ISSN:1359-7337    

DOI:10.1039/AC99633FX005

出版商:RSC    

年代:1996

数据来源: RSC

ISSN:1359-7337    

DOI:10.1039/AC99633BX007

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analytical Communications, February 1996, Vol33 (47-50) 47 Enzyme Assay for the Rapid Determination of Plasma Lysine Philip R. Beckett, Diane Wray-Cahen, Teresa A. Davis and Kenneth C. Copeland Department of Pediatrics, Texas Children's Hospital and United States Department of Agriculture, Agricultural Research Service, Children's Nutrition Research Center, Baylor College of Medicine, Houston, Texas 77030 The enzymic determination of lysine, using saccharopine dehydrogenase, has been modified to determine plasma lysine concentrations in approximately 5 min.The correlation between lysine concentration determined by using HPLC and this enzymic assay was high (r2 = 0.92). The main interfering compound was leucine which, at 250 pmol 1-1, inhibited the rate measured with lysine by 17%.One application of this method allows plasma lysine to be 'clamped' under hyperinsulinemic euglycemic conditions. The technique may be applicable to the bedside investigation of the roles of hormones and nutrition in the regulation of protein metabolism. Amino acid concentrations in physiological samples can be determined by using enzyme assays. Such methods have been developed for determining lysine concentrations; I-3 however, these methods are limited by their poor sensitivity over the physiological range in blood and/or long assay time.For some physiological applications (such as euaminoacidemic clampu) a method for determining amino acid concentrations within a few minutes is required. A rapid HPLC method has been developed,7 but it requires a specially adapted HPLC system and has a sample analysis time of at least 10 min.We have adapted the enzymic determination of lysine3 so that it is sensitive (i.e., requires small volumes of plasma) and can determine the lysine concentration in about 5 min. Experimental Reagents Saccharopine dehydrogenase (EC 1.5.1.7, SaDH) from baker's yeast, reduced nicotinamide adenine dinucleotide (NADH), a- ketoglutarate (Na2 salt), lysine HCl, pyruvate, L-leucine, L- isoleucine and L-valine were purchased from Sigma (St.Louis, MO, USA). SaDH (68 U mg-1 of protein; 1 U = 16.64 nkat) was diluted to 1 mg of protein per millilitre in 0.1 mol 1-1 potassium phosphate buffer, pH 6.8. Aliquots (2 Units) were stored at -20 "C. Each SaDH aliquot was diluted prior to use with phosphate buffer (0.1 mol l-l, pH 6.8, containing 1 mg ml-1 of bovine serum albumin).Assay buffer (0.1 moll-' potassium phosphate buffer, pH 6.8, containing 0.5% Nonidet P40) was used to dilute the reagents and to make up the final assay volume. Fresh solutions of SaDH, a-ketoglutarate and NADH were made daily. Assay buffer, serum and a-ketoglutar- ate were maintained at room temperature; NADH and SaDH were maintained at 5 "C.Assay Conditions The assay comprised 50 pl of sample, 100 pmol l-1 NADH, 5 mmol 1-1 a-ketoglutarate and assay buffer in a total volume of 290 pl in a 1 ml quartz cuvette. After mixing, the rate of change in absorbance was observed at 340 nm until it stabilized near zero. The rate of change in absorbance was then measured for 1 min (defined as the blank rate).After measurement of the blank rate, 0.04 U SaDH (10 pl) was added. After mixing, the rate was measured again for 1 min (defined as the assay rate). The blank rate was subtracted from the assay rate to obtain the enzyme rate. During development of the assay, the assay conditions were modified; therefore, some of the tests were performed using slightly different conditions.These conditions included increasing the assay volume from 200 to 300 p1, increasing the SaDH from 0.02 to 0.04 units, and decreasing NADH from 200 to 100 pmol 1-1. The differences are included where they occurred under each of the experiments described below. Spectrophometric measurements were made at room tem- perature using a Shimadzu UV16OU spectrophotometer (Kyoto, Japan) equipped