation and phase growth mechanism of the electropolymerization process.This detail is not apparent on very thin films, however, which simply follow the underlying surface. It is, therefore, important to specify both substrate surface roughness and film thickness when comparing the appearance of different films.In this, and our previous studies, the films were formed on microscopically rough surfaces. This was necessary since these polymer films adhere to gold electrodes primarily through mechanical interlock, and are prone to peeling from optically flat surfaces. In particular, the cleaning procedure used in the vapour sorption studies is highly effective as a means of removing conducting polymer films from polished gold electrodes, while films formed on the unpolished devices remain intact.The scanning electron micrographs for six of the polymers are shown in Fig. 4. Micrographs of PPY and PCPY have been published elsewhere.17,18,20,25 The corresponding film thicknesses, calculated from the in-air frequency shift due to film deposition, Dfs, are listed in Table 6. Generally, films of the Nsubstituted pyrrole derivatives show similar morphologies to the parent compound, although there are some differences. Both PMPY and PPPY, for example, exhibit a much more granular surface than PPY even for the same film thickness.The films of PBPY and PHPY are the most similar to PPY, while PCPY and PCbPY again show similar topography but with deep, open channels running into the bulk of the polymer. In this context, it should be noted that these channels disappear if the PCPY films are heavily oxidized by application of a large positive potential, becoming very similar in appearance to the as-prepared PBPY and PHPY films.The most different film in terms of morphology is PTHPY which, although as thick as the PPY and PPPY films, shows no apparent differences from the underlying gold electrode. These differences in morphology have a significant influence on the response of the coated vapour sensors. This occurs firstly through variations in effective surface area, which determines the extent of initial adsorption, and secondly through the film porosity, which affects both the absolute steady-state frequency response (extent of vapour absorption) and recovery time (vapour desorption). Table 5 Doping levels and number of electrons involved in electropolymerization –oxidation from the XPS and cyclic voltammetry (CV) data n Polymer Cl/N ratio Doping ratio XPS CV PPY 0.286 0.29 2.3 2.2 PMPY 0.232 0.23 2.2 2.3 PBPY 0.136 0.14 2.1 2.1 PCPY 0.289 0.15 2.2 2.6 PCbPY 0.219 0.22 2.2 2.7 PPPY 0.154 0.15 2.2 2.2 PHPY 0.205 0.21 2.2 2.5 PTHPY 0.488 0.49 2.5 2.6 Fig. 3 X-ray photoelectron spectra of PCbPY showing the C(1s), N(1s), O(1s) and Cl(2p) regions. 1134 Analyst, October 1997, Vol. 122Vapour Sorption Studies Since the magnitude of the observed frequency shift for vapour sorption (Dfv) varies with both vapour concentration (Cv) and film thickness, it is necessary to normalize the data before meaningful comparisons can be made. In keeping with our earlier work, we therefore calculate the sensor partition coefficient,33 K f f C S = D D nr V (2) where Dfv is in Hz, Dfs (kHz) is the frequency shift due to the polymer coating, r (g cm23) is the film density and Cv is in g cm23. Alternatively, the value of Dfv may be normalized to the shift that would have been obtained for a film having Dfs = 30 kHz.A series of organic vapours were chosen as probe molecules representing a variety of structural and functional group interactions between the vapour and the polymer film (Table 1). Hexane, for example, can interact solely through dispersion forces while toluene also exhibits p-electron overlap and polarizibility interactions.The remaining vapours were chosen to represent differing degrees of hydrogen bond acidity and basicity and dipole–dipole interactions, while a series of four primary alcohols was included to study the effects of chain length and analyte volatility. The solvatochromic parameters for all these vapours are listed in Table 7. While the corresponding Fig. 4 Scanning electron micrographs of six of the conducting polymer films. (a) PMPY; (b) PBPY; (c); PCbPY; (d) PPPY; (e) PHPY; and (f) PTHPY. Table 6 Film thickness and in-air frequency shifts corresponding to electropolymerization of the pyrrole derivatives on the TSM device Polymer PPY PMPY PCbPY PBPY PCPY PPPY PHPY PTHPY Thickness/mm 0.45 0.39 2.96 1.91 0.98 0.49 1.10 0.48 Dfs/kHz 28.8 21.3 104.5 99.0 50.9 26.0 55.2 22.0 Analyst, October 1997, Vol. 122 1135coating parameters are not known for the pyrrole derivatives used in this study, these values are nonetheless useful in comparing the response behaviour of the different vapour– coating combinations.The differing abilities of these combinations to interact through various mechanisms gives rise to different partial selectivities for each coating. This is reflected in the experimentally observed log K values, which are listed in Tables 8 and 9 and summarized graphically in Fig. 5. Comparing hexane and methanol, for example, shows that hexane generally gives much lower normalized frequency shifts ( < 2 kHz) than methanol (2–8 kHz), which reflects the difference in enthalpy of vaporization, and hence vapour concentration, between these two solvents (31.56 and 37.43 kJ mol21 at 25 °C, respectively).When the effect of vapour concentration is factored out by calculating log K, however, we see that the differences in response are actually minimal for the PMPY- and PBPY-coated sensors. On the other hand, fairly large differences in response in favour of methanol are observed for those coatings capable of forming hydrogen bonds, especially PCPY, PCbPY, PHPY and PPY.Water and toluene both have significantly higher boilingpoints and lower vapour pressures than hexane and methanol. Their normalized frequency shifts generally lie between the extremes observed for hexane and methanol, reflecting the fact that while their Cv values are lower, their rate of desorption from the film will also be slower resulting in a larger steadystate equilibrium vapour concentration within the film coatings.Comparing the log K values for this pair shows a lower extent of differentiation between water and toluene although, again, a large difference in favour of water vapour is observed for the two most polar coatings (PCPY and PCbPY). PBPY, PPY and PTHPY all show larger log K values for toluene than either water or methanol vapour.This is readily explained since toluene has the largest polarizibility, dipolar and dispersion solvatochromic parameters while the side chains of PBPY, PPY and PTHPY are all essentially non-polar in nature. PHPY, which has a relatively non-polar side chain in spite of its terminal hydroxyl group, also shows very similar response behaviour towards the same vapours. Comparing the remaining vapours, methanol and acetonitrile show very similar response patterns with the greatest differentiation observed for PTHPY.Triethylamine, on the other hand, gives a markedly lower response with the most polar coatings (PCPY and PCbPY) and reduced responses with PPY, PMPY, PPPY and PHPY. Butanal shows a similar pattern to methanol and acetonitrile, showing greater log K values for PPY, PMPY and PBPY. This reflects the larger dispersion parameter for butanal compared with methanol and acetonitrile. Neither butanal, acetonitrile nor triethylamine can function as Table 7 Solvatochromic parameters for the test liquids (from refs. 34 and 35). The parameters represent polarizability (R2), dipolar interactions (p2 *), hydrogen bond acidity and basicity (a2 H and b2 H, respectively) and dispersion interactions (log L16) Solvent R2 p2 * a2 H b2 H Log L16 Hexane 0.00 0.00 0.00 0.00 2.688 Toluene 0.601 0.55 0.00 0.14 3.344 Methanol 0.278 0.40 0.37 0.41 0.922 Water 0.00 0.43 0.33 0.65 0.267 Butanal 0.187 0.65 0.00 0.40 2.270 Acetonitrile 0.237 0.75 0.09 0.44 1.560 Triethylamine 0.101 0.15 0.00 0.67 3.077 Ethanol 0.246 0.40 0.33 0.44 1.485 Butan-1-ol 0.224 0.40 0.33 0.45 2.601 Hexan-1-ol 0.210 0.40 0.33 0.45 3.610 Nonan-1-ol 0.193 0.40 0.33 0.45 5.124 Table 8 Normalized frequency shifts and corresponding partition coefficients obtained for the coated TSM sensors on exposure to different test vapours Analyte Hexane Toluene Methanol Water Polymer Dfv/Hz Log K Dfv/Hz Log K Dfv/Hz Log K Dfv/Hz Log K PPY 279 ± 17 2.65 845 ± 19 3.70 2948 ± 82 3.56 2581 ± 53 4.01 PMPY 1038 ± 16 3.21 670 ± 83 3.51 1845 ± 97 3.35 1156 ± 45 3.57 PBPY 1860 ± 80 3.39 3462 ± 189 4.16 2230 ± 121 3.36 996 ± 65 3.43 PCPY 146 ± 9 2.34 83 ± 7 2.67 3870 ± 76 3.73 2863 ± 49 4.03 PCbPY 170 ± 9 2.41 281 ± 18 3.12 8019 ± 105 3.98 4252 ± 107 4.13 PPPY 987 ± 45 3.17 4030 ± 145 4.27 5331 ± 132 3.69 2954 ± 123 3.96 PHPY 441 ± 40 2.81 1461 ± 18 3.82 6757 ± 162 3.88 1824 ± 117 3.74 PTHPY 712 ± 75 2.97 2768 ± 87 4.05 4315 ± 152 3.64 1952 ± 35 3.72 Butanal Acetonitrile Triethylamine Dfv/Hz Log K Dfv/Hz Log K Dfv/Hz Log K 2078 ± 81 3.67 1875 ± 62 3.51 1065 ± 45 3.34 1873 ± 75 3.62 1564 ± 82 3.35 975 ± 21 3.20 3236 ± 83 3.78 2873 ± 46 3.54 2528 ± 138 3.54 3392 ± 91 3.86 4560 ± 152 3.88 573 ± 12 3.04 3775 ± 78 3.71 6556 ± 149 3.96 593 ± 42 2.97 2092 ± 11 3.65 4054 ± 52 3.75 1275 ± 63 3.30 4175 ± 39 3.93 4378 ± 121 3.76 1783 ± 103 3.43 1562 ± 50 3.46 1115 ± 16 3.13 2787 ± 33 3.58 Table 9 Normalized frequency shifts and corresponding partition coefficients obtained for the coated TSM sensors on exposure to a series of primary alcohol vapours Analyte Ethanol Butan-1-ol Hexan-1-ol Nonan-1-ol Polymer Dfv/Hz Log K Dfv/Hz Log K Dfv/Hz Log K Dfv/Hz Log K PPY 1299 ± 31 3.38 827 ± 43 3.69 714 ± 19 4.47 236 ± 11 4.59 PMPY 3373 ± 56 3.78 555 ± 19 3.51 521 ± 8 4.32 201 ± 15 4.51 PBPY 2673 ± 85 3.61 1421 ± 64 3.85 2661 ± 58 4.96 530 ± 28 4.86 PCPY 3959 ± 42 3.84 522 ± 27 3.47 437 ± 15 4.23 189 ± 8 4.47 PCbPY 4819 ± 67 3.93 2756 ± 94 4.12 1202 ± 71 4.68 300 ± 8 4.68 PPPY 4947 ± 79 3.93 3732 ± 153 4.32 2611 ± 39 5.01 161 ± 5 4.40 PHPY 2250 ± 28 3.57 1650 ± 25 3.95 2103 ± 46 4.90 393 ± 13 4.77 PTHPY 3613 ± 102 3.74 2930 ± 49 4.16 1827 ± 31 4.79 143 ± 9 4.29 1136 Analyst, October 1997, Vol. 122hydrogen bond donors, so one might expect lower responses to these vapours for coatings that are also unable to function as hydrogen bond donors (e.g., PMPY, PBPY and PPPY). This is in fact seen for butanal and acetonitrile, but dispersion interactions are also a strong factor for triethylamine, resulting in much higher responses with PBPY and PTHPY.The series of primary alcohols show the expected increase in log K with increasing carbon number, although it is non-linear. This is presumably a result of concentration effects arising from the high vapour concentrations used in this study compared with those found in typical gas chromatography experiments. Ethanol shows the least variation in log K between different coatings, while hexan-1-ol shows the greatest.The solvatochromic parameters for this series differ only in the magnitude of the polarizibility and dispersion parameters, reflecting the fact that only the chain length is changing. In this respect, it is interesting that the largest log K values are observed for the coatings having longer side chains such as PBPY and PHPY. One method to characterize the vapour sorption properties of the polymer coatings across the series of test vapours is to use hierarchical cluster analysis.In a previous paper, this technique was employed in order to examine similarities in response patterns (and therefore potential redundancy) for four longchain unsaturated aroma components.19 A similar analysis was performed on the log K values measured for the current set of analyte vapours using the Euclidean distance, single linkage (nearest neighbour) method without autoscaling. The cluster analysis was also repeated for various sub-sets of the data.Fig. 6 shows the dendrogram obtained using all the vapours listed in Table 1. Some differences were observed between the current and previous study, which employed only five- to nine-carbon chain unsaturated alcohols and aldehydes.19 For example, while the most polar films (PCPY and PCbPY) clustered together in both cases, different groupings were observed for the remaining polymers. In fact, the current results demonstrate a clustering pattern that much more closely reflects both the polarity and hydrogen bonding ability of the different polymeric coatings.This is clearly seen in the fact that the hydrogen bond donors PPY and PHPY cluster together, as do the longer side chain, non-hydrogen bond donors PPPY, PTHPY and PBPY. For comparison, in the previous study (where dispersion interactions were likely more significant than for the short chain analyte vapours used here) the clustering pattern showed a strong correlation with side-chain length so that PTHPY and PHPY were grouped together. Returning to the present study, the observed clustering of the polymeric coatings was further confirmed when the analysis was repeated on sub-sets of the data containing either polar or non-polar analytes.For the purposes of analysis, these were defined in two different ways: (1) on the presence or absence of polar functional groups (hydroxyl, carbonyl, etc.); and (2) on the arbitrary condition that, for ‘polar’ analytes, logL16 < 2.5.This second definition reflects the fact that longer chain analytes have a significant non-polar component, while the first definition yielded only two ‘non-polar’ cases. The resulting dendrograms (not shown) support the view that the magnitude of the response (and, therefore, the partial selectivity) for a given coating–vapour pair is a direct consequence of the nature and strength of the initial adsorption interaction.Since these interactions are also those involved in solvation processes, solvatochromic parameters such as those listed in Table 7 provide a useful tool for the rational choice of sensor coatings, even when the corresponding coating parameters are not known. Fig. 5 Response patterns expressed as log K for the different polymers and test vapours. Fig. 6 Hierarchical cluster analysis on the log K values for the 11 vapours listed in Table 4. Actual values are listed in Tables 8 and 9.Analyst, October 1997, Vol. 122 1137Conclusions Both the formation and properties of electropolymerized Nsubstituted pyrrole films have been characterized by various methods, including cyclic voltammetry, XPS, scanning electron microscopy and vapour sorption studies. In particular, the characterization of PBPY, PCbPY, PHPY and PTHPY by XPS is described. Peak potentials for electropolymerization and subsequent redox modification are shifted to more positive values for all the derivatives relative to the parent compound, reflecting the effect of the side chain on the susceptibility to oxidation.The number of electrons involved in the electropolymerization process varies from 2.1 to 2.7 according to both the XPS and cyclic voltammetry data. The response behaviour of the coated TSM devices may be explained by consideration of the different physical mechanisms responsible for reversible interactions between different vapours and the polymer films.These are conveniently quantified by the sensor partition coefficient and the relevant solvatochromic parameters. The results obtained both here and in previous studies show that polymeric films of N-substituted pyrrole derivatives may be employed as coatings for chemical vapour sensors. In particular, the influence of the side chain on the mechanism and extent of coating–vapour interactions renders these polymers well-suited to the production of chemical sensor arrays based on acoustic wave devices.We are grateful to the Natural Sciences and Engineering Research Council (Canada) for financial assistance. We also thank Professor R. H. Morris of the University of Toronto for the use of the cyclic voltammetry system. References 1 Handbook of Conducting Polymers, ed. Skotheim, T. A., Marcel Dekker, New York, 1986, vol. 1 and 2. 2 Conducting Polymers: Special Applications, ed. Alc�acer, L., Proceedings of the workshop held at Sintra, Portugal, July 28–31, 1986, Reidel, Dordrecht, 1987. 3 Aldissi, M., Inherently Conducting Polymers: Processing, Fabrication, Applications, Limitations, Noyes Data Corporation, Park Ridge, NJ, 1989. 4 Science and Applications of Conducting Polymers, ed. Salaneck, W. R., Clark, D. T., and Samuelsen, E. J., Papers from the 6th European Physical Society Industrial Workshop, Lofthus, Norway, May 28–31, 1990, Adam Hilger, Bristol, 1991. 5 Intrinsically Conducting Polymers: An Emerging Technology, ed.Aldissi, M., NATO ASI Series E: Applied Sciences Volume 246, Kluwer Academic Publishers, Dordrecht, 1991. 6 Josowicz, M., Analyst, 1995, 120, 1019. 7 Feldheim, D. L., and Elliott, C. M., J. Membr. Sci., 1992, 70, 9. 8 Schmidt, V. M., Tegtmeyer, D., and Heitbaum, H., Adv. Mater., 1992, 4, 428. 9 Ge, H., and Wallace, G. G., Anal. Chem., 1989, 61, 2391. 10 Deinhammer, R. S., Shimazu, K., and Porter, M. D., Anal. Chem., 1991, 63, 1889. 11 Kunugi, Y., Nigorikawa, K., Harima, Y., and Yamashita, K., J.Chem. Soc., Chem. Commun., 1994, 873. 12 Slater, J. M., Paynter, J., and Watt, E. J., Analyst, 1993, 118, 379. 13 Teasdale, P. R., and Wallace, G. G., Analyst, 1993, 118, 329. 14 Topart, P., and Josowicz, M., J. Phys. Chem., 1992, 96, 8662. 15 Bartlett, P. N., and Ling-Chung, S. K., Sens. Actuators, 1989, 20, 287. 16 Amrani, M. E. H., Ibrahim, M. S., and Persaud, K. C., Mater. Sci. Eng., 1993, C:1, 17. 17 Vigmond, S. J., Kallury, K. M.R., Ghaemmaghami, V., and Thompson, M., Talanta, 1992, 39, 449. 18 Deng, Z., Stone, D. C., and Thompson, M., Can. J. Chem., 1995, 73, 1427. 19 Deng, Z., Stone, D. C., and Thompson, M., Analyst, 1996, 121, 671. 20 Deng, Z., Stone, D. C., and Thompson, M., Analyst , 1996, 121, 1341. 21 Diaz, A. F., and Bargon, J., in Handbook of Conducting Polymers., ed. Skothem, T. A., Marcel Dekker, New York, 1986, vol. 1, pp. 81– 115. 22 Zhao, Z. S., and Pickup, P. G., J. Electroanal. Chem., 1996, 404, 55. 23 Adams, R. N., Acc. Chem. Res., 1969, 2, 175. 24 Waltman, R. J., and Bargon, J., Can. J. Chem., 1986, 64, 76. 25 Diaz, A. F., and Kanazawa, K. K., in Extended Linear Chain Compounds, ed. Miller, J. S., Plenum, New York, 1983, vol. 3, p. 417–441. 26 Pfluger, P., and Street, G. B., J. Chem. Phys., 1984, 80, 544. 27 Chan, H. S. O., Kang, E. T., Neoh, K. G., and Lim, Y. K., Synth. Met., 1989, 30, 189. 28 Atanasoska, L., Naoi, K., and Smyrl, W. H., Chem. Mater., 1992, 4, 988. 29 Lei, J., Cai, Z., and Martin, C. R., Synth. Met., 1992, 46, 53. 30 Kang, E. T., Neoh, K. G., Ong, Y. K., Tan, K. L., and Tan, B. T. G., Macromolecules, 1991, 24, 2822. 31 Toshima, N., and Tayanagi, J.-I., Chem. Lett., 1990, 1369. 32 Nicholson, R. S., and Shain, I., Anal. Chem., 1964, 36, 706. 33 Grate, J. W., Athur, S., Jr., Ballantine, D. S., Wohltjen, H., Abraham, M. H., McGill, R. A., and Sasson, P., Anal. Chem., 1988, 60, 869. 34 Abraham, M. H., Whiting, G. S., Doherty, R.M., and Shuely, W. J., J. Chem. Soc., Perkin Trans. 2, 1990, 2, 1451. 35 Abraham, M. H., Whiting, G. S., Doherty, R. M., and Shuely, W. J., J. Chem. Soc., Perkin Trans. 2, 1990, 2, 1851. Paper 7/03165C Received May 8, 1997 Accepted June 30, 1997 1138 Analyst, October 1997, Vol. 122 Characterization of Polymer Films of Pyrrole Derivatives for Chemical Sensing by Cyclic Voltammetry, X-ray Photoelectron Spectroscopy and Vapour Sorption Studies Zhiping Deng, David C.Stone and Michael Thompson Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario, Canada M5S 3H6 Eight different conducting polymer films formed from pyrrole and N-substituted pyrrole derivatives were characterized by cyclic voltammetry, X-ray photoelectron spectroscopy (XPS) and scanning electron microscopy. In particular, the XPS of poly[N-butylpyrrole], poly[N-(2-carboxyethyl)pyrrole], poly[N–(6–hydroxyhexyl)pyrrole] and poly[N-(6-tetrahydropyranylhexyl)pyrrole] is reported for the first time.The vapour sorption properties of these films were also examined by forming the films onto the electrodes of thickness-shear mode acoustic wave sensors. The influence of the pendant side chain is apparent in both the electrochemical behaviour, composition, doping level, morphology and the nature and extent of polymer–vapour interactions. The latter can be rationalized by consideration of vapour physical properties and solvatochromic parameters.Keywords: Cyclic voltammetry; X-ray photoelectron spectroscopy; poly(pyrrole) and derivatives; vapour sorption; thickness-shear mode; quartz crystal; acoustic wave chemical sensor Organic conducting polymers such as poly(pyrrole) have found extensive application in highly diverse fields such as biomaterials, chemistry, electronics, microfabrication, non-linear optics, sensors and textiles.1–5 In the area of analytical chemistry, for example, such materials have been used to form chemically modified electrodes,6 permeation membranes,7,8 liquid chromatographic stationary phases9,10 and chemical and biological sensors.11–16 These chemical applications are all based on the physico-chemical interactions that occur between the conducting polymer and the respective chemical species. Our own interest in these materials is their use as selective (or partially selective) coatings for thickness-shear mode (TSM) acoustic wave sensors.One reason for using conducting polymers in this way is the ease of film formation by direct electropolymerization onto the TSM electrode surface.Other advantages lie in the ease with which redox state and incorporated counter ions can be changed electrochemically, and the wide range of functionalities that can be introduced through chemical reaction with the pyrrole ring nitrogen. This in turn provides various mechanisms for modifying the extent of different reversible interactions between the polymer films and different analytes, thus affecting both the sensitivity and selectivity of the sensor system.In a series of papers,17–20 we have described the application of various N-substituted poly(pyrroles) to organic vapour sensing. These include the parent compound, poly(pyrrole) (PPY), and poly(N-methylpyrrole) (PMPY), poly(N-butylpyrrole) (PBPY), poly[N-(2-cyanoethyl)pyrrole] (PCPY), poly[N- (2-carboxyethyl)pyrrole] (PCbPY), poly(N-phenylpyrrole) (PPPY), poly[N–(6–hydroxyhexyl)pyrrole] (PHPY) and poly[N-(6-tetrahydropyranylhexyl)pyrrole] (PTHPY).In the preceding papers, we have described the preparation, characterization and effect of redox state of PCPY18,20 as well as the application of all eight polymers to the selective detection of aroma components.19 In the present paper, we provide a more detailed characterization and comparison of the eight conducting polymers, with an emphasis on their formation and application as coatings for organic vapour sensors.Experimental Reagents and Materials The solvents hexane, toluene, methanol, butanal, acetonitrile, triethylamine, ethanol, butan-1-ol, hexan-1-ol and nonan-1-ol (analytical-reagent grade, Aldrich, Milwaukee, WI, USA) were used as received. Pyrrole (98%), N-methypyrrole (99%) and N- (2-cyanoethyl)pyrrole (99%) (Aldrich) were vacuum-distilled before use. Both N-phenylpyrrole (99%) (Aldrich) and tetrabutylammonium perchlorate (TBAP) (electrochemical grade, Fluka, Buchs, Switzerland) were used as received.N-(2-Carboxyethyl) pyrrole, N-butylpyrrole, N-(6-tetrahydropyrylhexyl) pyrrole and N-(6-hydroxyhexyl)pyrrole were synthesized as described previously.19 All other chemicals were obtained from Aldrich and used as received. The piezoelectric crystals were rough or optically polished 9 MHz AT-cut quartz crystals with gold electrodes (International Crystal Manufacturing, Oklahoma City, OK, USA). A combined Pt–Ag/AgCl (3 mol l21 KCl) electrode (Metrohm, Herisau, Switzerland) was utilized as the counter and reference electrode, respectively.Polymerization Procedures One electrode of the TSM device was used as a working electrode for the electropolymerization of pyrrole and its derivatives using a cell as described previously.17 Unpolished crystals were used for vapour sorption experiments while optically polished crystals were used for the cyclic voltammetry studies. All devices were washed with acetone and dried under nitrogen before use.Polymerization was achieved using a solution containing monomer (0.1 mol l21) and TBAP (0.1 mol l21) in de-oxygenated acetonitrile using either constant potential or cyclic voltammetric deposition with a potentiostat/ galvanostat (Model 273, EG&G Princeton Applied Research, Princeton, NJ, USA). The polymerization potentials for the different monomers have been listed previously.19 The coated TSM sensors were then rinsed with acetonitrile, dried with nitrogen, washed with acetone, dried with nitrogen and finally placed in an oven for 1 h at 100 °C.The coating mass for each TSM device was determined from the in-air frequency difference before and after polymer deposition. Analyst, October 1997, Vol. 122 (1129–1138) 1129Cyclic Voltammetry Cyclic voltammograms for film deposition and growth on the electrodes of polished TSM devices were obtained for ten cycles at a sweep rate of 100 mV s21 for potentials varying between 20.8 and +1.2 V versus Ag/AgCl, depending on the monomer used.The resulting films were then studied using sweep rates of 20–100 mV s21 and potentials varying between 20.8 and +1.2 V versus Ag/AgCl using TBAP (0.1 mol l21) in acetonitrile as the supporting electrolyte. X-ray Photoelectron Spectroscopy (XPS) XPS was performed using a Leybold Max 200 XPS (Leybold, Cologne, Germany) instrument using an unmonochromatized Mg Ka source and an analysis area of 2 3 4 mm2.Survey and low resolution spectra were obtained using a pass energy of 192 eV; high resolution spectra were acquired with a pass energy of 48 eV. All spectra were satellite-subtracted and normalized using software and elemental sensitivity factors provided by the manufacturer. The binding energy scale was further calibrated to 285.0 eV for the main C (1s) feature in order to compensate for sample charging effects. Scanning Electron Microscopy Scanning electron micrographs were obtained using a Hitachi Model S-570 microscope and recorded using the Quartz PCI image capture system.In order to reduce charging effects, the coated TSM devices were sputter-coated with a thin gold film using an argon plasma vapour deposition system (Polaron Model E5100, Polaron, Watford, Hertfordshire, UK). All micrographs were acquired using an accelerating voltage of 18 kV. Vapour Sorption Studies Vapour generation was achieved by bubbling helium gas through the corresponding liquid at room temperature and pressure using the flow system described previously.18,19 The flow rates of the sample and purge streams were held at 30 ml min21 for all experiments.The frequency shift of the polymer-coated TSM sensors was measured using a universal frequency counter (Model HP5334B, Hewlett-Packard, Avondale, PA, USA). Data were collected, displayed and stored using a Macintosh II computer equipped with an IEEE 488 interface bus using software written in-house.The coated TSM sensors were initially purged using pure helium carrier gas for 1–2 h to allow the device to stabilize. Once a stable baseline had been obtained, the stream was switched to the sample vapour and returned to the purge gas once the steady-state frequency shift had been obtained. Measurements were performed in triplicate for each coating– vapour combination. Vapour concentrations were obtained using a liquid nitrogen trap for a fixed time interval and weighing the resulting condensate.The measured vapour concentrations are given in Table 1, together with the relevant physical properties for the compounds studied. Results and Discussion Cyclic Voltammetry The cyclic voltammograms for the electropolymerization of all eight monomers are fairly similar, the main differences being in the position and magnitude of the anodic and cathodic peaks. That for N-(2-carboxyethyl)pyrrole is shown in Fig. 1 as a representative example.As can be seen, the first oxidation peak occurs at a higher oxidation potential (+1.06 V) than that of pyrrole (+0.85 V). In fact, the oxidation potentials of all seven N-substituted pyrroles are higher than that of the parent compound (Table 2). This is attributable to the steric and electronic effects of the pendant side chain, the inductive effect rendering the oxidation process more difficult for the mono- Table 1 Physical data and concentration values for the test solvents and corresponding vapours Cv/mg l21 Boiling- Density/ Molecular (at 21 °C and Solvent point/°C g cm23 mass/g mol21 30 ml min21) Hexane 69 0.659 86.16 31 Toluene 110 0.865 110.6 10 Water 100 1.000 18.00 15 Acetonitrile 82 0.786 41.05 34 Triethylamine 88.8 0.726 101.19 30 Butanal 75 0.800 72.11 22 Methanol 64.7 0.791 32.04 40 Ethanol 78 0.785 46.07 27 Butan-1-ol 117.7 0.810 74.12 8.3 Hexan-1-ol 156.5 0.814 102.18 1.2 Nonan-1-ol 215 0.827 144.26 0.3 Fig. 1 Cyclic voltammetric polymerization of N-(2-carboxyethyl)pyrrole (0.1 mol l21 in 0.1 mol l21 TBAP–acetonitrile) at scan rate of 100 mV s21 versus Ag–AgCl (3 mol l21).Table 2 Anodic and cathodic peak potentials (versus Ag/AgCl) for the cyclic voltammetric polymerization of N-substituted pyrrole monomers (0.1 mol l21 monomer in 0.1 mol l21 TBAP–acetonitrile, 100 mV s21 scan rate) Epa Epa Epc Polymer (monomer)/V* (polymer)/V† (polymer)/V† PPY 0.85 20.08 20.23 PMPY 0.90 0.46 0.42 PBPY 1.08 0.65 0.47 PCPY 1.10 0.70 0.64 PCbPY 1.06 0.68 0.56 PPPY 1.19 0.67 0.58 PHPY 1.05 0.68 0.52 PTHPY 1.08 0.69 0.58 * From first scan.† From second scan. 1130 Analyst, October 1997, Vol. 122mers.21 When the potential is reversed on the first scan, the anodic current continues to increase for a short time before decreasing. The following cathodic current is also higher than that in the forward scan until the potential is +0.76 V, resulting in a loop in the voltammogram. This is indicative of a nucleation mechanism, and is commonly observed for the electropolymerization of conducting polymers.22 The effect is attributed to the fact that polymerization occurs faster at a polymeric nucleus than at the uncoated electrode surface.The anodic peak for the second scan occurs at a much lower potential (+0.68 V), and corresponds to polymer growth with incorporation of perchlorate counter ions (Table 2). The peak is broad because there will be a distribution of oligomer sizes, while diffusion of counter ions in and out of the film will be slow.Adams23 has concluded that whenever the second and subsequent sweeps of a cyclic voltammogram differ markedly from the first, a followup chemical reaction has occurred. These results, therefore, show the characteristics of an ECE reaction, in which electron transfer is followed by a chemical reaction and subsequent electron transfer reaction.24 Therefore, the overall polymerization mechanism for the pyrrole derivatives is the same as that for the parent compound.25 Another feature of the voltammograms is that, on successive scans, the anodic peak shifts to increasingly positive potential with a corresponding increase in current until a limiting value is reached.A similar effect is observed for the cathodic peak, except here the shift is towards a more negative potential. The shift is, at least in part, due to the polymer film resistance giving rise to a larger overpotential.The ratio of the cathodic to anodic peak currents is also less than one, confirming that the anodic peak consists of both reversible and irreversible contributions. The anodic peak potentials for the different films are very similar with the exception of PPY and PMPY. This may be explained by considering the effect of side-chain length on the porosity of the polymer. It is anticipated that as the length of the substituent chain increases, the polymer chains will be forced farther apart making it easier for counter ions to diffuse in and out of the film.This is confirmed by the cathodic peak potentials for the series PPY, PMPY, PBPY and PTHPY, which show an increasing shift towards positive potentials. Following film formation, the cyclic voltammogram for each polymer film was obtained at various scan rates using TBAP in acetonitrile without monomer as the supporting electrolyte. Well-defined, broad anodic and cathodic peaks were obtained in all cases (Fig. 2), which reflects the slow rate of counter ion transfer in and out of the film as well as interactions between electroactive sites and the electrochemical non-equivalence of these sites.26 The magnitude of the anodic and cathodic peak currents increases linearly with scan rate, while peak separation also increases. The polymer redox processes are, therefore, not diffusion-controlled. Table 3 summarizes the positions and separation of the anodic and cathodic peaks.Both the anodic (Epa) and cathodic (Epc) peak potentials become increasingly positive relative to the parent PPY as the chain length of the Nsubstituent increases. This trend arises from increasing distortion of inter-ring coplanarity, which would destabilize the cationic (oxidized) form of these polymers. One potential benefit of the positive shift in Epa is that films formed from Nsubstituted pyrrole derivatives will be less sensitive to atmospheric oxidation.This implies improved handling, storage and lifetime characteristics for chemical sensors fabricated using such films. Another observation is that the peak separation (DE) is considerably lower for PPY and PMPY compared with the other polymer films. In fact, the oxidation and reduction of both PPY and PMPY films are essentially reversible, while other irreversible processes must be involved for the remaining films. As an example, PHPY shows two anodic waves, the second appearing at Å 1.0 V for scan rates of 80 and 100 mV s21.This is likely associated with oxidation of the hydroxyl group, although further work is needed to confirm this. X-ray Photoelectron Spectroscopy Although the X-ray photoelectron spectra of PPY, PMPY, PPPY and PCPY films have been previously reported, 17,18,20,26–31 we have included the relevant data with those for PBPY, PCbPY, PHPY and PTHPY in Table 4 for ease of comparison. This lists the results of peak deconvolution on the C(1s), N(1s), O(1s) and Cl(2p) regions for all eight as-prepared polymer films.In addition, Fig. 3 shows the C(1s) spectrum obtained for PCbPY. All eight films show the expected principal C(1s) component at 285 eV arising from the pyrrole ring a and b carbons26 and methylene groups within the pendant side chain. These species are unresolvable under the experimental conditions used here. A second C(1s) component occurring between 286.1 and 286.4 eV can arise from a number of different sources. For PCPY, for example, it is attributed to the nitrile group as discussed previously.18 For PHPY, there is also a contribution from the side chain –CH2OH group, while PTHPY contains two ether groups.It is, therefore, not surprising that these three films show significant enhancement of the 286 eV component relative to the remaining polymers. The origin of this peak for PPY, PMPY, PBPY, PCbPY and PPPY is amenable to several interpretations. Pfluger and Street26 attribute this band in PPY to ‘disorder type’ carbons, which they define as being crosslinked, chain-terminating and non-a,aA bonded carbons as well as partially saturated rings.There will also be a small contribution from unavoidable hydrocarbon contaminants, while Atanasoska et al.28 have suggested that electrostatic interaction of ring a carbons with counter ions will also have an effect. A fourth possibility is the presence of covalently bound chlorine species, which we will consider shortly.The final C(1s) component between 288 and 289 eV is generally attributed to carbonyl or carboxyl species resulting from chain termination.21,28 This has been partly confirmed by FTIR studies of chemically polymerized PPY samples,29 which clearly show the presence of a small band at 1705 cm21 that scales in intensity with the 288 eV C(1s) component. For PCbPY, there is also the pendant carboxyl group (289.1 eV), which clearly increases the relative contribution of this component.Turning now to the N(1s) region, all the films show two peaks located at approximately 400 and 402 eV, with the major component being the lower binding energy signal. This is in general agreement with other XPS studies of different PPY species,26,28 a number of which also show a small shoulder at Å 397 eV.17,20,29,30 The main peak at 400 eV arises from the neutral pyrrole ring nitrogen, while the higher binding energy component is generally attributed to partially charged nitrogens within bipolaron sub-units.The presence of the shoulder at 397 eV depends, to some extent, on the experimental signal-to-noise ratio but also on the film preparation conditions since it is more pronounced for neutral than as-prepared or oxidized PPY and PCPY.17,20 Lei et al.29 have attributed it to –CNN– defects in the pyrrole backbone, while Vigmond et al.17 ascribed it to interchain hydrogen bonding effects. In this latter interpretation, equal intensity peaks on either side of the principal N(1s) line are expected due to electron donation from one nitrogen to another.This being the case, one would expect enhancement of the higher binding energy component in the as-prepared films for polymers in which electron donation from one ring nitrogen to an adjacent ring side chain is possible. Such an effect is, in fact, observed for all the polymers relative to the phenyl- and alkyl-substituted pyrroles (PMPY, PBPY and PPPY).The exception here is PCPY, although the situation in this case may Analyst, October 1997, Vol. 122 1131be masked by the additional nitrogen per pyrrole unit contributed by the nitrile group. The Cl(2p) spectra show peaks with the 2p3/2 component occurring at 207.6 and 200.7 eV. The higher binding energy peak is due to incorporated perchlorate ion, while the lower peak is assigned to covalently incorporated chlorine as discussed previously.20 Support for this assignment comes from the results of Kang et al.30 and Toshima and Tayanagi,31 who observed similar Cl(2p) spectra for chemically polymerized PPY samples.In this respect, it should be noted that the covalently bound chlorine will also result in an increase in the C(1s) component at 286 eV. It is possible to estimate the doping ratio of the as-prepared polymer films from the XPS data using the Cl/N ratio, although Fig. 2 Cyclic voltammograms for the polymer films in 0.1 mol l21 TBAP–acetonitrile without monomer at different scan rates (mV s21). 1132 Analyst, October 1997, Vol. 122this will not be as accurate as that calculated from the composition obtained by bulk elemental analysis since only a thin layer of an irregular surface is being analysed. The estimated doping ratio for each film is given in Table 5, together with the corresponding number of electrons per monomer unit (n) involved in polymerization and subsequent oxidation. It is also possible to estimate a value of n from the cyclic voltammetry data using Nicholson’s model for irreversible charge transfer32 Ip = nFAC*n1/2(pDanF/RT)1/2c(bt) (1) where F is Faraday’s constant, A is the area, C* is the surface concentration of electroactive species, v is the scan rate, D is the diffusion coefficient, a is the transfer coefficient and c(bt) is the current function for irreversible charge transfer.A difficulty here, however, is obtaining a reasonable estimate of the surface concentration for an amorphous, porous polymer film.One crude approach is to use the generally accepted value of n = 2.3–2.5 for PPY and use this to calculate the surface concentration, then assume an identical value for the Nsubstituted derivatives. Agreement between the XPS data and the values of n derived in this manner is surprisingly good, although it breaks down for PCPY and PCbPY. The O(1s) region can generally be deconvoluted into two components at 533.5 and 532.0 eV, representing singly and doubly bonded oxygen species, respectively.For the nonoxygen containing polymers, the only sources of oxygen in the final film will be incorporated counter ion and carboxy groups introduced by chain termination reactions.28 Note that perchlorates also show oxygen binding energies of Å 533 eV, which is unresolvable from the other species for the instrumental conditions used in this study. For these reasons, the O(1s) region has rarely been discussed in previous XPS studies of poly- (pyrroles).Simple composition calculations assuming that the perchlorate and carbonyl oxygen species overlap yield RNO: R– O ratios close to the theoretical 1 : 1 for PPY, PCbPY and PPPY, but break down for PMPY, PBPY, PCPY and PTHPY. This is anticipated since the relative amount of chlorine species other than perchlorate was not taken into account. What the results do show, however, is the expected progressive increase in R–O species for PHPY and PTHPY relative to the parent compound.One curious feature of the O(1s) deconvolution is the apparent presence of an additional oxygen species with a high binding energy of 535 eV. Such species can be observed on plasma-cleaned gold surfaces. This can be discounted, however, since no gold peaks are observed in any of the survey scans. Table 3 Anodic and cathodic peak potentials (versus Ag/AgCl) of Nsubstituted pyrrole polymers in 0.1 mol l21 TBAP–acetonitrile for a 20 mV s21 scan rate Polymer Epa/V Epc/V DE/V PPY 0.28 0.25 0.027 PMPY 0.51 0.48 0.030 PBPY 0.72 0.56 0.15 PCPY 0.82 0.72 0.10 PCbPY 0.88 0.76 0.12 PPPY 0.84 0.70 0.14 PHPY 0.81 0.64 0.16 PTHPY 0.86 0.76 0.10 Table 4 Polymer elemental composition (at.-%) and peak analysis from the XPS data C(1s) N(1s) O(1s) Cl(2p) Polymer BE*/eV at.-% BE*/eV at.-% BE*/eV at.-% BE*/eV at.-% PPY 66.9 11.6 18.2 3.3 285.0 76.9 400.2 71.3 532.4 87.5 207.6 100.0 286.3 18.4 402.1 28.7 533.7 12.5 288.1 4.7 PMPY 74.5 10.4 12.7 2.4 285.1 69.1 400.1 87.3 531.6 64.7 200.9 100.0 286.4 18.7 401.8 12.7 533.0 35.3 288.4 12.2 PBPY 71.9 13.4 12.8 1.9 285.0 75.7 398.5 87.9 531.8 52.9 201.1 78.6 286.4 16.6 400.3 12.1 533.0 47.1 208.1 21.4 288.2 7.7 PCPY 73.7 7.8 15.4 2.3 284.8 51.3 399.9 93.8 532.0 70.7 200.7 40.1 286.4 43.8 401.7 6.2 533.6 29.3 207.5 59.7 288.6 5.0 PCbPY 63.9 9.1 25.0 2.0 284.9 63.7 399.9 59.8 532.3 60.8 200.6 68.2 286.2 18.9 400.4 34.9 533.5 33.6 207.5 31.8 289.1 17.3 401.6 9.4 534.6 5.6 PPPY 73.3 8.1 16.7 1.3 284.9 75.0 400.7 92.0 532.6 63.0 200.4 100.0 286.1 21.0 402.4 8.0 533.5 37.0 287.9 4.0 PHPY 74.6 7.3 16.6 1.5 284.9 58.5 400.1 69.0 532.5 56.8 207.4 100.0 286.1 36.1 401.4 31.0 533.8 36.8 288.5 5.4 535.6 6.4 PTHPY 64.6 5.4 27.5 2.6 285.0 61.0 400.3 62.8 532.4 22.4 200.8 18.8 286.3 29.5 402.3 37.3 533.4 67.5 207.9 81.2 288.3 9.5 534.6 10.2 * BE = Binding energy.Analyst, October 1997, Vol. 122 1133There are several other possible explanations, including differential charging affecting the incorporated counter ion and the presence of an oxidized nitrogen species within the polymer.It is difficult to envisage the latter, however, since the peak is only observed for N-substituted polymers. A final possibility is that it is a fitting artifact arising from the unresolved perchlorate oxygen, since there are an infinite number of mathematical solutions that would equally well fit the O(1s) envelope.Further work is, therefore, needed to clarify this matter. Scanning Electron Microscopy In previous studies, we have found that surface morphology can have an effect on the response of the polymer-coated sensors to different vapours.17–20 Polymer morphology is, in turn, influenced to various degrees by substrate morphology, film thickness and level of oxidation, and the nature of the counter ion. PPY films are frequently described as having a ‘cauliflower- like’ appearance, which arises from the nucleation and phase growth mechanism of the electropolymerization process.This detail is not apparent on very thin films, however, which simply follow the underlying surface. It is, therefore, important to specify both substrate surface roughness and film thickness when comparing the appearance of different films. In this, and our previous studies, the films were formed on microscopically rough surfaces. This was necessary since these polymer films adhere to gold electrodes primarily through mechanical interlock, and are prone to peeling from optically flat surfaces.In particular, the cleaning procedure used in the vapour sorption studies is highly effective as a means of removing conducting polymer films from polished gold electrodes, while films formed on the unpolished devices remain intact. The scanning electron micrographs for six of the polymers are shown in Fig. 4. Micrographs of PPY and PCPY have been published elsewhere.17,18,20,25 The corresponding film thicknesses, calculated from the in-air frequency shift due to film deposition, Dfs, are listed in Table 6.Generally, films of the Nsubstituted pyrrole derivatives show similar morphologies to the parent compound, although there are some differences. Both PMPY and PPPY, for example, exhibit a much more granular surface than PPY even for the same film thickness. The films of PBPY and PHPY are the most similar to PPY, while PCPY and PCbPY again show similar topography but with deep, open channels running into the bulk of the polymer.In this context, it should be noted that these channels disappear if the PCPY films are heavily oxidized by application of a large positive potential, becoming very similar in appearance to the as-prepared PBPY and PHPY films. The most different film in terms of morphology is PTHPY which, although as thick as the PPY and PPPY films, shows no apparent differences from the underlying gold electrode.These differences in morphology have a significant influence on the response of the coated vapour sensors. This occurs firstly through variations in effective surface area, which determines the extent of initial adsorption, and secondly through the film porosity, which affects both the absolute steady-state frequency response (extent of vapour absorption) and recovery time (vapour desorption). Table 5 Doping levels and number of electrons involved in electropolymerization –oxidation from the XPS and cyclic voltammetry (CV) data n Polymer Cl/N ratio Doping ratio XPS CV PPY 0.286 0.29 2.3 2.2 PMPY 0.232 0.23 2.2 2.3 PBPY 0.136 0.14 2.1 2.1 PCPY 0.289 0.15 2.2 2.6 PCbPY 0.219 0.22 2.2 2.7 PPPY 0.154 0.15 2.2 2.2 PHPY 0.205 0.21 2.2 2.5 PTHPY 0.488 0.49 2.5 2.6 Fig. 3 X-ray photoelectron spectra of PCbPY showing the C(1s), N(1s), O(1s) and Cl(2p) regions. 