with an adjustable microcuvette cell holder (part no.206-14334). This cell holder was used to elevate the base of the cuvette so that the total volume of the assay could be reduced to 300 pl. Quartz microcuvette cells (1 ml) were used. Although direct comparisons were not made between serum and plasma, all studies were performed using fresh plasma except for those examining recovery of a lysine spike.Calculations Lysine concentration in plasma samples, [Lysplasma] was calculated as follows: [Lysplasma] = Aabsplasma/Aabsstd X [Lysstd] where Aabsplasma is the enzyme rate measured with the plasma sample; Aabsstd is the enzyme rate measured with the standard; and [Lysstd] is the concentration of lysine in the standard. Experiments Determination of enzyme stability An aliquot of SaDH was diluted to an enzyme protein concentration of 0.028 mg ml-l (4 U ml-1) with phosphate buffer (0.1 moll-1, pH 6.8) containing 1 mg ml-1 bovine serum albumin.The enzyme rate was measured as described under Assay Conditions except the total volume of the assay was 200 pl; the concentration of NADH was 200 pmol l-1, and 0.02 U SaDH were used.The rate was measured using 50 p1 of a 125 pmol l-1 lysine standard (concentration = 31.3 pmol l-1 in the assay) after 0, 2, 5, 5.5, 7.5, and 24 h. Determination of proportionality Serial dilutions of 10 mmol 1-l lysine were used to give standards of 1000,500, 250, 125,63,31 and 16 nmol ml-1 L- lysine in 18 MQ H20 to give lysine concentrations in the final assay of 250, 125, 63, 31, 16, 8 and 4 pmol 1-1, respectively.48 Analytical Communications, February 1996, Vol33 - I lz .- E 0.120- a, s n $0.080- n a 5 0.040- [r .The enzyme rate was measured for each lysine standard under the same conditions used to determine enzyme stability. 8 % a 5 ; 0.03- a 0.02- . Investigations of other interfering substances in serum The enzyme rate was measured with 25 p1 of human serum, with 8, 3 1, or 125 pmol 1-l of lysine, and with 25 pl serum spiked with 8, 31, or 125 pmol 1-1 of lysine.The assay was the same as described under Assay Conditions except that the total assay volume was 200 pl with 200 ymol 1-1 NADH and 0.02 units saccharopine dehydrogenase. Determination of Nrhether similar substrates can substitute .for lysine in the assay or inhibit the rate of convession oflysine to saccharopine The assay was performed using leucine, isoleucine, valine, and pyruvate.The enzyme rate was determined with 1667 pmol 1-1 leucine, isoleucine, valine, or pyruvate, or with 167, 83, 42, or 21 pmol 1-1 leucine, isoleucine, valine, or pyruvate plus 21 pmol 1-1 lysine, as described under Assay Conditions, except that the concentration of NADH was 200 pmol 1-I.Comparison of the assuy uith an established technique Samples of pig plasma were analysed by using both HPLC and spectrophotometry. For HPLC analysis, plasma was filtered through a 10 000 molecular mass filter with methionine sulfone added as an internal standard as described by Davis et u1.* Amino acids were precolumn derivatized with phenylisothio- cyanate and separated on a PicoTag reverse phase column (Waters, Milford, MA, USA). Derivatized amino acids were detected on-line spectrophotometrically, and quantities were calculated using a physiological amino acid standard (Pierce, Rockford, IL, USA).The spectrophometric analysis was performed as described under Assay Conditions. Statistical Analysis Enzyme stability and recoveries of the standards from the spike were analysed by one-way ANOVA (Minitab, State College, PA, USA).Multiple regression (Minitab) was used to compare lysine concentration determined by using HPLC and by the enzyme method and to determine the effect of leucine on their relationship. For the lysine assay, intra-assay variability was 2.5% and inter-assay variability was 3.7%.Values of p < 0.05 were considered significant. Results and Discussion Enzyme Stability For use in studies which require repeated analysis of plasma lysine over time, the rate of the enzymic reaction must be stable. The measured enzyme rate was stable for 7.5 h but the rate decreased by 20% from 0.0186 f 0.0002 at time 0 to 0.0146 f 0.0019 min-1 at 24 h (p < 0.07).Proportionality Since the rate of enzymic reaction is linearly related to substrate concentration, only at concentrations well below the K , of the enzyme, the assay was set up so that the K, of SaDH for lysine ( I 2 mmol 1-1) was at least one order of magnitude higher than the highest concentration of lysine expected. Actual physio- logical lysine concentrations usually range from 75 to 350 pmol 1- 1 in the human,g 40 to 250 pmol 1- 1 in the pig, lo 200 to 800 pmol 1-1 in the rat,* and 160 to 227 pmol l-1 in the dog.ll The enzyme rate was proportional to lysine concentration in lysine standards from 3 1 to 1000 pmol I-' lysine (r2 = 0.9992), or lysine concentration in the assay from 7.8 to 250 pmol 1-1 (Fig.1). NADH at concentrations of 120-280 pmol 1-I and p- ketoglutarate at concentrations of 3-7 mmol 1- l, much greater than their respective K , values, had no effect on the rate using 21 ymol 1-' lysine standard.Interfering Substances in Serum When lysine standards were added to serum, at concentrations of 8,31, or 125 pmol l-1, recoveries were 103, 109 and 10096, respectively, with a mean recovery of 103.7 k 5.47.Fig. 2 shows the enzyme rate with the serum alone, the standards alone, and with the serum spiked with the same standards. There were no differences in recovery among the three standards (p = 0.301), and the correlation by regression analysis was also consistent with proportionality, r2 = 0.997. Specifcity Leucine, isoleucine, valine or pyruvate at a concentration of 1667 ymol l-l in the assay caused no change in absorbance in the absence of lysine.With 21 pmol 1-1 lysine in the assay, however, the concurrent addition of leucine inhibited the rate of 0.164 0.000 , , I I , , 0 50 100 150 200 250 300 Lysine concentration/pmol I-' Fig. 1 Rate of change in absorbance was measured over 1 min at 340 nm in the presence of 0.02 U SaDH, 200 pmol 1-' NADH, 5 mmol 1-1 a - ketoglutarate with the addition of 50 wl L-lysine standards at concentrations of 4, 8, 16.31, 63, 125 or 250 pmol 1-1 lysine. Intercept, 0.00127; slope, 0.000141: and r2 = 0.992. 0.om n n 0 8 31 125 Lysine concentration in spike/pmol I-' Fig. 2 Recovery of lysine in human serum. The rate of disappearance of NADH was measured over 1 min at 340 nm in the presence of 0.02 U SaDH, 200 pmol I-' NADH and 5 mmol I--' a-ketoglutarate with either 25 p1 human serum or 8, 31, or 125 pmol 1-1 lysine standard or with 25 pl serum and the assay spiked with 8, 3 1, or 125 ymol l-1 lysine.The open bars are the rate measured with the lysine standards alone. The parr shaded bars are the rate measured with serum spiked with the lysine standard. The shaded portion is the rate that was measured with the serum alone.Recovery of the 8, 31, or 125 pmol 1-1 lysine standards were 103, 109 and loo%, respectively. When analysed by regression, r2 = 0.997. (Points are duplicates.)Analytical Communications, February 1996, Vol33 49 0.024 0.020 0.01 6 0.012 1 0.008 i. 0.004 0.000 4.004 0.024 0.020 I T 0.016 0.01 2 0.008 1 b 4 0.024 0.020 (r 0.016 0.01 2 0.008 0.004 1 0.000 -0.004 0.024 1 (a 0.020 1 i 0.008 0.004 0.000 4,004 10 100 1000 Concentration/pmol I-’ Fig.3 Interference of other substrates on the activity of SaDH. The rate of disappearance of NADH was measured over 1 min at 340 nm in the presence of 0.04 U SaDH, 200 pmol 1-l NADH and 5 mmol 1-l a- ketoglutarate, 21 pmol l-1 lysine and 0,21,42,83., or 167 pmol l-1 leucine (a), valine (b), isoleucine (c) or pyruvate (6).