1134 Analyst, October 1997, Vol. 122Vapour Sorption Studies Since the magnitude of the observed frequency shift for vapour sorption (Dfv) varies with both vapour concentration (Cv) and film thickness, it is necessary to normalize the data before meaningful comparisons can be made.In keeping with our earlier work, we therefore calculate the sensor partition coefficient,33 K f f C S = D D nr V (2) where Dfv is in Hz, Dfs (kHz) is the frequency shift due to the polymer coating, r (g cm23) is the film density and Cv is in g cm23. Alternatively, the value of Dfv may be normalized to the shift that would have been obtained for a film having Dfs = 30 kHz.A series of organic vapours were chosen as probe molecules representing a variety of structural and functional group interactions between the vapour and the polymer film (Table 1). Hexane, for example, can interact solely through dispersion forces while toluene also exhibits p-electron overlap and polarizibility interactions. The remaining vapours were chosen to represent differing degrees of hydrogen bond acidity and basicity and dipole–dipole interactions, while a series of four primary alcohols was included to study the effects of chain length and analyte volatility.The solvatochromic parameters for all these vapours are listed in Table 7. While the corresponding Fig. 4 Scanning electron micrographs of six of the conducting polymer films. (a) PMPY; (b) PBPY; (c); PCbPY; (d) PPPY; (e) PHPY; and (f) PTHPY. Table 6 Film thickness and in-air frequency shifts corresponding to electropolymerization of the pyrrole derivatives on the TSM device Polymer PPY PMPY PCbPY PBPY PCPY PPPY PHPY PTHPY Thickness/mm 0.45 0.39 2.96 1.91 0.98 0.49 1.10 0.48 Dfs/kHz 28.8 21.3 104.5 99.0 50.9 26.0 55.2 22.0 Analyst, October 1997, Vol. 122 1135coating parameters are not known for the pyrrole derivatives used in this study, these values are nonetheless useful in comparing the response behaviour of the different vapour– coating combinations.The differing abilities of these combinations to interact through various mechanisms gives rise to different partial selectivities for each coating. This is reflected in the experimentally observed log K values, which are listed in Tables 8 and 9 and summarized graphically in Fig. 5. Comparing hexane and methanol, for example, shows that hexane generally gives much lower normalized frequency shifts ( < 2 kHz) than methanol (2–8 kHz), which reflects the difference in enthalpy of vaporization, and hence vapour concentration, between these two solvents (31.56 and 37.43 kJ mol21 at 25 °C, respectively).When the effect of vapour concentration is factored out by calculating log K, however, we see that the differences in response are actually minimal for the PMPY- and PBPY-coated sensors. On the other hand, fairly large differences in response in favour of methanol are observed for those coatings capable of forming hydrogen bonds, especially PCPY, PCbPY, PHPY and PPY.Water and toluene both have significantly higher boilingpoints and lower vapour pressures than hexane and methanol. Their normalized frequency shifts generally lie between the extremes observed for hexane and methanol, reflecting the fact that while their Cv values are lower, their rate of desorption from the film will also be slower resulting in a larger steadystate equilibrium vapour concentration within the film coatings. Comparing the log K values for this pair shows a lower extent of differentiation between water and toluene although, again, a large difference in favour of water vapour is observed for the two most polar coatings (PCPY and PCbPY).PBPY, PPY and PTHPY all show larger log K values for toluene than either water or methanol vapour. This is readily explained since toluene has the largest polarizibility, dipolar and dispersion solvatochromic parameters while the side chains of PBPY, PPY and PTHPY are all essentially non-polar in nature.PHPY, which has a relatively non-polar side chain in spite of its terminal hydroxyl group, also shows very similar response behaviour towards the same vapours. Comparing the remaining vapours, methanol and acetonitrile show very similar response patterns with the greatest differentiation observed for PTHPY. Triethylamine, on the other hand, gives a markedly lower response with the most polar coatings (PCPY and PCbPY) and reduced responses with PPY, PMPY, PPPY and PHPY.Butanal shows a similar pattern to methanol and acetonitrile, showing greater log K values for PPY, PMPY and PBPY. This reflects the larger dispersion parameter for butanal compared with methanol and acetonitrile. Neither butanal, acetonitrile nor triethylamine can function as Table 7 Solvatochromic parameters for the test liquids (from refs. 34 and 35). The parameters represent polarizability (R2), dipolar interactions (p2 *), hydrogen bond acidity and basicity (a2 H and b2 H, respectively) and dispersion interactions (log L16) Solvent R2 p2 * a2 H b2 H Log L16 Hexane 0.00 0.00 0.00 0.00 2.688 Toluene 0.601 0.55 0.00 0.14 3.344 Methanol 0.278 0.40 0.37 0.41 0.922 Water 0.00 0.43 0.33 0.65 0.267 Butanal 0.187 0.65 0.00 0.40 2.270 Acetonitrile 0.237 0.75 0.09 0.44 1.560 Triethylamine 0.101 0.15 0.00 0.67 3.077 Ethanol 0.246 0.40 0.33 0.44 1.485 Butan-1-ol 0.224 0.40 0.33 0.45 2.601 Hexan-1-ol 0.210 0.40 0.33 0.45 3.610 Nonan-1-ol 0.193 0.40 0.33 0.45 5.124 Table 8 Normalized frequency shifts and corresponding partition coefficients obtained for the coated TSM sensors on exposure to different test vapours Analyte Hexane Toluene Methanol Water Polymer Dfv/Hz Log K Dfv/Hz Log K Dfv/Hz Log K Dfv/Hz Log K PPY 279 ± 17 2.65 845 ± 19 3.70 2948 ± 82 3.56 2581 ± 53 4.01 PMPY 1038 ± 16 3.21 670 ± 83 3.51 1845 ± 97 3.35 1156 ± 45 3.57 PBPY 1860 ± 80 3.39 3462 ± 189 4.16 2230 ± 121 3.36 996 ± 65 3.43 PCPY 146 ± 9 2.34 83 ± 7 2.67 3870 ± 76 3.73 2863 ± 49 4.03 PCbPY 170 ± 9 2.41 281 ± 18 3.12 8019 ± 105 3.98 4252 ± 107 4.13 PPPY 987 ± 45 3.17 4030 ± 145 4.27 5331 ± 132 3.69 2954 ± 123 3.96 PHPY 441 ± 40 2.81 1461 ± 18 3.82 6757 ± 162 3.88 1824 ± 117 3.74 PTHPY 712 ± 75 2.97 2768 ± 87 4.05 4315 ± 152 3.64 1952 ± 35 3.72 Butanal Acetonitrile Triethylamine Dfv/Hz Log K Dfv/Hz Log K Dfv/Hz Log K 2078 ± 81 3.67 1875 ± 62 3.51 1065 ± 45 3.34 1873 ± 75 3.62 1564 ± 82 3.35 975 ± 21 3.20 3236 ± 83 3.78 2873 ± 46 3.54 2528 ± 138 3.54 3392 ± 91 3.86 4560 ± 152 3.88 573 ± 12 3.04 3775 ± 78 3.71 6556 ± 149 3.96 593 ± 42 2.97 2092 ± 11 3.65 4054 ± 52 3.75 1275 ± 63 3.30 4175 ± 39 3.93 4378 ± 121 3.76 1783 ± 103 3.43 1562 ± 50 3.46 1115 ± 16 3.13 2787 ± 33 3.58 Table 9 Normalized frequency shifts and corresponding partition coefficients obtained for the coated TSM sensors on exposure to a series of primary alcohol vapours Analyte Ethanol Butan-1-ol Hexan-1-ol Nonan-1-ol Polymer Dfv/Hz Log K Dfv/Hz Log K Dfv/Hz Log K Dfv/Hz Log K PPY 1299 ± 31 3.38 827 ± 43 3.69 714 ± 19 4.47 236 ± 11 4.59 PMPY 3373 ± 56 3.78 555 ± 19 3.51 521 ± 8 4.32 201 ± 15 4.51 PBPY 2673 ± 85 3.61 1421 ± 64 3.85 2661 ± 58 4.96 530 ± 28 4.86 PCPY 3959 ± 42 3.84 522 ± 27 3.47 437 ± 15 4.23 189 ± 8 4.47 PCbPY 4819 ± 67 3.93 2756 ± 94 4.12 1202 ± 71 4.68 300 ± 8 4.68 PPPY 4947 ± 79 3.93 3732 ± 153 4.32 2611 ± 39 5.01 161 ± 5 4.40 PHPY 2250 ± 28 3.57 1650 ± 25 3.95 2103 ± 46 4.90 393 ± 13 4.77 PTHPY 3613 ± 102 3.74 2930 ± 49 4.16 1827 ± 31 4.79 143 ± 9 4.29 1136 Analyst, October 1997, Vol. 122hydrogen bond donors, so one might expect lower responses to these vapours for coatings that are also unable to function as hydrogen bond donors (e.g., PMPY, PBPY and PPPY). This is in fact seen for butanal and acetonitrile, but dispersion interactions are also a strong factor for triethylamine, resulting in much higher responses with PBPY and PTHPY. The series of primary alcohols show the expected increase in log K with increasing carbon number, although it is non-linear.This is presumably a result of concentration effects arising from the high vapour concentrations used in this study compared with those found in typical gas chromatography experiments. Ethanol shows the least variation in log K between different coatings, while hexan-1-ol shows the greatest. The solvatochromic parameters for this series differ only in the magnitude of the polarizibility and dispersion parameters, reflecting the fact that only the chain length is changing.In this respect, it is interesting that the largest log K values are observed for the coatings having longer side chains such as PBPY and PHPY. One method to characterize the vapour sorption properties of the polymer coatings across the series of test vapours is to use hierarchical cluster analysis. In a previous paper, this technique was employed in order to examine similarities in response patterns (and therefore potential redundancy) for four longchain unsaturated aroma components.19 A similar analysis was performed on the log K values measured for the current set of analyte vapours using the Euclidean distance, single linkage (nearest neighbour) method without autoscaling. The cluster analysis was also repeated for various sub-sets of the data.Fig. 6 shows the dendrogram obtained using all the vapours listed in Table 1. Some differences were observed between the current and previous study, which employed only five- to nine-carbon chain unsaturated alcohols and aldehydes.19 For example, while the most polar films (PCPY and PCbPY) clustered together in both cases, different groupings were observed for the remaining polymers. In fact, the current results demonstrate a clustering pattern that much more closely reflects both the polarity and hydrogen bonding ability of the different polymeric coatings.This is clearly seen in the fact that the hydrogen bond donors PPY and PHPY cluster together, as do the longer side chain, non-hydrogen bond donors PPPY, PTHPY and PBPY.For comparison, in the previous study (where dispersion interactions were likely more significant than for the short chain analyte vapours used here) the clustering pattern showed a strong correlation with side-chain length so that PTHPY and PHPY were grouped together. Returning to the present study, the observed clustering of the polymeric coatings was further confirmed when the analysis was repeated on sub-sets of the data containing either polar or non-polar analytes.For the purposes of analysis, these were defined in two different ways: (1) on the presence or absence of polar functional groups (hydroxyl, carbonyl, etc.); and (2) on the arbitrary condition that, for ‘polar’ analytes, logL16 < 2.5. This second definition reflects the fact that longer chain analytes have a significant non-polar component, while the first definition yielded only two ‘non-polar’ cases.The resulting dendrograms (not shown) support the view that the magnitude of the response (and, therefore, the partial selectivity) for a given coating–vapour pair is a direct consequence of the nature and strength of the initial adsorption interaction. Since these interactions are also those involved in solvation processes, solvatochromic parameters such as those listed in Table 7 provide a useful tool for the rational choice of sensor coatings, even when the corresponding coating parameters are not known.Fig. 5 Response patterns expressed as log K for the different polymers and test vapours. Fig. 6 Hierarchical cluster analysis on the log K values for the 11 vapours listed in Table 4. Actual values are listed in Tables 8 and 9. Analyst, October 1997, Vol. 122 1137Conclusions Both the formation and properties of electropolymerized Nsubstituted pyrrole films have been characterized by various methods, including cyclic voltammetry, XPS, scanning electron microscopy and vapour sorption studies.In particular, the characterization of PBPY, PCbPY, PHPY and PTHPY by XPS is described. Peak potentials for electropolymerization and subsequent redox modification are shifted to more positive values for all the derivatives relative to the parent compound, reflecting the effect of the side chain on the susceptibility to oxidation. The number of electrons involved in the electropolymerization process varies from 2.1 to 2.7 according to both the XPS and cyclic voltammetry data. The response behaviour of the coated TSM devices may be explained by consideration of the different physical mechanisms responsible for reversible interactions between different vapours and the polymer films.These are conveniently quantified by the sensor partition coefficient and the relevant solvatochromic parameters. The results obtained both here and in previous studies show that polymeric films of N-substituted pyrrole derivatives may be employed as coatings for chemical vapour sensors. In particular, the influence of the side chain on the mechanism and extent of coating–vapour interactions renders these polymers well-suited to the production of chemical sensor arrays based on acoustic wave devices. We are grateful to the Natural Sciences and Engineering Research Council (Canada) for financial assistance. We also thank Professor R. H. Morris of the University of Toronto for the use of the cyclic voltammetry system. References 1 Handbook of Conducting Polymers, ed. Skotheim, T. A., Marcel Dekker, New York, 1986, vol. 1 and 2. 2 Conducting Polymers: Special Applications, ed. Alc�acer, L., Proceedings of the workshop held at Sintra, Portugal, July 28–31, 1986, Reidel, Dordrecht, 1987. 3 Aldissi, M., Inherently Conducting Polymers: Processing, Fabrication, Applications, Limitations, Noyes Data Corporation, Park Ridge, NJ, 1989. 4 Science and Applications of Conducting Polymers, ed. Salaneck, W. R., Clark, D. T., and Samuelsen, E. J., Papers from the 6th European Physical Society Industrial Workshop, Lofthus, Norway, May 28–31, 1990, Adam Hilger, Bristol, 1991. 5 Intrinsically Conducting Polymers: An Emerging Technology, ed. Aldissi, M., NATO ASI Series E: Applied Sciences Volume 246, Kluwer Academic Publishers, Dordrecht, 1991. 6 Josowicz, M., Analyst, 1995, 120, 1019. 7 Feldheim, D. L., and Elliott, C. M., J. Membr. Sci., 1992, 70, 9. 8 Schmidt, V. M., Tegtmeyer, D., and Heitbaum, H., Adv. Mater., 1992, 4, 428. 9 Ge, H., and Wallace, G. G., Anal. Chem., 1989, 61, 2391. 10 Deinhammer, R. S., Shimazu, K., and Porter, M. D., Anal. Chem., 1991, 63, 1889. 11 Kunugi, Y., Nigorikawa, K., Harima, Y., and Yamashita, K., J. Chem. Soc., Chem. Commun., 1994, 873. 12 Slater, J. M., Paynter, J., and Watt, E. J., Analyst, 1993, 118, 379. 13 Teasdale, P. R., and Wallace, G. G., Analyst, 1993, 118, 329. 14 Topart, P., and Josowicz, M., J. Phys. Chem., 1992, 96, 8662. 15 Bartlett, P. N., and Ling-Chung, S. K., Sens. Actuators, 1989, 20, 287. 16 Amrani, M. E. H., Ibrahim, M. S., and Persaud, K. C., Mater. Sci. Eng., 1993, C:1, 17. 17 Vigmond, S. J., Kallury, K. M. R., Ghaemmaghami, V., and Thompson, M., Talanta, 1992, 39, 449. 18 Deng, Z., Stone, D. C., and Thompson, M., Can. J. Chem., 1995, 73, 1427. 19 Deng, Z., Stone, D. C., and Thompson, M., Analyst, 1996, 121, 671. 20 Deng, Z., Stone, D. C., and Thompson, M., Analyst , 1996, 121, 1341. 21 Diaz, A. F., and Bargon, J., in Handbook of Conducting Polymers., ed. Skothem, T. A., Marcel Dekker, New York, 1986, vol. 1, pp. 81– 115. 22 Zhao, Z. S., and Pickup, P. G., J. Electroanal. Chem., 1996, 404, 55. 23 Adams, R. N., Acc. Chem. Res., 1969, 2, 175. 24 Waltman, R. J., and Bargon, J., Can. J. Chem., 1986, 64, 76. 25 Diaz, A. F., and Kanazawa, K. K., in Extended Linear Chain Compounds, ed. Miller, J. S., Plenum, New York, 1983, vol. 3, p. 417–441. 26 Pfluger, P., and Street, G. B., J. Chem. Phys., 1984, 80, 544. 27 Chan, H. S. O., Kang, E. T., Neoh, K. G., and Lim, Y. K., Synth. Met., 1989, 30, 189. 28 Atanasoska, L., Naoi, K., and Smyrl, W. H., Chem. Mater., 1992, 4, 988. 29 Lei, J., Cai, Z., and Martin, C. R., Synth. Met., 1992, 46, 53. 30 Kang, E. T., Neoh, K. G., Ong, Y. K., Tan, K. L., and Tan, B. T. G., Macromolecules, 1991, 24, 2822. 31 Toshima, N., and Tayanagi, J.-I., Chem. Lett., 1990, 1369. 32 Nicholson, R. S., and Shain, I., Anal. Chem., 1964, 36, 706. 33 Grate, J. W., Athur, S., Jr., Ballantine, D. S., Wohltjen, H., Abraham, M. H., McGill, R. A., and Sasson, P., Anal. Chem., 1988, 60, 869. 34 Abraham, M. H., Whiting, G. S., Doherty, R. M., and Shuely, W. J., J. Chem. Soc., Perkin Trans. 2, 1990, 2, 1451. 35 Abraham, M. H., Whiting, G. S., Doherty, R. M., and Shuely, W. J., J. Chem. Soc., Perkin Trans. 2, 1990, 2, 1851. Paper 7/03165C Received May 8, 1997 Accepted June 30, 1997 1138 Analyst, October 1997, Vol. 1
ISSN:0003-2654
DOI:10.1039/a703165c
出版商:RSC
年代:1997
数据来源: RSC
|
24. |
Optimisation of the Experimental Conditions of a New Method, Based on a Quartz Crystal Microbalance, for the Determination of Cyanide |
|
Analyst,
Volume 122,
Issue 10,
1997,
Page 1139-1142
M. Teresa S. R. Gomes,
Preview
|
|
摘要:
Optimisation of the Experimental Conditions of a New Method, Based on a Quartz Crystal Microbalance, for the Determination of Cyanide M. Teresa S. R. Gomes*, A. Alexandre F. Silva, Armando C. Duarte and Jo�ao A. B. P. Oliveira Department of Chemistry, University of Aveiro, 3810 Aveiro, Portugal A new method based on a quartz crystal microbalance was developed for the determination of cyanide. As the sensitivity depends on pH, temperature and nitrogen flow rate, a modified simplex was used to optimise these experimental parameters.Two different versions of the proposed method were optimised. For the first version a sensitivity increase of 1.5 was observed after 27 runs, whereas for the second version a sensitivity increase of 1.7 was observed after 12 runs. Keywords: Simplex optimisation; quartz crystal microbalance; piezoelectric crystals; cyanide A new method based on a quartz crystal microbalance (QCM) was developed for the determination of cyanide.The method is based on the fact that the cyanide promotes the disproportionation of HgI:1 Hg2 2+ " Hg2+ + Hg0 (1) The addition of cyanide ion forces reaction (1) to the right, as it forms a strong complex with HgII. The complete reaction can be described by Hg2 2+ + 2CN2 " Hg0 + Hg(CN)2 (2) The amalgamation of the mercury vapour on the gold electrodes of a piezoelectric quartz crystal leads to a frequency decrease,2–5 which is a linear function of the cyanide content of the sample.If acid is added to the mercury solution, it suppresses the hydrolysis of HgII ions, as shown in reaction (3), which otherwise would drive the mercury disproportionation reaction to the right, creating a high background level of Hg0.1 Hg2+ + H2O " HgO + H+ (3) The frequency decrease for a specific sample depends on several experimental parameters such as the carrier gas flow rate, pH and temperature in the reaction cell. In order to reduce experimental errors, it is important to maximise the frequency changes.A modified simplex algorithm6 was used to optimise the experimental conditions for a solution with a cyanide concentration close to the limit established for industrial waste waters7 and to the centroid of the linear calibration curve. Experimental Apparatus The gas flow rate was controlled with a variable area flow meter (Cole Parmer, Chicago, IL, USA). The piezoelectric quartz crystals were 10 MHz (HC49/U; Euroquartz, Crewkerne, Somerset, UK) and the frequency was monitored with a frequency counter (PM6680, Philips, Eindhoven, The Netherlands).All the other equipment was laboratory made and has been described elsewhere.5 A nitrogen flow, controlled by a flow meter, enters the bottom of a thermostated glass cell, through a sintered glass plate. This cell contains an acidified mercury solution, and allows the introduction of the cyanide solution through a silicone-rubber septum at the top. The mercury vapour flows through a 3A molecular sieve column, where it is dried, and impacts both faces of a piezoelectric crystal with gold electrodes.The oscillation frequency of the crystal is monitored with a frequency counter and a frequency decrease, proportional to the added mass, is observed during the mercury amalgamation. In a second version, an extra cell was inserted between the existing one and the flow meter, for reasons explained in the Procedure section. Reagents Mercury nitrate (Panreac, Barcelona, Spain), nitric acid (Riedelde Ha�en, Hanover, Germany), potassium cyanide, (Merck, Darmstadt, Germany) and phosphoric acid (Fluka, Buchs, Switzerland) were all of analytical-reagent grade. Nitrogen was of R grade from ArL�ýquido (Porto, Portugal).Procedure Two different versions of the new method based on a QCM were developed for cyanide determination. In a first approach, 10.0 cm3 of a 4.0 31025 m HgNO3·H2O solution, acidified with HNO3, were introduced into a glass cell.Nitrogen flowing through the bottom of the cell bubbled through a sintered glass plate and carried the mercury vapour on to the gold electrodes of a piezoelectric crystal. After an initial decrease, the frequency stabilised, and then 5.0 cm3 of a cyanide solution were injected. A new frequency decrease, which was proportional to the cyanide content, was observed, because the formation of HgII complexes promotes mercury disproportionation. Chlorides and thiocyanides are often present in industrial waste waters and can also form strong complexes with HgII.Furthermore, the responses are strongly influenced by sample pH. Therefore, for applications where chlorides and thiocyanides are present, or low cyanide concentrations necessitate the introduction of large volumes of sample, the method needed to be changed. Another glass cell, inserted before the one that contains the mercury solution, allows the sample introduction over phosphoric acid and the formation of hydrocyanic acid.The hydrocyanic acid is then carried by the nitrogen flow into the mercury solution cell and, as before, a frequency decrease is observed. Samples with cyanide concentrations bracketing the Portuguese legal limit7 for industrial waste waters (0.5 ppm [CN2]) were analysed by both versions of the method, and a sample volume of 5.0 cm3 was experimentally found to be adequate and selected for all subsequent experiments. The experimental parameters were then optimised for the solution with a Analyst, October 1997, Vol. 122 (1139–1141) 1139concentration close to the centroid of the linear calibration curve, 0.389 ppm [CN2] and 0.761 ppm [CN2] for the first and second versions, respectively. After the evaluation of the responses of the first four sets of conditions, a computer program written in FORTRAN 77 calculated the next set of conditions to be investigated. The algorithm followed the rules of the modified simplex method of Nelder and Mead.6 Each measurement was performed just once, to keep the number of experiments to a minimum.However, if a vertex was retained in 3 + 1 simplexes, the response was re-evaluated. If a vertex corresponded to a negative quantity, or the frequency stabilisation before the cyanide introduction took more than 25 min, an arbitrary 0 Hz response was assigned. Before the optimisation procedure, the estimated relative standard deviation of the concentration corresponding to the centroid of the linear calibration curve of the method was 6%.Therefore, the search was halted when, for the latest simplex, the tolerance (defined as the ratio of the difference between the greater and smaller responses over their mean) was less than 0.060. The optimisation procedure for the second version started from the optimum conditions found for the first version. Results and Discussion Fig. 1 shows, for both versions, the evolution of the response during the experiments.For the first version, the first simplex vertex corresponded to a temperature of 26.9 °C, a flow rate of 103 cm3 min21 and a volume of 3.0 cm3 of HNO3 in 100 cm3 of the mercury solution. The observed frequency decrease with the introduction of 5.0 cm3 of a 0.389 ppm [CN2] solution was 1978 Hz. After the optimisation procedure, the frequency decrease obtained with the introduction of the same amount of cyanide solution increased to 3150 Hz, with a temperature of 35.0 °C, a flow rate of 60 cm3 min21 and a volume of 1.0 cm3 of HNO3 in the mercury solution.For the second version, and using the former optimum conditions, the introduction of 5.0 cm3 of a solution 0.761 ppm [CN2] produced a frequency decrease of 253 Hz, which was increased to 562 Hz with change in the parameters to temperature 36.0 °C, flow rate 31 cm3 min21 and 2.8 cm3 of HNO3. Fig. 2 shows that, for both versions, the response increased as the flow rate of the carrier gas decreased.For practical reasons, and as already mentioned, when the frequency stabilisation took longer than 25 min, the frequency decrease was set to zero. The time for attaining frequency stabilisation depends not only on the carrier gas flow rate but also on the quantity of HNO3. Fig. 3 shows, for the first and second versions, the frequency decrease observed versus the volume of 0.1 m HNO3 used in the preparation of 100.0 cm3 of Holution.For the first version, the optimum was obtained with small volumes of acid, whereas for the second version it corresponded to large volumes. This contradictory behaviour must result from the different way of supplying cyanide into the mercury solution: an alkaline solution of CN2 in the first version and hydrocyanic acid in the second one. Increasing the amount of HNO3 decreases the mercury disproportionation before sample introduction, according to eqn. (3), in the same way for both versions of the method.This contributes to an increase in the response to cyanide, as there is an increase in the amount of HgI present at the moment of sample introduction, in addition to a smaller amount of mercury already amalgamated on to the crystal electrodes. The introduction of the alkaline cyanide solution directly into the mercury solution, in the first version, contributes to an increase in Fig. 1 Evolution of the response during the experiments. Fig. 2 Frequency decrease versus nitrogen flow rate for the first and second versions of the method.The points with a zero frequency decrease correspond to a frequency stabilisation longer than 25 min. Fig. 3 Frequency decrease versus volume of HNO3 for the first and second versions of the method. The points with a zero frequency decrease correspond to a frequency stabilisation longer than 25 min. 1140 Analyst, October 1997, Vol. 122mercury disproportionation and a higher frequency decrease than in the second version, where no alkali was added to the mercury solution.Moreover, the CN2/HCN ratio, which depends on the acid concentration, can also influence the amount of mercury vapour that reaches the crystal, in each version of the method. The temperature seems to be a less important factor than the nitrogen flow rate or the amount of acid. As a general conclusion, the simplex method, with a small number of experiments (27 in the first version and 12 in the second), leads to a signal increase and a general improvement in the slope of the linear portion of the calibraiton curve.The sensitivity increased from 4897 to 7430 Hz ppm21 for the first version and from 562.4 to 969.2 Hz ppm21 for the second version. The linear dynamic working range for the first version was 0.10–0.78 and 0.05–0.50 ppm before and after optimization, respectively. For the second version it was 0.32–1.02 and 0.24–0.86 ppm, respectively. The first version of the method becomes the obvious choice in the absence of chlorides, since it has the highest sensitivity of the two versions.The equation of the calibration curve is DF = 7430 [CN2] + 89.9 (r2 = 0.997) for cyanide concentrations in the range 0.05–0.50 ppm for the first version and DF = 969.2[CN2] 2 184.0 (r2 = 0.990) for cyanide concentrations in the range 0.24–0.86 ppm for the second version, where DF is the observed frequency decrease in Hz and [CN2] is the concentration of cyanide in the analysed standard solutions in ppm.In addition to chlorides, the most common compounds in the waste waters from the electroplating industry, suspected possibly to interfere in determination of cyanide, were added to standard cyanide solutions with concentrations within the linear working range for both versions of the method. In the first version of the method, Cu, Zn and Ni were not suspected to interfere, as the stability constants of their cyanide complexes were smaller than that of [Hg2+(CN)4]22.8 As the formation constant of [Fe3+(CN)6]32 was larger than that for [Hg2+(CN)4]22, a solution of K3Fe(CN)6, with a concentration of cyanide close to the centroid of the linear calibration curve, was introduced and cyanide could not be detected, as no frequency decrease was observed.Although the formation constant of [Fe2+(CN)6]42 was smaller than that of the [Hg2+(CN)4]22 complex, a solution of K4Fe(CN)6, with a concentration of cyanide close to the centroid of the linear calibration curve, was introduced into the sample cell and no interference was observed, as the signal had a magnitude corresponding to the cyanide content.For the second version of the method, a solution of NaCl with a chloride concentration of 144.6 ppm was introduced into the second cell and, as expected, no response was observed. Solutions of Cu, Zn and Ni, with cyanide concentration within the linear working range, were prepared, and no interference was observed with concentration of those ions of 12.87, 1.25 and 39.4 ppm, respectively.With the introduction of solutions of K4Fe(CN)6 or K3Fe(CN)6, with a concentration of cyanide close to the centroid of the linear calibration curve, no signal was observed. The inability of the second version of the method to detect hexacyanoferrate(ii) or hexacyanoferrate(iii) was not considered to be an important disadvantage, as the cyanide in these chemical forms is not considered to be toxic. As H2O2 is usually added in electroplating plants to destroy cyanide, 5 cm3 of a solution 1% in H2O2 was introduced into the cell in the second version of the method, and no frequency decrease was observed.We are grateful to the Junta Nacional de Investigaç�ao Cient�ýfica e Tecnol�ogica for the financial support of A. A. F. Silva through a scholarship (BM/926/94). References 1 Marshall, G., and Midgley, D., Anal. Chem., 1981, 53, 1760. 2 Bristow, Q., J.Geochem. Explor., 1972, 1, 55. 3 Sheide, E. P., and Taylor, J. K., Environ. Sci. Technol., 1974, 8, 1087. 4 Suleiman, A. A., and Guilbault, G. G., Anal. Chem., 1984, 56, 2964. 5 Gomes, M. T., Rocha, T. A., Duarte, A. C., and Oliveira, J. O., Anal. Chem., 1996, 68, 1561. 6 Nelder, J. A., and Mead, R., Comput. J., 1965, 7, 308. 7 Portuguese Water Quality Policy Act, Decreto-Lei No. 74/90, Di�ario da Rep�ublica–I S�erie, Imprensa Nacional-Casa da Moeda, EP., Portugal, 1990 (in Portuguese). 8 Morel, F. M. M., Principles of Aquatic Chemistry, Wiley, New York, 1983. Paper 7/02660I Received April 18, 1997 Accepted August 5, 1997 Analyst, October 1997, Vol. 122 1141 Optimisation of the Experimental Conditions of a New Method, Based on a Quartz Crystal Microbalance, for the Determination of Cyanide M. Teresa S. R. Gomes*, A. Alexandre F. Silva, Armando C. Duarte and Jo�ao A. B. P. Oliveira Department of Chemistry, University of Aveiro, 3810 Aveiro, Portugal A new method based on a quartz crystal microbalance was developed for the determination of cyanide.As the sensitivity depends on pH, temperature and nitrogen flow rate, a modified simplex was used to optimise these experimental parameters. Two different versions of the proposed method were optimised. For the first version a sensitivity increase of 1.5 was observed after 27 runs, whereas for the second version a sensitivity increase of 1.7 was observed after 12 runs.Keywords: Simplex optimisation; quartz crystal microbalance; piezoelectric crystals; cyanide A new method based on a quartz crystal microbalance (QCM) was developed for the determination of cyanide. The method is based on the fact that the cyanide promotes the disproportionation of HgI:1 Hg2 2+ " Hg2+ + Hg0 (1) The addition of cyanide ion forces reaction (1) to the right, as it forms a strong complex with HgII. The complete reaction can be described by Hg2 2+ + 2CN2 " Hg0 + Hg(CN)2 (2) The amalgamation of the mercury vapour on the gold electrodes of a piezoelectric quartz crystal leads to a frequency decrease,2–5 which is a linear function of the cyanide content of the sample.If acid is added to the mercury solution, it suppresses the hydrolysis of HgII ions, as shown in reaction (3), which otherwise would drive the mercury disproportionation reaction to the right, creating a high background level of Hg0.1 Hg2+ + H2O " HgO + H+ (3) The frequency decrease for a specific sample depends on several experimental parameters such as the carrier gas flow rate, pH and temperature in the reaction cell.In order to reduce experimental errors, it is important to maximise the frequency changes. A modified simplex algorithm6 was used to optimise the experimental conditions for a solution with a cyanide concentration close to the limit established for industrial waste waters7 and d of the linear calibration curve.Experimental Apparatus The gas flow rate was controlled with a variable area flow meter (Cole Parmer, Chicago, IL, USA). The piezoelectric quartz crystals were 10 MHz (HC49/U; Euroquartz, Crewkerne, Somerset, UK) and the frequency was monitored with a frequency counter (PM6680, Philips, Eindhoven, The Netherlands). All the other equipment was laboratory made and has been described elsewhere.5 A nitrogen flow, controlled by a flow meter, enters the bottom of a thermostated glass cell, through a sintered glass plate.This cell contains an acidified mercury solution, and allows the introduction of the cyanide solution through a silicone-rubber septum at the top. The mercury vapour flows through a 3A molecular sieve column, where it is dried, and impacts both faces of a piezoelectric crystal with gold electrodes. The oscillation frequency of the crystal is monitored with a frequency counter and a frequency decrease, proportional to the added mass, is observed during the mercury amalgamation.In a second version, an extra cell was inserted between the existing one and the flow meter, for reasons explained in the Procedure section. Reagents Mercury nitrate (Panreac, Barcelona, Spain), nitric acid (Riedelde Ha�en, Hanover, Germany), potassium cyanide, (Merck, Darmstadt, Germany) and phosphoric acid (Fluka, Buchs, Switzerland) were all of analytical-reagent grade. Nitrogen was of R grade from ArL�ýquido (Porto, Portugal).Procedure Two different versions of the new method based on a QCM were developed for cyanide determination. In a first approach, 10.0 cm3 of a 4.0 31025 m HgNO3·H2O solution, acidified with HNO3, were introduced into a glass cell. Nitrogen flowing through the bottom of the cell bubbled through a sintered glass plate and carried the mercury vapour on to the gold electrodes of a piezoelectric crystal. After an initial decrease, the frequency stabilised, and then 5.0 cm3 of a cyanide solution were injected.A new frequency decrease, which was proportional to the cyanide content, was observed, because the formation of HgII complexes promotes mercury disproportionation. Chlorides and thiocyanides are often present in industrial waste waters and can also form strong complexes with HgII. Furthermore, the responses are strongly influenced by sample pH. Therefore, for applications where chlorides and thiocyanides are present, or low cyanide concentrations necessitate the introduction of large volumes of sample, the method needed to be changed.Another glass cell, inserted before the one that contains the mercury solution, allows the sample introduction over phosphoric acid and the formation of hydrocyanic acid. The hydrocyanic acid is then carried by the nitrogen flow into the mercury solution cell and, as before, a frequency decrease is observed. Samples with cyanide concentrations bracketing the Portuguese legal limit7 for industrial waste waters (0.5 ppm [CN2]) were analysed by both versions of the method, and a sample volume of 5.0 cm3 was experimentally found to be adequate and selected for all subsequent experiments.The experimental parameters were then optimised for the solution with a Analyst, October 1997, Vol. 122 (1139–1141) 1139concentration close to the centroid of the linear calibration curve, 0.389 ppm [CN2] and 0.761 ppm [CN2] for the first and second versions, respectively.After the evaluation of the responses of the first four sets of conditions, a computer program written in FORTRAN 77 calculated the next set of conditions to be investigated. The algorithm followed the rules of the modified simplex method of Nelder and Mead.6 Each measurement was performed just once, to keep the number of experiments to a minimum. However, if a vertex was retained in 3 + 1 simplexes, the response was re-evaluated. If a vertex corresponded to a negative quantity, or the frequency stabilisation before the cyanide introduction took more than 25 min, an arbitrary 0 Hz response was assigned.Before the optimisation procedure, the estimated relative standard deviation of the concentration corresponding to the centroid of the linear calibration curve of the method was 6%. Therefore, the search was halted when, for the latest simplex, the tolerance (defined as the ratio of the difference between the greater and smaller responses over their mean) was less than 0.060.The optimisation procedure for the second version started from the optimum conditions found for the first version. Results and Discussion Fig. 1 shows, for both versions, the evolution of the response during the experiments. For the first version, the first simplex vertex corresponded to a temperature of 26.9 °C, a flow rate of 103 cm3 min21 and a volume of 3.0 cm3 of HNO3 in 100 cm3 of the mercury solution. The observed frequency decrease with the introduction of 5.0 cm3 of a 0.389 ppm [CN2] solution was 1978 Hz.After the optimisation procedure, the frequency decrease obtained with the introduction of the same amount of cyanide solution increased to 3150 Hz, with a temperature of 35.0 °C, a flow rate of 60 cm3 min21 and a volume of 1.0 cm3 of HNO3 in the mercury solution. For the second version, and using the former optimum conditions, the introduction of 5.0 cm3 of a solution 0.761 ppm [CN2] produced a frequency decrease of 253 Hz, which was increased to 562 Hz with change in the parameters to temperature 36.0 °C, flow rate 31 cm3 min21 and 2.8 cm3 of HNO3.Fig. 2 shows that, for both versions, the response increased as the flow rate of the carrier gas decreased. For practical reasons, and as already mentioned, when the frequency stabilisation took longer than 25 min, the frequency decrease was set to zero. The time for attaining frequency stabilisation depends not only on the carrier gas flow rate but also on the quantity of HNO3.Fig. 3 shows, for the first and second versions, the frequency decrease observed versus the volume of 0.1 m HNO3 used in the preparation of 100.0 cm3 of HgNO3 solution. For the first version, the optimum was obtained with small volumes of acid, whereas for the second version it corresponded to large volumes. This contradictory behaviour must result from the different way of supplying cyanide into the mercury solution: an alkaline solution of CN2 in the first version and hydrocyanic acid in the second one.Increasing the amount of HNO3 decreases the mercury disproportionation before sample introduction, according to eqn. (3), in the same way for both versions of the method. This contributes to an increase in the response to cyanide, as there is an increase in the amount of HgI present at the moment of sample introduction, in addition to a smaller amount of mercury already amalgamated on to the crystal electrodes.The introduction of the alkaline cyanide solution directly into the mercury solution, in the first version, contributes to an increase in Fig. 1 Evolution of the response during the experiments. Fig. 2 Frequency decrease versus nitrogen flow rate for the first and second versions of the method. The points with a zero frequency decrease correspond to a frequency stabilisation longer than 25 min. Fig. 3 Frequency decrease versus volume of HNO3 for the first and second versions of the method.The points with a zero frequency decrease correspond to a frequency stabilisation longer than 25 min. 1140 Analyst, October 1997, Vol. 122mercury disproportionation and a higher frequency decrease than in the second version, where no alkali was added to the mercury solution. Moreover, the CN2/HCN ratio, which depends on the acid concentration, can also influence the amount of mercury vapour that reaches the crystal, in each version of the method.The temperature seems to be a less important factor than the nitrogen flow rate or the amount of acid. As a general conclusion, the simplex method, with a small number of experiments (27 in the first version and 12 in the second), leads to a signal increase and a general improvement in the slope of the linear portion of the calibraiton curve. The sensitivity increased from 4897 to 7430 Hz ppm21 for the first version and from 562.4 to 969.2 Hz ppm21 for the se version.The linear dynamic working range for the first version was 0.10–0.78 and 0.05–0.50 ppm before and after optimization, respectively. For the second version it was 0.32–1.02 and 0.24–0.86 ppm, respectively. The first version of the method becomes the obvious choice in the absence of chlorides, since it has the highest sensitivity of the two versions. The equation of the calibration curve is DF = 7430 [CN2] + 89.9 (r2 = 0.997) for cyanide concentrations in the range 0.05–0.50 ppm for the first version and DF = 969.2[CN2] 2 184.0 (r2 = 0.990) for cyanide concentrations in the range 0.24–0.86 ppm for the second version, where DF is the observed frequency decrease in Hz and [CN2] is the concentration of cyanide in the analysed standard solutions in ppm.In addition to chlorides, the most common compounds in the waste waters from the electroplating industry, suspected possibly to interfere in determination of cyanide, were added to standard cyanide solutions with concentrations within the linear working range for both versions of the method. In the first version of the method, Cu, Zn and Ni were not suspected to interfere, as the stability constants of their cyanide complexes were smaller than that of [Hg2+(CN)4]22.8 As the formation constant of [Fe3+(CN)6]32 was larger than that for [Hg2+(CN)4]22, a solution of K3Fe(CN)6, with a concentration of cyanide close to the centroid of the linear calibration curve, was introduced and cyanide could not be detected, as no frequency decrease was observed.Although the formation constant of [Fe2+(CN)6]42 was smaller than that of the [Hg2+(CN)4]22 complex, a solution of K4Fe(CN)6, with a concentration of cyanide close to the centroid of the linear calibration curve, was introduced into the sample cell and no interference was observed, as the signal had a magnitude corresponding to the cyanide content. For the second version of the method, a solution of NaCl with a chloride concentration of 144.6 ppm was introduced into the second cell and, as expected, no response was observed.Solutions of Cu, Zn and Ni, with cyanide concentration within the linear working range, were prepared, and no interference was observed with concentration of those ions of 12.87, 1.25 and 39.4 ppm, respectively. With the introduction of solutions of K4Fe(CN)6 or K3Fe(CN)6, with a concentration of cyanide close to the centroid of the linear calibration curve, no signal was observed. The inability of the second version of the method to detect hexacyanoferrate(ii) or hexacyanoferrate(iii) was not considered to be an important disadvantage, as the cyanide in these chemical forms is not considered to be toxic. As H2O2 is usually added in electroplating plants to destroy cyanide, 5 cm3 of a solution 1% in H2O2 was introduced into the cell in the second version of the method, and no frequency decrease was observed. We are grateful to the Junta Nacional de Investigaç�ao Cient�ýfica e Tecnol�ogica for the financial support of A. A. F. Silva through a scholarship (BM/926/94). References 1 Marshall, G., and Midgley, D., Anal. Chem., 1981, 53, 1760. 2 Bristow, Q., J. Geochem. Explor., 1972, 1, 55. 3 Sheide, E. P., and Taylor, J. K., Environ. Sci. Technol., 1974, 8, 1087. 4 Suleiman, A. A., and Guilbault, G. G., Anal. Chem., 1984, 56, 2964. 5 Gomes, M. T., Rocha, T. A., Duarte, A. C., and Oliveira, J. O., Anal. Chem., 1996, 68, 1561. 6 Nelder, J. A., and Mead, R., Comput. J., 1965, 7, 308. 7 Portuguese Water Quality Policy Act, Decreto-Lei No. 74/90, Di�ario da Rep�ublica–I S�erie, Imprensa Nacional-Casa da Moeda, EP., Portugal, 1990 (in Portuguese). 8 Morel, F. M. M., Principles of Aquatic Chemistry, Wiley, New York, 1983. Paper 7/02660I Received April 18, 1997 Accepted August 5, 1997 Analyst, October 1997, V
ISSN:0003-2654
DOI:10.1039/a702660i
出版商:RSC
年代:1997
数据来源: RSC
|
25. |
Highly Selective Iodide Poly(vinyl chloride) Membrane Electrode Based on a Nickel(II) Tetraazaannulene Macrocyclic Complex |
|
Analyst,
Volume 122,
Issue 10,
1997,
Page 1143-1146
Min Ying,
Preview
|
|
摘要:
Highly Selective Iodide Poly(vinyl chloride) Membrane Electrode Based on a Nickel(II) Tetraazaannulene Macrocyclic Complex Min Ying†, Ruo Yuan, Xiao-Min Zhang†, You-Qun Song, Zhi-Qiang Li, Guo-Li Shen and Ru-Qin Yu* Institute of Chemometrics and Chemical Sensing Technology, Chemistry and Chemical Engineering College, Hunan University, Changsha 410082, China A PVC membrane electrode based on a nickel(II) tetraazaannulene macrocyclic complex as the carrier is described. The electrode exhibits an anti-Hofmeister selectivity sequence with a preference for iodide at pH 3.0–4.0.It has a linear response to iodide from 8.0 3 1026 to 1 3 1021 mol dm23 with a slope of 254.7 ± 0.2 mV per decade (20 °C), a satisfactory reproducibility and a rapid response time. The pH dependence of the potential response is discussed. The response mechanism of the electrode was investigated by ac impedance, quartz crystal microbalance and spectroscopic techniques. The electrode can be used for the determination of iodide in drug preparations.Keywords: Iodide-selective electrode; nickel tetraazaannulene; macrocyclic complex; poly(vinyl chloride) membrane Research on anti-Hofmeister sensing materials for anions is an expeditiously expanding domain in chemical sensors. Recently, anion membrane electrodes, especially those based on vitamin B12 derivatives,1–3 metalloporphyrins,4,5 metallophthalocyanines, 6,7 Schiff base complexes8 and other complexes, 9 have been reported as possessing improved selectivity characteristics.Generally, the anti-Hofmeister anion selectivity of an electrode doped with organometallic complex is achieved by specific axial anion coordination with the central metal ion of the complex, and for such a coordination process the structure of the complex-forming carrier plays an important role in the final selectivity characteristics of the electrode in question.10 Metallotetraazaannulene-type macrocyclic complexes have received increasing attention in view of mimicking biological systems because of their chemical analogies to some naturally occurring porphyrins and corrin rings11–14 which are found in heme proteins, chlorophylls and metalloenzymes.Indeed, several complexes of this kind have been found to be effective as catalysts. In particular, a cobalt tetraazaannulene has been exploited as a catalyst for the reduction of molecular oxygen at the cathode in fuel cell;15 a nickel tetraazaannulene surfacemodified electrode has been used to catalyse the oxidation of water.16 The metallotetraazaannulenes have been studied as complexes similar to metalloporphyrins in structure.17,18 The unique saddle-like structures of this group of complexes make it of interest to investigate the potential application of the complexes as carriers for anion membrane electrodes, as the complex structure might favour the axial coordination of the central metals in these complexes.One would expect that the metallotetraazaannulene membrane electrodes would respond to some specific anions in an anti-Hofmeister fashion. The nickel complex of 5,7,12,14-tetramethyldibenzo- [b,i][1,4,8,11]tetraazacyclotetradecahexenate (Me4DBTA), easily derived in high yield from the template reaction of phenylene-1,2-diamine with pentane-2,4-dione in the presence of nickel(ii),13,19,20 was chosen as the first example of a metallotetraazaannulene carrier for the membrane electrode exhibiting desired potentiometric response characteristics.One outstanding property of Me4DBTA is its higher lipophilicity compared with planar macrocycles such as porphyrins and phthalocyanines resulting from its saddle-shaped deformation through steric interactions of the diiminato methyl group with the benzenoid rings.15 In this paper, we report the results of an investigation of the response characteristics of an Ni-Me4DBTA-based electrode towards iodide and the response mechanism in view of the axial coordination.The electrode has been preliminarily applied to the determination of iodide in some contrast media such as meglumine diatrizoate injection. Experimental Apparatus and Reagents Potentiometric and pH measurements were made with a Model PXS-125 Ionalyzer (Shanghai Analytical Instruments, Shanghai, China). The cell used was of the type Hg, Hg2Cl2, õKCl(saturated)õsample solutionõmembraneõ0.01 mol dm23 KClõ Ag, AgCl.Tested solutions were buffered with 0.1 mol dm23 H3PO4 and the expected pH was adjusted with NaOH solution. The activities of anions were calculated using the Debye–H�uckel equation. Before use, the electrode was conditioned in 0.01 mol dm23 KI solution for 1 d. Ni-Me4DBTA was synthesized with a yield of 40% by the template reaction as described by L’Eplattenier and Pugin20 and identified by elemental analysis and mass spectrometry: (MS) found, C 65.04, H 5.68, N 13.49; calculated, C 65.87, H 5.53, N 13.97%; MS, m/z 400.The structure of Ni-Me4DBTA is shown in Fig. 1. 2-Nitrophenyl octyl ether (o-NPOE) was prepared according to the literature method.21 Tetrahydrofuran, didecyl sebacate (DDS), didecyl phthalate (DDP) and chromatographicgrade poly(vinyl chloride) (PVC) were purchased from Shanghai Chemical (Shanghai, China). Meglumine diatrizoate injection was a pharmaceutical preparation from Xinyi Medicals (Shanghai, China). Re-distilled water and analytical-reagent grade reagents were used throughout.The membrane composition was optimized by an orthogonal experimental design with the linear response range and slope as the object functions. The optimum composition was found to be 1.5% (m/m) carrier, 32.5% (m/m) PVC and 66% (m/m) solvent mediator. The PVC membrane electrode was fabricated according to the literature method.22 The potentiometric selectivity coefficients, Ki,j pot, were determined in pH 3.0 buffer solutions containing different anions at 0.01 mol dm23 by the separate solution method.23 † On leave from the East China Geological College, Linchuan, Jiangxi 344000, China.Analyst, October 1997, Vol. 122 (1143–1146) 1143UV/VIS Absorption Spectra and Ac Impedance Experiment The absorption spectra were recorded on a Lambda 17 UV/VIS spectrophotometer (Perkin-Elmer, Norwalk, CT, USA). Chloroform was employed as solvent. Spectra of the Ni-Me4DBTA chloroform solutions were recorded before and after shaking them with 0.1 mol dm23 KI and blank solution, respectively.Quartz Crystal Microbalance Experiment The frequency shifts of three different chloroform solutions were measured on a CN 3165 Frequency Counter (Sampo Co., Taiwan). Two chloroform solutions containing 0.01 mol dm23 Ni-Me4DBTA were separately shaken with 0.1 mol dm23 I2 and blank buffered solution (pH 3.0) for 30 min and the organic phases were treated with 3% H2O2 in 1.0 mol dm23 H2SO4 medium; the gold-coated quartz crystal was AT-cut with a fundamental resonance frequency of about 9 MHz and used to adsorb I2 for 15 min from the two organic phases to measure the frequency shifts DF1 and DF2, respectively.24 The third solution of pure chloroform was treated with 0.1 mol dm23 I2 in the same manner as mentioned above to measure the frequency shift DF3.Results and Discussion Emf Response Characteristics and Selectivity The potential selectivity coefficients of the electrodes are summarized in Table 1, showing that the electrodes incorporating Ni-Me4DBTA possess high selectivity towards iodide ion.The anti-Hofmeister selectivity sequence was assumed to be associated with an interaction between the central metal of the carrier and iodide ion. The potentometric response characteristics of three differently plasticized electrodes towards I2 were compared in Table 2. The o-NPOE plasticized membrane electrode showed a near Nernstian potentiometric response for 0.1–8.0 3 1026 mol dm23 I2 with a detection limit of 5.0 ± 0.3 3 1026 mol dm23 and a slope of 254.7 ± 0.2 mV per decade in 0.1 mol dm23 phosphate buffer solution (pH 3.0).The dc resistance of the membrane was estimated to be 132.5 ± 0.4 kW (average of six determinations). The standard deviation of potential responses over a period of 6 h in 1023 mol dm23 I2 wa0.5 mV (n = 36) and the potential readings for the electrode dipped alternatively into stirred solutions of 1022 and 1023 mol dm23 I2 gave a standard deviation of 0.7 mV (n = 5).The dynamic responses to different concentrations of I2 recorded at an interval of 5 s and depicted in Fig. 2, indicated that the time for the electrode to reach 90% of the steady response and subsequently to recover 90% of the blank response was less than 5 and 30 s, respectively. The lifetime of the electrode continuously contacted with 1022 mol dm23 KI at ambient temperature was found to be at least 2 months.In order to exclude the possibility that the response originated from I2 produced from the slow oxidation of I2 in an acidic medium by air, the potential responses to various concentrations of I2 freshly prepared in oxygen-free solutions were examined and no differences in the dynamic response patterns were observed from those in corresponding oxygen-containing solutions. Effect of pH on Potential Response The pH dependence of the electrode potential response is shown in Fig. 3. The effect of pH on the response characteristics can be explained by coordination competition between I2 and OH2. The increasing competition of OH2 at higher pH resulted in a narrower linear response range with a concomitant decrease in the response slope. A similar effect of OH2 on the potential response for electrodes based on porphyrin and vitamin B12 Fig. 1 Structure of Ni-Me4DBTA. A, Planar structure showing bonding and peripheral substitutions; B, side view of the complex showing typical ligand distortion.Table 1 Potential selectivity coefficients, log Ki,j pot Anion Plasticizer I2 ClO42 SCN2 NO22 Br2 NO32 OAc2 Cl2 SO42 o-NPOE 0.00 22.30 22.76 23.02 23.48 23.87 24.00 24.06 24.95 DDP 0.00 22.95 21.75 22.91 23.10 23.58 23.41 23.56 24.55 DDS 0.00 22.46 21.78 22.78 23.25 23.58 23.62 23.74 24.81 Table 2 Response characteristics of Ni-Me4DBTA-based electrodes with different solvent mediators. Results obtained from six measurements with two electrodes Plasticizer Parameter o-NPOE DDP DDS Slope mV per decade 254.7 ± 0.2 249.4 ± 0.6 242.8 ± 0.5 Linear range/mol dm23 8.0 3 1026–1021 1025–1022 1025–1022 Regression coefficient 20.9999 20.9998 20.9993 Detection limit/mol dm23 5.0 ± 0.331026 7.6 ± 0.531026 8.8 ± 0.531026 1144 Analyst, October 1997, Vol. 122derivatives was observed.7,25 As shown in Fig. 3 (inset), the competitive effect of OH2 was negligible and the electrode therefore demonstrated a near-Nernstian response towards iodide.The deviations from the linear response for concentrations of I2 at pH < 2.0 were due to the relatively high degree of oxidation of I2. Mechanism of Iodide Response Me4DBTA is a completely conjugated 14-membered ligand and its metal complex has shorter metal–nitrogen distances than corresponding 16-membered metalloporphyrins. The inherent predisposition of Me4DBTA takes the metal out of the N4 coordination plane (saddle-shaped). Consequently, two axial ligand sites are non-equivalent and five-coordination is preferred. 26 The preferential response towards I2 is believed to be associated with the coordination of iodide with the central metal of the carrier.With UV/VIS spectra as illustrated in Fig. 4, it was possible to distinguish the interaction between the central metal and iodide. The substantial increase in absorbance at 394 nm after the contact of the carrier solution with the iodidecontaining phase suggested the absorbing species had increased in size and axial coordination was thought to take place.At the same time, the effects of two more lipophilic anions, ClO42 and SCN2, on the spectrum of the carrier were investigated and no detectable changes in the UV/VIS spectra were noted. Thus, the anti-Hofmeister behaviour of the electrode can be explained by the specific interaction between the central metal and iodide ion. The ac impedances of the carrier-containing and the carrierfree membranes are shown in Fig 5.Although both of the membranes showed well defined semi-cyclic impedances at high frequencies, the bulk resistance of the membrane incorporating carrier decreased with increasing I2 concentration whereas that of carrier-free membrane remained almost constant. This pheomenon was related to the transfer of I2 across the solvent membrane resulting from the interaction between I2 and the carrier molecule. Only for the carrier-containing membrane was Warburg impedance at low frequencies observed, clearly indicating that the transfer process of I2 across the solvent membrane was diffusion controlled.The quartz crystal microbalance technique was used to detect the existence of I2 resulting from the oxidation of I2, transferring from the aqueous phase into the organic phase, by the deliberate addition of hydrogen peroxide. The frequency shifts DF1, DF2 and DF3 corresponded to the interactions of carrier in chloroform with 0.1 mol dm23 I2 and with blank solution, and of pure chloroform with 0.1 mol dm23 I2, respectively. The values of DF1, DF2 and DF3 were measured to be 3874, 850 and 215 Hz, respectively. The much larger frequency shift DF1 than DF2 and DF3 indirectly confirmed the transfer process of I2 across the organic–water interface.From the results mentioned above, it seems the metallotetraazaannulene- type carriers are promising as a new class of ion carriers for anions. By introducing different substituents on the phenylene and diimine carbon atoms as reported in the literature,27,28 one can prepare a whole spectrum of new carriers Fig. 2 Typical dynamic response of o-NPOE plasticized electrode towards different concentrations of I2. Fig. 3 Potential response curves of Ni-Me4DBTA-based electrode with o- NPOE as solvent mediator at different pH values. [I2]: 1026 (2); 1025 (.); 1024 (8); 1023 (½);1022 (Ó); and 1021 mol dm23 (5). Inset: potential response versus log [I2] at pH 3.0.Fig. 4 UV/VIS absorption spectra of Ni-Me4DBTA in chloroform treated with A, pH 3.0 blank solution and B, 0.1 mol dm21 KI solution. Fig. 5 Impedance plots of Ni-Me4DBTA-containing membrane (Ç) and Ni-Me4DBTA-free membrane (2) plasticized by o-NPOE. (a) 1 3 1024 and (b) 1 3 1023 mol dm21 KI. Analyst, October 1997, Vol. 122 1145with different possible electronic configurations. Further studies in this regard are in progress. Preliminary Application The resulting electrode was applied to the determination of iodide in a drug preparation (meglumine diatrizoate injection).The iodide content was determined by the standard additions method. An appropriate amount of sample was refluxed in concentrated NaOH solution in the presence of zinc powder for 30 min. After cooling, the reaction mixture was filtered and washed with water three times. The filtrate was acidified with H2SO4 to pH 3.0–4.0 and diluted to 200 ml with water.The resulting sample solution was determined using the present electrode and the precipitation method with AgNO3 as the titrant and tetrabromophenolphthalein as the indicator.29 The results listed in Table 3 were in agreement with those obtained by the precipitation method. It was concluded that there was a coordination interaction between I2 and Ni-Me4DBTA, which played an important role in improving the response selectivity, and for the saddle-shaped macrocyclic complex of Ni-Me4DBTA five-coordination was adopted to describe the coordination mechanism.Although the present electrode was pH dependent, it demonstrated a preferential, rapid and reproductible response towards iodide ion in pH 3.0–4.0 buffer solution and was successfully applied to the determination of iodide in drug preparations. This work was supported by the Foundations of the National Education Commission, the Machinery Ministry and the Hunan Scientific Commission of China. References 1 Daunert, S., and Bachas, L.G., Anal. Chem., 1989, 61, 499. 2 Schulthess, P., Ammann, D., Kr�autler, B., Caderas, C., Step�ýnek, R., and Simon, W., Anal. Chem., 1985, 57, 1397. 3 Step�ýnek, R., Kr�autler, B., Schulthess, P., Lindemann, B., Ammann, D., and Simon, W., Anal. Chim. Act 182, 83. 4 Gao, D., Li, J.-Z., and Yu, R.-Y., Anal. Chem., 1994, 66, 2245. 5 Daunert, S., Wallace, S., Florido, A., and Bachas, L. G., Anal. Chem., 1991, 63, 1676. 6 Li, J.-Z., Wu, X.-C., Yuan, R., Lin, H.-G., and Yu, R.-Q., Analyst, 1994, 119, 1363. 7 Chaniotakis, N. A., Park, S. B., and Meyerhoff, M. E., Anal. Chem., 1989, 61, 566. 8 Yuan, R., Chai, Y.-Q., Liu, D., Gao, D., Li, J.-Z., and Yu, R.-Q., Anal. Chem., 1993, 65, 2572. 9 Rothmaier, M., and Simon, W., Anal. Chim. Acta, 1993, 271, 135. 10 Glazier, S. A., and Arnold, M. A., Anal. Chem., 1991, 63, 754. 11 Weiss, M. C., and Goedken, V. L., J. Am. Chem. Soc., 1976, 98, 3389. 12 Nafie, L. A., Pastor, R. W., Dabrowiak, J.C., and Woodruff, W. H., J. Am. Chem. Soc., 1976, 98, 8007. 13 Place, D. A., Ferrara, G. P., Harland, J. J., and Dabrowiak, J. C., J. Heterocycl. Chem., 1980, 17, 439. 14 Hanke, R., and Breitmaier, E., Chem. Ber., 1982, 115, 1657. 15 Clauberg, B. W., and Sandstede, G., J. Electroanal. Chem. Interfacial Electrochem., 1976, 74, 393. 16 Issahary, D. A., Ginzburg, G., Polak, M., and Meyerstein, B., J. Chem. Soc., Chem. Commun., 1984, 441. 17 Weiss, M. C., Bursten, B., Peg, S.-M., and Geodken, V.L., J. Am. Chem. Soc., 1976, 98, 8021. 18 Woodruff, W. H., Pastor, R. W., and Dabrowiak, J. C., J. Am. Chem. Soc., 1976, 98, 7999. 19 Cutler, A. R., Alleyne, C. S., and Dolphin, D., Inorg. Chem., 1985, 24, 2276. 20 L’Eplattenier, F. A., and Pugin, A., Helv. Chim. Acta, 1975, 58, 101. 21 Horning, E. C., Org. Synth. Coll. Vol., 1955, 3, 140. 22 Moody, G. J., Oke, R. B., and Tomas, J. D. R., Analyst, 1970, 95, 910. 23 Guilbault, G. G., Durst, R. A., Frant, M.S., Freiser, H., Hansen, E. H., Light, T. S., Pungor, E., Rechnitz, G., Rice, N. M., Rohm, T. J., Simon, W., and Thomas, J. D. R., Pure Appl. Chem., 1976, 48, 127. 24 Nie, L.-H., Chen, B., and Yao, S.-Z., Acta Pharm. Sin., 1986, 21, 605. 25 O’Reilly, S. A., Daunert, S., and Bachas, G. L., Anal. Chem., 1991, 63, 1278. 26 Goedken, V. L., Pluth, J. J., Peg, S.-M., and Bursten, B., J. Am. Chem. Soc., 1976, 98, 8014 27 Sakata, K., Saitoh, Y., Kawakami, K., Nakamura, N., and Hashimoto, M., Synth.React. Inorg. Met.-Org. Chem., 1995, 25, 1279. 28 Sakata, K., Tagami, H., and Hashimoto, M., J. Heterocycl. Chem., 1989, 26, 805. 29 Pharmacopoeia Committee of the Ministry of Health of China, Chinese Pharmacopoeia, Chinese Health Press, Beijing, 1990, vol. 2. Paper 7/00544J Received January 23, 1997 Accepted May 30, 1997 Table 3 Determination of iodide in meglumine diatrizoate injection (%, m/v) Analysis No. Method 1 2 3 4 Mean Present electrode 35.23 35.89 35.09 35.95 35.42 ± 0.35 Precipitation method29 34.92 35.54 35.81 35.07 35.34 ± 0.41 1146 Analyst, October 1997, Vol. 122 Highly Selective Iodide Poly(vinyl chloride) Membrane Electrode Based on a Nickel(II) Tetraazaannulene Macrocyclic Complex Min Ying†, Ruo Yuan, Xiao-Min Zhang†, You-Qun Song, Zhi-Qiang Li, Guo-Li Shen and Ru-Qin Yu* Institute of Chemometrics and Chemical Sensing Technology, Chemistry and Chemical Engineering College, Hunan University, Changsha 410082, China A PVC membrane electrode based on a nickel(II) tetraazaannulene macrocyclic complex as the carrier is described.The electrode exhibits an anti-Hofmeister selectivity sequence with a preference for iodide at pH 3.0–4.0. It has a linear response to iodide from 8.0 3 1026 to 1 3 1021 mol dm23 with a slope of 254.7 ± 0.2 mV per decade (20 °C), a satisfactory reproducibility and a rapid response time. The pH dependence of the potential response is discussed. The response mechanism of the electrode was investigated by ac impedance, quartz crystal microbalance and spectroscopic techniques.The electrode can be used for the determination of iodide in drug preparations. Keywords: Iodide-selective electrode; nickel tetraazaannulene; macrocyclic complex; poly(vinyl chloride) membrane Research on anti-Hofmeister sensing materials for anions is an expeditiously expanding domain in chemical sensors. Recently, anion membrane electrodes, especially those based on vitamin B12 derivatives,1–3 metalloporphyrins,4,5 metallophthalocyanines, 6,7 Schiff base complexes8 and other complexes, 9 have been reported as possessing improved selectivity characteristics. Generally, the anti-Hofmeister anion selectivity of an electrode doped with organometallic complex is achieved by specific axial anion coordination with the central metal ion of the complex, and for such a coordination process the structure of the complex-forming carrier plays an important role in the final selectivity characteristics of the electrode in question.10 Metallotetraazaannulene-type macrocyclic complexes have received increasing attention in view of mimicking biological systems because of their chemical analogies to some naturally occurring porphyrins and corrin rings11–14 which are found in heme proteins, chlorophylls and metalloenzymes.Indeed, several complexes of this kind have been found to be effective as catalysts. In particular, a cobalt tetraazaannulene has been exploited as a catalyst for the reduction of molecular oxygen at the cathode in fuel cell;15 a nickel tetraazaannulene surfacemodified electrode has been used to catalyse the oxidation of water.16 The metallotetraazaannulenes have been studied as complexes similar to metalloporphyrins in structure.17,18 The unique saddle-like structures of this group of complexes make it of interest to investigate the potential application of the complexes as carriers for anion membrane electrodes, as the complex structure might favour the axial coordination of the central metals in these complexes.One would expect that the metallotetraazaannulene membrane electrodes would respond to some specific anions in an anti-Hofmeister fashion. The nickel complex of 5,7,12,14-tetramethyldibenzo- [b,i][1,4,8,11]tetraazacyclotetradecahexenate (Me4DBTA), easily derived in high yield from the template reaction of phenylene-1,2-diamine with pentane-2,4-dione in the presence of nickel(ii),13,19,20 was chosen as the first example of a metallotetraazaannulene carrier for the membrane electrode exhibiting desired potentiometric response characteristics.One outstanding property of Me4DBTA is its higher lipophilicity compared with planar macrocycles such as porphyrins and phthalocyanines resulting from its saddle-shaped deformation through steric interactions of the diiminato methyl group with the benzenoid rings.15 In this paper, we report the results of an investigation of the response characteristics of an Ni-Me4DBTA-based electrode towards iodide and the response mechanism in view of the axial coordination.The electrode has been preliminarily applied to the determination of iodide in some contrast media such as meglumine diatrizoate injection. Experimental Apparatus and Reagents Potentiometric and pH measurements were made with a Model PXS-125 Ionalyzer (Shanghai Analytical Instruments, Shanghai, China).The cell used was of the type Hg, Hg2Cl2, õKCl(saturated)õsample solutionõmembraneõ0.01 mol dm23 KClõ Ag, AgCl. Tested solutions were buffered with 0.1 mol dm23 H3PO4 and the expected pH was adjusted with NaOH solution. The activities of anions were calculated using the Debye–H�uckel equation. Before use, the electrode was conditioned in 0.01 mol dm23 KI solution for 1 d. Ni-Me4DBTA was synthesized with a yield of 40% by the template reaction as described by L’Eplattenier and Pugin20 and identified by elemental analysis and mass spectrometry: (MS) found, C 65.04, H 5.68, N 13.49; calculated, C 65.87, H 5.53, N 13.97%; MS, m/z 400.The structure of Ni-Me4DBTA is shown in Fig. 1. 2-Nitrophenyl octyl ether (o-NPOE) was prepared according to the literature method.21 Tetrahydrofuran, didecyl sebacate (DDS), didecyl phthalate (DDP) and chromatographicgrade poly(vinyl chloride) (PVC) were purchased from Shanghai Chemical (Shanghai, China).Meglumine diatrizoate injection was a pharmaceutical preparation from Xinyi Medicals (Shanghai, China). Re-distilled water and analytical-reagent grade reagents were used throughout. The membra composition was optimized by an orthogonal experimental design with the linear response range and slope as the object functions. The optimum composition was found to be 1.5% (m/m) carrier, 32.5% (m/m) PVC and 66% (m/m) solvent mediator. The PVC membrane electrode was fabricated according to the literature method.22 The potentiometric selectivity coefficients, Ki,j pot, were determined in pH 3.0 buffer solutions containing different anions at 0.01 mol dm23 by the separate solution method.23 † On leave from the East China Geological College, Linchuan, Jiangxi 344000, China.Analyst, October 1997, Vol. 122 (1143–1146) 1143UV/VIS Absorption Spectra and Ac Impedance Experiment The absorption spectra were recorded on a Lambda 17 UV/VIS spectrophotometer (Perkin-Elmer, Norwalk, CT, USA).Chloroform was employed as solvent. Spectra of the Ni-Me4DBTA chloroform solutions were recorded before and after shaking them with 0.1 mol dm23 KI and blank solution, respectively. Quartz Crystal Microbalance Experiment The frequency shifts of three different chloroform solutions were measured on a CN 3165 Frequency Counter (Sampo Co., Taiwan). Two chloroform solutions containing 0.01 mol dm23 Ni-Me4DBTA were separately shaken with 0.1 mol dm23 I2 and blank buffered solution (pH 3.0) for 30 min and the organic phases were treated with 3% H2O2 in 1.0 mol dm23 H2SO4 medium; the gold-coated quartz crystal was AT-cut with a fundamental resonance frequency of about 9 MHz and used to adsorb I2 for 15 min from the two organic phases to measure the frequency shifts DF1 and DF2, respectively.24 The third solution of pure chloroform was treated with 0.1 mol dm23 I2 in the same manner as mentioned above to measure the frequency shift DF3.Results and Discussion Emf Response Characteristics and Selectivity The potential selectivity coefficients of the electrodes are summarized in Table 1, showing that the electrodes incorporating Ni-Me4DBTA possess high selectivity towards iodide ion. The anti-Hofmeister selectivity sequence was assumed to be associated with an interaction between the central metal of the carrier and iodide ion. The potentometric response characteristics of three differently plasticized electrodes towards I2 were compared in Table 2.The o-NPOE plasticized membrane electrode showed a near Nernstian potentiometric response for 0.1–8.0 3 1026 mol dm23 I2 with a detection limit of 5.0 ± 0.3 3 1026 mol dm23 and a slope of 254.7 ± 0.2 mV per decade in 0.1 mol dm23 phosphate buffer solution (pH 3.0). The dc resistance of the membrane was estimated to be 132.5 ± 0.4 kW (average of six determinations). The standard deviation of potential responses over a period of 6 h in 1023 mol dm23 I2 was 0.5 mV (n = 36) and the potential readings for the electrode dipped alternatively into stirred solutions of 1022 and 1023 mol dm23 I2 gave a standard deviation of 0.7 mV (n = 5).The dynamic responses to different concentrations of I2 recorded at an interval of 5 s and depicted in Fig. 2, indicated that the time for the electrode to reach 90% of the steady response and subsequently to recover 90% of the blank response was less than 5 and 30 s, respectively. The lifetime of the electrode continuously contacted with 1022 mol dm23 KI at ambient temperature was found to be at least 2 months.In order to exclude the possibility that the response originated from I2 produced from the slow oxidation of I2 in an acidic medium by air, the potential responses to various concentrations of I2 freshly prepared in oxygen-free solutions were examined and no differences in the dynamic response patterns were observed from those in corresponding oxygen-containing solutions.Effect of pH on Potential Response The pH dependence of the electrode potential response is shown in Fig. 3. The effect of pH on the response characteristics can be explained by coordination competition between I2 and OH2. The increasing competition of OH2 at higher pH resulted in a narrower linear response range with a concomitant decrease in the response slope. A similar effect of OH2 on the potential response for electrodes based on porphyrin and vitamin B12 Fig. 1 Structure of Ni-Me4DBTA. A, Planar structure showing bonding and peripheral substitutions; B, side view of the complex showing typical ligand distortion. Table 1 Potential selectivity coefficients, log Ki,j pot Anion Plasticizer I2 ClO42 SCN2 NO22 Br2 NO32 OAc2 Cl2 SO42 o-NPOE 0.00 22.30 22.76 23.02 23.48 23.87 24.00 24.06 24.95 DDP 0.00 22.95 21.75 22.91 23.10 23.58 23.41 23.56 24.55 DDS 0.00 22.46 21.78 22.78 23.25 23.58 23.62 23.74 24.81 Table 2 Response characteristics of Ni-Me4DBTA-based electrodes with different solvent mediators.Results obtained from six measurements with two electrodes Plasticizer Parameter o-NPOE DDP DDS Slope mV per decade 254.7 ± 0.2 249.4 ± 0.6 242.8 ± 0.5 Linear range/mol dm23 8.0 3 1026–1021 1025–1022 1025–1022 Regression coefficient 20.9999 20.9998 20.9993 Detection limit/mol dm23 5.0 ± 0.331026 7.6 ± 0.531026 8.8 ± 0.531026 1144 Analyst, October 1997, Vol. 122derivatives was observed.7,25 As shown in Fig. 3 (inset), the competitive effect of OH2 was negligible and the electrode therefore demonstrated a near-Nernstian response towards iodide.The deviations from the linear response for concentrations of I2 at pH < 2.0 were due to the relatively high degree of oxidation of I2. Mechanism of Iodide Response Me4DBTA is a completely conjugated 14-membered ligand and its metal complex has shorter metal–nitrogen distances than corresponding 16-membered metalloporphyrins. The inherent predisposition of Me4DBTA takes the metal out of the N4 coordination plane (saddle-shaped).Consequently, two axial ligand sites are non-equivalent and five-coordination is preferred. 26 The preferential response towards I2 is believed to be associated with the coordination of iodide with the central metal of the carrier. With UV/VIS spectra as illustrated in Fig. 4, it was possible to distinguish the interaction between the central metal and iodide. The substantial increase in absorbance at 394 nm after the contact of the carrier solution with the iodidecontaining phase suggested the absorbing species had increased in size and axial coordination was thought to take place.At the same time, the effects of two more lipophilic anions, ClO42 and SCN2, on the spectrum of the carrier were investigated and no detectable changes in the UV/VIS spectra were noted. Thus, the anti-Hofmeister behaviour of the electrode can be explained by the specific interaction between the central metal and iodide ion.The ac impedances of the carrier-containing and the carrierfree membranes are shown in Fig 5. Although both of the membranes showed well defined semi-cyclic impedances at high frequencies, the bulk resistance of the membrane incorporating carrier decreased with increasing I2 concentration whereas that of carrier-free membrane remained almost constant. This pheomenon was related to the transfer of I2 across the solvent membrane resulting from the interaction between I2 and the carrier molecule.Only for the carrier-containing membrane was Warburg impedance at low frequencies observed, clearly indicating that the transfer process of I2 across the solvent membrane was diffusion controlled. The quartz crystal microbalance technique was used to detect the existence of I2 resulting from the oxidation of I2, transferring from the aqueous phase into the organic phase, by the deliberate addition of hydrogen peroxide.The frequency shifts DF1, DF2 and DF3 corresponded to the interactions of carrier in chloroform with 0.1 mol dm23 I2 and with blank solution, and of pure chloroform with 0.1 mol dm23 I2, respectively. The values of DF1, DF2 and DF3 were measured to be 3874, 850 and 215 Hz, respectively. The much larger frequency shift DF1 than DF2 and DF3 indirectly confirmed the transfer process of I2 across the organic–water interface. From the results mentioned above, it seems the metallotetraazaannulene- type carriers are promising as a new class of ion carriers for anions.By introducing different substituents on the phenylene and diimine carbon atoms as reported in the literature,27,28 one can prepare a whole spectrum of new carriers Fig. 2 Typical dynamic response of o-NPOE plasticized electrode towards different concentrations of I2. Fig. 3 Potential response curves of Ni-Me4DBTA-based electrode with o- NPOE as solvent mediator at different pH values.[I2]: 1026 (2); 1025 (.); 1024 (8); 1023 (½);1022 (Ó); and 1021 mol dm23 (5). Inset: potential response versus log [I2] at pH 3.0. Fig. 4 UV/VIS absorption spectra of Ni-Me4DBTA in chloroform treated with A, pH 3.0 blank solution and B, 0.1 mol dm21 KI solution. Fig. 5 Impedance plots of Ni-Me4DBTA-containing membrane (Ç) and Ni-Me4DBTA-free membrane (2) plasticized by o-NPOE. (a) 1 3 1024 and (b) 1 3 1023 mol dm21 KI. Analyst, October 1997, Vol. 122 1145with different possible electronic configurations.Further studies in this regard are in progress. Preliminary Application The resulting electrode was applied to the determination of iodide in a drug preparation (meglumine diatrizoate injection). The iodide content was determined by the standard additions method. An appropriate amount of sample was refluxed in concentrated NaOH solution in the presence of zinc powder for 30 min. After cooling, the reaction mixture was filtered and washed with water three times.The filtrate was acidified with H2SO4 to pH 3.0–4.0 and diluted to 200 ml with water. The resulting sample solution was determined using the present electrode and the precipitation method with AgNO3 as the titrant and tetrabromophenolphthalein as the indicator.29 The results listed in Table 3 were in agreement with those obtained by the precipitation method. It was concluded that there was a coordination interaction between I2 and Ni-Me4DBTA, which played an important role in improving the response selectivity, and for the saddle-shaped macrocyclic complex of Ni-Me4DBTA five-coordination was adopted to describe the coordination mechanism. Although the present electrode was pH dependent, it demonstrated a preferential, rapid and reproductible response towards iodide ion in pH 3.0–4.0 buffer solution and was successfully applied to the determination of iodide in drug preparations.This work was supported by the Foundations of the National Education Commission, the Machinery Ministry and the Hunan Scientific Commission of China.References 1 Daunert, S., and Bachas, L. G., Anal. Chem., 1989, 61, 499. 2 Schulthess, P., Ammann, D., Kr�autler, B., Caderas, C., Step�ýnek, R., and Simon, W., Anal. Chem., 1985, 57, 1397. 3 Step�ýnek, R., Kr�autler, B., Schulthess, P., Lindemann, B., Ammann, D., and Simon, W., Anal. Chim. Acta, 1986, 182, 83. 4 Gao, D., Li, J.-Z., and Yu, R.-Y., Anal.Chem., 1994, 66, 2245. 5 Daunert, S., Wallace, S., Florido, A., and Bachas, L. G., Anal. Chem., 1991, 63, 1676. 6 Li, J.-Z., Wu, X.-C., Yuan, R., Lin, H.-G., and Yu, R.-Q., Analyst, 1994, 119, 1363. 7 Chaniotakis, N. A., Park, S. B., and Meyerhoff, M. E., Anal. Chem., 1989, 61, 566. 8 Yuan, R., Chai, Y.-Q., Liu, D., Gao, D., Li, J.-Z., and Yu, R.-Q., Anal. Chem., 1993, 65, 2572. 9 Rothmaier, M., and Simon, W., Anal. Chim. Acta, 1993, 271, 135. 10 Glazier, S.A., and Arnold, M. A., Anal. Chem., 1991, 63, 754. 11 Weiss, M. C., and Goedken, V. L., J. Am. Chem. Soc., 1976, 98, 3389. 12 Nafie, L. A., Pastor, R. W., Dabrowiak, J. C., and Woodruff, W. H., J. Am. Chem. Soc., 1976, 98, 8007. 13 Place, D. A., Ferrara, G. P., Harland, J. J., and Dabrowiak, J. C., J. Heterocycl. Chem., 1980, 17, 439. 14 Hanke, R., and Breitmaier, E., Chem. Ber., 1982, 115, 1657. 15 Clauberg, B. W., and Sandstede, G., J. Electroanal. Chem. Interfacial Electrochem., 1976, 74, 393. 16 Issahary, D. A., Ginzburg, G., Polak, M., and Meyerstein, B., J. Chem. Soc., Chem. Commun., 1984, 441. 17 Weiss, M. C., Bursten, B., Peg, S.-M., and Geodken, V. L., J. Am. Chem. Soc., 1976, 98, 8021. 18 Woodruff, W. H., Pastor, R. W., and Dabrowiak, J. C., J. Am. Chem. Soc., 1976, 98, 7999. 19 Cutler, A. R., Alleyne, C. S., and Dolphin, D., Inorg. Chem., 1985, 24, 2276. 20 L’Eplattenier, F. A., and Pugin, A., Helv. Chim. Acta, 1975, 58, 101. 21 Horning, E. C., Org. Synth. Coll. Vol., 1955, 3, 140. 22 Moody, G. J., Oke, R. B., and Tomas, J. D. R., Analyst, 1970, 95, 910. 23 Guilbault, G. G., Durst, R. A., Frant, M. S., Freiser, H., Hansen, E. H., Light, T. S., Pungor, E., Rechnitz, G., Rice, N. M., Rohm, T. J., Simon, W., and Thomas, J. D. R., Pure Appl. Chem., 1976, 48, 127. 24 Nie, L.-H., Chen, B., and Yao, S.-Z., Acta Pharm. Sin., 1986, 21, 605. 25 O’Reilly, S. A., Daunert, S., and Bachas, G. L., Anal. Chem., 1991, 63, 1278. 26 Goedken, V. L., Pluth, J. J., Peg, S.-M., and Bursten, B., J. Am. Chem. Soc., 1976, 98, 8014 27 Sakata, K., Saitoh, Y., Kawakami, K., Nakamura, N., and Hashimoto, M., Synth. React. Inorg. Met.-Org. Chem., 1995, 25, 1279. 28 Sakata, K., Tagami, H., and Hashimoto, M., J. Heterocycl. Chem., 1989, 26, 805. 29 Pharmacopoeia Committee of the Ministry of Health of China, Chinese Pharmacopoeia, Chinese Health Press, Beijing, 1990, vol. 2. Paper 7/00544J Received January 23, 1997 Accepted May 30, 1997 Table 3 Determination of iodide in meglumine diatrizoate injection (%, m/v) Analysis No. Method 1 2 3 4 Mean Present electrode 35.23 35.89 35.09 35.95 35.42 ± 0.35 Precipitation method29 34.92 35.54 35.81 35.07 35.34 ± 0.41 1146 Analyst, October 1997,
ISSN:0003-2654
DOI:10.1039/a700544j
出版商:RSC
年代:1997
数据来源: RSC
|
26. |
Comparison of the Gold Reduction and Stripping Processes at Platinum, Rhodium, Iridium, Gold and Glassy Carbon Micro- and Macrodisk Electrodes |
|
Analyst,
Volume 122,
Issue 10,
1997,
Page 1147-1152
Alan M. Bond,
Preview
|
|
摘要:
Comparison of the Gold Reduction and Stripping Processes at Platinum, Rhodium, Iridium, Gold and Glassy Carbon Micro- and Macrodisk Electrodes Alan M. Bonda, Steven Kratsisa, Shelly Mitchellb and Jan Mocakc a Department of Chemistry, Monash University, Clayton, Victoria 3168, Australia b Department of Chemistry, Latrobe University, Bundoora, Victoria 3083, Australia c Department of Analytical Chemistry, Slovak Technical University, SK-81237 Bratislava, Slovak Republic The gold AuIII + 3e2 ? Au0 reduction and Au0 ? AuIII + 3e2 oxidation stripping processes in dilute aqua regia electrolyte (0.1 m HCl + 0.32 m HNO3) were examined at platinum, rhodium, iridium, gold and glassy carbon disk electrodes. After ascertaining that the preferred material was platinum, the effect of electrode size was evaluated by using nine different platinum disk electrodes having diameters ranging from 2 to 2000 mm.The optimum analytical response was obtained with a 50 mm diameter platinum disk electrode.With this electrode diameter, a sharp symmetrical gold stripping peak was obtained and the deposition process occurred predominantly under conditions of radial diffusion so that stirring of the solution was not required. In contrast, larger sized platinum electrodes produced a broader, asymmetric stripping response for the gold oxidation peak, whereas electrodes of smaller diameter provided poorer signal-to-noise ratios. The limit of detection and limit of quantification were calculated to be 4.4 3 1027 m (86 ppb) and 13.1 3 1027 m (258 ppb), respectively, at the 50 mm diameter platinum disk electrode under conditions of linear sweep stripping voltammetry at a scan rate of 200 mV s21 and a 140 s deposition time.The optimum electrode gave a very well defined gold oxidation signal with negligible background current when applied to the determination of gold in a gold ore sample. Keywords: Gold determination; voltammetry; optimum electrode size and electrode material There are numerous permutations and combinations of electrode materials and size available for use in analytical applications of voltammetry.Although the nature of the electrode material is often carefully considered in the development of a voltammetric method of analysis, there are relatively few studies in which the performance has been studied as a function of electrode size. However, the importance of electrode size is now emerging more frequently as a significant variable with the commercial availability of microdisk electrodes having diameters of around 10 mm.In contrast, common commercially available macrodisk electrodes typically have diameters in the millimetre range. Numerous analytical methods are available for the determination of the precious metals and many of them have recently been reviewed by Qu.1 A wide range of carbon, metal and chemically modified electrodes have been proposed for the voltammetric determination of gold.1,2 However, the specific dependence of the analytical response on the nature and size of the electrode material has yet to be determined.In this paper, the reduction of gold(iii): AuIII + 3e2 ? Au0 (1) and the stripping of gold: Au0 ? AuIII + 3e2 (2) in dilute aqua regia electrolyte (0.1 m HCl + 0.32 m HNO3) has been studied at platinum, rhodium, iridium, gold and glassy carbon disk electrodes to determine the optimum electrode surface in the analytical sense.Having ascertained that platinum is the preferred electrode material, the response at a range of electrode diameters varying from 2 to 2000 mm was examined to determine the optimum electrode size. The limit of detection (LOD) and limit of quantification (LOQ) were calculated for a typical set of conditions and the optimum electrode was applied to the determination of gold in a gold ore sample. Experimental Chemicals The dilute aqua regia electrolyte (0.1 m HCl + 0.32 m HNO3) was prepared by dilution of analytical-reagent grade nitric and hydrochloric acids.Dilution of the acids was effected with high-purity water (17 MWcm; Nanopure, Barnstead, IA, USA). Tests for interferences from iron and copper were undertaken using analytical-reagent grade FeSO4·7H2O and CuSO4·5H2O, respectively. Gold standard solutions were prepared by dilution of a 1000 ppm Gold Spectrosol solution (BDH, Poole, Dorset, UK) with the dilute aqua regia electrolyte.All glassware was cleaned in a detergent solution, followed by boiling nitric acid and then rinsing with water prior to drying at 200 °C. Preparation of the Gold Ore Sample The method of sample preparation used prior to the voltammetric analysis of gold ore samples, donated by Unichema International, was as follows: 1. The ore sample was dried at 105 °C and crushed to less than 100 mm particle size. 2. A 25 g amount of the sample was accurately weighed into a 250 ml beaker and 50 ml of 5 m HCl were added.The resultant mixture was left to stand at room temperature for 30 min to destroy sulfides. 3. The solution was heated on a hot-plate at approximately 120 °C for 2 h, with the beaker being covered with a watchglass. 4. After the solution had cooled, 10 ml of concentrated HCl and 10 ml of concentrated HNO3 were added. The solution was then heated for 4 h, allowing the liquid volume to decrease but not to go to dryness. 5. The solution was allowed to cool and 25 ml of 8 m HCl were added.After stirring, the solution was filtered into a separating funnel with the residues being rinsed with 2 m HCl. 6. The filtrate was extracted three times with diethyl ether (3 3 20 ml). Analyst, October 1997, Vol. 122 (1147–1152) 11477. The ether extract was washed with 0.1% HCl to remove iron. 8. The ether extracts were evaporated over 10% HCl using a heater. When the ether had volatilised, the gold remained in the acid solution.The sample was then added to dilute aqua regia electrolyte (3 : 1 ratio) and was ready for voltammetric analysis. The use of diethyl ether solvent extraction to remove interferences has been reported,3,4 as have other forms of solvent extraction.5–9 Equipment All voltammetric measurements were undertaken at 20 °C with a BAS 100A Electrochemical Analyzer (Bioanalytical Systems, West Lafayette, IN, USA) equipped with a BAS PA-1 preamplifier and Faraday cage.Where necessary, oxygen was removed by degassing solutions with nitrogen. The standard three-electrode arrangement was used with a platinum wire counter electrode, an Ag/AgCl (3 m KCl) reference electrode and the following working electrodes: 2000, 500, 250, 100, 70, 50, 25, 10 and 2 mm diameter platinum disk electrodes, 500 mm diameter glassy carbon, rhodium and iridium disk electrodes and a 2000 mm diameter gold disk electrode. The working electrodes were made via methods related to those described in the literature10 using metal wires from Goodfellow Metals (Cambridge, UK).Prior to each voltammetric experiment, the working electrode surface was polished on a LECO polishing pad with 0.6 mm alumina slurry, then rinsed with distilled water and carefully dried. Method for the Evaluation of Detection Limits According to the IUPAC definition, the limit of detection is the lowest concentration of an analyte that an analytical process can reliably detect.Originating from this definition, the signal counterpart of the limit of detection (LOD) is located 3sb (three times the standard deviation) or 6sb above the gross blank signal or, equivalently, 3sb or 6sb above zero when using the net signal. The use of 6sb instead of 3sb is substantiated by the tendency to diminish type II error11 causing an inadequate judgment of a true signal as the blank, although this is not important in our further discussion of the problem. If the calibration plot is a perfect straight line passing through the origin (net signals are assumed), then the limit of detection is simply determined by projecting the intercept of 3sb with the calibration plot on to the concentration axis.By analogy, the limit of quantification (LOQ) refers to the lowest concentration, or another quantity, which can be quantitatively measured with reasonable reliability by a given procedure. Its signal counterpart was defined as 10sb above the gross blank signal, or 10sb above zero when the net signals are used.In ref. 11, a 9sb (nine times the population standard deviation) approach is advocated instead. Also in this case, the use of a calibration plot is necessary for the calculation of the LOQ. When using the described standard method of obtaining the LOD, three main factors can considerably influence the reported result. The first occurs commonly because the number of measurements made on the blank and/or the sample itself is not sufficiently large to fulfil the conditions required to ensure that the normal distribution is valid for the measured signal.This problem may be overcome by replacing the value 3sb by the corresponding product t(n, a)sb where t(n,a) is the critical value of the t-distribution, n is the number of degrees of freedom and a is the confidence level.11 The second problem is related to inconsistency of the position of the zero net signal and the intercept of the calibration plot, denoted q0.The accuracy of the evaluated LOD depends on how positive or negative q0 is. When q0 is negative, the projection through the calibration plot leads to an over-estimation of the LOD, whereas a positive value leads to an under-estimation or even in extreme cases to a negative LOD value. If the intercept q0 is ignored, the LOD signal is projected through a line parallel to the net signal calibration plot, having the same slope q1 but a zero intercept (which is equivalent to calculation of the LOD as 3sb/q1) via this procedure, the negative LOD values are prevented and the over- and under-estimation proceed again but in the opposite way.The third problem is associated with the uncertainty of the position of the calibration line itself. A line of best fit could be shifted by chance or even by a statistical manipulation in such a way that a desired (usually lower) LOD is obtained owing to the improper values of regression parameters.Flaws in the ways in which the IUPAC definition of the LOD is commonly interpreted can be minimised by using a method known as the upper limit approach (ULA),11 which also is used in this paper. Using the ULA, the upper one-sided confidence limits of individual signal observations are calculated, which depend on the errors in regression and also on the type of the calibration model. If a one-parameter straight-line calibration model is statistically correct (the straight line passing through the origin), then the upper confidence limit for any chosen concentration c0 is distant from the regression line by the value of t(n,a)s, where n is now n 2 1 and s is defined by the equation s = (1 + c0 2/Sci 2)1/2 sy (3) with sy denoting the residual standard deviation.For the calculation of the limit of detection by the ULA, it is necessary to consider the case where c0 = 0 in eqn. (3). Under these conditions, the product in eqn. (3) refers to the upper limit signal at zero concentration.Further, when c0 = 0 and thus s = sy, the limit of detection can be simply calculated as LOD = t(n,a) sy/q1 (4) where n = n 2 1 and q1 is the slope of the calibration plot. Differences in the standard approach and the ULA method are illustrated in Fig. 1. Results and Discussion Selection of the Optimum Electrode Material The initial investigations concerned the selection of the optimum working electrode material for the voltammetric determination of gold(iii) in 0.1 m HCl + 0.32 m HNO3 electrolyte, which for convenience is refered to as dilute aqua regia.For this study, 2000 mm diameter platinum and 500 mm diameter platinum, glassy carbon, rhodium and iridium electrodes and a 2000 mm diameter gold electrode were chosen. Platinum macrodisk electrodes Cyclic voltammograms at 2000 and 500 mm diameter platinum macrodisk electrodes in dilute aqua regia showed a well defined response with the AuIII + 3e2?Au0 reduction and (Au0 ? AuIII + 3e2) oxidation peak potentials of 0.550 and 0.900 V versus Ag/AgCl, respectively, giving a peak-to-peak separation of 0.350 V with the 500 mm diameter electrode under the conditions in Fig. 2(a). An important feature with the use of this electrode material is the sharp gold oxidation peak which is required for the sensitive stripping method. Two closely spaced stripping peaks have been observed in other work,12 but under the conditions of the present study only a single stripping peak was found.At the platinum macrodisk electrodes, both the reduction and oxidation peaks when using cyclic voltammetry varied linearly with gold concentration over the range 2 3 1148 Analyst, October 1997, Vol. 1221025–2 3 1023 m and the limit of detection was found to be 2 3 1026 m and the limit of quantification 2 3 1025 m for the reduction process at a scan rate of 375 mV s21. Further, the reduction peak current was found to be dependent on the square root of the scan rate over the scan rate range 20–1000 mV s21 with a slope of 0.51 being obtained from a log(scan rate) versus log (peak height) plot.These data are consistent with a (linear) diffusion-controlled reduction process. The stripping process gave a slope of 0.76 when plotting log(scan rate) versus log (peak height) when the switching potential was 400 mV versus Ag/AgCl. Iridium electrode The first scan at the iridium working electrode exhibited similar characteristics to that found at the platinum electrode, with a sharp gold oxidation peak, reduction and oxidation peaks of 0.530 and 0.870 V versus Ag/AgCl, respectively, and a peak-topeak separation of 0.340 V under the conditions in Fig. 2(b). However, on repetitive cycling, the nature of the response changes as dissolution of the electrode occurs. With repetitive cycling, the gold reduction peak increases in current and shifts to more negative potentials (0.05 V shift by the fourth scan).Conversely, the gold stripping peak decreases in height with each cycle, but does not shift in potential [Fig. 2(b)]. The changes in the voltammetry with repetitive cycling are attributed to dissolution of the iridium metal electrode which is accompanied by the development of an IrIII–IrIV redox couple. 13,14. The dissolution is detected voltammetrically by the development of an oxidation current at 0.700 V versus Ag/ AgCl. Rhodium electrode Potential cycling of a gold(iii) solution using a rhodium electrode in dilute aqua regia also was accompanied by rhodium metal electrode dissolution [Fig. 2(c)]. Problems associated with dissolution of a particular electrode material are detected via an enhanced background current when monitoring the voltammetry of dilute gold solutions. For example, stripping voltammograms obtained from 3 31026 m gold(iii) solutions at platinum, rhodium and iridium electrodes with 2 3 1026 m standard additions using a 140 s deposition time and a scan rate of 200 mV s21 are compared in Fig. 3. The background with the platinum electrode where no dissolution occurs is very small in comparison with that observed with the rhodium and iridium electrodes. Glassy carbon electrode Under the conditions in Fig. 2(d), the cyclic voltammogram of gold(iii) with a glassy carbon working electrode showed reduction and oxidation peak potentials of 0.400 and 0.950 V versus Ag/AgCl, respectively, and a large peak-to-peak separation of 0.550 V for the gold(iii) reduction and gold metal reoxidation.However, apart from the considerable irreversibility of the process, the response is well defined [Fig. 2(d)]. At other forms of carbon electrode15 two gold stripping peaks have been reported, although only a single response was observed under the conditions of this study. Gold electrode A gold electrode can be used for the determination of gold via use of the gold reduction process. As shown in Fig. 2(e), the Fig. 1 Illustration of differences in determining the limit of detection by (a) the upper limit approach for the case when the calibration model is represented by a straight line passing through the origin (ULA1) and (b) the standard approach (SA) based on the IUPAC definition, calculating LOD1 by means of the calibration line and LOD2 by using an auxiliary line parallel to the calibration line (the same slope but a zero intercept). Fig. 2 Cyclic voltammograms of a 3.5 3 1024 m gold(iii) solution obtained in dilute aqua regia electrolyte at a scan rate 20 mV s21 using (a) platinum (500 mm), (b) iridium (500 mm), (c) rhodium (500 mm), (d) glassy carbon (500 mm) and (e) gold (2000 mm) working electrodes.Analyst, October 1997, Vol. 122 1149peak potential for reduction of gold when using a gold electrode was 0.550 V versus Ag/AgCl, which is comparable to values obtained at platinum and iridium electrodes. Obviously, stripping voltammetry of gold at a gold electrode is not possible.Further Considerations Concerning the Choice of Working Electrode Material The above data suggest that either platinum and glassy carbon may be the preferred electrode materials for the determination of gold. However, at low gold(iii) concentrations, problems with the background current arise when using the glassy carbon electrode. Various oxygen-containing functional groups have been identified by various workers which can be present on the carbon surface (e.g., phenol, carbonyl and quinone groups).16,17 Scanning to positive potentials oxidises these functional groups, forming a multilayer oxide which increases the background current.An additional problem with the glassy carbon electrode arises from the fact that the reduction peak potential for gold(iii) is almost 200 mV more negative than found at a platinum electrode. Therefore, when using a stripping technique for the gold determination, a considerably more negative deposition potential has to be used, which can cause problems with impurities such as iron, copper and lead, which are reduced at these more negative potentials.Hence, these elements, which are often present at relatively high levels in gold ore samples, may cause interference in the determination of gold at a glassy carbon electrode, but not at platinum. The problems with the presence of iron and the use of a glassy carbon electrode can be seen by comparing voltammograms obtained at platinum and glassy carbon electrodes for a 10 mg l21 gold(iii) solution spiked with 15 mg l21 [Fig. 4(a) and (b)] and 150 mg l21 iron(iii) [Fig. 4(c) and (d)]. The platinum electrode also is favourable for minimising interferences from copper because a deposition potential of 250 mV versus Ag/ AgCl can be used, which is more positive than the potential for the reduction of copper. Fig. 4(e) shows a stripping voltammogram of a 10 mg l21 gold(iii) solution spiked with 15 mg l21 copper(ii).Further details on intereferences likely to be encountered in the voltammetric determination of gold are available.18–20 Glassy carbon electrodes have been widely used for the determination of gold;5,6 in contrast and perhaps suprisingly, according to our findings, platinum indicating electrodes have been only rarely used.12,18 Determination of the Optimum Platinum Electrode Size for the Determination of Gold Evaluation of the optimum size platinum electrode was undertaken by comparing voltammograms obtained with a 3.5 3 1024 m gold(iii) solution using nine different platinum disk electrode sizes ranging from 2000 to 2 mm in diameter at a scan rate of 20 mVs21 (Fig. 5). At large electrode sizes (diameters of 2000, 500 and 250 mm) with this scan rate the reduction process exhibits a peak-shaped response whereas at smaller diameters (100–2 mm) a sigmoidal-shaped response is observed as the diffusion of gold(iii) to the electrode surface changes from linear at the larger diameter macrodisk electrodes to radial at the smaller diameter microdisk electrodes.21 The gold oxidation stripping peak also changes in shape from relatively broad and slightly asymmetric at larger diameters to sharp and symmetrical with a well defined baseline for diameters @50 mm.Further, with the smaller electrode sizes, stirring of the solution is not required during the deposition step22 so that microelectrodes may offer advantages over conventionally sized electrodes when using stripping methods.The change in voltammetry as a function of electrode diameter at a constant scan rate can be represented graphically by plotting Fig. 3 Stripping voltammograms obtained in dilute aqua regia electrolyte at 500 mm diameter (a) platinum, (b) rhodium and (c) iridium electrodes using a 140 s deposition time and a scan rate of 200 mV s21 for 3 3 1026 m gold(iii) solutions with 2 3 1026 m gold(iii) standard additions.Fig. 4 Stripping voltammograms obtained in dilute aqua regia electrolyte using a 140 s deposition and a scan rate of 200 mV s21 for a 10 mg l21 gold(iii) solution after addition of 10 mg l21 iron (iii) at (a) platinum (500 mm) and (b) glassy carbon (500 mm) electrodes and after addition of 150 mg l21 iron(iii) containing the same (c) platinum and (d) glassy carbon electrodes. (e) Stripping voltammogram obtained at a platinum (500 mm) electrode for 10 mg l21 gold(iii) solution containing 15 mg l21 copper(ii). 1150 Analyst, October 1997, Vol. 122the ratio of stripping peak current to the reduction peak or limiting current versus the logarithm of the area of the electrode. From this form of data analysis it can be observed that initially there is an increase in the ratio of oxidation to reduction currents as the size of the electrode is decreased until a diameter of about 50 mm is reached (Fig. 6). At very small sizes, the reproducibility for the gold process decreased, presumably owing to variable electrode areas resulting from greater difficulty in reproducibly polishing the smaller sized electrodes.Therefore, a 50 mm diameter electrode was selected as optimum on the basis of possessing a sharp, symmetrical gold oxidation peak, efficient electrolysis (oxidation/reduction value) and a favourable signalto- noise ratio. In practice, the 50 mm diameter electrode also represents close to the largest size electrode available for achieving near steady-state behaviour in practical analytical voltammetry.Hence this size of electrode is also about the largest for which stirring of the solution is not required to enhance the gold deposition process. Determination of the Limits of Detection and Quantification Using a 50 mm Diameter Platinum Microdisk Electrode The limits of detection and quantification under a typical set of conditions were determined using the 50 mm diameter platinum disk electrode and linear-sweep stripping voltammetry.Using a 140 s deposition time and a scan rate of 200 mV s21, a small but reproducible signal could be obtained with a 4 3 1027 m gold(iii) solution. A calibration plot of peak current versus concentration was constructed using 4 3 1027 m as the lowest analyte concentration and 4 3 1027 m standard additions (Fig. 7). Using the upper limit approach, the LOD was determined to be 4.4 3 1027 m (86 ppb) and the LOQ 13.1 3 1027 m (258 ppb).Obviously, the LOD and LOQ are strongly influenced by scan rate, deposition time and deposition potential. However, the conditions used above certainly provide adequate sensitivity for the determination of gold in gold ore samples, which was the problem of practical interest in this study. Determination of Gold in a Gold Ore Sample The 50 mm platinum microdisk electrode was used to determine gold in ore samples provided by Unichema International (Port Fig. 5 Cyclic voltammograms obtained in dilute aqua regia electrolyte at a scan rate of 20 mV s21 for 3.5 31024 m gold(iii) solutions using platinum electrodes with sizes varying from (a) 2000 to 70 mm and (b) 50 to 2 mm in diameter.Fig. 6 Variation in the ratio of anodic to cathodic peak or limiting currents versus log (electrode area) for the gold oxidation and gold(iii) reduction processes. Experimental conditions as in Fig. 6. Fig. 7 Stripping voltammogram obtained in dilute aqua regia electrolyte for a 140 s deposition time and a scan rate of 200 mV s21 at a 50 mm platinum disk electrode for a 4 3 1027 m gold(iii) solution with 4 x 1027 m standard additions of gold(iii).Analyst, October 1997, Vol. 122 1151Melbourne, Victoria, Australia). The stated concentration in the sample considered in Fig. 8 was 32 mg l21. A well defined stripping response was observed with a 140 s deposition time and a scan rate of 200 mV s21 . Standard additions of 13 mg l21 gold(iii) were made to the solution (Fig. 8) and a plot of peak current versus concentration of added gold(iii) was constructed. The gold concentration was calculated from the intercept and found to be 28 mg l21, which compares favourably with the stated value of 32 mg l21. Equally, well defined gold stripping voltammograms were obtained for other gold ore samples provided by Unichema International. Differential-pulse Stripping Voltammetry Differential-pulse stripping voltammetry is widely used in trace analysis.However, for the determination of gold at a platinum electrode, stripping studies at the 1026 m concentration level produced broader and more complex signals of lower reproducibility than obtained under linear-sweep conditions. The lack of improvement associated with the use of the differential-pulse method is probably associated with the extreme irreversibility of the gold redox chemistry. Consequently, the linear-sweep method is preferred for the determination of gold at platinum disk electrodes.Conclusions The preferred electrode material for the voltammetric determination of gold in 0.1 m HCl + 0.32 m HNO3 electrolyte was found to be platinum. Rhodium, iridium, glassy carbon and gold electrode materials exhibited less well defined responses or also exhibited surface interferences or dissolution problems in the acid electrolyte. A 50 mm diameter electrode was selected as an ideal size for a platinum disk electrode on the basis of generating the maximum (stripping) peak to reduction current ratio, possessing a sharp, symmetrical gold oxidation peak and having advantages associated with microelectrode (radial diffusion) characteristics.An excellent response was obtained when the optimum electrode was used to determine gold in gold ore samples. The authors thank T. Hughes and G. Scollary for numerous helpful discussions and Unichema International for financial assistance and the provision of the gold ore samples.References 1 Qu, Y. B., Analyst, 1996, 121, 139. 2 Turyan, I., and Mandler, D., Anal. Chem., 1993, 65, 2089. 3 Lintern, M., Mann, A., and Longman, D., Anal. Chim. Acta, 1988, 209, 193. 4 Kaplin, A. A., Pichugina, V. M., and Filichkina, O. G., Zavod. Lab., 1988, 54, 4. 5 Hall, G. E. M., and Vaive, J. E., Chem. Geol., 1992, 102, 41. 6 Jakubec, K., and Sir, Z., Anal. Chim. Acta, 1985, 172, 359. 7 Brainina, Kh. Z., Gornostaeva, T. D., and Pronin, V.A., Anal. Chem. (USSR), 1979, 34, 831; Zh. Anal. Khim., 1979, 34, 1081. 8 Gornostaeva, T. D., and Pronin, V. A., Anal. Chem. (USSR), 1971, 26, 1549; Zh. Anal. Khim., 1971, 26, 1736. 9 Larkins, P. L., Anal. Chim. Acta, 1985, 173, 77. 10 Koppenol, M., Cooper, J. B., and Bond, A. M., Am. Lab., 1994, 26, July, 25. 11 Mocak, J., Bond A. M., Mitchell, S., and Scollary, G., Pure Appl. Chem., 1997, 69, 297. 12 Bruk, B. S., Pozina, M. I., and Rozenfeld, E. I., Anal. Chem. (USSR), 1979, 34, 842, Zh.Anal. Khim., 1979, 34, 1095. 13 Bard, A. J., Encyclopedia of Electrochemistry of the Elements, Marcel Dekker, New York, 1976, vol. 6, p. 232. 14 Llopis, J., Catal. Rev., 1968, 2, 161. 15 Vasileva, L. N., and Koroleva, T. A., Anal. Chem. (USSR), 1973, 28, 1875; Zh. Anal. Khim., 1973, 28, 2107. 16 McCreery, R. N., in Electroanalytical Chemistry. A Series of Advances, ed. Bard, A. J., Marcel Dekker, New York, 1991, vol. 17, p. 259. 17 R. E. Panzer and P. R. Elving, Electrochim.Acta, 1975, 20, 635. 18 Huiliang, H., Jagner, D., and Renman, L., Anal. Chim. Acta, 1988, 208, 301. 19 Alexander, R., Kinsella, B., and Middleton, A., J. Electroanal. Chem., 1978, 93, 19. 20 Gao, Z., Li, P., Dong, S., and Zhao, Z., Anal. Chim. Acta, 1990, 232, 367. 21 Bond, A. M., Analyst, 1994, 119 (11) 1R. 22 Brainina, Kh. Z., and Bond, A. M., Anal. Chem., 1995, 67, 2586. Paper 7/02632C Received April 17, 1997 Accepted June 18, 1997 Fig. 8 Determination of gold(iii) in a gold ore sample by using linearsweep stripping voltammetry in dilute aqua regia electrolyte at a 50 mm platinum disk electrode with 4 3 13 mg l21 standard additions of gold(iii) with a deposition time of 140 s and a scan rate of 200 mV s21. 1152 Analyst, October 1997, Vol. 122 Comparison of the Gold Reduction and Stripping Processes at Platinum, Rhodium, Iridium, Gold and Glassy Carbon Micro- and Macrodisk Electrodes Alan M. Bonda, Steven Kratsisa, Shelly Mitchellb and Jan Mocakc a Department of Chemistry, Monash University, Clayton, Victoria 3168, Australia b Department of Chemistry, Latrobe University, Bundoora, Victoria 3083, Australia c Department of Analytical Chemistry, Slovak Technical University, SK-81237 Bratislava, Slovak Republic The gold AuIII + 3e2 ? Au0 reduction and Au0 ? AuIII + 3e2 oxidation stripping processes in dilute aqua regia electrolyte (0.1 m HCl + 0.32 m HNO3) were examined at platinum, rhodium, iridium, gold and glassy carbon disk electrodes.After ascertaining that the preferred material was platinum, the effect of electrode size was evaluated by using nine different platinum disk electrodes having diameters ranging from 2 to 2000 mm. The optimum analytical response was obtained with a 50 mm diameter platinum disk electrode. With this electrode diameter, a sharp symmetrical gold stripping peak was obtained and the deposition process occurred predominantly under conditions of radial diffusion so that stirring of the solution was not required.In contrast, larger sized platinum electrodes produced a broader, asymmetric stripping response for the gold oxidation peak, whereas electrodes of smaller diameter provided poorer signal-to-noise ratios. The limit of detection and limit of quantification were calculated to be 4.4 3 1027 m (86 ppb) and 13.1 3 1027 m (258 ppb), respectively, at the 50 mm diameter platinum disk electrode under conditions of linear sweep stripping voltammetry at a scan rate of 200 mV s21 and a 140 s deposition time.The optimum electrode gave a very well defined gold oxidation signal with negligible background current when applied to the determination of gold in a gold ore sample. Keywords: Gold determination; voltammetry; optimum electrode size and electrode material There are numerous permutations and combinations of electrode materials and size available for use in analytical applications of voltammetry.Although the nature of the electrode material is often carefully considered in the development of a voltammetric method of analysis, there are relatively few studies in which the performance has been studied as a function of electrode size. However, the importance of electrode size is now emerging more frequently as a significant variable with the commercial availability of microdisk electrodes having diameters of around 10 mm. In contrast, common commercially available macrodisk electrodes typically have diameters in the millimetre range.Numerous analytical methods are available for the determination of the precious metals and many of them have recently been reviewed by Qu.1 A wide range of carbon, metal and chemically modified electrodes have been proposed for the voltammetric determination of gold.1,2 However, the specific dependence of the analytical response on the nature and size of the electrode material has yet to be determined.In this paper, the reduction of gold(iii): AuIII + 3e2 ? Au0 (1) and the stripping of gold: Au0 ? AuIII + 3e2 (2) in dilute aqua regia electrolyte (0.1 m HCl + 0.32 m HNO3) has been studied at platinum, rhodium, iridium, gold and glassy carbon disk electrodes to determine the optimum electrode surface in the analytical sense. Having ascertained that platinum is the preferred electrode material, the response at a range of electrode diameters varying from 2 to 2000 mm was examined to determine the optimum electrode size.The limit of detection (LOD) and limit of quantification (LOQ) were calculated for a typical set of conditions and the optimum electrode was applied to the determination of gold in a gold ore sample. Experimental Chemicals The dilute aqua regia electrolyte (0.1 m HCl + 0.32 m HNO3) was prepared by dilution of analytical-reagent grade nitric and hydrochloric acids. Dilution of the acids was effected with high-purity water (17 MWcm; Nanopure, Barnstead, IA, USA).Tests for interferences from iron and copper were undertaken using analytical-reagent grade FeSO4·7H2O and CuSO4·5H2O, respectively. Gold standard solutions were prepared by dilution of a 1000 ppm Gold Spectrosol solution (BDH, Poole, Dorset, UK) with the dilute aqua regia electrolyte. All glassware was cleaned in a detergent solution, followed by boiling nitric acid and then rinsing with water prior to drying at 200 °C. Preparation of the Gold Ore Sample The method of sample preparation used prior to the voltammetric analysis of gold ore samples, donated by Unichema International, was as follows: 1.The ore sample was dried at 105 °C and crushed to less than 100 mm particle size. 2. A 25 g amount of the sample was accurately weighed into a 250 ml beaker and 50 ml of 5 m HCl were added. The resultant mixture was left to stand at room temperature for 30 min to destroy sulfides. 3. The solution was heated on a hot-plate at approximately 120 °C for 2 h, with the beaker being covered with a watchglass. 4. After the solution had cooled, 10 ml of concentrated HCl and 10 ml of concentrated HNO3 were added. The solution was then heated for 4 h, allowing the liquid volume to decrease but not to go to dryness. 5. The solution was allowed to cool and 25 ml of 8 m HCl were added. After stirring, the solution was filtered into a separating funnel with the residues being rinsed with 2 m HCl. 6. The filtrate was extracted three times with diethyl ether (3 3 20 ml).Analyst, October 1997, Vol. 122 (1147–1152) 11477. The ether extract was washed with 0.1% HCl to remove iron. 8. The ether extracts were evaporated over 10% HCl using a heater. When the ether had volatilised, the gold remained in the acid solution. The sample was then added to dilute aqua regia electrolyte (3 : 1 ratio) and was ready for voltammetric analysis. The use of diethyl ether solvent extraction to remove interferences has been reported,3,4 as have other forms of solvent extraction.5–9 Equipment All voltammetric measurements were undertaken at 20 °C with a BAS 100A Electrochemical Analyzer (Bioanalytical Systems, West Lafayette, IN, USA) equipped with a BAS PA-1 preamplifier and Faraday cage.Where necessary, oxygen was removed by degassing solutions with nitrogen. The standard three-electrode arrangement was used with a platinum wire counter electrode, an Ag/AgCl (3 m KCl) reference electrode and the following working electrodes: 2000, 500, 250, 100, 70, 50, 25, 10 and 2 mm diameter platinum disk electrodes, 500 mm diameter glassy carbon, rhodium and iridium disk electrodes and a 2000 mm diameter gold disk electrode. The working electrodes were made via methods related to those described in the literature10 using metal wires from Goodfellow Metals (Cambridge, UK). Prior to each voltammetric experiment, the working electrode surface was polished on a LECO polishing pad with 0.6 mm alumina slurry, then rinsed with distilled water and carefully dried.Method for the Evaluation of Detection Limits According to the IUPAC definition, the limit of detection is the lowest concentration of an analyte that an analytical process can reliably detect. Originating from this definition, the signal counterpart of the limit of detection (LOD) is located 3sb (three times the standard deviation) or 6sb above the gross blank signal or, equivalently, 3sb or 6sb above zero when using the net signal. The use of 6sb instead of 3sb is substantiated by the tendency to diminish type II error11 causing an inadequate judgment of a true signal as the blank, although this is not important in our further discussion of the problem.If the calibration plot is a perfect straight line passing through the origin (net signals are assumed), then the limit of detection is simply determined by projecting the intercept of 3sb with the calibration plot on to the concentration axis. By analogy, the limit of quantification (LOQ) refers to the lowest concentration, or another quantity, which can be quantitatively measured with reasonable reliability by a given procedure.Its signal counterpart was defined as 10sb above the gross blank signal, or 10sb above zero when the net signals are used. In ref. 11, a 9sb (nine times the population standard deviation) approach is advocated instead. Also in this case, the use of a calibration plot is necessary for the calculation of the LOQ.When using the described standard method of obtaining the LOD, three main factors can considerably influence the reported result. The first occurs commonly because the number of measurements made on the blank and/or the sample itself is not sufficiently large to fulfil the conditions required to ensure that the normal distribution is valid for the measured signal. This problem may be overcome by replacing the value 3sb by the corresponding product t(n, a)sb where t(n,a) is the critical value of the t-distribution, n is the number of degrees of freedom and a is the confidence level.11 The second problem is related to inconsistency of the position of the zero net signal and the intercept of the calibration plot, denoted q0.The accuracy of the evaluated LOD depends on how positive or negative q0 is. When q0 is negative, the projection through the calibration plot leads to an over-estimation of the LOD, whereas a positive value leads to an under-estimation or even in extreme cases to a negative LOD value.If the intercept q0 is ignored, the LOD signal is projected through a line parallel to the net signal calibration plot, having the same slope q1 but a zero intercept (which is equivalent to calculation of the LOD as 3sb/q1) via this procedure, the negative LOD values are prevented and the over- and under-estimation proceed again but in the opposite way.The third problem is associated with the uncertainty of the position of the calibration line itself. A line of best fit could be shifted by chance or even by a statistical manipulation in such a way that a desired (usually lower) LOD is obtained owing to the improper values of regression parameters. Flaws in the ways in which the IUPAC definition of the LOD is commonly interpreted can be minimised by using a method known as the upper limit approach (ULA),11 which also is used in this paper.Using the ULA, the upper one-sided confidence limits of individual signal observations are calculated, which depend on the errors in regression and also on the type of the calibration model. If a one-parameter straight-line calibration model is statistically correct (the straight line passing through the origin), then the upper confidence limit for any chosen concentration c0 is distant from the regression line by the value of t(n,a)s, where n is now n 2 1 and s is defined by the equation s = (1 + c0 2/Sci 2)1/2 sy (3) with sy denoting the residual standard deviation.For the calculation of the limit of detection by the ULA, it is necessary to consider the case where c0 = 0 in eqn. (3). Under these conditions, the product in eqn. (3) refers to the upper limit signal at zero concentration. Further, when c0 = 0 and thus s = sy, the limit of detection can be simply calculated as LOD = t(n,a) sy/q1 (4) where n = n 2 1 and q1 is the slope of the calibration plot.Differences in the standard approach and the ULA method are illustrated in Fig. 1. Results and Discussion Selection of the Optimum Electrode Material The initial investigations concerned the selection of the optimum working electrode material for the voltammetric determination of gold(iii) in 0.1 m HCl + 0.32 m HNO3 electrolyte, which for convenience is refered to as dilute aqua regia. For this study, 2000 mm diameter platinum and 500 mm diameter platinum, glassy carbon, rhodium and iridium electrodes and a 2000 mm diameter gold electrode were chosen.Platinum macrodisk electrodes Cyclic voltammograms at 2000 and 500 mm diameter platinum macrodisk electrodes in dilute aqua regia showed a well defined response with the AuIII + 3e2?Au0 reduction and (Au0 ? AuIII + 3e2) oxidation peak potentials of 0.550 and 0.900 V versus Ag/AgCl, respectively, giving a peak-to-peak separation of 0.350 V with the 500 mm diameter electrode under the conditions in Fig. 2(a). An important feature with the use of this electrode material is the sharp gold oxidation peak which is required for the sensitive stripping method. Two closely spaced stripping peaks have been observed in other work,12 but under the conditions of the present study only a single stripping peak was found. At the platinum macrodisk electrodes, both the reduction and oxidation peaks when using cyclic voltammetry varied linearly with gold concentration over the range 2 3 1148 Analyst, October 1997, Vol. 1221025–2 3 1023 m and the limit of detection was found to be 2 3 1026 m and the limit of quantification 2 3 1025 m for the reduction process at a scan rate of 375 mV s21. Further, the reduction peak current was found to be dependent on the square root of the scan rate over the scan rate range 20–1000 mV s21 with a slope of 0.51 being obtained from a log(scan rate) versus log (peak height) plot. These data are consistent with a (linear) diffusion-controlled reduction process. The stripping process gave a slope of 0.76 when plotting log(scan rate) versus log (peak height) when the switching potential was 400 mV versus Ag/AgCl. Iridium electrode The first scan at the iridium working electrode exhibited similar characteristics to that found at the platinum electrode, with a sharp gold oxidation peak, reduction and oxidation peaks of 0.530 and 0.870 V versus Ag/AgCl, respectively, and a peak-topeak separation of 0.340 V under the conditions in Fig. 2(b). However, on repetitive cycling, the nature of the response changes as dissolution of the electrode occurs. With repetitive cycling, the gold reduction peak increases in current and shifts to more negative potentials (0.05 V shift by the fourth scan). Conversely, the gold stripping peak decreases in height with each cycle, but does not shift in potential [Fig. 2(b)]. The changes in the voltammetry with repetitive cycling are attributed to dissolution of the iridium metal electrode which is accompanied by the development of an IrIII–IrIV redox couple. 13,14. The dissolution is detected voltammetrically by the development of an oxidation current at 0.700 V versus Ag/ AgCl. Rhodium electrode Potential cycling of a gold(iii) solution using a rhodium electrode in dilute aqua regia also was accompanied by rhodium metal electrode dissolution [Fig. 2(c)]. Problems associated with dissolution of a particular electrode material are detected via an enhanced background current when monitoring the voltammetry of dilute gold solutions.For example, stripping voltammograms obtained from 3 31026 m gold(iii) solutions at platinum, rhodium and iridium electrodes with 2 3 1026 m standard additions using a 140 s deposition time and a scan rate of 200 mV s21 are compared in Fig. 3. The background with the platinum electrode where no dissolution occurs is very small in comparison with that observed with the rhodium and iridium electrodes.Glassy carbon electrode Under the conditions in Fig. 2(d), the cyclic voltammogram of gold(iii) with a glassy carbon working electrode showed reduction and oxidation peak potentials of 0.400 and 0.950 V versus Ag/AgCl, respectively, and a large peak-to-peak separation of 0.550 V for the gold(iii) reduction and gold metal reoxidation. However, apart from the considerable irreversibility of the process, the response is well defined [Fig. 2(d)].At other forms of carbon electrode15 two gold stripping peaks have been reported, although only a single response was observed under the conditions of this study. Gold electrode A gold electrode can be used for the determination of gold via use of the gold reduction process. As shown in Fig. 2(e), the Fig. 1 Illustration of differences in determining the limit of detection by (a) the upper limit approach for the case when the calibration model is represented by a straight line passing through the origin (ULA1) and (b) the standard approach (SA) based on the IUPAC definition, calculating LOD1 by means of the calibration line and LOD2 by using an auxiliary line parallel to the calibration line (the same slope but a zero intercept).Fig. 2 Cyclic voltammograms of a 3.5 3 1024 m gold(iii) solution obtained in dilute aqua regia electrolyte at a scan rate 20 mV s21 using (a) platinum (500 mm), (b) iridium (500 mm), (c) rhodium (500 mm), (d) glassy carbon (500 mm) and (e) gold (2000 mm) working electrodes.Analyst, October 1997, Vol. 122 1149peak potential for reduction of gold when using a gold electrode was 0.550 V versus Ag/AgCl, which is comparable to values obtained at platinum and iridium electrodes. Obviously, stripping voltammetry of gold at a gold electrode is not possible. Further Considerations Concerning the Choice of Working Electrode Material The above data suggest that either platinum and glassy carbon may be the preferred electrode materials for the determination of gold.However, at low gold(iii) concentrations, problems with the background current arise when using the glassy carbon electrode. Various oxygen-containing functional groups have been identified by various workers which can be present on the carbon surface (e.g., phenol, carbonyl and quinone groups).16,17 Scanning to positive potentials oxidises these functional groups, forming a multilayer oxide which increases the background current.An additional problem with the glassy carbon electrode arises from the fact that the reduction peak potential for gold(iii) is almost 200 mV more negative than found at a platinum electrode. Therefore, when using a stripping technique for the gold determination, a considerably more negative deposition potential has to be used, which can cause problems with impurities such as iron, copper and lead, which are reduced at these more negative potentials.Hence, these elements, which are often present at relatively high levels in gold ore samples, may cause interference in the determination of gold at a glassy carbon electrode, but not at platinum. The problems with the presence of iron and the use of a glassy carbon electrode can be seen by comparing voltammograms obtained at platinum and glassy carbon electrodes for a 10 mg l21 gold(iii) solution spiked with 15 mg l21 [Fig. 4(a) and (b)] and 150 mg l21 iron(iii) [Fig. 4(c) and (d)].The platinum electrode also is favourable for minimising interferences from copper because a deposition potential of 250 mV versus Ag/ AgCl can be used, which is more positive than the potential for the reduction of copper. Fig. 4(e) shows a stripping voltammogram of a 10 mg l21 gold(iii) solution spiked with 15 mg l21 copper(ii). Further details on intereferences likely to be encountered in the voltammetric determination of gold are available.18–20 Glassy carbon electrodes have been widely used for the determination of gold;5,6 in contrast and perhaps suprisingly, according to our findings, platinum indicating electrodes have been only rarely used.12,18 Determination of the Optimum Platinum Electrode Size for the Determination of Gold Evaluation of the optimum size platinum electrode was undertaken by comparing voltammograms obtained with a 3.5 3 1024 m gold(iii) solution using nine different platinum disk electrode sizes ranging from 2000 to 2 mm in diameter at a scan rate of 20 mVs21 (Fig. 5). At large electrode sizes (diameters of 2000, 500 and 250 mm) with this scan rate the reduction process exhibits a peak-shaped response whereas at smaller diameters (100–2 mm) a sigmoidal-shaped response is observed as the diffusion of gold(iii) to the electrode surface changes from linear at the larger diameter macrodisk electrodes to radial at the smaller diameter microdisk electrodes.21 The gold oxidation stripping peak also changes in shape from relatively broad and slightly asymmetric at larger diameters to sharp and symmetrical with a well defined baseline for diameters @50 mm.Further, with the smaller electrode sizes, stirring of the solution is not required during the deposition step22 so that microelectrodes may offer advantages over conventionally sized electrodes when using stripping methods. The change in voltammetry as a function of electrode diameter at a constant scan rate can be represented graphically by plotting Fig. 3 Stripping voltammograms obtained in dilute aqua regia electrolyte at 500 mm diameter (a) platinum, (b) rhodium and (c) iridium electrodes using a 140 s deposition time and a scan rate of 200 mV s21 for 3 3 1026 m gold(iii) solutions with 2 3 1026 m gold(iii) standard additions. Fig. 4 Stripping voltammograms obtained in dilute aqua regia electrolyte using a 140 s deposition and a scan rate of 200 mV s21 for a 10 mg l21 gold(iii) solution after addition of 10 mg l21 iron (iii) at (a) platinum (500 mm) and (b) glassy carbon (500 mm) electrodes and after addition of 150 mg l21 iron(iii) containing the same (c) platinum and (d) glassy carbon electrodes. (e) Stripping voltammogram obtained at a platinum (500 mm) electrode for 10 mg l21 gold(iii) solution containing 15 mg l21 copper(ii). 1150 Analyst, October 1997, Vol. 122the ratio of stripping peak current to the reduction peak or limiting current versus the logarithm of the area of the electrode. From this form of data analysis it can be observed that initially there is an increase in the ratio of oxidation to reduction currents as the size of the electrode is decreased until a diameter of about 50 mm is reached (Fig. 6). At very small sizes, the reproducibility for the gold process decreased, presumably owing to variable electrode areas resulting from greater difficulty in reproducibly polishing the smaller sized electrodes. Therefore, a 50 mm diameter electrode was selected as optimum on the basis of possessing a sharp, symmetrical gold oxidation peak, efficient electrolysis (oxidation/reduction value) and a favourable signalto- noise ratio.In practice, the 50 mm diameter electrode also represents close to the largest size electrode available for achieving near steady-state behaviour in practical analytical voltammetry. Hence this size of electrode is also about the largest for which stirring of the solution is not required to enhance the gold deposition process.Determination of the Limits of Detection and Quantification Using a 50 mm Diameter Platinum Microdisk Electrode The limits of detection and quantification under a typical set of conditions were determined using the 50 mm diameter platinum disk electrode and linear-sweep stripping voltammetry. Using a 140 s deposition time and a scan rate of 200 mV s21, a small but reproducible signal could be obtained with a 4 3 1027 m gold(iii) solution.A calibration plot of peak current versus concentration was constructed using 4 3 1027 m as the lowest analyte concentration and 4 3 1027 m standard additions (Fig. 7). Using the upper limit approach, the LOD was determined to be 4.4 3 1027 m (86 ppb) and the LOQ 13.1 3 1027 m (258 ppb). Obviously, the LOD and LOQ are strongly influenced by scan rate, deposition time and deposition potential. However, the conditions used above certainly provide adequate sensitivity for the determination of gold in gold ore samples, which was the problem of practical interest in this study.Determination of Gold in a Gold Ore Sample The 50 mm platinum microdisk electrode was used to determine gold in ore samples provided by Unichema International (Port Fig. 5 Cyclic voltammograms obtained in dilute aqua regia electrolyte at a scan rate of 20 mV s21 for 3.5 31024 m gold(iii) solutions using platinum electrodes with sizes varying from (a) 2000 to 70 mm and (b) 50 to 2 mm in diameter.Fig. 6 Variation in the ratio of anodic to cathodic peak or limiting currents versus log (electrode area) for the gold oxidation and gold(iii) reduction processes. Experimental conditions as in Fig. 6. Fig. 7 Stripping voltammogram obtained in dilute aqua regia electrolyte for a 140 s deposition time and a scan rate of 200 mV s21 at a 50 mm platinum disk electrode for a 4 3 1027 m gold(iii) solution with 4 x 1027 m standard additions of gold(iii). Analyst, October 1997, Vol. 122 1151Melbourne, Victoria, Australia). The stated concentration in the sample considered in Fig. 8 was 32 mg l21. A well defined stripping response was observed with a 140 s deposition time and a scan rate of 200 mV s21 . Standard additions of 13 mg l21 gold(iii) were made to the solution (Fig. 8) and a plot of peak current versus concentration of added gold(iii) was constructed. The gold concentration was calculated from the intercept and found to be 28 mg l21, which compares favourably with the stated value of 32 mg l21. Equally, well defined gold stripping voltammograms were obtained for other gold ore samples provided by Unichema International.Differential-pulse Stripping Voltammetry Differential-pulse stripping voltammetry is widely used in trace analysis. However, for the determination of gold at a platinum electrode, stripping studies at the 1026 m concentration level produced broader and more complex signals of lower reproducibility than obtained under linear-sweep conditions. The lack of improvement associated with the use of the differential-pulse method is probably associated with the extreme irreversibility of the gold redox chemistry.Consequently, the linear-sweep method is preferred for the determination of gold at platinum disk electrodes. Conclusions The preferred electrode material for the voltammetric determination of gold in 0.1 m HCl + 0.32 m HNO3 electrolyte was found to be platinum.Rhodium, iridium, glassy carbon and gold electrode materials exhibited less well defined responses or also exhibited surface interferences or dissolution problems in the acid electrolyte. A 50 mm diameter electrode was selected as an ideal size for a platinum disk electrode on the basis of generating the maximum (stripping) peak to reduction current ratio, possessing a sharp, symmetrical gold oxidation peak and having advantages associated with microelectrode (radial diffusion) characteristics. An excellent response was obtained when the optimum electrode was used to determine gold in gold ore samples. The authors thank T. Hughes and G. Scollary for numerous helpful discussions and Unichema International for financial assistance and the provision of the gold ore samples. References 1 Qu, Y. B., Analyst, 1996, 121, 139. 2 Turyan, I., and Mandler, D., Anal. Chem., 1993, 65, 2089. 3 Lintern, M., Mann, A., and Longman, D., Anal. Chim. Acta, 1988, 209, 193. 4 Kaplin, A. A., Pichugina, V. M., and Filichkina, O. G., Zavod. Lab., 1988, 54, 4. 5 Hall, G. E. M., and Vaive, J. E., Chem. Geol., 1992, 102, 41. 6 Jakubec, K., and Sir, Z., Anal. Chim. Acta, 1985, 172, 359. 7 Brainina, Kh. Z., Gornostaeva, T. D., and Pronin, V. A., Anal. Chem. (USSR), 1979, 34, 831; Zh. Anal. Khim., 1979, 34, 1081. 8 Gornostaeva, T. D., and Pronin, V. A., Anal. Chem. (USSR), 1971, 26, 1549; Zh. Anal. Khim., 1971, 26, 1736. 9 Larkins, P. L., Anal. Chim. Acta, 1985, 173, 77. 10 Koppenol, M., Cooper, J. B., and Bond, A. M., Am. Lab., 1994, 26, July, 25. 11 Mocak, J., Bond A. M., Mitchell, S., and Scollary, G., Pure Appl. Chem., 1997, 69, 297. 12 Bruk, B. S., Pozina, M. I., and Rozenfeld, E. I., Anal. Chem. (USSR), 1979, 34, 842, Zh. Anal. Khim., 1979, 34, 1095. 13 Bard, A. J., Encyclopedia of Electrochemistry of the Elements, Marcel Dekker, New York, 1976, vol. 6, p. 232. 14 Llopis, J., Catal. Rev., 1968, 2, 161. 15 Vasileva, L. N., and Koroleva, T. A., Anal. Chem. (USSR), 1973, 28, 1875; Zh. Anal. Khim., 1973, 28, 2107. 16 McCreery, R. N., in Electroanalytical Chemistry. A Series of Advances, ed. Bard, A. J., Marcel Dekker, New York, 1991, vol. 17, p. 259. 17 R. E. Panzer and P. R. Elving, Electrochim. Acta, 1975, 20, 635. 18 Huiliang, H., Jagner, D., and Renman, L., Anal. Chim. Acta, 1988, 208, 301. 19 Alexander, R., Kinsella, B., and Middleton, A., J. Electroanal. Chem., 1978, 93, 19. 20 Gao, Z., Li, P., Dong, S., and Zhao, Z., Anal. Chim. Acta, 1990, 232, 367. 21 Bond, A. M., Analyst, 1994, 119 (11) 1R. 22 Brainina, Kh. Z., and Bond, A. M., Anal. Chem., 1995, 67, 2586. Paper 7/02632C Received April 17, 1997 Accepted June 18, 1997 Fig. 8 Determination of gold(iii) in a gold ore sample by using linearsweep stripping voltammetry in dilute aqua regia electrolyte at a 50 mm platinum disk electrode with 4 3 13 mg l21 standard additions of gold(iii) with a deposition time of 140 s and a scan rate of 200 mV s21. 1152 Analyst, October 1997, Vol. 122
ISSN:0003-2654
DOI:10.1039/a702632c
出版商:RSC
年代:1997
数据来源: RSC
|
27. |
Determination of Ascorbic Acid in Foodstuffs by Microdialysis Sampling and Liquid Chromatography With Electrochemical Detection |
|
Analyst,
Volume 122,
Issue 10,
1997,
Page 1153-1154
S. Mannino,
Preview
|
|
摘要:
Determination of Ascorbic Acid in Foodstuffs by Microdialysis Sampling and Liquid Chromatography With Electrochemical Detection S. Mannino* and M. S. Cosio Dipartimento di Scienze e Tecnologie Alimentari e Microbiologiche, Universit`a di Milano, Via Celoria 2, 20133 Milan, Italy A method for the determination of ascorbic acid in foodstuffs utilising microdialysis sampling coupled with LC with electrochemical detection is described. The application of the method to liquid food samples such as milk, yoghurt and fruit juices was evaluated. Compared with the classical method, the proposed procedure features straightforward sample preparation and improved sensitivity and selectivity.A sampling rate of 20 h21 is possible, depending on the particular sample under study. Keywords: Ascorbic acid; microdialysis sampling; liquid chromatography; electrochemical detection; milk products; fruit juices Ascorbic acid (AA) is a powerful antioxidant naturally present in many foods, especially fruits and vegetables, that plays an important role in the prevention of infectious diseases.Apart from its vitamin activity, ascorbic acid is frequently used in the food industry as an antioxidant to prevent undesirable changes in colour, taste and odour. Because of its biological and technological importance, it is of great interest in the food field to have rapid and sensitive methods for its routine and reliable determination. Ascorbic acid is measured in food samples after different sample preparation procedures, depending on the matrix, either by forming a stable derivative that can be measured spectrophotometrically or by chromatographic separation of the analyte and its subsequent monitoring by UV or electrochemical detection (ED).1,2 However, the accurate measurement of ascorbic acid in food products depends on the development of extraction procedures and analytical methods that do not permit its oxidation and photodecomposition.In fact, l-ascorbic acid is unstable, reacting very easily with oxygen, especially in the presence of heavy metal ions and light, forming l-dehydroascorbic acid and further degradation products. This instability is a serious problem for its quantification in complex samples such as milk and carbonated beverages since they require extensive and carefully sample preparation.3–5 It has been reported recently that significant losses of ascorbic acid can occur during the preparation of beer samples even when very gentle agitation is used to release dissolved CO2.6 This paper reports a procedure for the rapid and accurate determination of AA in food based on microdialysis sampling on line coupled with HPLC and amperometric detection that does not require any sample pre-treatment or the addition of antioxidants and chelating agents in order to stabilize AA during the course of its determination.Experimental Reagents All reagents were of analytical-reagent grade.A stock standard solution (10 g l21) of l-ascorbic acid (Merck, Darmstadt, Germany) was prepared in Milli-Q-purified water containing 3% m/v of metaphosphoric acid. This solution was stored in the dark at 4 °C and was stable for at least 7 d. Fresh working standard solutions were prepared daily by dilution. A mobile phase of 0.001 m sulfuric acid (Merck) was used for the isocratic elution of AA. Apparatus The microdialysis sampling system consisted of a CMA/100 microinjection pump (CMA/Microdialysis, Stockholm, Sweden) connected to an HPLC apparatus through a CMA/160 online injector (CMA/Microdialysis). A diagram of the apparatus was given elsewhere.7 The microdialysis probe was prepared manually according to the procedure of Moscone and Mascini.8 The microdialysis fibre used was a Filtral 12 AN 69 HF (id 200 mm and molecular mass cut-off approximately 20 000 Da) obtained from Hospal Industrie (Meyzief, France).The perfusion liquid was de-ionized water pumped at various flow rates according to the sample being analysed.The HPLC system consisted of a Model 880 PU pump (Jasco, Tokyo, Japan) and a Model 400 thin-layer amperometric detector EG&G PAR (Princeton, NJ, USA) equipped with a glassy carbon electrode and reference (Ag/AgCl, saturated) and counter electrodes. The chromatographic conditions were as follows: column, Fruit Quality Analysis (100 3 7.8 mm id) (Bio-Rad Labs., Richmond, CA, USA); eluting solution, 0.001 m sulfuric acid flow rate, 0.7 ml min21; electrochemical detection at +750 mV versus Ag/AgCl; and current range, usually 0.5 mA.The sample injection volume was 20 ml. Integration of peak areas and retention times was performed with Borwin v. 1.2 software (Jmbs Developments, Le Fontanil, France). Results and Discussion As the performance of the microdialysis probe depends on several factors such as flow rate, probe length and type of analyte, preliminary assays were carried out in order to find the best extraction conditions for the type of samples under study.9 It was found that probes of 5 mm length and a flow rate of the perfusion liquid of 20 ml min21 were the most suitable parameters for our purpose and they were therefore selected for all subsequent experimental work. Prior to use, all the probes were tested in order to verify their performance, linearity range and repeatability under the adopted experimental conditions. The repeatability was determined from 10 replicate analyses of standard solutions of 100 mg l21 of ascorbic acid.Fig. 1 shows response peaks for a series of 10 repetitive microdialysis –LC–ED experiments. As can be seen, well defined peaks are observed with a half-width of 30 s, allowing a sample throughput of about 20 injections h21. A high repeatability response is observed with a RSD of 0.9% (n = 10) and a mean of 143 nA. Such precision indicates that the same probe can be Analyst, October 1997, Vol. 122 (1153–1154) 1153used for prolonged periods without losing its extraction efficiency. Using the system under the conditions described under Experimental, the linear range for AA extended to 1000 mg l21 with an r2 value for the linear regression of 0.9995. A typical calibration curve for the working range is y = 2226x + 4130, where y is the AA peak area and x the concentration of the analyte in mg l21. However, with complex samples such as milk and dairy products the standard additions technique is recommended in order to compensate for any variability in the recovery of the microdialysis probe.The detection limit calculated using the linear regression technique from Miller and Miller10 was 3.2 mg l21. The proposed procedure was applied to the analysis of milk products and fruit juices and beverages. Selected sample chromatograms of milk and grapefruit nectar are shown in Fig. 2. As can be seen, no matrix interference was observed, indicating that the microdialysis sampling is also effective for sample clean-up.No sample preparation or bubble removal was necessary for fruit juices and carbonated beverages but dilution (1 + 1) with water for milk and yoghurt was necessary in order to have a well homogenized representative sample and to facilitate the standard additions procedure. The ascorbic acid content of selected food samples determined by the proposed method is illustrated in Table 1.Also reported are the results obtained by analysing the same samples using the sample preparation procedures proposed by Ashoor et al.11 followed by LC–ED analysis as described here. As can be seen from Table 1, the results obtained for all samples examined with the two procedures are not comparable, systematically lower values being obtained with the traditional procedure owing to the complexity of the prior sample preparation. Fruit juices showed higher recovery values than milk products as compared with the proposed method.In order to verify the accuracy of the proposed method, a sample of yoghurt and a sample of orange juice, gently boiled in order to destroy all AA, were spiked with known amounts of AA. The recovery of AA ranged from 97 to 102%. The over-all precision for an orangeade sample spiked at the level of 50 ppm of AA was found to be 1.1 mg l21 (n = 7). Conclusions The determination of AA based on microdialysis sampling and LC–ED has been demonstrated to be a robust method that can be advantageously applied to fruit juices, soft drinks and milk products.In contrast to other methods, the sample does not require any pre-treatments and no loss of AA occurs. Furthermore, the separation power of LC coupled with the inherent specificity of the electrochemical detector showed no interference from the other components of the matrices examined. References 1 Bajaj, K. L., and Kaur, G., Analyst, 1981, 106, 117. 2 Tsao, C. S., and Young, M., J. Chromatogr., 1985, 330, 408. 3 Koshiishi, I., and Imanari, T., Anal. Chem., 1997, 69, 216. 4 Rose, R. C., and Nahrwold, D. L., Anal. Biochem.,1981, 114, 140. 5 Margolis, S., and Biack, I., J. Assoc. Off. Anal. Chem., 1987, 7, 5. 6 Madigan, D., McMurrough, I., and Smyth, M. R., Anal. Commun., 1996, 33, 9. 7 Mannino, S., Cosio, M. S., and Zimei, P., Electroanalysis, 1996, 8, 4. 8 Moscone, D., and Mascini, M., Ann. Biol., 1992, 50, 123. 9 Mannino, S., and Cosio, M.S., Ital. Food Sci., 1996, 4, 11. 10 Miller, J. C., and Miller, J. N., Statistics for Analytical Chemists, Ellis Horwood Chichester, 3rd edn., 1993, p. 115. 11 Ashoor, S. A., Monte, W. C., and Welty, J., J. Assoc. Off. Anal. Chem.,1984, 67, 78. Paper 7/02224G Received April 2, 1997 Accepted July 9, 1997 Fig. 1 Chromatographic responses of standard solutions of 100 mg l21 of AA after microdialysis sampling. Conditions: applied potential, 0.75 V; flow rate, 20 ml min21; current range, 0.5 mA. Fig. 2 HPLC traces for (a) partially skimmed milk and (b) grapefruit nectar. Flow rate: (a) 20 and (b) 40 ml min21; other conditions as in Fig. 1. Calibration equations: (a) y = 2200x + 48333 (r2 = 0.9973); (b) y = 825x + 217500 (r2 = 0.9973) (y = peak area; x = AA concentration in mg l21). Table 1 Ascorbic acid in foods samples determined by the proposed and reference methods Ascorbic acid/mg l21 Proposed Reference Sample procedure* method† Grapefruit nectar 271.5 ± 5.78 238.9 ± 7.20 Orangeade 25.5 ± 0.26 23.0 ± 0.53 Orange juice 374 ± 6.27 321.6 ± 8.53 Pear juice 115.8 ± 1.45 98.4 ± 1.72 Peach juice 169.6 ± 2.14 140.8 ± 2.53 Pineapple juice 299.0 ± 5.40 272.1 ± 6.87 UHT whole milk 8.2 ± 0.65 5.7 ± 0.60 Partially skimmed milk (source A) 25.5 ± 0.22 18.4 ± 0.40 Partially skimmed milk (source B) 47.5 ± 0.52 35.6 ± 0.96 Skimmed milk 65.5 ± 1.22 51.1 ± 1.24 Apricot yoghurt 25.5 ± 0.38 17.6 ± 0.70 Strawberry yoghurt 35.7 ± 0.56 24.0 ± 1.42 * Average of five determinations.† Average of three determinations see text for explanation of method. 1154 Analyst, October 1997, Vol. 122 Determination of Ascorbic Acid in Foodstuffs by Microdialysis Sampling and Liquid Chromatography With Electrochemical Detection S. Mannino* and M. S. Cosio Dipartimento di Scienze e Tecnologie Alimentari e Microbiologiche, Universit`a di Milano, Via Celoria 2, 20133 Milan, Italy A method for the determination of ascorbic acid in foodstuffs utilising microdialysis sampling coupled with LC with electrochemical detection is described.The application of the method to liquid food samples such as milk, yoghurt and fruit juices was evaluated. Compared with the classical method, the proposed procedure features straightforward sample preparation and improved sensitivity and selectivity. A sampling rate of 20 h21 is possible, depending on the particular sample under study. Keywords: Ascorbic acid; microdialysis sampling; liquid chromatography; electrochemical detection; milk products; fruit juices Ascorbic acid (AA) is a powerful antioxidant naturally present in many foods, especially fruits and vegetables, that plays an important role in the prevention of infectious diseases. Apart from its vitamin activity, ascorbic acid is frequently used in the food industry as an antioxidant to prevent undesirable changes in colour, taste and odour.Because of its biological and technological importance, it is of great interest in the food field to have rapid and sensitive methods for its routine and reliable determination.Ascorbic acid is measured in food samples after different sample preparation procedures, depending on the matrix, either by forming a stable derivative that can be measured spectrophotometrically or by chromatographic separation of the analyte and its subsequent monitoring by UV or electrochemical detection (ED).1,2 However, the accurate measurement of ascorbic acid in food products depends on the development of extraction procedures and analytical methods that do not permit its oxidation and photodecomposition. In fact, l-ascorbic acid is unstable, reacting very easily with oxygen, especially in the presence of heavy metal ions and light, forming l-dehydroascorbic acid and further degradation products.This instability is a serious problem for its quantification in complex samples such as milk and carbonated beverages since they require extensive and carefully sample preparation.3–5 It has been reported recently that significant losses of ascorbic acid can occur during the preparation of beer samples even when very gentle agitation is used to release dissolved CO2.6 This paper reports a procedure for the rapid and accurate determination of AA in food based on microdialysis sampling on line coupled with HPLC and amperometric detection that does not require any sample pre-treatment or the addition of antioxidants and chelating agents in order to stabilize AA during the course of its determination.Experimental Reagents All reagents were of analytical-reagent grade. A stock standard solution (10 g l21) of l-ascorbic acid (Merck, Darmstadt, Germany) was prepared in Milli-Q-purified water containing 3% m/v of metaphosphoric acid. This solution was stored in the dark at 4 °C and was stable for at least 7 d. Fresh working standard solutions were prepared daily by dilution.A mobile phase of 0.001 m sulfuric acid (Merck) was used for the isocratic elution of AA. Apparatus The microdialysis sampling system consisted of a CMA/100 microinjection pump (CMA/Microdialysis, Stockholm, Sweden) connected to an HPLC apparatus through a CMA/160 online injector (CMA/Microdialysis). A diagram of the apparatus was given elsewhere.7 The microdialysis probe was prepared manually according to the procedure of Moscone and Mascini.8 The microdialysis fibre used was a Filtral 12 AN 69 HF (id 200 mm and molecular mass cut-off approximately 20 000 Da) obtained from Hospal Industrie (Meyzief, France).The perfusion liquid was de-ionized water pumped at various flow rates according to the sample being analysed. The HPLC system consisted of a Model 880 PU pump (Jasco, Tokyo, Japan) and a Model 400 thin-layer amperometric detector EG&G PAR (Princeton, NJ, USA) equipped with a glassy carbon electrode and reference (Ag/AgCl, saturated) and counter electrodes. The chromatographic conditions were as follows: column, Fruit Quality Analysis (100 3 7.8 mm id) (Bio-Rad Labs., Richmond, CA, USA); eluting solution, 0.001 m sulfuric acid flow rate, 0.7 ml min21; electrochemical detection at +750 mV versus Ag/AgCl; and current range, usually 0.5 mA.The sample injection volume was 20 ml. Integration of peak areas and retention times was performed with Borwin v. 1.2 software (Jmbs Developments, Le Fontanil, France).Results and Discussion As the performance of the microdialysis probe depends on several factors such as flow rate, probe length and type of analyte, preliminary assays were carried out in order to find the best extraction conditions for the type of samples under study.9 It was found that probes of 5 mm length and a flow rate of the perfusion liquid of 20 ml min21 were the most suitable parameters for our purpose and they were therefore selected for all subsequent experimental work.Prior to use, all the probes were tested in order to verify their performance, linearity range and repeatability under the adopted experimental conditions. The repeatability was determined from 10 replicate analyses of standard solutions of 100 mg l21 of ascorbic acid. Fig. 1 shows response peaks for a series of 10 repetitive microdialysis –LC–ED experiments. As can be seen, well defined peaks are observed with a half-width of 30 s, allowing a sample throughput of about 20 injections h21.A high repeatability response is observed with a RSD of 0.9% (n = 10) and a mean of 143 nA. Such precision indicates that the same probe can be Analyst, October 1997, Vol. 122 (1153–1154) 1153used for prolonged periods without losing its extraction efficiency. Using the system under the conditions described under Experimental, the linear range for AA extended to 1000 mg l21 with an r2 value for the linear regression of 0.9995. A typical calibration curve for the working range is y = 2226x + 4130, where y is the AA peak area and x the concentration of the analyte in mg l21. However, with complex samples such as milk and dairy products the standard additions technique is recommended in order to compensate for any variability in the recovery of the microdialysis probe.The detection limit calculated using the linear regression technique from Miller and Miller10 was 3.2 mg l21. The proposed procedure was applied to the analysis of milk products and fruit juices and beverages.Selected sample chromatograms of milk and grapefruit nectar are shown in Fig. 2. As can be seen, no matrix interference was observed, indicating that the microdialysis sampling is also effective for sample clean-up. No sample preparation or bubble removal was necessary for fruit juices and carbonated beverages but dilution (1 + 1) with water for milk and yoghurt was necessary in order to have a well homogenized representative sample and to facilitate the standard additions procedure.The ascorbic acid content of selected food samples determined by the proposed method is illustrated in Table 1. Also reported are the results obtained by analysing the same samples using the sample preparation procedures proposed by Ashoor et al.11 followed by LC–ED analysis as described here. As can be seen from Table 1, the results obtained for all samples examined with the two procedures are not comparable, systematically lower values being obtained with the traditional procedure owing to the complexity of the prior sample preparation.Fruit juices showed higher recovery values than milk products as compared with the proposed method. In order to verify the accuracy of the proposed method, a sample of yoghurt and a sample of orange juice, gently boiled in order to destroy all AA, were spiked with known amounts of AA. The recovery of AA ranged from 97 to 102%.The over-all precision for an orangeade sample spiked at the level of 50 ppm of AA was found to be 1.1 mg l21 (n = 7). Conclusions The determination of AA based on microdialysis sampling and LC–ED has been demonstrated to be a robust method that can be advantageously applied to fruit juices, soft drinks and milk products. In contrast to other methods, the sample does not require any pre-treatments and no loss of AA occurs. Furthermore, the separation power of LC coupled with the inherent specificity of the electrochemical detector showed no interference from the other components of the matrices examined.References 1 Bajaj, K. L., and Kaur, G., Analyst, 1981, 106, 117. 2 Tsao, C. S., and Young, M., J. Chromatogr., 1985, 330, 408. 3 Koshiishi, I., and Imanari, T., Anal. Chem., 1997, 69, 216. 4 Rose, R. C., and Nahrwold, D. L., Anal. Biochem.,1981, 114, 140. 5 Margolis, S., and Biack, I., J. Assoc. Off. Anal. Chem., 1987, 7, 5. 6 Madigan, D., McMurrough, I., and Smyth, M.R., Anal. Commun., 1996, 33, 9. 7 Mannino, S., Cosio, M. S., and Zimei, P., Electroanalysis, 1996, 8, 4. 8 Moscone, D., and Mascini, M., Ann. Biol., 1992, 50, 123. 9 Mannino, S., and Cosio, M. S., Ital. Food Sci., 1996, 4, 11. 10 Miller, J. C., and Miller, J. N., Statistics for Analytical Chemists, Ellis Horwood Chichester, 3rd edn., 1993, p. 115. 11 Ashoor, S. A., Monte, W. C., and Welty, J., J. Assoc. Off. Anal. Chem.,1984, 67, 78. Paper 7/02224G Received April 2, 1997 Accepted July 9, 1997 Fig. 1 Chromatographic responses of standard solutions of 100 mg l21 of AA after microdialysis sampling. Conditions: applied potential, 0.75 V; flow rate, 20 ml min21; current range, 0.5 mA. Fig. 2 HPLC traces for (a) partially skimmed milk and (b) grapefruit nectar. Flow rate: (a) 20 and (b) 40 ml min21; other conditions as in Fig. 1. Calibration equations: (a) y = 2200x + 48333 (r2 = 0.9973); (b) y = 825x + 217500 (r2 = 0.9973) (y = peak area; x = AA concentration in mg l21). Table 1 Ascorbic acid in foods samples determined by the proposed and reference methods Ascorbic acid/mg l21 Proposed Reference Sample procedure* method† Grapefruit nectar 271.5 ± 5.78 238.9 ± 7.20 Orangeade 25.5 ± 0.26 23.0 ± 0.53 Orange juice 374 ± 6.27 321.6 ± 8.53 Pear juice 115.8 ± 1.45 98.4 ± 1.72 Peach juice 169.6 ± 2.14 140.8 ± 2.53 Pineapple juice 299.0 ± 5.40 272.1 ± 6.87 UHT whole milk 8.2 ± 0.65 5.7 ± 0.60 Partially skimmed milk (source A) 25.5 ± 0.22 18.4 ± 0.40 Partially skimmed milk (source B) 47.5 ± 0.52 35.6 ± 0.96 Skimmed milk 65.5 ± 1.22 51.1 ± 1.24 Apricot yoghurt 25.5 ± 0.38 17.6 ± 0.70 Strawberry yoghurt 35.7 ± 0.56 24.0 ± 1.42 * Average of five determinations. † Average of three determinations see text for explanation of method. 1154 Analyst, October 1997, Vol. 122
ISSN:0003-2654
DOI:10.1039/a702224g
出版商:RSC
年代:1997
数据来源: RSC
|
28. |
A DANREF Certified Reference Material for Chromate in Cement |
|
Analyst,
Volume 122,
Issue 10,
1997,
Page 1155-1159
Jesper Kristiansen,
Preview
|
|
摘要:
A DANREF Certified Reference Material for Chromate in Cement Jesper Kristiansen*, Jytte Molin Christensen and Kirsten Byrialsen National Institute of Occupational Health, Lersø Parkall�e 105, DK-2100 Copenhagen, Denmark. E-mail: jkr@ami.dk Two candidate reference materials for chromate in cement were produced in the DANREF network and certified in an interlaboratory study. Fifteen laboratories participated in the interlaboratory study and six different analytical methods were used.The certified values were estimated as the consensus mean of laboratory mean values (outliers excluded). Only results from laboratories using methods relying on chromate speciation were accepted. The certified values (±95% confidence limits) were 0.678 (±0.075) mg CrVI kg21 dry cement for the low level and 6.04 (±0.28) mg CrVI kg21 dry cement for the high level. Methods based on total chromium determination gave on average results that were 6.5% higher (both levels). However, the difference between speciation and non-speciation results was significant at the high concentration level only.Keywords: Hexavalent chromium; cement; interlaboratory study; certified reference material Hexavalent chromium is toxic to humans, in whom it may cause bronchial cancer,1–3 skin allergy4,5 and probably also asthma1 and renal injury.6,7 Portland cement contains 1–40 mg of watersoluble hexavalent chromium per kilogram of dry cement8 and the potential risk of working with cement has been recognized by occupational health services in several countries.In Denmark, a Working Environment Service Order9 has decreed a maximum content of 2 mg of extractable chromate per kilogram of dry cement, and this limit has also been established in other countries, including Sweden, Norway and Germany. Chromate† dissolves when cement is mixed with water, and hand contact with wet cement may lead to a skin diseases such as ‘cement eczema’ or ‘chromate ulcers’, which are serious, often debilitating, allergic skin reactions.1,10–13 Unfortunately, chromate cannot be removed from the raw cement, but must be reduced to chromium(iii) after dissolution.Chromium(iii) is insoluble in the strongly alkaline solution formed by adding water to cement.10 Reducing agents such as iron(ii) are therefore added to the cement, but sometimes other, and unfortunately less effective, reducing agents are used, e.g., sulfite. Sulfite does not reduce chromium(vi) in alkaline solution, but when the cement extract is acidified during standard routine analysis, chromium(vi) is reduced, giving rise to a false-negative result.In view of this, the Danish Working Environment Service demanded a revision of the Danish standard method for the determination of extractable chromate in cement. In the course of this work, the Danish National Institute of Occupational Health initiated an intercomparison between three Danish laboratories. The results of this intercomparison indicated systematic differences between the laboratories, indicating a need to improve the quality of the analysis. Hence candidate certified reference materials for chromate in cement were produced within the auspices of DANREF, which is a network of Danish laboratories that produces and supports the use of certified reference materials.14 In order to establish the certified values, an interlaboratory collaborative study was launched with the participation of 15 laboratories from seven European countries.This paper presents the results of this interlaboratory study. Experimental Reagents and Materials Cement was prepared by Aalborg Portland (Aalborg, Denmark). Potassium dichromate (pro analysi), 1,5-diphenylcarbohydrazide (pro analysi) and sulfuric acid (pro analysi) were purchased from Merck (Darmstadt, Germany). The indicator solution was prepared by dissolving 0.125 g of 1,5-diphenylcarbohydrazide in 25 ml of ethanol (r = 0.79 g ml21) in a 50 ml calibrated flask and diluting to volume with water.A chromate stock standard solution (50 mg l21 Cr6+) was prepared by dissolving 0.1414 g of dried potassium dichromate (K2Cr2O7) in water in a 1000 ml calibrated flask and diluting to volume with water. A chromate working standard solution (5 mg l21 Cr6+) was prepared by diluting 50.0 ml of the stock standard solution with water in a 500 ml calibrated flask. All water used was Milli-Q filtered water (Millipore, Molsheim, France).Cement was stored in polypropylene flasks fitted with screwcaps (100 ml or 1 l), that were acid cleaned before use. Extraction of cement with water took place in glass test-tubes (40 mm diameter). A filter funnel mounted on a 100 ml suction flask connected to a suction pump was used to separate aqueous extract from cement. Polypropylene filters (pore diameter 5 mm) were used. Pipettes (5 and 25 ml) and calibrated flasks (50 ml) were used for dilutions. Preparation of the candidate reference material The candidate reference material was produced by mixing two cements with low and high chromate levels.The total amount prepared at each level was approximately 1 kg. The cement was mixed in a ball mill with rubber pebbles for 30 min, and thereafter transferred to a 1 l polypropylene flask and stored under nitrogen. This stock cement was thoroughly remixed immediately before the preparation of candidate reference material samples.The samples were prepared by filling 100 ml polypropylene flasks with approximately 30 g of cement and flushing with nitrogen. The flasks were kept in tightly closed plastic bags to prevent uptake of water by the cement. The homogeneity was checked afterwards by chemical analysis. Apparatus The absorption of the diphenylcarbohydrazide–chromate complex at 540 nm was measured using a Hewlett-Packard (Avondale, PA, USA) Model 8450A UV/VIS spectrophotometer. Total chromium in the extracts was determined by † In the following, the expression ‘chromate in cement’ will be used synonymously with ‘extractable chromate in cement’, well knowing that not all chromate of the cement is extracted under normal working conditions.10 Analyst, October 1997, Vol. 122 (1155–1159) 1155atomic absorption spectrometry using a Perkin-Elmer (Norwalk, CT, USA) Z-5100 atomic absorption spectrometer with Zeeman-effect background correction and equipped with a PE- 60 autosampler.Determination of CrVI The Danish Standard DS 1020 describes a method for determination of chromate in cement based on the formation of a colored complex between the chromate ion and 1,5-diphenylcarbohydrazide. 15 The following is a brief summary of DS 1020. Several standard methods for chromate determinations issued in other countries, e.g., in Sweden and Germany,16 are similar in their principles to DS 1020. Extraction Cement (25 g, weighed exactly) was mixed with 25.0 ml of water.The mixture was shaken or stirred vigorously for 15 min. The aqueous extract was separated from cement by filtering the suspension through a dry filter funnel placed on a dry suction flask. Measurement A 5.00 ml volume of the filtrate was transferred into a 50 ml calibrated flask, 5 ml of sulfuric acid (1.8 mol l21) were added and the solution was diluted to approximately 40 ml with water. The solution was allowed to cool to room temperature, then 5.0 ml of indicator solution were added, water was added to the mark and the solution was thoroughly mixed.The absorbance was measured 15–30 min after addition of the indicator solution at 540 nm. A 5 ml volume of water treated as the filtrate served as blank. Calibration Volumes of 1.0, 2.0, 5.0, 10.0 and 15.0 ml of chromate standard solution were transferred into 50 ml calibrated flasks and 5 ml of sulfuric acid (1.8 mol l21) and water were added as described above.The calibration solutions contained 5, 10, 25, 50 and 75 mg of Cr6+, respectively. Analytical variation The between-laboratory standard deviation of the DS 1020 standard method was 0.5 mg of chromium(vi) per kilogram of dry cement at a concentration of 5.3 mg kg21 in an interlaboratory comparison.16 Homogeneity and Stability Homogeneity of the candidate reference material was coirmed before the material was distributed to laboratories in the intercomparison study by analyzing three samples at each level.The homogeneity was investigated in more detail in the intercomparison study17 by testing the pooled repeatability standard deviation (expressing the between-vial variability) versus the analytical standard deviation (expressing the analytical variability). The stability of the material at room temperature was indicated by a preliminary study of another cement batch. The stability of the candidate reference material will be monitored at regular 6 month intervals.Initial data from two study occasions showed no indications of instability. Intercomparison Study Two concentration levels were distributed to the laboratories participating in the study. In measuring the samples, the laboratories were advised to follow a common extraction procedure (as described above). All laboratories received written instructions for handling the materials and for extraction and analytical procedures. However, the laboratories were free to choose the analytical method for the determination of chromate in the extract.The analytical methods used are described briefly in Table 1. Three samples at each concentration level were sent to all laboratories. After determining the chromium(vi) content the laboratories returned the results on a standardized report sheet together with information about the analytical method, including extraction and calibration. Data Evaluation Data from the intercomparison study were evaluated using the DANREF PC program.14 This computer program is designed specifically for statistical evaluation of reference material data and documentation of the quality of reference materials in accordance with international guidelines.14,17,18 The steps in the statistical evaluation of the intercomparison results were as follows (in brief).First, the goodness of fit of the laboratory mean values to a normal distribution was tested by the Kolmogorov–Smirnov–Lilliefors test.19 Failure to fit a normal distribution may be due to the presence of outlying values among the laboratory results.Therefore, the second step in the evaluaiton was to detect extreme values among the laboratory mean values and variances by the Grubbs and Cochran test20 according to the IUPAC 1994 harmonized procedure for outlier removal.21 The significance level was 99%. Laboratories with extreme values were excluded from the data sets before further processing. Third, each data set was analyzed by a one-way analysis of variance (ANOVA) to test if all laboratory mean values estimate the same (consensus) mean value, and to estimate the within- and between-laboratory variation.The fourth and final step in the evaluation was testing for normal distribution of laboratory means with laboratories with extreme values excluded. Repeatability and reproducibility were estimated according to ISO 5725.20 For more general statistical analysis of data, the Minitab statistical software was used.Results and Discussion Interlaboratory Study A brief description of the analytical principles used in the intercalibration study is given in Table 1. Individual results, given as laboratory mean values with 95% confidence limits, are given in Fig. 1. All laboratories used 10–25 g of cement, a cement : water ratio of 1.00 and an extraction time of 15 min. As can be seen Table 1 Description of the methods used in the intercomparison study on chromium(vi) in cement Method ID Analytical principle Color Direct determination of chromate color DPC Reaction with 1,5-diphenylcarboxyhydrazide and determination of colored complex ETAAS Determination of total chromium in cement extracts by ETAAS FAAS Determination of total chromium in cement extracts by FAAS ICP1 Separation of chromium species on an aluminium oxide column and determination of chromate by ICP-AES ICP2 Determination of total chromium in cement extracts by ICP-AES 1156 Analyst, October 1997, Vol. 122from Fig. 1, 13 laboratories used a method identical with or similar to, the Danish standard method (method DPC). Two laboratories determined chromate spectrophotometrically by directly measuring the chromate color at 372.5 nm in the extract (method Color). One laboratory used an aluminum oxide column to separate chromate from other anionic and cationic species and detection of chromate by ICP-MS (method ICP1).22 Three methods did not include speciation of CrVI.Instead they relied on the assumption that CrIII precipitates in the alkaline Fig. 1 Results from the laboratory intercomparison study (a) Low level; (b) high level. Laboratory means and 95% confidence intervals. Methods: see Table 1. Methods marked with asterisks indicate total chromium determinations; results marked Cochran indicate Cochran outlier; results marked Grubbs indicate Grubbs outlier (see text). Summary of results [mean value ± half-width of the 95% confidence interval (number of accepted results)]: (a) low level, all methods, 0.704 ± 0.069 mg kg21 (n = 17); chromate-speciating methods, 0.678 ± 0.075 mg kg21 (n = 13); total chromium methods, 0.79 ± 0.22 mg kg21 (n = 4); (b) high level, all methods, 6.21 ± 0.27 mg kg21 (n = 17); chromate-speciating method, 6.04 ± 0.28 mg kg21 (n = 13); total chromium methods, 6.77 ± 0.25 mg kg21 (n = 4); the last value is significantly different from the mean of chromate-speciating methods, P < 0.01.Analyst, October 1997, Vol. 122 1157cement–water suspension. Two laboratories used FAAS, one laboratory used ETAAS and one laboratory ICP-AES (method ICP2) to determine the chromium content in the cement extract. All laboratories that participated in the study used aqueous standards prepared from potassium dichromate (purity > 99.9%) for calibration. Six sets of results from four laboratories did not pass the outlier test described in the Experimental section (Fig. 1). In five cases the reason was a too high analytical standard deviation (Cochran outlier), i.e., these laboratories reported significantly less precise results than the other laboratories. Laboratory L6 (low level) was identified as having an extreme mean value by the Grubbs test, but also in this case the variance was fairly large (Fig. 1). Consensus mean values were estimated after outlier exclusion for the whole data set and for subsets hereof, namely methods based on speciation (DPC, Color, ICP1) and nonspeciation methods (FAAS, ETAAS, ICP2) (Fig. 1). Although the difference between consensus mean values of speciating and non-speciating methods is significant only at the high concentration level, it could not be excluded that methods relying on the determination of total chromium in cement extracts in general yield higher results compared with methods that rely on speciation. Four laboratories reported results by both speciating and non-speciating methods, and the average relative deviation between their speciation and non-speciation results was 6.5% (SEM 8.2%) at both levels.Homogeneity The laboratory repeatability standard deviation contains variability from both analytical uncertainty and between-vial differences. In Table 2, the pooled repeatability standard deviation obtained in the intercomparison study is compared with the analytical repeatability standard deviation obtained by replicate analysis of cement extracts. The ratio of the variances is not significantly different from unity (F-test, P > 0.05).Hence the between-vial differences are not significant in comparison with the analytical standard deviation. Certified Values and Their Uncertainties As differences between speciation methods and total chromium methods could not be excluded, it was decided to apply the consensus means of speciation methods as the certified values. Hence the certified values should be 0.678 mg kg21 at the low level and 6.04 mg kg21 at the high level.For the candidate materials to be accepted as certified reference materials, the certified values must be relevant to the users, and established with an uncertainty estimate that encompasses all relevant uncertainty contributions.23 The ‘naturally’ occurring values of extractable chromate in cement are often in the range 1–15 mg CrVI kg21 dry cement, but occasionally cement may contain up to 40 mg kg21.8 Hence the consensus means of the two materials, 0.678 and 6.04 mg CrVI kg21 dry cement, are at the low end of the naturally occurring range.However, they fall on each side of the value 2 mg CrVI kg21 cement, which is a statutory limit value in several countries. Moreover, mixing of the two materials allows the evaluation of an analytical method in the range between 0.678 and 6.04 mg kg21. Hence the values are relevant with respect both to naturally occurring contents of chromate in cement and to the need in occupational health measurements. The uncertainties of the certified values were estimated as the half-width of the 95% confidence interval of the consensus value.Since the variability includes contributions from different vials and from the passage of time, the uncertainty encompasses both stability and homogeneity uncertainty components, in addition to analytical variability. The uncertainty is 0.075 mg kg21 (relative uncertainty 11%) at the low level and 0.28 mg kg21 (relative uncertainty 4.6%) at the high level.These values are based on speciation methods only, and do not include uncertainty from method-dependent variations. As mentioned above, methods measuring total chromium have a tendency to give slightly higher results (+6.5%). To give an acceptable range of the certified value for these methods, the uncertainty must be recalculated in order to account for methoddependent effects. Because method-dependent differences constitute an uncorrected systematic effect, the systematic effect is added linearly to the uncertainty.23 As the systematic effect (+6.5%) amounts to 0.044 mg kg21 at the low level and 0.39 mg kg21, at the high level, the recalculated uncertainties are 0.12 mg kg21 (low level) and 0.67 mg kg21 (high level).These uncertainties should be applied only when a total chromium method is used for the determination of chromate in cement and when the correction for systematic effects due to non-speciation is unknown.The certified values and the uncertainties apply only to the extraction procedure described in this paper. In conclusion, the certified reference materials may be useful in validating analytical methods for determination of chromate in cement and in demonstrating compliance with occupational exposure norms. The availability of certified reference materials is important in harmonizing chromate regulations in the occupational health field.Financial support from the Danish Agency for Trade and Industry is gratefully acknowledged. We are grateful to Aalborg Portland (Aalborg Denmark) for providing the cement material. M. Toxværd and G. S. Nielsen are acknowledged for carrying out chemical analyses. We are greatly indebted to the following laboratories and persons for their participation in this interlaboratory collaborative study: Arbeitssicherheit und Umweltschutz, Betriebsorganisation und Arbeitsstudium eV, Berlin, Germany (Dr.Kieburg), Berufsgenossenschaftliches Institut f�ur Arbeitssicherheit, Sankt Augustin, Germany (D. Breuer), Euroc Research AB, Slite, Sweden (K. Nyberg), Finncement AB, Pargas, Finland (P. Tuohiniemi), Force Institute, Brøndby, Denmark (O. Petersen), Henkel Bautechnik GmbH, Unna, Germany (K. Ehlermann), Murværkscentret DTI, Hasselager, Denmark (H. Gram Pedersen), Norcem AS, Brevik, Norway (E. Stoltenberg-Hansson), Sheffield Hallam University, UK (C. McCleod), SINTEF, Trondheim, Norway (I.Meland), Swedish National Testing and Research Institute, Borås, Sweden (J. Winblad), University for Horticulture and Food Industry, Hungary (P. Fodor), Aalborg Portland A/S, Aalborg, Denmark (J. Almeborg) and National Institute of Occupational Health, Oslo, Norway (Y. Thomassen). References 1 Environmental Health Criteria 61, World Health Organization, Geneva, 1988. Tale 2 Testing of homogeneity of the cement materials. sr = Pooled repeatability standard deviation obtained in the intercomparison; s0 = analytical standard deviation; df = degrees of freedom; NS = between-vial differences not significant (P > 0.05) sr, pooled s0, pooled repeatability SD analytical SD Cement (between vials)/ (within vial)/ F-statistic material mg kg21 mg kg21 ( = sr,2/s0 2) Low level 0.040 (df = 21) 0.031 (df = 16) 1.655 (NS) High level 0.26 (df = 18) 0.172 (df = 16) 2.285 (NS) 1158 Analyst, October 1997, Vol. 1222 IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, IARC Monographs, Vol. 49, International Agency for Research on Cancer, Lyon, 1990. 3 Aitio, A., and Tomatis, L., in Trace Elements in Health and Disease, ed. Aitio, A., Aro, A., J�arvisalo, J., and Vainio, H., Royal Society of Chemistry, Cambridge, 1991. 4 Basketter, D. A., Bratico-Vangosa, G., Kaestner, W., Lally, C., and Bontinck, W. J., Contact Dermatitis, 1993, 28, 15. 5 Menn�e, T., Veien, N., Sjølin, K. E., and Maibach, H. I., Am.J. Contact Dermatitis, 1994, 5, 1. 6 Environmental Health Criteria 119, World Health Organization, Geneva, 1991. 7 Petersen, R., Mikkelsen, S., and Thomsen, O. F., Occup. Environ. Med., 1994, 51, 259. 8 Fregert, S., and Gruvberger, B., Berufsdermatosen, 1972, 20, 238. 9 Danish Working Environment Service Order, Water-Soluble Chromate in Cement, The Danish Working Environment Service, Copenhagen, 1983 (in Danish). 10 Fregert, S., Gruvberger, B., and Sandahl, E., Contact Dermatitis, 1979, 3, 39. 11 Goh, C. L., Gan, S. L., and Ngui, S. J., Contact Dermatitis, 1986, 15, 235. 12 Avnstorp, C., Contact Dermatitis, 1991, 25, 81. 13 Irvine, C., Pugh, C. E., Hansen, E. J., and Rycroft, R. J. G., Occup. Med., 1994, 44, 17. 14 Kristiansen, J., Christensen, J. M., Lillemark, L., Linde, S. A., Merry, J., Nyeland, B., and Petersen, O., Fresenius’ J. Anal. Chem., 1955, 352, 157. 15 Cement—Water-Soluble Chromate—Test Method, DS 1020, Danish Standard, Copenhagen, 1st edn., 1984 (in Danish). 16 Technische Regeln f�ur Gefahrstoffe 613, Bundesarbeitsblatt, 1993, 4, 63. 17 ISO Guide 35, Certification of Reference Materials, General and Statistical Principles, International Organization for Standardization, Geneva, 1985. 18 European Commission, Guidelines for the Production and Certification of BCR Reference Materials, Doc. BCR/48/93, Commission of the European Community, Brussels, 1994. 19 Conover, W. J., Practical Nonparametric Statistics, Wiley, New York, 2nd edn., 1981, p. 357. 20 ISO Standard 5725-2, Accuracy (Trueness and Precision) of Measurement Methods and Results—Part 2: Basic Method for the Determination of the Repeatability and Reproducibility of a Standard Measurement Method, International Organization for Standardization, Geneva, 1994. 21 Horwitz, W., Pure Appl. Chem., 1995, 67, 331. 22 Cox, A. G., Cook, I. G., and McLeod, C. W., Analyst, 1985, 110, 331. 23 Bureau International des Poid et Mesures, International Electrotechnical Commission, International Federation of Clinical Chemistry, International Organization for Standardization, International Union of Pure and Applied Chemistry, International Union of Pure and Applied Physics and International Organization of Legal Metrology, Guide to the Expression of Uncertainty in Measurement, International Organization for Standardization, Geneva, 1993.Paper 7/01619K Received March 7, 1997 Accepted June 23, 1997 Analyst, October 1997, Vol. 122 1159 A DANREF Certified Reference Material for Chromate in Cement Jesper Kristiansen*, Jytte Molin Christensen and Kirsten Byrialsen National Institute of Occupational Health, Lersø Parkall�e 105, DK-2100 Copenhagen, Denmark. E-mail: jkr@ami.dk Two candidate reference materials for chromate in cement were produced in the DANREF network and certified in an interlaboratory study. Fifteen laboratories participated in the interlaboratory study and six different analytical methods were used.The certified values were estimated as the consensus mean of laboratory mean values (outliers excluded). Only results from laboratories using methods relying on chromate speciation were accepted. The certified values (&plusfidence limits) were 0.678 (±0.075) mg CrVI kg21 dry cement for the low level and 6.04 (±0.28) mg CrVI kg21 dry cement for the high level. Methods based on total chromium determination gave on average results that were 6.5% higher (both levels).However, the difference between speciation and non-speciation results was significant at the high concentration level only. Keywords: Hexavalent chromium; cement; interlaboratory study; certified reference material Hexavalent chromium is toxic to humans, in whom it may cause bronchial cancer,1–3 skin allergy4,5 and probably also asthma1 and renal injury.6,7 Portland cement contains 1–40 mg of watersoluble hexavalent chromium per kilogram of dry cement8 and the potential risk of working with cement has been recognized by occupational health services in several countries.In Denmark, a Working Environment Service Order9 has decreed a maximum content of 2 mg of extractable chromate per kilogram of dry cement, and this limit has also been established in other countries, including Sweden, Norway and Germany. Chromate† dissolves when cement is mixed with water, and hand contact with wet cement may lead to a skin diseases such as ‘cement eczema’ or ‘chromate ulcers’, which are serious, often debilitating, allergic skin reactions.1,10–13 Unfortunately, chromate cannot be removed from the raw cement, but must be reduced to chromium(iii) after dissolution.Chromium(iii) is insoluble in the strongly alkaline solution formed by adding water to cement.10 Reducing agents such as iron(ii) are therefore added to the cement, but sometimes other, and unfortunately less effective, reducing agents are used, e.g., sulfite. Sulfite does not reduce chromium(vi) in alkaline solution, but when the cement extract is acidified during standard routine analysis, chromium(vi) is reduced, giving rise to a false-negative result.In view of this, the Danish Working Environment Service demanded a revision of the Danish standard method for the determination of extractable chromate in cement. In the course of this work, the Danish National Institute of Occupational Health initiated an intercomparison between three Danish laboratories.The results of this intercomparison indicated systematic differences between the laboratories, indicating a need to improve the quality of the analysis. Hence candidate certified reference materials for chromate in cement were produced within the auspices of DANREF, which is a network of Danish laboratories that produces and supports the use of certified reference materials.14 In order to establish the certified values, an interlaboratory collaborative study was launched with the participation of 15 laboratories from seven European countries.This paper presents the results of this interlaboratory study. Experimental Reagents and Materials Cement was prepared by Aalborg Portland (Aalborg, Denmark). Potassium dichromate (pro analysi), 1,5-diphenylcarbohydrazide (pro analysi) and sulfuric acid (pro analysi) were purchased from Merck (Darmstadt, Germany). The indicator solution was prepared by dissolving 0.125 g of 1,5-diphenylcarbohydrazide in 25 ml of ethanol (r = 0.79 g ml21) in a 50 ml calibrated flask and diluting to volume with water.A chromate stock standard solution (50 mg l21 Cr6+) was prepared by dissolving 0.1414 g of dried potassium dichromate (K2Cr2O7) in water in a 1000 ml calibrated flask and diluting to volume with water. A chromate working standard solution (5 mg l21 Cr6+) was prepared by diluting 50.0 ml of the stock standard solution with water in a 500 ml calibrated flask. All water used was Milli-Q filtered water (Millipore, Molsheim, France).Cement was stored in polypropylene flasks fitted with screwcaps (100 ml or 1 l), that were acid cleaned before use. Extraction of cement with water took place in glass test-tubes (40 mm diameter). A filter funnel mounted on a 100 ml suction flask connected to a suction pump was used to separate aqueous extract from cement. Polypropylene filters (pore diameter 5 mm) were used. Pipettes (5 and 25 ml) and calibrated flasks (50 ml) were used for dilutions.Preparation of the candidate reference material The candidate reference material was produced by mixing two cements with low and high chromate levels. The total amount prepared at each level was approximately 1 kg. The cement was mixed in a ball mill with rubber pebbles for 30 min, and thereafter transferred to a 1 l polypropylene flask and stored under nitrogen. This stock cement was thoroughly remixed immediately before the preparation of candidate reference material samples.The samples were prepared by filling 100 ml polypropylene flasks with approximately 30 g of cement and flushing with nitrogen. The flasks were kept in tightly closed plastic bags to prevent uptake of water by the cement. The homogeneity was checked afterwards by chemical analysis. Apparatus The absorption of the diphenylcarbohydrazide–chromate complex at 540 nm was measured using a Hewlett-Packard (Avondale, PA, USA) Model 8450A UV/VIS spectrophotometer.Total chromium in the extracts was determined by † In the following, the expression ‘chromate in cement’ will be used synonymously with ‘extractable chromate in cement’, well knowing that not all chromate of the cement is extracted under normal working conditions.10 Analyst, October 1997, Vol. 122 (1155–1159) 1155atomic absorption spectrometry using a Perkin-Elmer (Norwalk, CT, USA) Z-5100 atomic absorption spectrometer with Zeeman-effect background correction and equipped with a PE- 60 autosampler.Determination of CrVI The Danish Standard DS 1020 describes a method for determination of chromate in cement based on the formation of a colored complex between the chromate ion and 1,5-diphenylcarbohydrazide. 15 The following is a brief summary of DS 1020. Several standard methods for chromate determinations issued in other countries, e.g., in Sweden and Germany,16 are similar in their principles to DS 1020. Extraction Cement (25 g, weighed exactly) was mixed with 25.0 ml of water. The mixture was shaken or stirred vigorously for 15 min.The aqueous extract was separated from cement by filtering the suspension through a dry filter funnel placed on a dry suction flask. Measurement A 5.00 ml volume of the filtrate was transferred into a 50 ml calibrated flask, 5 ml of sulfuric acid (1.8 mol l21) were added and the solution was diluted to approximately 40 ml with water. The solution was allowed to cool to room temperature, then 5.0 ml of indicator solution were added, water was added to the mark and the solution was thoroughly mixed.The absorbance was measured 15–30 min after addition of the indicator solution at 540 nm. A 5 ml volume of water treated as the filtrate served as blank. Calibration Volumes of 1.0, 2.0, 5.0, 10.0 and 15.0 ml of chromate standard solution were transferred into 50 ml calibrated flasks and 5 ml of sulfuric acid (1.8 mol l21) and water were added as described above.The calibration solutions contained 5, 10, 25, 50 and 75 mg of Cr6+, respectively. Analytical variation The between-laboratory standard deviation of the DS 1020 standard method was 0.5 mg of chromium(vi) per kilogram of dry cement at a concentration of 5.3 mg kg21 in an interlaboratory comparison.16 Homogeneity and Stability Homogeneity of the candidate reference material was confirmed before the material was distributed to laboratories in the intercomparison study by analyzing three samples at each level.The homogeneity was investigated in more detail in the intercomparison study17 by testing the pooled repeatability standard deviation (expressing the between-vial variability) versus the analytical standard deviation (expressing the analytical variability). The stability of the material at room temperature was indicated by a preliminary study of another cement batch. The stability of the candidate reference material will be monitored at regular 6 month intervals.Initial data from two study occasions showed no indications of instability. Intercomparison Study Two concentration levels were distributed to the laboratories participating in the study. In measuring the samples, the laboratories were advised to follow a common extraction procedure (as described above). All laboratories received written instructions for handling the materials and for extraction and analytical procedures. However, the laboratories were free to choose the analytical method for the determination of chromate in the extract.The analytical methods used are described briefly in Table 1. Three samples at each concentration level were sent to all laboratories. After determining the chromium(vi) content the laboratories returned the results on a standardized report sheet together with information about the analytical method, including extraction and calibration. Data Evaluation Data from the intercomparison study were evaluated using the DANREF PC program.14 This computer program is designed specifically for statistical evaluation of reference material data and documentation of the quality of reference materials in accordance with international guidelines.14,17,18 The steps in the statistical evaluation of the intercomparison results were as follows (in brief). First, the goodness of fit of the laboratory mean values to a normal distribution was tested by the Kolmogorov–Smirnov–Lilliefors test.19 Failure to fit a normal distribution may be due to the presence of outlying values among the laboratory results.Therefore, the second step in the evaluaiton was to detect extreme values among the laboratory mean values and variances by the Grubbs and Cochran test20 according to the IUPAC 1994 harmonized procedure for outlier removal.21 The significance level was 99%. Laboratories with extreme values were excluded from the data sets before further processing.Third, each data set was analyzed by a one-way analysis of variance (ANOVA) to test if all laboratory mean values estimate the same (consensus) mean value, and to estimate the within- and between-laboratory variation. The fourth and final step in the evaluation was testing for normal distribution of laboratory means with laboratories with extreme values excluded. Repeatability and reproducibility were estimated according to ISO 5725.20 For more general statistical analysis of data, the Minitab statistical software was used.Results and Discussion Interlaboratory Study A brief description of the analytical principles used in the intercalibration study is given in Table 1. Individual results, given as laboratory mean values with 95% confidence limits, are given in Fig. 1. All laboratories used 10–25 g of cement, a cement : water ratio of 1.00 and an extraction time of 15 min. As can be seen Table 1 Description of the methods used in the intercomparison study on chromium(vi) in cement Method ID Analytical principle Color Direct determination of chromate color DPC Reaction with 1,5-diphenylcarboxyhydrazide and determination of colored complex ETAAS Determination of total chromium in cement extracts by ETAAS FAAS Determination of total chromium in cement extracts by FAAS ICP1 Separation of chromium species on an aluminium oxide column and determination of chromate by ICP-AES ICP2 Determination of total chromium in cement extracts by ICP-AES 1156 Analyst, October 1997, Vol. 122from Fig. 1, 13 laboratories used a method identical with or similar to, the Danish standard method (method DPC). Two laboratories determined chromate spectrophotometrically by directly measuring the chromate color at 372.5 nm in the extract (method Color). One laboratory used an aluminum oxide column to separate chromate from other anionic and cationic species and detection of chromate by ICP-MS (method ICP1).22 Three methods did not include speciation of CrVI. Instead they relied on the assumption that CrIII precipitates in the alkaline Fig. 1 Results from the laboratory intercomparison study (a) Low level; (b) high level. Laboratory means and 95% confidence intervals. Methods: see Table 1. Methods marked with asterisks indicate total chromium determinations; results marked Cochran indicate Cochran outlier; results marked Grubbs indicate Grubbs outlier (see text). Summary of results [mean value ± half-width of the 95% confidence interval (number of accepted results)]: (a) low level, all methods, 0.704 ± 0.069 mg kg21 (n = 17); chromate-speciating methods, 0.678 ± 0.075 mg kg21 (n = 13); total chromium methods, 0.79 ± 0.22 mg kg21 (n = 4); (b) high level, all methods, 6.21 ± 0.27 mg kg21 (n = 17); chromate-speciating method, 6.04 ± 0.28 mg kg21 (n = 13); total chromium methods, 6.77 ± 0.25 mg kg21 (n = 4); the last value is significantly different from the mean of chromate-speciating methods, P < 0.01.Analyst, October 1997, Vol. 122 1157cement–water suspension. Two laboratories used FAAS, one laboratory used ETAAS and one laboratory ICP-AES (method ICP2) to determine the chromium content in the cement extract. All laboratories that participated in the study used aqueous standards prepared from potassium dichromate (purity > 99.9%) for calibration. Six sets of results from four laboratories did not pass the outlier test described in the Experimental section (Fig. 1). In five cases the reason was a too high analytical standard deviation (Cochran outlier), i.e., these laboratories reported significantly less precise results than the other laboratories. Laboratory L6 (low level) was identified as having an extreme mean value by the Grubbs test, but also in this case the variance was fairly large (Fig. 1). Consensus mean values were estimated after outlier exclusion for the whole data set and for subsets hereof, namely methods based on speciation (DPC, Color, ICP1) and nonspeciation methods (FAAS, ETAAS, ICP2) (Fig. 1). Although the difference between consensus mean values of speciating and non-speciating methods is significant only at the high concentration level, it could not be excluded that methods relying on the determination of total chromium in cement extracts in general yield higher results compared with methods that rely on speciation. Four laboratories reported results by both speciating and non-speciating methods, and the average relative deviation between their speciation and non-speciation results was 6.5% (SEM 8.2%) at both levels.Homogeneity The laboratory repeatability standard deviation contains variability from both analytical uncertainty and between-vial differences. In Table 2, the pooled repeatability standard deviation obtained in the intercomparison study is compared with the analytical repeatability standard deviation obtained by replicate analysis of cement extracts.The ratio of the variances is not significantly different from unity (F-test, P > 0.05). Hence the between-vial differences are not significant in comparison with the analytical standard deviation. Certified Values and Their Uncertainties As differences between speciation methods and total chromium methods could not be excluded, it was decided to apply the consensus means of speciation methods as the certified values. Hence the certified values should be 0.678 mg kg21 at the low level and 6.04 mg kg21 at the high level.For the candidate materials to be accepted as certified reference materials, the certified values must be relevant to the users, and established with an uncertainty estimate that encompasses all relevant uncertainty contributions.23 The ‘naturally’ occurring values of extractable chromate in cement are often in the range 1–15 mg CrVI kg21 dry cement, but occasionally cement may contain up to 40 mg kg21.8 Hence the consensus means of the two materials, 0.678 and 6.04 mg CrVI kg21 dry cement, are at the low end of the naturally occurring range.However, they fall on each side of the value 2 mg CrVI kg21 cement, which is a statutory limit value in several countries. Moreover, mixing of the two materials allows the evaluation of an analytical method in the range between 0.678 and 6.04 mg kg21. Hence the values are relevant with respect both to naturally occurring contents of chromate in cement and to the need in occupational health measurements.The uncertainties of the certified values were estimated as the half-width of the 95% confidence interval of the consensus value. Since the variability includes contributions from different vials and from the passage of time, the uncertainty encompasses both stability and homogeneity uncertainty components, in addition to analytical variability. The uncertainty is 0.075 mg kg21 (relative uncertainty 11%) at the low level and 0.28 mg kg21 (relative uncertainty 4.6%) at the high level.These values are based on speciation methods only, and do not include uncertainty from method-dependent variations. As mentioned above, methods measuring total chromium have a tendency to give slightly higher results (+6.5%). To give an acceptable range of the certified value for these methods, the uncertainty must be recalculated in order to account for methoddependent effects. Because method-dependent differences constitute an uncorrected systematic effect, the systematic effect is added linearly to the uncertainty.23 As the systematic effect (+6.5%) amounts to 0.044 mg kg21 at the low level and 0.39 mg kg21, at the high level, the recalculated uncertainties are 0.12 mg kg21 (low level) and 0.67 mg kg21 (high level).These uncertainties should be applied only when a total chromium method is used for the determination of chromate in cement and when the correction for systematic effects due to non-speciation is unknown.The certified values and the uncertainties apply only to the extraction procedure described in this paper. In conclusion, the certified reference materials may be useful in validating analytical methods for determination of chromate in cement and in demonstrating compliance with occupational exposure norms. The availability of certified reference materials is important in harmonizing chromate regulations in the occupational health field.Financial support from the Danish Agency for Trade and Industry is gratefully acknowledged. We are grateful to Aalborg Portland (Aalborg Denmark) for providing the cement material. M. Toxværd and G. S. Nielsen are acknowledged for carrying out chemical analyses. We are greatly indebted to the following laboratories and persons for their participation in this interlaboratory collaborative study: Arbeitssicherheit und Umweltschutz, Betriebsorganisation und Arbeitsstudium eV, Berlin, Germany (Dr.Kieburg), Berufsgenossenschaftliches Institut f�ur Arbeitssicherheit, Sankt Augustin, Germany (D. Breuer), Euroc Research AB, Slite, Sweden (K. Nyberg), Finncement AB, Pargas, Finland (P. Tuohiniemi), Force Institute, Brøndby, Denmark (O. Petersen), Henkel Bautechnik GmbH, Unna, Germany (K. Ehlermann), Murværkscentret DTI, Hasselager, Denmark (H. Gram Pedersen), Norcem AS, Brevik, Norway (E. Stoltenberg-Hansson), Sheffield Hallam University, UK (C.McCleod), SINTEF, Trondheim, Norway (I. Meland), Swedish National Testing and Research Institute, Borås, Sweden (J. Winblad), University for Horticulture and Food Industry, Hungary (P. Fodor), Aalborg Portland A/S, Aalborg, Denmark (J. Almeborg) and National Institute of Occupational Health, Oslo, Norway (Y. Thomassen). References 1 Environmental Health Criteria 61, World Health Organization, Geneva, 1988. Tale 2 Testing of homogeneity of the cement materials. sr = Pooled repeatability standard deviation obtained in the intercomparison; s0 = analytical standard deviation; df = degrees of freedom; NS = between-vial differences not significant (P > 0.05) sr, pooled s0, pooled repeatability SD analytical SD Cement (between vials)/ (within vial)/ F-statistic material mg kg21 mg kg21 ( = sr,2/s0 2) Low level 0.040 (df = 21) 0.031 (df = 16) 1.655 (NS) High level 0.26 (df = 18) 0.172 (df = 16) 2.285 (NS) 1158 Analyst, October 1997, Vol. 1222 IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, IARC Monographs, Vol. 49, International Agency for Research on Cancer, Lyon, 1990. 3 Aitio, A., and Tomatis, L., in Trace Elements in Health and Disease, ed. Aitio, A., Aro, A., J�arvisalo, J., and Vainio, H., Royal Society of Chemistry, Cambridge, 1991. 4 Basketter, D. A., Bratico-Vangosa, G., Kaestner, W., Lally, C., and Bontinck, W. J., Contact Dermatitis, 1993, 28, 15. 5 Menn�e, T., Veien, N., Sjølin, K. E., and Maibach, H. I., Am. J. Contact Dermatitis, 1994, 5, 1. 6 Environmental Health Criteria 119, World Health Organization, Geneva, 1991. 7 Petersen, R., Mikkelsen, S., and Thomsen, O. F., Occup. Environ. Med., 1994, 51, 259. 8 Fregert, S., and Gruvberger, B., Berufsdermatosen, 1972, 20, 238. 9 Danish Working Environment Service Order, Water-Soluble Chromate in Cement, The Danish Working Environment Service, Copenhagen, 1983 (in Danish). 10 Fregert, S., Gruvberger, B., and Sandahl, E., Contact Dermatitis, 1979, 3, 39. 11 Goh, C. L., Gan, S. L., and Ngui, S. J., Contact Dermatitis, 1986, 15, 235. 12 Avnstorp, C., Contact Dermatitis, 1991, 25, 81. 13 Irvine, C., Pugh, C. E., Hansen, E. J., and Rycroft, R. J. G., Occup. Med., 1994, 44, 17. 14 Kristiansen, J., Christensen, J. M., Lillemark, L., Linde, S. A., Merry, J., Nyeland, B., and Petersen, O., Fresenius’ J. Anal. Chem., 1955, 352, 157. 15 Cement—Water-Soluble Chromate—Test Method, DS 1020, Danish Standard, Copenhagen, 1st edn., 1984 (in Danish). 16 Technische Regeln f�ur Gefahrstoffe 613, Bundesarbeitsblatt, 1993, 4, 63. 17 ISO Guide 35, Certification of Reference Materials, General and Statistical Principles, International Organization for Standardization, Geneva, 1985. 18 European Commission, Guidelines for the Production and Certification of BCR Reference Materials, Doc. BCR/48/93, Commission of the European Community, Brussels, 1994. 19 Conover, W. J., Practical Nonparametric Statistics, Wiley, New York, 2nd edn., 1981, p. 357. 20 ISO Standard 5725-2, Accuracy (Trueness and Precision) of Measurement Methods and Results—Part 2: Basic Method for the Determination of the Repeatability and Reproducibility of a Standard Measurement Method, International Organization for Standardization, Geneva, 1994. 21 Horwitz, W., Pure Appl. Chem., 1995, 67, 331. 22 Cox, A. G., Cook, I. G., and McLeod, C. W., Analyst, 1985, 110, 331. 23 Bureau International des Poid et Mesures, International Electrotechnical Commission, International Federation of Clinical Chemistry, International Organization for Standardization, International Union of Pure and Applied Chemistry, International Union of Pure and Applied Physics and International Organization of Legal Metrology, Guide to the Expression of Uncertainty in Measurement, International Organization for Standardization, Geneva, 1993. Paper 7/01619K Received March 7, 1997 Accepted June 23, 1997 Analyst, October 1997, Vol.
ISSN:0003-2654
DOI:10.1039/a701619k
出版商:RSC
年代:1997
数据来源: RSC
|
29. |
Sorption–Catalytic Determination of Manganese Directly on a Paper-based Chelating Sorbent |
|
Analyst,
Volume 122,
Issue 10,
1997,
Page 1161-1166
M. K. Beklemishev,
Preview
|
|
摘要:
H2N Me Me Me NH2 Me N N OH N O OH O OH O OH O ( a) ( b) Sorption–Catalytic Determination of Manganese Directly on a Paper-based Chelating Sorbent M. K. Beklemishev, T. A. Stoyan and I. F. Dolmanova* Analytical Chemistry Division, Department of Chemistry, Moscow State University, Moscow 119899 GSP, Russia. E-mail: mkb@analyt.chem.msu.su A hybrid technique based on a catalytic reaction carried out on the surface of a paper-based sorbent is proposed. It is shown that MnII exhibits its catalytic action in the oxidation of 3,3A,5,5A-tetramethylbenzidine with periodate in aqueous solution as well as on filter-paper with or without attached diethylenetriaminetetraacetate (DETATA) groups.Optimum conditions differ for the reaction in solution and on filter-paper. For equal catalyst concentrations, higher initial reaction rates are attainable on the filter-papers. Preconcentration of manganese on the DETATA sorbent combined with the subsequent catalytic reaction improves the selectivity, reduces the limit of determination down to 5 3 1026 mg l21 (as compared with 6 3 1025 mg l21 in solution) and expands the linear range by an order of magnitude (to 5 3 1026–2.5 3 1023 mg l21).The precision of manganese determination on the sorbent is high (the RSDs are @5% for !6 3 1024 mg Mn). Samples of tap and river water were analyzed by use of the proposed sorption–catalytic technique. A rapid procedure for the determination of manganese with visual detection was also developed. Keywords: Catalytic kinetic method of analysis; chelating sorbent; diethylenetriaminetetraacetate; preconcentration of manganese Catalytic procedures are appreciated by analytical chemists for their sensitivity and simplicity in realization.Catalytic indicator reactions can be applied to the determination of a large number of compounds including the catalyst, inhibitors, activators, and compounds which convert the catalyst into an active state (oxidizers, ligands1), and catalytically inactive metal ions through the use of the competitive complexation principle.Catalytic methods become especially powerful when combined with analyte separation/preconcentration, which allows the selectivity to be increased and the detection limits to be lowered. One of the approaches involves adsorption of metal ions on chelating sorbents with subsequent desorption and catalytic determination.2 However, this type of procedure, while being convenient in flow systems,2 is tedious in other cases owing to the need for desorption of the analyte.Here, we propose to eliminate the desorption stage by conducting the catalytic indicator reaction directly on the sorbent used for preconcentration purposes. Little information has been published to date on the foregoing principle.3 This novel area requires studies of the peculiarities of catalytic indicator reactions as they occur on the surfaces of sorbents, including kinetics, concentration effects, and the metrological characteristics of the procedures.The main purpose of this study was the practical realization of the aforementioned sorption–catalytic principle. For the investigation, the preconcentration of manganese on diethylenetriaminetetraacetate (DETATA) paper-based chelating sorbent was chosen as it combines effectiveness of metal sorption and simplicity of operation.2,4 As an indicator reaction, the oxidation of 3,3A,5,5A-tetramethylbenzidine (TMB) by potassium periodate (KIO4) catalyzed by MnII was selected.Various amines other than TMB were previously studied in this reaction;5–7 however, TMB is more stable in air, less carcinogenic and its oxidation provides higher absorbing products. In the present work, the catalytic effect of MnII on the reaction between TMB and KIO4 in solution was studied. The aim was to employ this system to ascertain the analytical possibilities of the sorption–catalytic approach for the determination of manganese with both instrumental and visual detection.A potential benefit of the sorption–catalytic approach may be realized in ‘rapid-test’ techniques, such as spot-tests, field tests and similar procedures,8 in which a visible signal appears on the surface of paper strips or other supports. The development of rapid tests for manganese may be a subject for further work originating from this study. Experimental Reagents, Solutions and Apparatus TMB ‘for analysis’ was obtained from Riedel-de H�aen, Hannover, Germany; boric, hydrochloric and sulfuric acids (‘special purity’) from Reakhim (Moscow, Russia) were used.All other reagents were of analytical-reagent grade. Humic and fulvic acids were obtained from Moscow river water by adsorption on poly(styrene–divinylbenzene)sorbent (XAD) with subsequent desalination. All aqueous solutions were prepared using distilled water. Ethanolic solutions of TMB (0.025 mol l21) were prepared every 10 d; less concentrated solutions were obtained by dilution with ethanol when necessary.A stock solution of periodate in water (1 g l21 KIO4) was diluted to an appropriate concentration (in most cases 0.1 g l21) every 4–5 d. A stock solution of anhydrous MnSO4 (1 g l21 Mn) was standardized by titration.9 Solutions with lower contents of the metal were prepared every 3 d (10 and 1 mg l21) or daily (0.01 mg l21) by dilution of the stock solution. All the manganese solutions were acidified with H2SO4 to pH 1.8–1.9; it was found that such acidification was necessary in order to obtain reproducible results.Hydrochloric acid was used to provide pH values ranging from 2 to 4; it was purified by isothermal distillation of the concentrated acid (reagent grade). Borate buffer of pH 6.8 was prepared by dropwise addition of Scheme 1 (a) TMB; (b) DETATA groups attached to filter paper. Analyst, October 1997, Vol. 122 (1161–1165) 11610.2 mol l21 KOH to a solution of 3.3 g of boric acid in 1000 ml of water until the desired pH value was reached.Buffers for the pH range 4–6 (0.1 mol l21 in sodium) were prepared from acetic acid and sodium acetate, and buffers for pH values of 7–11 (0.1 mol l21 in borate) were prepared from boric acid, Na2- B4O7·10H2O and KOH, by using standard procedures.10 The other buffer solutions with pH > 4 were purified by pumping through two consecutive DETATA filters placed in a special polyethylene holder (‘Biospectr’, St.Petersburg, Russia) at a rate of 8–10 ml min21, which removed transition metal impurities. The chelating sorbent was synthesized as described previously by chemical attachment of aminocarboxylic (diethylenetriaminetetraacetate, DETATA) groups to cellulose filterpaper. 11 The filter-paper was 2.5 cm in diameter. Water was distilled in a commercial apparatus (D3-4-2M, Ioshkar-Ola, Russia). A KFK-3 spectrophotometer (ZOMZ, Zagorsk, Russia) was used for measurements of absorbance. To spray the filter-papers with periodate solution, a hand-operated sprinkler for thin-layer chromatograms was used.Reaction in Solution Preliminary studies of various mixing orders of the components of the indicator reaction showed that maximum absorbance was achieved when MnII and borate buffer (pH 6.8) were mixed first, followed by addition of TMB and finally of KIO4. Based on this, the following procedure was used.In a 15 ml glass test-tube were placed the reagents in the following sequence (the amounts are given for the final procedure for manganese determination): 3.9 ml of the buffer (pH 6.8), 0.5 ml of the solution, containing MnII at a pH of 1.8–1.9, 0.1 ml of a 2.5 3 1024 mol l21 ethanolic solution of TMB and 0.5 ml of 4.3 3 1022 mol l21 periodate. The addition of periodate was taken as the time of the start of the reaction. After agitation, the reaction mixture was transferred into a 0.5 cm cell and the absorbance of the oxidation product was measured at 650 nm against water.The absorbance value at 3 min (A3) was taken as the analytical signal. The experiments were performed at ambient temperature, which was kept at 24 ± 1 °C. Reaction on Filter-paper In preliminary experiments the optimum mixing order of the reactants was sought witthe use of visual detection. Coloration of the filter-paper was most intense when TMB and the buffered manganese solution were mixed first and periodate was added at the last step.The role of the humidity of the filter-papers was also studied (the filter-papers were stored for up to 3 d in a desiccator with controlled 75–95% humidity); no effect of humidity on the reaction rate was observed. The final procedure for the reaction on filter-paper (with or without attached DETATA groups) is given below. MnII was first applied onto a filter-paper either (a) in the form of a 0.02 ml aliquot of a solution (0.4 mg l21 Mn) added to the centre of the filter-paper by an Eppendorf pipette, or (b) by pumping a buffered manganese solution of pH 6.8 (5 3 1025–0.05 mg l21 MnII, 0.08 mol l21 in borate) through the filter-paper by use of a peristaltic pump at a rate of 8 ml min21.The filter-paper was then dried with a gentle flow of pressurized air until no traces of moisture could be detected, after which TMB was pipetted onto the filter-paper (0.02 ml of a 4 3 1024 mol l21 solution) followed by an identical drying procedure.The oxidant was added by sprinkling the filter-paper with a 4.3 3 1024 mol l21 solution of KIO4, which was taken to be the beginning of the reaction. The coloured reaction zone was about 20 mm in diameter. Sprinkling rather than pipetting of periodate was necessary in order to obtain uniform coloration of the filterpapers; the sprinkling technique was applicable owing to a weak dependence of the analytical signal on the amount of KIO4 on the filter-paper, as shown below.The absorbance of a filter-paper was measured in its wet state against the same filter-paper without added periodate (i.e., not coloured with the reaction products) using the following procedure. On addition of TMB solution, but before drying, the filter-paper was placed between two glass plates fixed together with a clip and then placed in the cell compartment of the spectrophotometer perpendicular to the light beam. The absorbance of this filter-paper was taken as zero. The filter-paper was then dried, sprayed with KIO4 as described above, and the wet specimen was again placed between the glass plates for the measurements. The absorbance at 2 min (A2) was taken as the analytical signal.Calculation of Metrological Characteristics The limit of determination was defined as the concentration of manganese for which the RSD did not exceed 33%; this concentration was taken as the lower limit of the linear range.The limit of detection (cmin) was calculated as 3a/b; for the logarithmic plot y = a + bx, where y is absorbance and x = log(cMn), there are some complications. The following procedure was used: (1) shift of the graph y = a + bx to the origin (x = y = 0), i.e., transformation of the graph to the form Y = A + bX where A = 0 and b is the previous value. This required a recalculation: X = x + (a/b) [for example, X = log(cMn) + 4.82 for the reaction on DETATA filters]; (2) calculation of the regression parameters for the new graph Y = A + bX, including sA; (3) calculation of log(cmin) as 3sA/b.Results and Discussion Reaction Products At least two different coloured products are observed in the course of the studied reaction catalyzed by MnII in solution. A bluish green product (absorption maxima 370 and 650 nm) is formed under conditions where there is a deficiency of the oxidant (less than 1 3 1024 mol l21 KIO4 for a TMB concentration of 2.5 3 1024 mol l21, Fig. 1). According to the literature,12 this species may be a dimeric dication (lmax = 380 Fig. 1 Absorbance of the indicator reaction products as a function of KIO4 concentration. Curves 1 and 2, reaction in solution; curves 3 and 4, reaction on filter-paper with attached DETATA groups (KIO4 applied by sprinkling the solution onto the filter-paper). For the reaction in solution (curve 1): 2.5 3 1024 mol l21 TMB, 1 3 1023 mg l21 MnII; the measurements in solution (A3) were made at 650 nm for cKIO4 < 5 3 1025 mol l21 (bluish green product) or at 460 nm for cKIO4 > 5 3 1025 mol l21 (orange product). For the reaction on DETATA filter-paper (curve 3): 8 3 1028 mol TMB, 8 3 1023 mg l21 MnII; all the measurements of the filter-paper absorbances (A2) were made at 650 nm (bluish green product).For curves 2 and 4, the conditions are the same as for curves 1 and 3, respectively, but with no MnII. 1162 Analyst, October 1997, Vol. 122and 650 nm), probably in a mixture with a meriquinone (lmax = 655 nm).An increase in periodate concentration over 1 3 1023 mol l21 results in an orange product (lmax = 465 nm), which is probably the result of a more extensive oxidation and may be ascribed a quinonediimine structure12 (lmax = 465 nm). Intermediate concentrations of the oxidant provide a brown mixture with absorbance maxima at 380, 650 and 465 nm. As can be observed from Fig. 1, the maximum difference in the rates of MnII-catalyzed and non-catalytic reactions in solution corresponds to the formation of the bluish green product, whereas for higher periodate concentrations (and, consequently, for orange product formation) the difference decreases.For the determination of manganese, the bluish green product (370, 650 nm) was used. Kinetic Curves Kinetic curves for the TMB–MnII–periodate reaction are depicted in Fig. 2. In order to compare the data for the homogeneous and heterogeneous variants of the reaction, the catalyst concentrations should be presented in units which are common to both the solution and the filter-paper.For instance, the mass of manganese per square of the cross-section of the spectrophotometer light beam (mg cm22) would be a common value for both the sorbent and a cell with the solution. If the kinetic curves for the same concentration of manganese (in mg cm22, Fig. 2) are compared, it can be seen that the slope of the ascending portion of the curve is markedly higher for the reaction on filter-paper (with or without DETATA groups) than for that in solution.The optimum conditions for the reaction in solution and on the filter-papers differ, so it can only be noted that the highest attainable initial rate on the filter-papers is higher than that in solution. The absorbance of the filter-paper specimen (which corresponds to the formation of the bluish green product) increases rapidly for the first 0.5 min (counting from the start of the reaction), after which a slow decrease in absorption follows.In solution this decrease leads to the formation of a scarcelycoloured final product which requires a few hours. On the dry filter-paper the bluish green product is more stable (the coloration is not diminished during 24 h). The slow increase in the absorption of filter-paper specimens at 6–10 min is likely to be caused by processes in the cellulose filter-paper itself, e.g., swelling (wet filters without the reagents also exhibit an Fig. 2 Kinetic curves for the KIO4–MnII–TMB reaction at pH 6.8 in solution (1, 4), on filter-paper with attached DETATA groups (2, 5) and on filter-paper (3, 6). Curve 1, 2.5 3 1024 mol l21 TMB, 2.6 3 1025 mol l21 KIO4, 2.6 ng cm22 MnII; 2, 8 3 1028 mol TMB, 4.3 3 1024 mol l21 KIO4, 2.6 ng cm22 MnII. For curve 3, Mn solution (20 ml) was applied by pipetting. For curve 2, Mn solution (20 ml) was pumped through the DETATA filterpaper. The solution of KIO4 was applied onto the filter-paper (curves 2, 3, 5, 6) by sprinkling.For curves 4, 5 and 6, the conditions are the same as for curves 1, 2 and 3, respectively, but without MnII. The measurements were made at 650 nm. Fig. 3 Absorbance of the reaction products as a function of pH in solution (3, 4) (A3) and on filter-paper with attached DETATA groups (1, 2) (A2). Curve 1, 2.5 3 1024 mol l21 TMB, 2.6 3 1025 mol l21 KIO4, 1 ng ml21 MnII; 3, 8 3 1028 mol TMB, 4.3 3 1024 mol l21 KIO4, 8 ng ml21 MnII.For curves 2 and 4, the conditions are the same as for curves 1 and 3, respectively, but without MnII. The measurements were made at 650 nm. To study the effect of pH, manganese solution (0.04 mg l21) of the appropriate pH value was pumped through the DETATA filter-paper and the reaction was carried out as described under Reaction on Filter-paper. Fig. 4 Absorbance of the reaction products as a function of amount of TMB in solution (2, 4) (A3) and on filter-paper with attached DETATA groups (1, 3) (A2).Curve 1, 4.3 3 1024 mol l21 KIO4, 8 ng ml21 MnII; 2, 2.6 31025 mol l21 KIO4, 1 ng ml21 MnII. For curves 3 and 4, the conditions are the same as for curves 1 and 2, respectively, but without MnII. The measurements were made at 650 nm and at pH 6.8. Fig. 5 Relative standard deviations of absorbance of the reaction products as a function of manganese concentration for the reaction carried out in solution (1) and on filter-paper with attached DETATA groups (2).Curve 1, MnII amount denotes cMn/mg cm22 in solution (reaction conditions: 2.5 3 1024 mol l21 TMB, 2.6 3 1025 mol l21 KIO4); curve 2, MnII amounts were calculated as cMn30.02, where cMn is the MnII concentration in the solution that was pumped through the DETATA filter-paper and 0.02 l is the solution volume (conditions for the reaction on the filter-paper: 8 3 1028 mol TMB, 4.3 3 1024 mol l21 KIO4). The measurements were made at 650 nm and at pH 6.5.Analyst, October 1997, Vol. 122 1163increase in absorption). If the latter is not taken into account, the shape of the kinetic curves (Fig. 1) may be tentatively explained13 by reversible sequential–parallel reactions such as TMB"Product I (370, 650 nm)"Product II (colourless) or a set of two reversible reactions TMB"Product I (370, 650 nm) TMB"Product II (colourless) If one of these schemes is true, the steady-state portion of the absorbance–time plot will correspond to an equilibrium of the bluish green product with a colourless product.Effect of pH and Reagent Concentrations It was interesting to study whether the influence of reagent concentrations (TMB, KIO4) and pH on the signal differs for the reaction on filter-paper as opposed to in solution. As shown in Figs. 3–5, considerable differences exist. As can be seen from the pH curve (Fig. 3), the maximum amount of the reaction products is formed at pH 3.1 and 6.8 both in solution and on the DETATA filter-papers, but only on the sorbent is there a catalytic effect of manganese at pH 3.1.The signal on the DETATA filter-papers is higher at pH 3.1 than at pH 6.8 but the precision is poorer, viz., RSDs of 8 and 2%, respectively, are obtained (for 8 31028 mol TMB, 4.3 31024 mol l21 KIO4 and 8 ng ml21 MnII pumped through the DETATA filter-papers). One reason for the low reproducibility at pH 3.1 may be incomplete sorption of MnII at this pH value.Subsequent reactions, both in solution and on the filter-papers, were carried out at pH 6.8. A study of the effect of the periodate concentration on the reaction on the filter-papers showed that only the bluish green product is formed even at high concentrations of the oxidant. In solution (Fig. 1), the orange product was found at KIO4 :TMB ratios !1 : 1, whereas on the filter-papers it was never obtained. This implies some sort of stabilization by the filter-paper of the bluish green oxidation product; a possible explanation is the reducing properties of the filter-paper. Another property of the reaction on the filter-papers is a virtual absence of the effect of KIO4 concentration on the difference in absorbances for catalytic and non-catalytic reactions (Fig. 1). The signal is only slightly affected by periodate concentration in the range 1024–1023 mol l21. This permits the oxidant to be applied onto the filter-papers without strict control of the amount; sprinkling of the filter-papers with KIO4 was used.Sensitivity and Precision of Manganese Determination The absorbance of the TMB–KIO4 reaction products was found to be proportional to the logarithm of the manganese concentration for the reaction both in solution and on DETATA filterpapers. The metrological characteristics of the determination procedure are given in Table 1. In solution, the limit of determination (6 31025 mg l21) is close to that reported for the most sensitive reactions for manganese: viz., oxidation with periodate of N,N-diethylaniline [1 3 1025 (ref. 14) or 1 3 1024 mg l21 (ref. 15)], p-phenetidine [1 3 1024 mg l21 (ref. 16)] and o-dianisidine [2 3 1024 mg l21 (ref. 15)]. Preconcentration of MnII on DETATA filter-papers with the determination directly on the filter-paper makes it feasible not only to decrease the detection limit but also to expand the linear range for MnII from 1.5 orders (in solution) to over 2.5 orders of magnitude (on DETATA filter-papers) (Table 1).As regards the precision of the determination, it was thought that it would be fairly low on the filter-papers because of both the additional preconcentration operation and irregularities in the paper structure (hence, irregular colouring of the filter-papers). However, the RSD values for the reaction on the filter-papers are close to those in solution (Fig. 5), i.e., the precision of the determination remains fairly high. Interferences The criterion for interference was taken as a change of ±5% in the absorbance for 0.001 (reaction in solution) or 5 3 1025 (reaction on DETATA filter-papers) mg l21 of manganese in the analyzed aqueous solution.No interference results from the presence of a 1000-fold molar ratio of various ions (Table 2) or of 1 mg l21 humic and fulvic acids from river water. The selectivity for manganese in the reaction on DETATA filterpapers is higher than that in solution, and both procedures are no Table 1 Equations for the calibration graphs and linear ranges for the determination of MnII by oxidation of TMB with KIO4 in solution and on DETATA filter-papers (preconcentration of MnII from 20 ml of solution for DETATA filter-papers) Linear range/ Reaction a* sa b sb r cmin/mg l21 3 mg l21 In solution 0.455 0.037 0.104 0.024 0.987 1.531025 631025–231023 On DETATA filter-paper 0.385 0.015 0.077 0.006 0.996 2.531025 531026–2.531023 * For the equation y = a + bx, where x = log (cMn) and y = A3 (absorbance measured 3 min after the start of the reaction; blank absorbance was not subtracted).Table 2 Tolerance limits for foreign ions (cion : cMnII) in the determination of MnII by use of the catalytic reaction of TMB with KIO4 in solution and on filter-paper with chelating DETATA groups DETATA filter-paper (1 ng Mn; Solution (0.001 preconcentration Foreign ion mg l21 Mn) from 20 ml) FeII 5 50 ZnII 50 150 Cl2 500 500 FeII* 500 !1000 K, Na, Ca, Mg, Al, 700 700 FeIII, CuII, Br2, SO4 22, acetate !1000 !1000 * In the presence of 0.2 mg l21 KF.Table 3 Effect of FeII and fluoride on the absorbance of the products of the TMB–MnII–KIO4 reaction in solution 3 min after the start of the reaction (A3). [TMB] = 2.5 31024 mol l21; [KIO4] = 2.6 31025 mol l21; l = 650 nm; l = 0.5 cm Concentration of MnII/mg l21 KF/mg l21 FeII/mg l21 0.0002 0.002 0.02 0 0 0.056 0.204 0.223 0 0.28 0.053 0.086 0.097 0.2 0.28 0.058 0.206 0.220 1164 Analyst, October 1997, Vol. 122less (sometimes more) selective than those using the reactions of periodate oxidation of other amines.14–16 The only exception for the TMB–KIO4 reaction is FeII, which significantly interferes by decreasing the reaction rate (tolerance limit is 5 : 1 FeII :MnII for the reaction in solution): in p-phenetidine oxidation,16 a 100-fold amount of FeII was tolerated. At the same time, FeIII does not interfere in large amounts. The effect of FeII can be removed by adding 0.2 mg l21 potassium fluoride (Table 3).The mechanism of fluoride action is not clear, neither is the mechanism of FeII interference itself. Fluoride is not able to complex strongly with FeII ions; however, FeIII may be formed in situ, while fluoride can change the redox potentials of the pairs FeIII–FeII and MnIII–MnII simultaneously in such a manner that iron may no longer participate in the reaction. Rapid Determination of Manganese With Visual Detection One of the potential advantages of sorption–catalytic techniques is the feasibility of rapid determinations of analytes directly on the sorbents with no use of instrumental detection.The TMB– MnII–KIO4 reaction was conducted on DETATA filter-papers after preconcentration of manganese from 20 ml of aqueous solution, using the same procedure as for quantitative measurements (see under Reaction on Filter-papers). Instead of measuring the absorbance of the filter-paper after sprinkling it with periodate, it was dried with a stream of air (which required about 3 min) and the colour was observed visually.The colour remains stable in air for not less than 6 h. Various concentrations in the range from 1 3 1022 to 100 ng of manganese in 20 ml of solution were studied. It was found that confident discrimination of the colour intensities can be made for manganese concentrations which differ by not less than half an order of magnitude (i.e., 3 times).The determination is reliable for 0.1–10 ng of manganese (5 31026–5 31024 mg l21 for a pumped volume of 20 ml), which allows a colour scale to be constructed for the semiquantitative determination of manganese in this range. The whole procedure requires 6–7 min, starting with the pumping of the manganese solution through the DETATA filter-paper. Analysis of Tap and River Water For analysis, an aliquot of the sample (1.0 ml of tap water or 0.10 ml of river water preserved by adding sulfuric acid to pH 1.85 immediately after sampling) with 0.2 ml of KF (20 mg l21) added was diluted to 20 ml with borate buffer (pH 6.8).The analyses were performed as described under Reaction in solution and Reaction on Filter-paper. The results agreed with those obtained by spectrophotometry17 and/or flame atomic absorption spectrometry (Table 4). The high values obtained with the catalytic method in solution and by spectrophotometry might be due to the lower selectivity of these techniques.When manganese is preconcentrated on the DETATA sorbent, it is separated from interfering species and the results obtained agree with those obtained by another selective technique (atomic absorption). The authors thank Dr. G.I. Tsysin for providing the DETATA filter-papers and for fruitful discussions, Dr. N.M. Sorokina for AAS measurements, Dr. T.V. Polenova for the humic acid preparation, and the Russian Foundation for Basic Research for financial support (grant No. 96-03-08854). References 1 Dolmanova, I. F., and Peshkova, V. M., Vestn. Mosk. Gosud. Univ., Ser. 2: Khim., 1977, 18, 599. 2 Kolotyrkina, I. Ya., Shpigun, L. K., Zolotov, Yu. A., and Tsysin, G. I., Analyst, 1991, 116, 707. 3 Tikhonova, L. P., Bakay, E. A., Prokhorenko, E. P., Tarkovskaya, I. A., and Svarkovskaya, I. P., presented at the 5th International Symposium on Kinetics in Analytical Chemistry, September 25–28, 1995, Moscow, Russia; Abstracts of Papers, Nauka, Moscow, 1995, L24. 4 Varshal, G. M., Velyukhanova, T. K., Pavlutskaya, V. I., Starshinova, N. P., Formanovsky, A. A., Seregina, I. F., Shilnikov, A. M., Tsysin, G. I., and Zolotov, Yu. A., Int. J. Environ. Anal. Chem., 1994, 57, 107. 5 Naylor, F. J., and Saunders, B. C., J. Chem. Soc., 1950, 3519. 6 Hester, R. E., and Williams, K. P. J., J. Chem. Soc., Faraday Trans. II, 1981, 77, 541. 7 Makemoto, K., and Maysunaka, M., Bull. Chem. Soc. Jpn., 1968, 41, 764. 8 Zolotov, Yu.A., Zh. Anal. Khim., 1994, 49, 149. 9 Pöribil, R., Analytical Application of EDTA and Related Compounds, Mir, Moscow, 1975, p. 200 (in Russian). 10 Lurye, Yu., Handbook in Analytical Chemistry, Khimiya, Moscow, 1989 (in Russian). 11 Tsysin, G. I., Mikhura, I. V., Formanovsky, A. A., and Zolotov, Yu. A., Mikrochim. Acta, 1991, III, 53. 12 Saunders, B. C., and Watson, G. M. R., Biochem. J., 1950, 46, 629. 13 Denisov, E. T., Kinetics of Homogeneous Chemical Reactions, Vysshaya Shkola, Moscow, 1988, p. 57. 14 Nikolesis, D. P., Anal. Chem., 1978, 50, 205. 15 Dolmanova, I. F., and Yatsimirskaya, N. T., Zh. Anal. Khim., 1971, 26, 1540. 16 Gragorovich, F. G., Fresenius’ Z. Anal. Chem., 1974, 271, 5, 354. 17 Alimarin, I. P., Practical Recommendations on Physico-Chemical Methods for Analysis, Moskovskii Gosudarstvennyi Universitet, Moscow, 1987, p. 58 (in Russian). Paper 7/02595E Received April 16, 1997 Accepted June 30, 1997 Table 4 Concentrations of manganese in water (mg l21) found by using the TMB–KIO4 reaction and reference techniques.The RSDs were obtained from five parallel runs Catalytic method Reaction on Atomic DETATA absorption Sample Reaction in solution* filter-paper† spectrometry Spectrophotometry‡ Tap water (1.1 ± 0.3)31023 (0.7 ± 0.1)31023 (0.6 ± 0.1)31023 — River water (1.2 ± 0.1)31021 (0.9 ± 0.2)31021 (0.76 ± 0.03)31021 (1.4 ± 0.1)31021 * [TMB] = 2.5 3 1024 mol21; [KIO4] = 2.6 3 1025 mol l21; sample volume = 0.1–1 ml; [KF] = 0.2 mg l21; pH, 6.8; l = 650 nm, l = 0.5 cm.† The analyzed solution with added buffer (pH 6.8) and 0.2 mg l21 KF was pumped through the DETATA filter-paper and the reaction was carried out as described under Reaction on Filter-paper. ‡ Determined with formaldoxime.17 Analyst, October 1997, Vol. 122 1165 H2N Me Me Me NH2 Me N N OH N O OH O OH O OH O ( a) ( b) Sorption–Catalytic Determination of Manganese Directly on a Paper-based Chelating Sorbent M. K. Beklemishev, T.A. Stoyan and I. F. Dolmanova* Analytical Chemistry Division, Department of Chemistry, Moscow State University, Moscow 119899 GSP, Russia. E-mail: mkb@analyt.chem.msu.su A hybrid technique based on a catalytic reaction carried out on the surface of a paper-based sorbent is proposed. It is shown that MnII exhibits its catalytic action in the oxidation of 3,3A,5,5A-tetramethylbenzidine with periodate in aqueous solution as well as on filter-paper with or without attached diethylenetriaminetetraacetate (DETATA) groups.Optimum conditions differ for the reaction in solution and on filter-paper. For equal catalyst concentrations, higher initial reaction rates are attainable on the filter-papers. Preconcentration of manganese on the DETATA sorbent combined with the subsequent catalytic reaction improves the selectivity, reduces the limit of determination down to 5 3 1026 mg l21 (as compared with 6 3 1025 mg l21 in solution) and expands the linear range by an order of magnitude (to 5 3 1026–2.5 3 1023 mg l21).The precision of manganese determination on the sorbent is high (the RSDs are @5% for !6 3 1024 mg Mn). Samples of tap and river water were analyzed by use of the proposed sorption–catalytic technique. A rapid procedure for the determination of manganese with visual detection was also developed. Keywords: Catalytic kinetic method of analysis; chelating sorbent; diethylenetriaminetetraacetate; preconcentration of manganese Catalytic procedures are appreciated by analytical chemists for their sensitivity and simplicity in realization.Catalytic indicator reactions can be applied to the determination of a large number of compounds including the catalyst, inhibitors, activators, and compounds which convert the catalyst into an active state (oxidizers, ligands1), and catalytically inactive metal ions through the use of the competitive complexation principle. Catalytic methods become especially powerful when combined with analyte separation/preconcentration, which allows the selectivity to be increased and the detection limits to be lowered.One of the approaches involves adsorption of metal ions on chelating sorbents with subsequent desorption and catalytic determination.2 However, this type of procedure, while being convenient in flow systems,2 is tedious in other cases owing to the need for desorption of the analyte. Here, we propose to eliminate the desorption stage by conducting the catalytic indicator reaction directly on the sorbent used for preconcentration purposes. Little information has been published to date on the foregoing principle.3 This novel area requires studies of the peculiarities of catalytic indicator reactions as they occur on the surfaces of sorbents, including kinetics, concentration effects, and the metrological characteristics of the procedures. The main purpose of this study was the practical realization of the aforementioned sorption–catalytic principle.For the investigation, the preconcentration of manganese on diethylenetriaminetetraacetate (DETATA) paper-based chelating sorbent was chosen as it combines effectiveness of metal sorption and simplicity of operation.2,4 As an indicator reaction, the oxidation of 3,3A,5,5A-tetramethylbenzidine (TMB) by potassium periodate (KIO4) catalyzed by MnII was selected. Various amines other than TMB were previously studied in this reaction;5–7 however, TMB is more stable in air, less carcinogenic and its oxidation provides higher absorbing products.In the present work, the catalytic effect of MnII on the reaction between TMB and KIO4 in solution was studied. The aim was to employ this system to ascertain the analytical possibilities of the sorption–catalytic approach for the determination of manganese with both instrumental and visual detection. A potential benefit of the sorption–catalytic approach may be realized in ‘rapid-test’ techniques, such as spot-tests, field tests and similar procedures,8 in which a visible signal appears on the surface of paper strips or other supports.The development of rapid tests for manganese may be a subject for further work originating from this study. Experimental Reagents, Solutions and Apparatus TMB ‘for analysis’ was obtained from Riedel-de H�aen, Hannover, Germany; boric, hydrochloric and sulfuric acids (‘special purity’) from Reakhim (Moscow, Russia) were used.All other reagents were of analytical-reagent grade. Humic and fulvic acids were obtained from Moscow river water by adsorption on poly(styrene–divinylbenzene)sorbent (XAD) with subsequent desalination. All aqueous solutions were prepared using distilled water. Ethanolic solutions of TMB (0.025 mol l21) were prepared every 10 d; less concentrated solutions were obtained by dilution with ethanol when necessary. A stock solution of periodate in water (1 g l21 KIO4) was diluted to an appropriate concentration (in most cases 0.1 g l21) every 4–5 d.A stock solution of anhydrous MnSO4 (1 g l21 Mn) was standardized by titration.9 Solutions with lower contents of the metal were prepared every 3 d (10 and 1 mg l21) or daily (0.01 mg l21) by dilution of the stock solution. All the manganese solutions were acidified with H2SO4 to pH 1.8–1.9; it was found that such acidification was necessary in order to obtain reproducible results.Hydrochloric acid was used to provide pH values ranging from 2 to 4; it was purified by isothermal distillation of the concentrated acid (reagent grade). Borate buffer of pH 6.8 was prepared by dropwise addition of Scheme 1 (a) TMB; (b) DETATA groups attached to filter paper. Analyst, October 1997, Vol. 122 (1161–1165) 11610.2 mol l21 KOH to a solution of 3.3 g of boric acid in 1000 ml of water until the desired pH value was reached. Buffers for the pH range 4–6 (0.1 mol l21 in sodium) were prepared from acetic acid and sodium acetate, and buffers for pH values of 7–11 (0.1 mol l21 in borate) were prepared from boric acid, Na2- B4O7·10H2O and KOH, by using standard procedures.10 The other buffer solutions with pH > 4 were purified by pumping through two consecutive DETATA filters placed in a special polyethylene holder (‘Biospectr’, St.Petersburg, Russia) at a rate of 8–10 ml min21, which removed transition metal impurities. The chelating sorbent was synthesized as described previously by chemical attachment of aminocarboxylic (diethylenetriaminetetraacetate, DETATA) groups to cellulose filterpaper. 11 The filter-paper was 2.5 cm in diameter.Water was distilled in a commercial apparatus (D3-4-2M, Ioshkar-Ola, Russia). A KFK-3 spectrophotometer (ZOMZ, Zagorsk, Russia) was used for measurements of absorbance. To spray the filter-papers with periodate solution, a hand-operated sprinkler for thin-layer chromatograms was used.Reaction in Solution Preliminary studies of various mixing orders of the components of the indicator reaction showed that maximum absorbance was achieved when MnII and borate buffer (pH 6.8) were mixed first, followed by addition of TMB and finally of KIO4. Based on this, the following procedure was used. In a 15 ml glass test-tube were placed the reagents in the following sequence (the amounts are given for the final procedure for manganese determination): 3.9 ml of the buffer (pH 6.8), 0.5 ml of the solution, containing MnII at a pH of 1.8–1.9, 0.1 ml of a 2.5 3 1024 mol l21 ethanolic solution of TMB and 0.5 ml of 4.3 3 1022 mol l21 periodate.The addition of periodate was taken as the time of the start of the reaction. After agitation, the reaction mixture was transferred into a 0.5 cm cell and the absorbance of the oxidation product was measured at 650 nm against water. The absorbance value at 3 min (A3) was taken as the analytical signal.The experiments were performed at ambient temperature, which was kept at 24 ± 1 °C. Reaction on Filter-paper In preliminary experiments the optimum mixing order of the reactants was sought with the use of visual detection. Coloration of the filter-paper was most intense when TMB and the buffered manganese solution were mixed first and periodate was added at the last step. The role of the humidity of the filter-papers was also studied (the filter-papers were stored for up to 3 d in a desiccator with controlled 75–95% humidity); no effect of humidity on the reaction rate was observed.The final procedure for the reaction on filter-paper (with or without attached DETATA groups) is given below. MnII was first applied onto a filter-paper either (a) in the form of a 0.02 ml aliquot of a solution (0.4 mg l21 Mn) added to the centre of the filter-paper by an Eppendorf pipette, or (b) by pumping a buffered manganese solution of pH 6.8 (5 3 1025–0.05 mg l21 MnII, 0.08 mol l21 in borate) through the filter-paper by use of a peristaltic pump at a rate of 8 ml min21.The filter-paper was then dried with a gentle flow of pressurized air until no traces of moisture could be detected, after which TMB was pipetted onto the filter-paper (0.02 ml of a 4 3 1024 mol l21 solution) followed by an identical drying procedure. The oxidant was added by sprinkling the filter-paper with a 4.3 3 1024 mol l21 solution of KIO4, which was taken to be the beginning of the reaction.The coloured reaction zone was about 20 mm in diameter. Sprinkling rather than pipetting of periodate was necessary in order to obtain uniform coloration of the filterpapers; the sprinkling technique was applicable owing to a weak dependence of the analytical signal on the amount of KIO4 on the filter-paper, as shown below. The absorbance of a filter-paper was measured in its wet state against the same filter-paper without added periodate (i.e., not coloured with the reaction products) using the following procedure.On addition of TMB solution, but before drying, the filter-paper was placed between two glass plates fixed together with a clip and then placed in the cell compartment of the spectrophotometer perpendicular to the light beam. The absorbance of this filter-paper was taken as zero. The filter-paper was then dried, sprayed with KIO4 as described above, and the wet specimen was again placed between the glass plates for the measurements.The absorbance at 2 min (A2) was taken as the analytical signal. Calculation of Metrological Characteristics The limit of determination was defined as the concentration of manganese for which the RSD did not exceed 33%; this concentration was taken as the lower limit of the linear range. The limit of detection (cmin) was calculated as 3a/b; for the logarithmic plot y = a + bx, where y is absorbance and x = log(cMn), there are some complications.The following procedure was used: (1) shift of the graph y = a + bx to the origin (x = y = 0), i.e., transformation of the graph to the form Y = A + bX where A = 0 and b is the previous value. This required a recalculation: X = x + (a/b) [for example, X = log(cMn) + 4.82 for the reaction on DETATA filters]; (2) calculation of the regression parameters for the new graph Y = A + bX, including sA; (3) calculation of log(cmin) as 3sA/b. Results and Discussion Reaction Products At least two differe coloured products are observed in the course of the studied reaction catalyzed by MnII in solution.A bluish green product (absorption maxima 370 and 650 nm) is formed under conditions where there is a deficiency of the oxidant (less than 1 3 1024 mol l21 KIO4 for a TMB concentration of 2.5 3 1024 mol l21, Fig. 1). According to the literature,12 this species may be a dimeric dication (lmax = 380 Fig. 1 Absorbance of the indicator reaction products as a function of KIO4 concentration.Curves 1 and 2, reaction in solution; curves 3 and 4, reaction on filter-paper with attached DETATA groups (KIO4 applied by sprinkling the solution onto the filter-paper). For the reaction in solution (curve 1): 2.5 3 1024 mol l21 TMB, 1 3 1023 mg l21 MnII; the measurements in solution (A3) were made at 650 nm for cKIO4 < 5 3 1025 mol l21 (bluish green product) or at 460 nm for cKIO4 > 5 3 1025 mol l21 (orange product). For the reaction on DETATA filter-paper (curve 3): 8 3 1028 mol TMB, 8 3 1023 mg l21 MnII; all the measurements of the filter-paper absorbances (A2) were made at 650 nm (bluish green product).For curves 2 and 4, the conditions are the same as for curves 1 and 3, respectively, but with no MnII. 1162 Analyst, October 1997, Vol. 122and 650 nm), probably in a mixture with a meriquinone (lmax = 655 nm). An increase in periodate concentration over 1 3 1023 mol l21 results in an orange product (lmax = 465 nm), which is probably the result of a more extensive oxidation and may be ascribed a quinonediimine structure12 (lmax = 465 nm).Intermediate concentrations of the oxidant provide a brown mixture with absorbance maxima at 380, 650 and 465 nm. As can be observed from Fig. 1, the maximum difference in the rates of MnII-catalyzed and non-catalytic reactions in solution corresponds to the formation of the bluish green product, whereas for higher periodate concentrations (and, consequently, for orange product formation) the difference decreases.For the determination of manganese, the bluish green product (370, 650 nm) was used. Kinetic Curves Kinetic curves for the TMB–MnII–periodate reaction are depicted in Fig. 2. In order to compare the data for the homogeneous and heterogeneous variants of the reaction, the catalyst concentrations should be presented in units which are common to both the solution and the filter-paper. For instance, the mass of manganese per square of the cross-section of the spectrophotometer light beam (mg cm22) would be a common value for both the sorbent and a cell with the solution.If the kinetic curves for the same concentration of manganese (in mg cm22, Fig. 2) are compared, it can be seen that the slope of the ascending portion of the curve is markedly higher for the reaction on filter-paper (with or without DETATA groups) than for that in solution. The optimum conditions for the reaction in solution and on the filter-papers differ, so it can only be noted that the highest attainable initial rate on the filter-papers is higher than that in solution.The absorbance of the filter-paper specimen (which corresponds to the formation of the bluish green product) increases rapidly for the first 0.5 min (counting from the start of the reaction), after which a slow decrease in absorption follows. In solution this decrease leads to the formation of a scarcelycoloured final product which requires a few hours.On the dry filter-paper the bluish green product is more stable (the coloration is not diminished during 24 h). The slow increase in the absorption of filter-paper specimens at 6–10 min is likely to be caused by processes in the cellulose filter-paper itself, e.g., swelling (wet filters without the reagents also exhibit an Fig. 2 Kinetic curves for the KIO4–MnII–TMB reaction at pH 6.8 in solution (1, 4), on filter-paper with attached DETATA groups (2, 5) and on filter-paper (3, 6).Curve 1, 2.5 3 1024 mol l21 TMB, 2.6 3 1025 mol l21 KIO4, 2.6 ng cm22 MnII; 2, 8 3 1028 mol TMB, 4.3 3 1024 mol l21 KIO4, 2.6 ng cm22 MnII. For curve 3, Mn solution (20 ml) was applied by pipetting. For curve 2, Mn solution (20 ml) was pumped through the DETATA filterpaper. The solution of KIO4 was applied onto the filter-paper (curves 2, 3, 5, 6) by sprinkling. For curves 4, 5 and 6, the conditions are the same as for curves 1, 2 and 3, respectively, but without MnII.The measurements were made at 650 nm. Fig. 3 Absorbance of the reaction products as a function of pH in solution (3, 4) (A3) and on filter-paper with attached DETATA groups (1, 2) (A2). Curve 1, 2.5 3 1024 mol l21 TMB, 2.6 3 1025 mol l21 KIO4, 1 ng ml21 MnII; 3, 8 3 1028 mol TMB, 4.3 3 1024 mol l21 KIO4, 8 ng ml21 MnII. For curves 2 and 4, the conditions are the same as for curves 1 and 3, respectively, but without MnII.The measurements were made at 650 nm. To study the effect of pH, manganese solution (0.04 mg l21) of the appropriate pH value was pumped through the DETATA filter-paper and the reaction was carried out as described under Reaction on Filter-paper. Fig. 4 Absorbance of the reaction products as a function of amount of TMB in solution (2, 4) (A3) and on filter-paper with attached DETATA groups (1, 3) (A2). Curve 1, 4.3 3 1024 mol l21 KIO4, 8 ng ml21 MnII; 2, 2.6 31025 mol l21 KIO4, 1 ng ml21 MnII.For curves 3 and 4, the conditions are the same as for curves 1 and 2, respectively, but without MnII. The measurements were made at 650 nm and at pH 6.8. Fig. 5 Relative standard deviations of absorbance of the reaction products as a function of manganese concentration for the reaction carried out in solution (1) and on filter-paper with attached DETATA groups (2). Curve 1, MnII amount denotes cMn/mg cm22 in solution (reaction conditions: 2.5 3 1024 mol l21 TMB, 2.6 3 1025 mol l21 KIO4); curve 2, MnII amounts were calculated as cMn30.02, where cMn is the MnII concentration in the solution that was pumped through the DETATA filter-paper and 0.02 l is the solution volume (conditions for the reaction on the filter-paper: 8 3 1028 mol TMB, 4.3 3 1024 mol l21 KIO4).The measurements were made at 650 nm and at pH 6.5. Analyst, October 1997, Vol. 122 1163increase in absorption). If the latter is not taken into account, the shape of the kinetic curves (Fig. 1) may be tentatively explained13 by reversible sequential–parallel reactions such as TMB"Product I (370, 650 nm)"Product II (colourless) or a set of two reversible reactions TMB"Product I (370, 650 nm) TMB"Product II (colourless) If one of these schemes is true, the steady-state portion of the absorbance–time plot will correspond to an equilibrium of the bluish green product with a colourless product. Effect of pH and Reagent Concentrations It was interesting to study whether the influence of reagent concentrations (TMB, KIO4) and pH on the signal differs for the reaction on filter-paper as opposed to in solution.As shown in Figs. 3–5, considerable differences exist. As can be seen from the pH curve (Fig. 3), the maximum amount of the reaction products is formed at pH 3.1 and 6.8 both in solution and on the DETATA filter-papers, but only on the sorbent is there a catalytic effect of manganese at pH 3.1. The signal on the DETATA filter-papers is higher at pH 3.1 than at pH 6.8 but the precision is poorer, viz., RSDs of 8 and 2%, respectively, are obtained (for 8 31028 mol TMB, 4.3 31024 mol l21 KIO4 and 8 ng ml21 MnII pumped through the DETATA filter-papers). One reason for the low reproducibility at pH 3.1 may be incomplete sorption of MnII at this pH value.Subsequent reactions, both in solution and on the filter-papers, were carried out at pH 6.8. A study of the effect of the periodate concentration on the reaction on the filter-papers showed that only the bluish green product is formed even at high concentrations of the oxidant.In solution (Fig. 1), the orange product was found at KIO4 :TMB ratios !1 : 1, whereas on the filter-papers it was never obtained. This implies some sort of stabilization by the filter-paper of the bluish green oxidation product; a possible explanation is the reducing properties of the filter-paper. Another property of the reaction on the filter-papers is a virtual absence of the effect of KIO4 concentration on the difference in absorbances for catalytic and non-catalytic reactions (Fig. 1). The signal is only slightly affected by periodate concentration in the range 1024–1023 mol l21. This permits the oxidant to be applied onto the filter-papers without strict control of the amount; sprinkling of the filter-papers with KIO4 was used. Sensitivity and Precision of Manganese Determination The absorbance of the TMB–KIO4 reaction products was found to be proportional to the logarithm of the manganese concentration for the reaction both in solution and on DETATA filterpapers.The metrological characteristics of the determination procedure are given in Table 1. In solution, the limit of determination (6 31025 mg l21) is close to that reported for the most sensitive reactions for manganese: viz., oxidation with periodate of N,N-diethylaniline [1 3 1025 (ref. 14) or 1 3 1024 mg l21 (ref. 15)], p-phenetidine [1 3 1024 mg l21 (ref. 16)] and o-dianisidine [2 3 1024 mg l21 (ref. 15)]. Preconcentration of MnII on DETATA filter-papers with the determination directly on the filter-paper makes it feasible not only to decrease the detection limit but also to expand the linear range for MnII from 1.5 orders (in solution) to over 2.5 orders of magnitude (on DETATA filter-papers) (Table 1). As regards the precision of the determination, it was thought that it would be fairly low on the filter-papers because of both the additional preconcentration operation and irregularities in the paper structure (hence, irregular colouring of the filter-papers). However, the RSD values for the reaction on the filter-papers are close to those in solution (Fig. 5), i.e., the precision of the determination remains fairly high. Interferences The criterion for interference was taken as a change of ±5% in the absorbance for 0.001 (reaction in solution) or 5 3 1025 (reaction on DETATA filter-papers) mg l21 of manganese in the analyzed aqueous solution.No interference results from the presence of a 1000-fold molar ratio of various ions (Table 2) or of 1 mg l21 humic and fulvic acids from river water. The selectivity for manganese in the reaction on DETATA filterpapers is higher than that in solution, and both procedures are no Table 1 Equations for the calibration graphs and linear ranges for the determination of MnII by oxidation of TMB with KIO4 in solution and on DETATA filter-papers (preconcentration of MnII from 20 ml of solution for DETATA filter-papers) Linear range/ Reaction a* sa b sb r cmin/mg l21 3 mg l21 In solution 0.455 0.037 0.104 0.024 0.987 1.531025 631025–231023 On DETATA filter-paper 0.385 0.015 0.077 0.006 0.996 2.531025 531026–2.531023 * For the equation y = a + bx, where x = log (cMn) and y = A3 (absorbance measured 3 min after the start of the reaction; blank absorbance was not subtracted).Table 2 Tolerance limits for foreign ions (cion : cMnII) in the determination of MnII by use of the catalytic reaction of TMB with KIO4 in solution and on filter-paper with chelating DETATA groups DETATA filter-paper (1 ng Mn; Solution (0.001 preconcentration Foreign ion mg l21 Mn) from 20 ml) FeII 5 50 ZnII 50 150 Cl2 500 500 FeII* 500 !1000 K, Na, Ca, Mg, Al, 700 700 FeIII, CuII, Br2, SO4 22, acetate !1000 !1000 * In the presence of 0.2 mg l21 KF.Table 3 Effect of FeII and fluoride on the absorbance of the products of the TMB–MnII–KIO4 reaction in solution 3 min after the start of the reaction (A3).[TMB] = 2.5 31024 mol l21; [KIO4] = 2.6 31025 mol l21; l = 650 nm; l = 0.5 cm Concentration of MnII/mg l21 KF/mg l21 FeII/mg l21 0.0002 0.002 0.02 0 0 0.056 0.204 0.223 0 0.28 0.053 0.086 0.097 0.2 0.28 0.058 0.206 0.220 1164 Analyst, October 1997, Vol. 122less (sometimes more) selective than those using the reactions of periodate oxidation of other amines.14–16 The only exception for the TMB–KIO4 reaction is FeII, which significantly interferes by decreasing the reaction rate (tolerance limit is 5 : 1 FeII :MnII for the reaction in solution): in p-phenetidine oxidation,16 a 100-fold amount of FeII was tolerated.At the same time, FeIII does not interfere in large amounts. The effect of FeII can be removed by adding 0.2 mg l21 potassium fluoride (Table 3). The mechanism of fluoride action is not clear, neither is the mechanism of FeII interference itself.Fluoride is not able to complex strongly with FeII ions; however, FeIII may be formed in situ, while fluoride can change the redox potentials of the pairs FeIII–FeII and MnIII–MnII simultaneously in such a manner that iron may no longer participate in the reaction. Rapid Determination of Manganese With Visual Detection One of the potential advantages of sorption–catalytic techniques is the feasibility of rapid determinations of analytes directly on the sorbents with no use of instrumental detection.The TMB– MnII–KIO4 reaction was conducted on DETATA filter-papers after preconcentration of manganese from 20 ml of aqueous solution, using the same procedure as for quantitative measurements (see under Reaction on Filter-papers). Instead of measuring the absorbance of the filter-paper after sprinkling it with periodate, it was dried with a stream of air (which required about 3 min) and the colour was observed visually.The colour remains stable in air for not less than 6 h. Various concentrations in the range from 1 3 1022 to 100 ng of manganese in 20 ml of solution were studied. It was found that confident discrimination of the colour intensities can be made for manganese concentrations which differ by not less than half an order of magnitude (i.e., 3 times). The determination is reliable for 0.1–10 ng of manganese (5 31026–5 31024 mg l21 for a pumped volume of 20 ml), which allows a colour scale to be constructed for the semiquantitative determination of manganese in this range.The whole procedure requires 6–7 min, starting with the pumping of the manganese solution through the DETATA filter-paper. Analysis of Tap and River Water For analysis, an aliquot of the sample (1.0 ml of tap water or 0.10 ml of river water preserved by adding sulfuric acid to pH 1.85 immediately after sampling) with 0.2 ml of KF (20 mg l21) added was diluted to 20 ml with borate buffer (pH 6.8).The analyses were performed as described under Reaction in solution and Reaction on Filter-paper. The results agreed with those obtained by spectrophotometry17 and/or flame atomic absorption spectrometry (Table 4). The high values obtained with the catalytic method in solution and by spectrophotometry might be due to the lower selectivity of these techniques. When manganese is preconcentrated on the DETATA sorbent, it is separated from interfering species and the results obtained agree with those obtained by another selective technique (atomic absorption). The authors thank Dr.G.I. Tsysin for providing the DETATA filter-papers and for fruitful discussions, Dr. N.M. Sorokina for AAS measurements, Dr. T.V. Polenova for the humic acid preparation, and the Russian Foundation for Basic Research for financial support (grant No. 96-03-08854). References 1 Dolmanova, I. F., and Peshkova, V. M., Vestn. Mosk. Gosud. Univ., Ser. 2: Khim., 1977, 18, 599. 2 Kolotyrkina, I. Ya., Shpigun, L. K., Zolotov, Yu. A., and Tsysin, G. I., Analyst, 1991, 116, 707. 3 Tikhonova, L. P., Bakay, E. A., Prokhorenko, E. P., Tarkovskaya, I. A., and Svarkovskaya, I. P., presented at the 5th International Symposium on Kinetics in Analytical Chemistry, September 25–28, 1995, Moscow, Russia; Abstracts of Papers, Nauka, Moscow, 1995, L24. 4 Varshal, G. M., Velyukhanova, T. K., Pavlutskaya, V. I., Starshinova, N. P., Formanovsky, A. A., Seregina, I. F., Shilnikov, A. M., Tsysin, G. I., and Zolotov, Yu. A., Int. J. Environ. Anal. Chem., 1994, 57, 107. 5 Naylor, F. J., and Saunders, B. C., J. Chem. Soc., 1950, 3519. 6 Hester, R. E., and Williams, K. P. J., J. Chem. Soc., Faraday Trans. II, 1981, 77, 541. 7 Makemoto, K., and Maysunaka, M., Bull. Chem. Soc. Jpn., 1968, 41, 764. 8 Zolotov, Yu. A., Zh. Anal. Khim., 1994, 49, 149. 9 Pöribil, R., Analytical Application of EDTA and Related Compounds, Mir, Moscow, 1975, p. 200 (in Russian). 10 Lurye, Yu., Handbook in Analytical Chemistry, Khimiya, Moscow, 1989 (in Russian). 11 Tsysin, G. I., Mikhura, I. V., Formanovsky, A. A., and Zolotov, Yu. A., Mikrochim. Acta, 1991, III, 53. 12 Saunders, B. C., and Watson, G. M. R., Biochem. J., 1950, 46, 629. 13 Denisov, E. T., Kinetics of Homogeneous Chemical Reactions, Vysshaya Shkola, Moscow, 1988, p. 57. 14 Nikolesis, D. P., Anal. Chem., 1978, 50, 205. 15 Dolmanova, I. F., and Yatsimirskaya, N. T., Zh. Anal. Khim., 1971, 26, 1540. 16 Gragorovich, F. G., Fresenius’ Z. Anal. Chem., 1974, 271, 5, 354. 17 Alimarin, I. P., Practical Recommendations on Physico-Chemical Methods for Analysis, Moskovskii Gosudarstvennyi Universitet, Moscow, 1987, p. 58 (in Russian). Paper 7/02595E Received April 16, 1997 Accepted June 30, 1997 Table 4 Concentrations of manganese in water (mg l21) found by using the TMB–KIO4 reaction and reference techniques. The RSDs were obtained from five parallel runs Catalytic method Reaction on Atomic DETATA absorption Sample Reaction in solution* filter-paper† spectrometry Spectrophotometry‡ Tap water (1.1 ± 0.3)31023 (0.7 ± 0.1)31023 (0.6 ± 0.1)31023 — River water (1.2 ± 0.1)31021 (0.9 ± 0.2)31021 (0.76 ± 0.03)31021 (1.4 ± 0.1)31021 * [TMB] = 2.5 3 1024 mol21; [KIO4] = 2.6 3 1025 mol l21; sample volume = 0.1–1 ml; [KF] = 0.2 mg l21; pH, 6.8; l = 650 nm, l = 0.5 cm. † The analyzed solution with added buffer (pH 6.8) and 0.2 mg l21 KF was pumped through the DETATA filter-paper and the reaction was carried out as described under Reaction on Filter-paper. ‡ Determined with formaldoxime.17 Analyst, October 1997, Vol. 122 1165
ISSN:0003-2654
DOI:10.1039/a702595e
出版商:RSC
年代:1997
数据来源: RSC
|
30. |
Application of Gas–Liquid Chromatography to the Analysis of Essential OilsPart XVII.† Fingerprinting of Essential Oils by Temperature-programmed Gas–Liquid Chromatography Using Capillary Columns With Non-polar Stationary Phases |
|
Analyst,
Volume 122,
Issue 10,
1997,
Page 1167-1174
Analytical Methods Committee,
Preview
|
|
摘要:
Application of Gas–Liquid Chromatography to the Analysis of Essential Oils Part XVII.† Fingerprinting of Essential Oils by Temperature-programmed Gas–Liquid Chromatography Using Capillary Columns With Non-polar Stationary Phases Analytical Methods Committee‡ The Royal Society of Chemistry, Burlington House, Piccadilly, London, UK W1V 0BN Problems in obtaining reproducible results when ‘fingerprinting’ essential oils by temperature-programmed gas–liquid chromatography have been reported on in Parts VII and VIII of this series.Those reports were concerned with the general problems and the use of packed columns. This report is concerned with the use of capillary columns and non-polar stationary phases. A collaborative study using capillary columns with non-polar stationary phases has resulted in a method which specifies the ‘g-pack value’ of a column and gives reproducible relative retention indices for the test compounds limonene, acetophenone, linalol, naphthalene, linalyl acetate and cinnamyl alcohol. The method has been applied successfully to the examination of oil of rosemary.A recommended method is given for the reproducible temperature-programmed gas–liquid chromatographic fingerprinting of essential oils using capillary columns with non-polar stationary phases. Keywords: Essential oil analysis; gas–liquid chromatography; fingerprinting; non-polar capillary column The Analytical Methods Committee (AMC) has received and has approved for publication the following report from its Essential Oils Sub-Committee.Report The constitution of the Sub-Committee responsible for the preparation of this report was: Mr. M. J. Milchard (Chairman), Mr. A. M. Humphrey (Chairman to October 1993), Mr. N. Boley, Mr. B. Conway, Mr. R. Esdale, Ms. M. Flowerdew, Miss D. M. Michalkiewicz, Mr. D. A. Moyler, Mr. A. Osbiston, Mr. D. Powis, Mr. A. Sherlock, Mr. R. Smith, Mr. S. Smith, Mr. B. Starr and Mr. T. M. Stevens, with Mr.J. J. Wilson as Secretary. The Sub-Committee would like to thank the following manufacturers of capillary chromatography columns for the interest which they have shown in this work and their willingness to participate in the collaborative trial: Alltech Associates Inc., Chrompack UK Ltd., Hewlett-Packard Ltd., J&W Scientific/Jones Chromatography, Quadrex Corporation, Restek/Thames Chromatography and SGE (UK) Ltd. Their assistance is gratefully acknowledged.Introduction The development of gas–liquid chromatography (GLC) in the early 1950s began a new era in the analysis of essential oils. Until that time the qualities, purities and origins of oils were assessed by physical measurements and chemical assays. This new method gave improved results over the older methods, many of which we now know gave precise but inaccurate results. The technique was soon applied to the accurate determination of major and other components of interest.2–11 However, the concept of ‘fingerprinting’ oils had not been addressed successfully.The chemical nature of essential oils makes them particularly suitable for analysis by GLC. If temperature-programmed operation is used, a very high proportion of the total number of components present can be resolved. Many attempts have been made to establish libraries of chromatograms from temperatureprogrammed GLC analysis. This would allow sample and reference tracings to be compared and the authenticity and quality of the sample to be determined. However, it was quickly found that the conditions of the GLC analysis had to be strictly controlled, but even then reproducibility was poor, particularly among different laboratories.It was apparent that the temperature programming and the nature of the column itself caused the greatest variation in the results. This has led to individual libraries being established, which makes comparison difficult. The Sub-Committee has been studying the fingerprinting of essential oils over many years.Initial work on packed columns established that the lack of reproducibility of results was due to the lack of reproducibility of the columns themselves. This in turn was attributed to problems with coating the support and the subsequent ageing of the packing with use. There was, therefore, a requirement for the standardisation of the column efficiency and of its selectivity without using one as a factor of the other.A publication by van den Dool12 described a method for the characterisation of GLC columns using the relative retentionindices (RRIs) of a group of six test compounds. By calculating the RRIs of the compounds in the mixture on a particular column and then applying a series of further calculations to these RRIs, van den Dool obtained a figure representing the polarity factor of that particular column which he called the g- † For Part XVI, see ref. 1. ‡ Correspondence should be addressed to the Secretary, Analytical Methods Committee, Analytical Division, The Royal Society of Chemistry, Burlington House, Piccadilly, London, UK W1V 0BN. Analyst, October 1997, Vol. 122 (1167–1174) 1167pack value. This value will vary with the different types of stationary phase and their condition. Van den Dool’s work was based exclusively on packed columns and concentrated on two particular stationary phases: SE-30 as an example of a non-polar type and Carbowax 20M as a moderately polar type.However, the g-pack concept can be applied to any stationary phase. The mixture of compounds used in the determination, known as the NC (Netherlands Committee) mixture, was chosen to represent a range of compounds with functional groups similar to those found in essential oils. Also, the mixture was chosen so that there are two pairs of compounds in which the components of each pair elute close together on the two different phases, which can be used to give a measure of the resolving power of the column. However, the greatly increased efficiency of capillary columns over packed columns means that this property of the mixture is not so significant except in cases of extreme column degradation.The NC mixture consists of limonene, linalol, linalyl acetate, acetophenone, naphthalene and cinnamyl alcohol. The Sub-Committee has published several papers, based on the work of van den Dool, defining the methodology to be used in obtaining these fingerprints13–15 and standard fingerprint traces of selected oils.1,16,17 During these studies on the application of the g-pack concept it was found that the value for a packed column was unaffected by changes in operating parameters such as carrier gas flow rate, temperature programming rate, initial temperature and final temperature hold.However, the value could be decreased by loss of stationary phase over a period of time, due to column bleeding, and could be increased by modification of the stationary phase due to oxidation.The g-pack value can, therefore, give a good indication of the condition of the column. With the wider availability and advances in capillary column technology the Sub-Committee decided that the technique should be updated to make use of this technology and set about determining the optimum operating parameters. It was appreciated that the application to capillary columns was not likely to be as straightforward as with packed columns.A major difference was in the control of the stationary phase. For packed columns, good agreement among laboratories was only obtained when there was control over the preparation of the stationary phase and packing of the column. This involved developing a method for coating the stationary phase onto the support (the absorption coating technique), which led to columns giving good reproducibility and high efficiencies. While it is possible to prepare and coat capillary columns, in reality laboratories buy in their columns from one of the specialist manufacturers.This means that individual laboratories have no control over the preparation of the columns and that different methods of preparation between manufacturers could lead to slightly different performances of nominally the same stationary phase. It was concluded, therefore, that any method for fingerprinting essential oils on capillary columns would have to be based on commercially available columns and be robust enough to cope with a range of operating parameters.If the method is to be widely applicable it is unrealistic to expect laboratories to buy columns from a specific manufacturer or to change carrier gas. Experimental It was decided that the initial examination would be carried out on a specified stationary phase on columns which the members of the committee had available in their laboratories. A non-polar phase was chosen equivalent to SE-30.A protocol was provided specifying the samples to be examined and the initial column temperature, temperature programme rate and final column temperature. The choice of carrier gas was left to individuals with the proviso that the flow rate was optimised for the column configuration. The three samples examined were: a mixture of the NC mix and a series of even numbered carbon aliphatic hydrocarbons (C8–C24) known as the NC–HC mixture; Spanish rosemary oil; and Spanish rosemary oil plus the hydrocarbon mixture.The initial results showed that there were differences between different column manufacturers but very similar results were obtained between laboratories using similar columns from the same manufacturer. It was, therefore, decided to approach the major capillary column manufacturers and invite them to examine the same samples on their columns. All of the manufacturers were very willing to collaborate in this examination with the given protocol.With their inclusion a total of 16 results were obtained. Results The results of the collaborative study using the NC mixture with the proposed procedure are given in Table 1 in ascending order of g-pack values and are summarised in Table 2. It can be seen from Table 2 that the mean RRIs for the test compounds obtained in this examination compare well with results previously obtained on packed columns. Table 3 shows the elution temperatures of the test compounds and n-alkane hydrocarbons.There is a much wider variation in elution temperatures with capillary columns than was found with packed columns. The results from the examination of oil of rosemary are given in Tables 4 and 5. The results obtained were considered satisfactory by the Sub-Committee. Discussion Throughout all of the collaborative exercises, the Sub-Committee was conscious of the fact that many of the parameters examined were chosen with arbitrary limits and that some justification for the choices should be made.In particular, this applies to the nature of the test compounds in the NC mixture and the use of the RRI system. This system is widely used and most gas chromatographers are acquainted with it. It is a simple system and, although it is relative rather than absolute, it allows for a choice of reference compounds which can be made according to other requirements. For the purpose of this and the previous investigation it was felt that the homologous series of n-alkanes was the most suitable as reference compounds.They are readily available in pure form, are extremely stable and are the least likely compounds to exhibit chromatographic anomalies. In addition, they formed the basis of the work by van den Dool. The elution of a variety of homologous series under temperature-programmed conditions is not linear and the departure from linearity increases as the programming rate decreases.18 Also, this departure is greater for homologous series of polar compounds than for the n-alkanes and this was another reason for the choice of the latter.This non-linearity of elution raises the question of the method of application of the system and if an assumption of linearity between successive n-alkanes introduces errors. Van den Dool recommends the use of both even and odd numbered carbon nalkanes to reduce any error, whereas all of our collaborative work has been carried out with the even numbered carbon nalkanes making the assumption of linearity between each pair.Comparisons have been made of the effect on the calculated RRIs of the six NC text compounds when using all of the n- 1168 Analyst, October 1997, Vol. 122alkanes against using only the even numbered carbon ones and also by calculating on the assumption of linearity as well as by a graphical method to obtain more accurate figures. The differences that were found were so small (in some instances zero) that they were considered to be insignificant.Therefore, our recommended procedure uses only the even numbered carbon n-alkanes and assumes linearity between them. The six compounds in the NC mixture were selected by van den Dool on the basis of their similarity to the types of compounds found in essential oil analyses. Their choice could perhaps be criticised if used in conjunction with other sample types. However, the NC mixture is used for the calibration of the column irrespective of its ultimate use.As the mathematical treatment of the results has already been worked out by van den Dool, there seemed to be no advantage in changing the mixture. The choice of six compounds is a reasonable compromise between having an excessive number of interfering peaks and mathematical calculations and a reduced amount of data leading to a less accurate result. With packed columns, more control was exercised over the operating conditions by arranging for the C24 alkane to elute at the upper temperature of the temperature programme run by adjustment of the carrier gas flow rate.This approach is not relevant for capillary columns as, to obtain satisfactory results, it is necessary to optimise the flow rate, which depends on the dimensions of the individual columns and the nature of the carrier gas. The superior resolving power of capillary columns led to the conclusion that measurement of the resolution between the limonene and acetophenone peaks would not be meaningful as there was baseline separation in all cases.The column would be in poor condition or the operating conditions be significantly different from those recommended for these compounds not to be resolved. Developments in gas chromatographic instrumentation have led to discussions on the relative merits of operating the carrier gas system under conditions of constant flow or constant pressure. The NC–HC mixture was chromatographed on a column in an instrument equipped to run under both conditions.As shown below, the RRIs were virtually identical under both conditions. However, the chromatogram of the sample run under constant pressure conditions showed an increase in the retention times of the components compared with those run under constant flow conditions. This is because of a reduction in flow rate during temperature-programmed operation under constant pressure. RRI Compound Constant pressure Constant flow Limonene 1020 1020 Acetophenone 1034 1034 Linalol 1084 1084 Naphthalene 1158 1156 Linalyl acetate 1242 1241 Cinnamyl alcohol 1274 1273 g-pack 0.999 This demonstrates that either method of carrier gas control should be suitable for undertaking fingerprint determinations using this procedure.The RRI and area % composition of the 17 most abundant components in the sample of oil of rosemary were calculated from chromatograms run under the same conditions as for the HC–NC mixture.From the data in Tables 4 and 5, the Sub- Committee considers that the collated results of the collaborative trials show good agreement. The aromatic compounds show greater variability in their RRIs than the non-aromatic compounds. This effect was noted in the previous publications on fingerprinting by the Sub-Committee on packed columns. 13,14 It is attributed to the greater variability in RRI of the aromatic compounds with changes in elution temperature when compared with the non-aromatics.This observation is being studied and will be reported on separately. The data in Table 5 are given as an indication of the composition of oil of rosemary. They assume that all of the compounds have the same response to the flame-ionisation detector and that the oil does not contain any non-volatile material. Also, different integrator parameters will affect the percentage composition data. Conclusion The Sub-Committee recommends the procedure given in the Appendix for the reproducible fingerprinting of essential oils by temperature-programmed GLC using capillary columns with non-polar stationary phases.Although the procedures have been developed and investigated for the analyses of essential oils, it is felt that they have a Table 1 Results of collaborative study on the NC mixture Laboratory 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Stationary phase CPSil5 DB1 BP1 AT1 BP1 007-1 HP1 BP1 DB1 BP1 Rtx1 CPSil5 HP1 CPSil5 HP1 DB1 Column length/m 25 30 25 25 25 25 30 25 60 50 30 25 50 25 25 60 Internal diameter/mm 0.32 0.25 0.22 0.32 0.22 0.25 0.25 0.22 0.25 0.22 0.25 0.25 0.20 0.25 0.32 0.32 Film thickness/ mm 0.25 0.25 0.25 0.30 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.50 0.25 1.05 0.25 Carrier gas He H2 H2 He He He He N2 He H2 He He He He He He Test compounds, RRI values— Limonene 1014 1015 1018 1019 1021 1020 1020 1021 1024 1025 1027 1028 1028 1029 1028 1029 Acetophenone 1025 1026 1030 1032 1034 1036 1034 1034 1037 1038 1042 1043 1044 1047 1048 1046 Linalol 1076 1079 1083 1084 1085 1086 1084 1085 1085 1087 1088 1089 1090 1090 1094 1091 Naphthalene 1139 1144 1153 1155 1158 1158 1156 1159 1167 1168 1173 1176 1176 1179 1176 1177 Linalyl acetate 1240 1240 1241 1242 1242 1243 1241 1242 1242 1242 1243 1242 1244 1245 1248 1238 Cinnamyl alcohol 1270 1268 1271 1274 1277 1279 1273 1276 1276 1279 1283 1281 1284 1290 1288 1268 g-pack 0.992 0.994 0.997 0.998 0.999 0.999 0.999 1.000 1.001 1.002 1.004 1.005 1.005 1.006 1.007 1.007 Analyst, October 1997, Vol. 122 1169much wider application and should find use in many other fields of GLC analysis. Appendix Recommended Method for the Reproducible Fingerprinting of Essential Oils by Temperature-programmed Gas–Liquid Chromatography Using Non-polar Stationary Phases The column The preparation of the types of capillary column available today requires considerable experience and expertise. Chromatographers, therefore, have to rely on specialist manufacturers as few have the required skills.However, this study has shown that columns obtained ‘off-the-shelf’ from the major manufacturers were all suitable for use with this method. The stationary phases referred to in Table 1 are the manufacturers’ trade names for polysiloxane phases with no modifications. The dimensions of the column are not critical to the successful application of this method but the following are recommended.Column length, 25–30 m; internal diameter, 0.22–0.25 mm; film thickness, 0.25 mm. A film thickness of more than 3 mm should not be used as different effects are observed. Columns are usually delivered ready for use but any manufacturers’ instructions concerning conditioning should be heeded. Gas chromatographic conditions The temperature control in the ovens of modern gas chromatographs is very accurate. However, if older instruments are used or if any doubt exists, oven settings should be checked using a thermometer of known accuracy.The temperature range for this procedure is 50–250 °C. The programme rate is very important and must be checked to ensure that it is linear. In this case a rate of 4 °C min21 is used. Any of the three commonly used carrier gases, helium, hydrogen and nitrogen, may be used. However, it is most important that the linear gas velocity is adjusted so that the column is operating at optimum efficiency. The actual velocity will depend on the carrier gas and the dimensions of the column.Column manufacturers will advise on this. Carrier gas control may be by either constant flow or constant pressure. Preparation of test mixtures Prepare a mixture of equal masses of the even carbon numbered n-alkanes from C8 to C24. Prepare a mixture of 1.00 part of limonene, 1.37 parts of linalol, 1.60 parts of linalyl acetate, 1.40 parts of acetophenone, 1.13 parts of naphthalene and 1.80 parts of cinnamyl alcohol (NC mixture).Prepare a mixture of 55% m/ m of the NC mixture and 45% m/m of the hydrocarbon mixture. Each n-alkane is then 5% and each NC component is then approximately 10% of the total mixture. These mixtures are now available commercially. Test chromatogram Set up the chromatograph, with the prepared column, for temperature-programmed operation between 50 and 250 °C at 4 °C min21. Inject the combined test mixture and start the programme and integrator. Continue heating at 250 °C until a stable baseline is obtained.Repeat the run if necessary, adjusting the attenuation to bring all the peaks on-scale. A typical chromatogram is shown in Fig. 1. The sample size, Table 2 Summary of results on the NC mixture (Table 1) and comparison with those on packed columns previously examined Relative Capillary columns— Standard standard Test compound Mean RRI deviation deviation (%) Limonene 1023 4.87 0.48 Acetophenone 1037 6.98 0.67 Linalol 1086 4.36 0.40 Naphthalene 1163 12.03 1.03 Linalyl acetate 1242 2.19 0.18 Cinnamyl alcohol 1277 6.48 0.51 Relative Packed columns*— Standard standard Test compound Mean RRI deviation deviation (%) Limonene 1027 1.63 0.16 Acetophenone 1041 2.99 0.29 Linalol 1086 1.80 0.17 Naphthalene 1172 3.55 0.30 Linalyl acetate 1241 1.90 0.15 Cinnamyl alcohol 1280 3.80 0.30 * See ref. 14. Table 3 Elution temperatures of the NC–HC mixture (°C) Laboratory 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Limonene 72 76 87 93 96 91 94 96 117 112 124 125 127 112 122 145 Acetophenone 73 77 89 95 98 93 96 98 119 114 126 127 130 114 125 148 Linalol 79 84 96 103 106 100 102 106 127 122 133 134 137 121 132 155 Naphthalene 86 92 106 113 116 111 112 117 139 134 147 147 150 134 144 169 Linalyl acetate 98 105 118 125 128 123 123 129 151 146 158 157 159 144 154 180 Cinnamyl alcohol 102 108 123 130 133 128 127 133 155 151 164 162 165 150 159 185 g-pack 0.992 0.994 0.997 0.998 0.999 0.999 0.999 1.000 1.001 1.002 1.004 1.005 1.005 1.006 1.007 1.007 C8 59 59 64 69 71 66 70 70 86 80 90 89 92 82 88 112 C10 71 74 85 91 93 88 91 93 113 108 119 120 123 107 118 140 C12 93 99 113 120 122 117 118 123 145 139 151 151 153 137 147 173 C14 118 126 140 148 151 145 143 151 173 169 181 178 181 165 174 209 C16 143 150 165 173 176 170 167 177 199 195 208 202 206 190 198 1170 Analyst, October 1997, Vol. 122dilution and split ratio should be such that the capacity of the column is not exceeded.Calculation of results When a satisfactory chromatogram has been obtained with baseline separation of all peaks, calculate the RRIs of the NC components assuming a linear span between adjacent hydrocarbon peaks. Using the values obtained, tabulate the results as in Table 6 and calculate the g-pack value for the column following the given worked example. This calculation can easily be performed by using a computer spreadsheet. Calculation of relative retention indices (RRIs) RRI = � - + - + [ ( )] ( ) 200 2 100 Rtc Rtn Rt n Rtn n where n = carbon number of 1st hydrocarbon; Rtc = retention time of the compound; Rtn = retention time of 1st hydrocarbon; and Rt(n + 2) = retention time of 2nd hydrocarbon.Table 4 Results of collaborative study on rosemary oil: relative retention indices Laboratory Compound 1 2 3 4 5 6 7 8 9 10 11 12 13 Mean s s (%) a-Pinene 906 909 916 918 922 918 920 928 928 930 937 936 935 923 9.61 1.04 Camphene 918 921 929 930 935 932 934 942 942 946 951 949 950 937 10.41 1.11 Sabinene —* —* 956 957 959 959 959 —* 963 966 966 964 970 962 4.35 0.45 b-Pinene 949 952 960 960 964 963 964 971 971 975 978 976 977 966 9.07 0.94 Myrcene 972 973 977 977 979 979 979 981 981 982 983 982 983 979 3.41 0.35 p-Cymene 1005 1006 1009 1009 1011 1010 1011 1014 1014 1017 1018 1016 1017 1012 4.10 0.41 Limonene/ cineol 1012 1014 1017 1018 1022 1019 1019 1024 1024 1027 1030 1029 1033 1022 6.14 0.60 Linalol 1074 1076 1082 1081 1084 1083 1083 1084 1086 1088 1088 1087 1092 1084 4.70 0.43 Camphor 1100 1106 1114 1117 1121 1118 1118 1127 1127 1132 1136 1135 1140 1122 11.43 1.02 Borneol 1130 1135 1142 1143 1147 1146 1147 1153 1155 1160 1161 1159 1163 1149 9.89 0.86 Terpinen-4-ol 1143 1148 1157 1156 1159 1159 1160 11 1166 1171 1171 1169 1174 1161 8.84 0.76 a-Terpineol 1157 1161 1161 1167 1171 1170 1171 1175 1177 1181 1181 1179 1184 1172 7.85 0.67 Verbenone 1162 1166 1173 1175 1178 1178 1178 1184 1186 1193 1192 1190 1200 1181 10.57 0.90 Bornyl acetate 1255 1259 1265 1265 1268 1269 1269 1276 1275 1280 1280 1278 1283 1271 8.25 0.65 b-Caryophyllene 1392 1433 1409 1410 1414 1414 1416 1430 1427 1434 1437 1434 1433 1422 13.07 0.92 a-Humulene 1425 1467 1444 1443 1447 1447 1450 1464 1461 1468 1471 1467 1467 1455 13.27 0.91 g-pack 0.992 0.994 0.997 0.998 0.999 0.999 1.000 1.001 1.002 1.004 1.005 1.005 1.007 * Compound not resolved.Table 5 Results of collaborative study on rosemary oil: area % of selected components Laboratory Compound 1 2 3 4 5 6 7 8 9 Mean s s (%) a-Pinene 22.5 18.6 20.2 24.3 19.8 20.4 21.4 20.0 20.3 20.8 1.59 7.6 Camphene 9.2 9.4 8.4 9.8 8.2 8.3 9.1 8.6 8.6 8.8 0.52 5.9 Sabinene 1.3 1.4 1.3 1.4 1.1 1.3 1.2 1.2 1.3 0.10 7.7 b-Pinene 4.5 3.2 2.9 3.2 2.9 2.8 2.9 2.9 2.9 3.1 0.50 16.1 Myrcene 4.2 4.2 4.2 4.3 4.1 3.9 4.1 4.0 3.9 4.1 0.13 3.2 p-Cymene 2.3 2.5 1.9 2.2 1.8 1.8 2.3 1.8 2.1 2.1 0.25 11.9 Limonene/cineol 26.5 26.9 25.2 26.6 25.4 25.6 26.0 26.0 26.0 26.0 0.53 2.0 Linalol 0.8 1.2 1.1 0.9 1.1 1.0 1.1 1.5 1.1 1.1 0.19 17.3 Camphor 17.5 21.2 18.7 16.9 18.9 18.8 18.3 19.4 20.1 18.9 1.22 6.5 Borneol 2.8 3.7 3.2 2.7 3.2 2.8 3.0 2.9 3.2 3.1 0.29 9.4 Terpinen-4-ol 0.7 0.9 0.8 0.7 0.8 0.7 0.8 0.7 0.8 0.8 0.07 8.8 a-Terpineol 1.2 1.5 1.4 1.1 1.4 1.3 1.3 1.4 1.5 1.3 0.13 10.0 Verbenone 1.5 1.9 1.7 1.4 1.6 1.5 1.6 1.6 1.7 1.6 0.14 8.8 Bornyl acetate 0.8 1.2 0.9 0.7 0.9 0.9 0.9 1.0 1.0 0.9 0.13 14.4 b-Caryophyllene 1.8 2.3 2.2 1.5 2.2 2.1 1.9 2.2 2.2 2.0 0.25 12.5 a-Humulene 0.6 0.7 0.7 0.5 0.7 0.6 0.6 0.7 0.7 0.6 0.07 11.7 g-pack 0.994 0.997 0.999 0.999 1.000 1.002 1.004 1.005 1.007 Analyst, October 1997, Vol. 122 1171For example, if the retention times of the C10 hydrocarbon, limonene and C12 hydrocarbon are 17.7, 18.6 and 25.2 min, respectively, then n = 10 and RRI for limonene = [200 (18.6 - 17.7)] 25.2 - 17.7 � + = 1000 1024 The operating conditions and chromatographic system may be considered satisfactory if the results lie within the following specification: g-pack 1.0 ± 0.005 RRI values: Limonene 1018–1028 Acetophenone 1030–1044 Linalol 1082–1090 Naphthalene 1151–1175 Linalyl acetate 1240–1244 Cinnamyl alcohol 1271–1283 Standardised chromatograms of essential oils When the performance of a column within a chromatographic system has been satisfactorily established according to the procedure given above, it can be used under similar conditions for analysing essential oils.The performance of the system should be checked regularly. A sample of the essential oil should be injected and run under the conditions established above (first chromatogram). On completion of the run, a sample of a mixture of the essential oil and the n-alkane hydrocarbon mixture should be run under Fig. 1 Typical chromatogram of NC–hydrocarbon mixture. Table 6 Calculation of g-pack value Test compound RRI Y factor X factor Z factor Limonene 1020 (RRI30.14) + 2 = 1.0629 1.05843Y = 1.1250 1.058423Y = 1.1907 136.23 Acetophenone 1034 (RRI30.14) + 2 = 1.2216 1.33503Y = 1.6308 1.335023Y = 2.1772 120.14 Linalol 1084 (RRI30.14) + 2 = 0.9969 1.02183Y = 1.0186 1.021823Y = 1.0408 154.24 Naphthalene 1156 (RRI30.14) + 2 = 1.2784 1.33613Y = 1.7081 1.336123Y = 2.2822 128.16 Linalyl acetate 1241 (RRI30.14) + 2 = 0.8954 0.87973Y = 0.7877 0.879723Y = 0.6929 196.28 Cinnamyl alcohol 1273 (RRI30.14) + 2 = 1.3432 1.49933Y = 2.0139 1.499323Y = 3.0194 134.17 Sum SY = 6.7984 SX = 8.2841 SZ = 10.4032 3factor f f1SY = 6.79843 f2SX = 8.28413 f3SZ = 10.40323 1.07977 = 7.3407 2.88734 = 23.9190 1.49758 = 15.5796 g-pack value = f2SX2f1SY2f3SZ = 23.919027.3407215.5796 = 0.9987 1172 Analyst, October 1997, Vol. 122identical conditions (second chromatogram). Comparison of the two chromatograms should show identical retention times and comparisons of the peak heights should be consistent with any dilution due to the n-alkane mixture. These comparisons allow for a check on the reproducibility of the system and also enable the position of the n-alkanes to be transferred from the second chromatogram (Fig. 3) to the first (Fig. 2) in such a manner that visual interference is avoided. In instances where a component of the essential oil overlaps or obscures an n-alkane peak, its position can be determined by comparison with the chromatogram obtained for the g-pack calculation (Fig. 1). Thus, it is possible to determine RRIs for any of the peaks of interest in the essential oil chromatogram.References 1 Analytical Methods Committee, Analyst, 1993, 118, 1089. 2 Analytical Methods Committee, Analyst, 1971, 96, 887. 3 Analytical Methods Committee, Analyst, 1973, 98, 616. 4 Analytical Methods Committee, Analyst, 1973, 98, 823. 5 Analytical Methods Committee, Analyst, 1975, 100, 593. 6 Analytical Methods Committee, Analyst, 1977, 102, 607. Fig. 2 Typical chromatogram of oil of rosemary Spanish. Fig. 3 Typical chromatogram of oil of rosemary–hydrocarbon mixture.Analyst, October 1997, Vol. 122 11737 Analytical Methods Committee, Analyst, 1978, 103, 375. 8 Analytical Methods Committee, Analyst, 1981, 106, 456. 9 Analytical Methods Committee, Analyst, 1987, 112, 1315. 10 Analytical Methods Committee, Analyst, 1988, 113, 657. 11 Analytical Methods Committee, Analyst, 1990, 115, 459. 12 van den Dool, H., Standardisation of G.C. Analysis of Essential Oils, Proefschrift, Rijksuniversiteit te Groningen, Rotterdam, 1974. 13 Analytical Methods Committee, Analyst, 1980, 105, 262. 14 Analytical Methods Committee, Analyst, 1981, 106, 448. 15 Analytical Methods Committee, Analyst, 1984, 109, 1339. 16 Analytical Methods Committee, Analyst, 1984, 109, 1343. 17 Analytical Methods Committee, Analyst, 1988, 113, 1125. 18 Grant, D. W., and Hollis, M. G., J. Chromatogr., 1978, 158, 3. Paper 7/04651K Accepted July 2, 1997 1174 Analyst, October 1997, Vol. 122 Application of Gas–Liquid Chromatography to the Analysis of Essential Oils Part XVII.† Fingerprinting of Essential Oils by Temperature-programmed Gas–Liquid Chromatography Using Capillary Columns With Non-polar Stationary Phases Analytical Methods Committee‡ The Royal Society of Chemistry, Burlington House, Piccadilly, London, UK W1V 0BN Problems in obtaining reproducible results when ‘fingerprinting’ essential oils by temperature-programmed gas–liquid chromatography have been reported on in Parts VII and VIII of this series.Those reports were concerned with the general problems and the use of packed columns. This report is concerned with the use of capillary columns and non-polar stationary phases. A collaborative study using capillary columns with non-polar stationary phases has resulted in a method which specifies the ‘g-pack value’ of a column and gives reproducible relative retention indices for the test compounds limonene, acetophenone, linalol, naphthalene, linalyl acetate and cinnamyl alcohol.The method has been applied successfully to the examination of oil of rosemary. A recommended method is given for the reproducible temperature-programmed gas–liquid chromatographic fingerprinting of essential oils using capillary columns with non-polar stationary phases. Keywords: Essential oil analysis; gas–liquid chromatography; fingerprinting; non-polar capillary column The Analytical Methods Committee (AMC) has received and has approved for publication the following report from its Essential Oils Sub-Committee.Report The constitution of the Sub-Committee responsible for the preparation of this report was: Mr. M. J. Milchard (Chairman), Mr. A. M. Humphrey (Chairman to October 1993), Mr. N. Boley, Mr. B. Conway, Mr. R. Esdale, Ms. M. Flowerdew, Miss D. M. Michalkiewicz, Mr. D. A. Moyler, Mr. A. Osbiston, Mr. D. Powis, Mr. A. Sherlock, Mr. R. Smith, Mr. S. Smith, Mr. B. Starr and Mr. T. M.Stevens, with Mr. J. J. Wilson as Secretary. The Sub-Committee would like to thank the following manufacturers of capillary chromatography columns for the interest which they ha shown in this work and their willingness to participate in the collaborative trial: Alltech Associates Inc., Chrompack UK Ltd., Hewlett-Packard Ltd., J&W Scientific/Jones Chromatography, Quadrex Corporation, Restek/Thames Chromatography and SGE (UK) Ltd. Their assistance is gratefully acknowledged.Introduction The development of gas–liquid chromatography (GLC) in the early 1950s began a new era in the analysis of essential oils. Until that time the qualities, purities and origins of oils were assessed by physical measurements and chemical assays. This new method gave improved results over the older methods, many of which we now know gave precise but inaccurate results. The technique was soon applied to the accurate determination of major and other components of interest.2–11 However, the concept of ‘fingerprinting’ oils had not been addressed successfully.The chemical nature of essential oils makes them particularly suitable for analysis by GLC. If temperature-programmed operation is used, a very high proportion of the total number of components present can be resolved. Many attempts have been made to establish libraries of chromatograms from temperatureprogrammed GLC analysis. This would allow sample and reference tracings to be compared and the authenticity and quality of the sample to be determined.However, it was quickly found that the conditions of the GLC analysis had to be strictly controlled, but even then reproducibility was poor, particularly among different laboratories. It was apparent that the temperature programming and the nature of the column itself caused the greatest variation in the results. This has led to individual libraries being established, which makes comparison difficult. The Sub-Committee has been studying the fingerprinting of essential oils over many years.Initial work on packed columns established that the lack of reproducibility of results was due to the lack of reproducibility of the columns themselves. This in turn was attributed to problems with coating the support and the subsequent ageing of the packing with use. There was, therefore, a requirement for the standardisation of the column efficiency and of its selectivity without using one as a factor of the other.A publication by van den Dool12 described a method for the characterisation of GLC columns using the relative retentionindices (RRIs) of a group of six test compounds. By calculating the RRIs of the compounds in the mixture on a particular column and then applying a series of further calculations to these RRIs, van den Dool obtained a figure representing the polarity factor of that particular column which he called the g- † For Part XVI, see ref. 1. ‡ Correspondence should be addressed to the Secretary, Analytical Methods Committee, Analytical Division, The Royal Society of Chemistry, Burlington House, Piccadilly, London, UK W1V 0BN.Analyst, October 1997, Vol. 122 (1167–1174) 1167pack value. This value will vary with the different types of stationary phase and their condition. Van den Dool’s work was based exclusively on packed columns and concentrated on two particular stationary phases: SE-30 as an example of a non-polar type and Carbowax 20M as a moderately polar type. However, the g-pack concept can be applied to any stationary phase.The mixture of compounds used in the determination, known as the NC (Netherlands Committee) mixture, was chosen to represent a range of compounds with functional groups similar to those found in essential oils. Also, the mixture was chosen so that there are two pairs of compounds in which the components of each pair elute close together on the two different phases, which can be used to give a measure of the resolving power of the column.However, the greatly increased efficiency of capillary columns over packed columns means that this property of the mixture is not so significant except in cases of extreme column degradation. The NC mixture consists of limonene, linalol, linalyl acetate, acetophenone, naphthalene and cinnamyl alcohol. The Sub-Committee has published several papers, based on the work of van den Dool, defining the methodology to be used in obtaining these fingerprints13–15 and standard fingerprint traces of selected oils.1,16,17 During these studies on the application of the g-pack concept it was found that the value for a packed column was unaffected by changes in operating parameters such as carrier gas flow rate, temperature programming rate, initial temperature and final temperature hold.However, the value could be decreased by loss of stationary phase over a period of time, due to column bleeding, and could be increased by modification of the stationary phase due to oxidation. The g-pack value can, therefore, give a good indication of the condition of the column.With the wider availability and advances in capillary column technology the Sub-Committee decided that the technique should be updated to make use of this technology and set about determining the optimum operating parameters. It was appreciated that the application to capillary columns was not likely to be as straightforward as with packed columns.A major difference was in the control of the stationary phase. For packed columns, good agreement among laboratories was only obtained when there was control over the preparation of the stationary phase and packing of the column. This involved developing a method for coating the stationary phase onto the support (the absorption coating technique), which led to columns giving good reproducibility and high efficiencies. While it is possible to prepare and coat capillary columns, in reality laboratories buy in their columns from one of the specialist manufacturers. This means that individual laboratories have no control over the preparation of the columns and that different methods of preparation between manufacturers could lead to slightly different performances of nominally the same stationary phase.It was concluded, therefore, that any method for fingerprinting essential oils on capillary columns would have to be based on commercially available columns and be robust enough to cope with a range of operating parameters. If the method is to be widely applicable it is unrealistic to expect laboratories to buy columns from a specific manufacturer or to change carrier gas.Experimental It was decided that the initial examination would be carried out on a specified stationary phase on columns which the members of the committee had available in their laboratories. A non-polar phase was chosen equivalent to SE-30.A protocol was provided specifying the samples to be examined and the initial column temperature, temperature programme rate and final column temperature. The choice of carrier gas was left to individuals with the proviso that the flow rate was optimised for the column configuration. The three samples examined were: a mixture of the NC mix and a series of even numbered carbon aliphatic hydrocarbons (C8–C24) known as the NC–HC mixture; Spanish rosemary oil; and Spanish rosemary oil plus the hydrocarbon mixture.The initial results showed that there were differences between different column manufacturers but very similar results were obtained between laboratories using similar columns from the same manufacturer. It was, therefore, decided to approach the major capillary column manufacturers and invite them to examine the same samples on their columns. All of the manufacturers were very willing to collaborate in this examination with the given protocol.With their inclusion a total of 16 results were obtained. Results The results of the collaborative study using the NC mixture with the proposed procedure are given in Table 1 in ascending order of g-pack values and are summarised in Table 2. It can be seen from Table 2 that the mean RRIs for the test compounds obtained in this examination compare well with results previously obtained on packed columns. Table 3 shows the elution temperatures of the test compounds and n-alkane hydrocarbons.There is a much wider variation in elution temperatures with capillary columns than was found with packed columns. The results from the examination of oil of rosemary are given in Tables 4 and 5. The results obtained were considered satisfactory by the Sub-Committee. Discussion Throughout all of the collaborative exercises, the Sub-Committee was conscious of the fact that many of the parameters examined were chosen with arbitrary limits and that some justification for the choices should be made.In particular, this applies to the nature of the test compounds in the NC mixture and the use of the RRI system. This system is widely used and most gas chromatographers are acquainted with it. It is a simple system and, although it is relative rather than absolute, it allows for a choice of reference compounds which can be made according to other requirements. For the purpose of this and the previous investigation it was felt that the homologous series of n-alkanes was the most suitable as reference compounds.They are readily available in pure form, are extremely stable and are the least likely compounds to exhibit chromatographic anomalies. In addition, they formed the basis of the work by van den Dool. The elution of a variety of homologous series under temperature-programmed conditions is not linear and the departure from linearity increases as the programming rate decreases.18 Also, this departure is greater for homologous series of polar compounds than for the n-alkanes and this was another reason for the choice of the latter.This non-linearity of elution raises the question of the method of application of the system and if an assumption of linearity between successive n-alkanes introduces errors. Van den Dool recommends the use of both even and odd numbered carbon nalkanes to reduce any error, whereas all of our collaborative work has been carried out with the even numbered carbon nalkanes making the assumption of linearity between each pair.Comparisons have been made of the effect on the calculated RRIs of the six NC text compounds when using all of the n- 1168 Analyst, October 1997, Vol. 122alkanes against using only the even numbered carbon ones and also by calculating on the assumption of linearity as well as by a graphical method to obtain more accurate figures. The differences that were found were so small (in some instances zero) that they were considered to be insignificant.Therefore, our recommended procedure uses only the even numbered carbon n-alkanes and assumes linearity between them. The six compounds in the NC mixture were selected by van den Dool on the basis of their similarity to the types of compounds found in essential oil analyses. Their choice could perhaps be criticised if used in conjunction with other sample types. However, the NC mixture is used for the calibration of the column irrespective of its ultimate use.As the mathematical treatment of the results has already been worked out by van den Dool, there seemed to be no advantage in changing the mixture. The choice of six compounds is a reasonable compromise between having an excessive number of interfering peaks and mathematical calculations and a reduced amount of data leading to a less accurate result. With packed columns, more control was exercised over the operating conditions by arranging for the C24 alkane to elute at the upper temperature of the temperature programme run by adjustment of the carrier gas flow rate.This approach is not relevant for capillary columns as, to obtain satisfactory results, it is necessary to optimise the flow rate, which depends on the dimensions of the individual columns and the nature of the carrier gas. The superior resolving power of capillary columns led to the conclusion that measurement of the resolution between the limonene and acetophenone peaks would not be meaningful as there was baseline separation in all cases. The column would be in poor condition or the operating conditions be significantly different from those recommended for these compounds not to be resolved. Developments in gas chromatographic instrumentation have led to discussions on the relative merits of operating the carrier gas system under conditions of constant flow or constant pressure.The NC–HC mixture was chromatographed on a column in an instrument equipped to run under both conditions.As shown below, the RRIs were virtually identical under both conditions. However, the chromatogram of the sample run under constant pressure conditions showed an increase in the retention times of the components compared with those run under constant flow conditions. This is because of a reduction in flow rate during temperature-programmed operation under constant pressure.RRI Compound Constant pressure Constant flow Limonene 1020 1020 Acetophenone 1034 1034 Linalol 1084 1084 Naphthalene 1158 1156 Linalyl acetate 1242 1241 Cinnamyl alcohol 1274 1273 g-pack 0.999 This demonstrates that either method of carrier gas control should be suitable for undertaking fingerprint determinations using this procedure. The RRI and area % composition of the 17 most abundant components in the sample of oil of rosemary were calculated from chromatograms run under the same conditions as for the HC–NC mixture.From the data in Tables 4 and 5, the Sub- Committee considers that the collated results of the collaborative trials show good agreement. The aromatic compounds show greater variability in their RRIs than the non-aromatic compounds. This effect was noted in the previous publications on fingerprinting by the Sub-Committee on packed columns. 13,14 It is attributed to the greater variability in RRI of the aromatic compounds with changes in elution temperature when compared with the non-aromatics.This observation is being studied and will be reported on separately. The data in Table 5 are given as an indication of the composition of oil of rosemary. They assume that all of the compounds have the same response to the flame-ionisation detector and that the oil does not contain any non-volatile material. Also, different integrator parameters will affect the percentage composition data. Conclusion The Sub-Committee recommends the procedure given in the Appendix for the reproducible fingerprinting of essential oils by temperature-programmed GLC using capillary columns with non-polar stationary phases.Although the procedures have been developed and investigated for the analyses of essential oils, it is felt that they have a Table 1 Results of collaborative study on the NC mixture Laboratory 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Stationary phase CPSil5 DB1 BP1 AT1 BP1 007-1 HP1 BP1 DB1 BP1 Rtx1 CPSil5 HP1 CPSil5 HP1 DB1 Column length/m 25 30 25 25 25 25 30 25 60 50 30 25 50 25 25 60 Internal diameter/mm 0.32 0.25 0.22 0.32 0.22 0.25 0.25 0.22 0.25 0.22 0.25 0.25 0.20 0.25 0.32 0.32 Film thickness/ mm 0.25 0.25 0.25 0.30 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.50 0.25 1.05 0.25 Carrier gas He H2 H2 He He He He N2 He H2 He He He He He He Test compounds, RRI values— Limonene 1014 1015 1018 1019 1021 1020 1020 1021 1024 1025 1027 1028 1028 1029 1028 1029 Acetophenone 1025 1026 1030 1032 1034 1036 1034 1034 1037 1038 1042 1043 1044 1047 1048 1046 Linalol 1076 1079 1083 1084 1085 1086 1084 1085 1085 1087 1088 1089 1090 1090 1094 1091 Naphthalene 1139 1144 1153 1155 1158 1158 1156 1159 1167 1168 1173 1176 1176 1179 1176 1177 Linalyl acetate 1240 1240 1241 1242 1242 1243 1241 1242 1242 1242 1243 1242 1244 1245 1248 1238 Cinnamyl alcohol 1270 1268 1271 1274 1277 1279 1273 1276 1276 1279 1283 1281 1284 1290 1288 1268 g-pack 0.992 0.994 0.997 0.998 0.999 0.999 0.999 1.000 1.001 1.002 1.004 1.005 1.005 1.006 1.007 1.007 Analyst, October 1997, Vol. 122 1169much wider application and should find use in many other fields of GLC analysis. Appendix Recommended Method for the Reproducible Fingerprinting of Essential Oils by Temperature-programmed Gas–Liquid Chromatography Using Non-polar Stationary Phases The column The preparation of the types of capillary column available today requires considerable experience and expertise. Chromatographers, therefore, have to rely on specialist manufacturers as few have the required skills.However, this study has shown that columns obtained ‘off-the-shelf’ from the major manufacturers were all suitable for use with this method. The stationary phases referred to in Table 1 are the manufacturers’ trade names for polysiloxane phases with no modifications. The dimensions of the column are not critical to the successful application of this method but the following are recommended.Column length, 25–30 m; internal diameter, 0.22–0.25 mm; film thickness, 0.25 mm. A film thickness of more than 3 mm should not be used as different effects are observed. Columns are usually delivered ready for use but any manufacturers’ instructions concerning conditioning should be heeded. Gas chromatographic conditions The temperature control in the ovens of modern gas chromatographs is very accurate. However, if older instruments are used or if any doubt exists, oven settings should be checked using a thermometer of known accuracy.The temperature range for this procedure is 50–250 °C. The programme rate is very important and must be checked to ensure that it is linear. In this case a rate of 4 °C min21 is used. Any of the three commonly used carrier gases, helium, hydrogen and nitrogen, may be used. However, it is most important that the linear gas velocity is adjusted so that the column is operating at optimum efficiency.The actual velocity will depend on the carrier gas and the dimensions of the column. Column manufacturers will advise on this. Carrier gas control may be by either constant flow or constant pressure. Preparation of test mixtures Prepare a mixture of equal masses of the even carbon numbered n-alkanes from C8 to C24. Prepare a mixture of 1.00 part of limonene, 1.37 parts of linalol, 1.60 parts of linalyl acetate, 1.40 parts of acetophenone, 1.13 parts of naphthalene and 1.80 parts of cinnamyl alcohol (NC mixture).Prepare a mixture of 55% m/ m of the NC mixture and 45% m/m of the hydrocarbon mixture. Each n-alkane is then 5% and each NC component is then approximately 10% of the total mixture. These mixtures are now available commercially. Test chromatogram Set up the chromatograph, with the prepared column, for temperature-programmed operation between 50 and 250 °C at 4 °C min21. Inject the combined test mixture and start the programme and integrator.Continue heating at 250 °C until a stable baseline is obtained. Repeat the run if necessary, adjusting the attenuation to bring all the peaks on-scale. A typical chromatogram is shown in Fig. 1. The sample size, Table 2 Summary of results on the NC mixture (Table 1) and comparison with those on packed columns previously examined Relative Capillary columns— Standard standard Test compound Mean RRI deviation deviation (%) Limonene 1023 4.87 0.48 Acetophenone 1037 6.98 0.67 Linalol 1086 4.36 0.40 Naphthalene 1163 12.03 1.03 Linalyl acetate 1242 2.19 0.18 Cinnamyl alcohol 1277 6.48 0.51 Relative Packed columns*— Standard standard Test compound Mean RRI deviation deviation (%) Limonene 1027 1.63 0.16 Acetophenone 1041 2.99 0.29 Linalol 1086 1.80 0.17 Naphthalene 1172 3.55 0.30 Linalyl acetate 1241 1.90 0.15 Cinnamyl alcohol 1280 3.80 0.30 * See ref. 14. Table 3 Elution temperatures of the NC–HC mixture (°C) Laboratory 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Limonene 72 76 87 93 96 91 94 96 117 112 124 125 127 112 122 145 Acetophenone 73 77 89 95 98 93 96 98 119 114 126 127 130 114 125 148 Linalol 79 84 96 103 106 100 102 106 127 122 133 134 137 121 132 155 Naphthalene 86 92 106 113 116 111 112 117 139 134 147 147 150 134 144 169 Linalyl acetate 98 105 118 125 128 123 123 129 151 146 158 157 159 144 154 180 Cinnamyl alcohol 102 108 123 130 133 128 127 133 155 151 164 162 165 150 159 185 g-pack 0.992 0.994 0.997 0.998 0.999 0.999 0.999 1.000 1.001 1.002 1.004 1.005 1.005 1.006 1.007 1.007 C8 59 59 64 69 71 66 70 70 86 80 90 89 92 82 88 112 C10 71 74 85 91 93 88 91 93 113 108 119 120 123 107 118 140 C12 93 99 113 120 122 117 118 123 145 139 151 151 153 137 147 173 C14 118 126 140 148 151 145 143 151 173 169 181 178 181 165 174 209 C16 143 150 165 173 176 170 167 177 199 195 208 202 206 190 198 1170 Analyst, October 1997, Vol. 122dilution and split ratio should be such that the capacity of the column is not exceeded.Calculation of results When a satisfactory chromatogram has been obtained with baseline separation of all peaks, calculate the RRIs of the NC components assuming a linear span between adjacent hydrocarbon peaks. Using the values obtained, tabulate the results as in Table 6 and calculate the g-pack value for the column following the given worked example. This calculation can easily be performed by using a computer spreadsheet. Calculation of relative retention indices (RRIs) RRI = � - + - + [ ( )] ( ) 200 2 100 Rtc Rtn Rt n Rtn n where n = carbon number of 1st hydrocarbon; Rtc = retention time of the compound; Rtn = retention time of 1st hydrocarbon; and Rt(n + 2) = retention time of 2nd hydrocarbon.Table 4 Results of collaborative study on rosemary oil: relative retention indices Laboratory Compound 1 2 3 4 5 6 7 8 9 10 11 12 13 Mean s s (%) a-Pinene 906 909 916 918 922 918 920 928 928 930 937 936 935 923 9.61 1.04 Camphene 918 921 929 930 935 932 934 942 942 946 951 949 950 937 10.41 1.11 Sabinene —* —* 956 957 959 959 959 —* 963 966 966 964 970 962 4.35 0.45 b-Pinene 949 952 960 960 964 963 964 971 971 975 978 976 977 966 9.07 0.94 Myrcene 972 973 977 977 979 979 979 981 981 982 983 982 983 979 3.41 0.35 p-Cymene 1005 1006 1009 1009 1011 1010 1011 1014 1014 1017 1018 1016 1017 1012 4.10 0.41 Limonene/ cineol 1012 1014 1017 1018 1022 1019 1019 1024 1024 1027 1030 1029 1033 1022 6.14 0.60 Linalol 1074 1076 1082 1081 1084 1083 1083 1084 1086 1088 1088 1087 1092 1084 4.70 0.43 Camphor 1100 1106 1114 1117 1121 1118 1118 1127 1127 1132 1136 1135 1140 1122 11.43 1.02 Borneol 1130 1135 1142 1143 1147 1146 1147 1153 1155 1160 1161 1159 1163 1149 9.89 0.86 Terpinen-4-ol 1143 1148 1157 1156 1159 1159 1160 1165 1166 1171 1171 1169 1174 1161 8.84 0.76 a-Terpineol 1157 1161 1161 1167 1171 1170 1171 1175 1177 1181 1181 1179 1184 1172 7.85 0.67 Verbenone 1162 1166 1173 1175 1178 1178 1178 1184 1186 1193 1192 1190 1200 1181 10.57 0.90 Bornyl acetate 1255 1259 1265 1265 1268 1269 1269 1276 1275 1280 1280 1278 1283 1271 8.25 0.65 b-Caryophyllene 1392 1433 1409 1410 1414 1414 1416 1430 1427 1434 1437 1434 1433 1422 13.07 0.92 a-Humulene 1425 1467 1444 1443 1447 1447 1450 1464 1461 1468 1471 1467 1467 1455 13.27 0.91 g-pack 0.992 0.994 0.997 0.998 0.999 0.999 1.000 1.001 1.002 1.004 1.005 1.005 1.007 * Compound not resolved. Table 5 Results of collaborative study on rosemary oil: area % of selected components Laboratory Compound 1 2 3 4 5 6 7 8 9 Mean s s (%) a-Pinene 22.5 18.6 20.2 24.3 19.8 20.4 21.4 20.0 20.3 20.8 1.59 7.6 Camphene 9.2 9.4 8.4 9.8 8.2 8.3 9.1 8.6 8.6 8.8 0.52 5.9 Sabinene 1.3 1.4 1.3 1.4 1.1 1.3 1.2 1.2 1.3 0.10 7.7 b-Pinene 4.5 3.2 2.9 3.2 2.9 2.8 2.9 2.9 2.9 3.1 0.50 16.1 Myrcene 4.2 4.2 4.2 4.3 4.1 3.9 4.1 4.0 3.9 4.1 0.13 3.2 p-Cymene 2.3 2.5 1.9 2.2 1.8 1.8 2.3 1.8 2.1 2.1 0.25 11.9 Limonene/cineol 26.5 26.9 25.2 26.6 25.4 25.6 26.0 26.0 26.0 26.0 0.53 2.0 Linalol 0.8 1.2 1.1 0.9 1.1 1.0 1.1 1.5 1.1 1.1 0.19 17.3 Camphor 17.5 21.2 18.7 16.9 18.9 18.8 18.3 19.4 20.1 18.9 1.22 6.5 Borneol 2.8 3.7 3.2 2.7 3.2 2.8 3.0 2.9 3.2 3.1 0.29 9.4 Terpinen-4-ol 0.7 0.9 0.8 0.7 0.8 0.7 0.8 0.7 0.8 0.8 0.07 8.8 a-Terpineol 1.2 1.5 1.4 1.1 1.4 1.3 1.3 1.4 1.5 1.3 0.13 10.0 Verbenone 1.5 1.9 1.7 1.4 1.6 1.5 1.6 1.6 1.7 1.6 0.14 8.8 Bornyl acetate 0.8 1.2 0.9 0.7 0.9 0.9 0.9 1.0 1.0 0.9 0.13 14.4 b-Caryophyllene 1.8 2.3 2.2 1.5 2.2 2.1 1.9 2.2 2.2 2.0 0.25 12.5 a-Humulene 0.6 0.7 0.7 0.5 0.7 0.6 0.6 0.7 0.7 0.6 0.07 11.7 g-pack 0.994 0.997 0.999 0.999 1.000 1.002 1.004 1.005 1.007 Analyst, October 1997, Vol. 122 1171For example, if the retention times of the C10 hydrocarbon, limonene and C12 hydrocarbon are 17.7, 18.6 and 25.2 min, respectively, then n = 10 and RRI for limonene = [200 (18.6 - 17.7)] 25.2 - 17.7 � + = 1000 1024 The operating conditions and chromatographic system may be considered satisfactory if the results lie witthe following specification: g-pack 1.0 ± 0.005 RRI values: Limonene 1018–1028 Acetophenone 1030–1044 Linalol 1082–1090 Naphthalene 1151–1175 Linalyl acetate 1240–1244 Cinnamyl alcohol 1271–1283 Standardised chromatograms of essential oils When the performance of a column within a chromatographic system has been satisfactorily established according to the procedure given above, it can be used under similar conditions for analysing essential oils.The performance of the system should be checked regularly.A sample of the essential oil should be injected and run under the conditions established above (first chromatogram). On completion of the run, a sample of a mixture of the essential oil and the n-alkane hydrocarbon mixture should be run under Fig. 1 Typical chromatogram of NC–hydrocarbon mixture. Table 6 Calculation of g-pack value Test compound RRI Y factor X factor Z factor Limonene 1020 (RRI30.14) + 2 = 1.0629 1.05843Y = 1.1250 1.058423Y = 1.1907 136.23 Acetophenone 1034 (RRI30.14) + 2 = 1.2216 1.33503Y = 1.6308 1.335023Y = 2.1772 120.14 Linalol 1084 (RRI30.14) + 2 = 0.9969 1.02183Y = 1.0186 1.021823Y = 1.0408 154.24 Naphthalene 1156 (RRI30.14) + 2 = 1.2784 1.33613Y = 1.7081 1.336123Y = 2.2822 128.16 Linalyl acetate 1241 (RRI30.14) + 2 = 0.8954 0.87973Y = 0.7877 0.879723Y = 0.6929 196.28 Cinnamyl alcohol 1273 (RRI30.14) + 2 = 1.3432 1.49933Y = 2.0139 1.499323Y = 3.0194 134.17 Sum SY = 6.7984 SX = 8.2841 SZ = 10.4032 3factor f f1SY = 6.79843 f2SX = 8.28413 f3SZ = 10.40323 1.07977 = 7.3407 2.88734 = 23.9190 1.49758 = 15.5796 g-pack value = f2SX2f1SY2f3SZ = 23.919027.3407215.5796 = 0.9987 1172 Analyst, October 1997, Vol. 122identical conditions (second chromatogram). Comparison of the two chromatograms should show identical retention times and comparisons of the peak heights should be consistent with any dilution due to the n-alkane mixture. These comparisons allow for a check on the reproducibility of the system and also enable the position of the n-alkanes to be transferred from the second chromatogram (Fig. 3) to the first (Fig. 2) in such a manner that visual interference is avoided. In instances where a component of the essential oil overlaps or obscures an n-alkane peak, its position can be determined by comparison with the chromatogram obtained for the g-pack calculation (Fig. 1). Thus, it is possible to determine RRIs for any of the peaks of interest in the essential oil chromatogram. References 1 Analytical Methods Committee, Analyst, 1993, 118, 1089. 2 Analytical Methods Committee, Analyst, 1971, 96, 887. 3 Analytical Methods Committee, Analyst, 1973, 98, 616. 4 Analytical Methods Committee, Analyst, 1973, 98, 823. 5 Analytical Methods Committee, Analyst, 1975, 100, 593. 6 Analytical Methods Committee, Analyst, 1977, 102, 607. Fig. 2 Typical chromatogram of oil of rosemary Spanish. Fig. 3 Typical chromatogram of oil of rosemary–hydrocarbon mixture. Analyst, October 1997, Vol. 122 11737 Analytical Methods Committee, Analyst, 1978, 103, 375. 8 Analytical Methods Committee, Analyst, 1981, 106, 456. 9 Analytical Methods Committee, Analyst, 1987, 112, 1315. 10 Analytical Methods Committee, Analyst, 1988, 113, 657. 11 Analytical Methods Committee, Analyst, 1990, 115, 459. 12 van den Dool, H., Standardisation of G.C. Analysis of Essential Oils, Proefschrift, Rijksuniversiteit te Groningen, Rotterdam, 1974. 13 Analytical Methods Committee, Analyst, 1980, 105, 262. 14 Analytical Methods Committee, Analyst, 1981, 106, 448. 15 Analytical Methods Committee, Analyst, 1984, 109, 1339. 16 Analytical Methods Committee, Analyst, 1984, 109, 1343. 17 Analytical Methods Committee, Analyst, 1988, 113, 1125. 18 Grant, D. W., and Hollis, M. G., J. Chromatogr., 1978, 158, 3. Paper 7/04651K Accepted July 2, 1997 1174 Analyst, October 1997, Vol. 122
ISSN:0003-2654
DOI:10.1039/a704651k
出版商:RSC
年代:1997
数据来源: RSC
|
|