(Points are duplicates.) NADH disappearance [Fig. 3(a)]. Isoleucine, valine and pyruvate at physiological concentrations had no effect on the rate of NADH disappearance when the lysine standard was present (Fig. 3). Comparison with HPLC Samples of pig serum analysed both by an enzymic technique and by using HPLC correlated closely (r2 = 0.92, p < 0.001, Fig.4) although the slope was 0.66. Multiple regression analysis with inclusion of the concentration of plasma leucine from HPLC did not influence correlation of the two methods (r2 = 0.93) but did normalize the slope to 0.96. Thus, it should be emphasized that in the presence of high physiological leucine concentrations, the assay will underestimate lysine concentra- tion.As long as lysine and leucine concentrations remain //../ I I I 1 0 50 100 150 200 Serum lysine by using HPLC/pmol I-’ Fig. 4 Comparison of plasma lysine concentration determined by using HPLC and by using an enzymic technique. HPLC was performed on filtered plasma samples derivatized by PITC, separated by reverse phase chroma- tography and detected spectrophometrically using the Pico.Tag chemistry and a Waters column against a Pierce standard.The enzyme assay was used to measure the rate of disappearance of NADH over 1 min at 340 nm in the presence of 0.04 U SaDH, 200 pmol 1-* NADH, 5 mmol 1-I a- ketoglutarate and 50 pl plasma. Lysine concentration was calculated based on the rate measured with a 125 pmol 1-1 lysine standard.Intercept, 16.6; slope, 0.66; 1-2 = 0.92; p < 0.001. proportional and concordant, the assay should serve as a reliable indicator of lysine concentration, even at high leucine con- centrations. Conclusion This report describes the modification and validation of a rapid and accurate enzymic determination of plasma lysine. This method is simple, reliable and applicable for use as a bedside method for determining acute changes in plasma lysine concentrations. The method appears to be limited primarily by very high concentrations of leucine.The techniques described can be used for clamping amino acids during experimental, hormonal and nutritional manipulations, and may also be clinically applicable to certain pathological conditions. We thank P. J. Reeds for helpful discussions.Funding has been provided from the U.S. Department of Agriculture, Agricultural Research Service, under Cooperative Agreement No. 58-6250-1-003. The contents of this publication do not necessarily reflect the views or policies of the U.S. Department of Agriculture, nor does the mention of trade names, commer- cial products, or organizations imply endorsement by the U.S.Government. References 1 Tran, N. D., Romette, J. L., and Thomas, D., Biotechnol. Bioeng. , 1983, 25, 329. 2 Romette, J. L., Yang, J. S . , Kusakabe, H., and Thomas, D., Biotechnol. Bioeng., 1983, 25, 2557. 3 Nakatani, Y., Fujioka, M., and Higashino, K., Anal. Biochem., 1972, 49, 225. 4 Frexes-Steed, M., Warner, M. L., Bulus, N., Flakoll, P., and Abumrad, N. N., Am. J . Physiol., 1990, 258, E907. 5 Russell-Jones, D. L., Umpleby, A. M., Hennessy, T. R., Bowes, S. B., Shojaee-Moradie, F., Hophns, K. D., Jackson, N. C . , Kelly, J. M., Jones, R. H., and Sonksen, P. H., Am. J . Physiol., 1994, 267, E591. 6 Flakoll, P. J., Kulaylat, M., Frexes-Steed, M., Hill, J. O., and Abumrad, N. N., J . Parenter. Enteral Nutr., 1991, 15, 123. 7 Brown, L. L., Williams, P. E., Becker, T. A., Ensley, R. J., May, M. E., and Abumrad, N. N., J . Chromatogr., 1988, 426, 370.50 Analytical Communications, February 1996, Vol33 8 Davis, T. A,, Fiorotto, M. L., Nguyen, H. V., and Reeds, P. J., Am. .I. Schaeffer, M. C., Rogers, Q. R., Leung, P. M., Wolfe, B. M., and Physiol., 1993, 265, R334. Strombeck, D. R., Life Sci., 1991, 48,2215. 9 Motil, K. J., Opekun, A. R., Montandon, C. M., Berthold, H. K., Davis, T. A., Klein, P. D., and Reeds, P. J., J. Nutr., 1994, 124,41. Paper 5107851 B 10 Ebner, S., Schoknecht, P., Reeds, P., and Bunin, D., Am. J . Physiol., Received December 4, 1995 1994,266, R1736. Accepted January 2,1996 11

ISSN:1359-7337    

DOI:10.1039/AC9963300047

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analytical Communications, February 1996, Vol33 (51 -52) 51 Nitrate and Nitrite Determination in Complex Matrices by Gradient Ion Chromatography David Crowthera, John M. Monaghanb, Kenneth Cookc and David Garac University, Sheffield, UK S1 I WB t) Health Research Institute, Chemistry Division, Sheflield Hailam University, Sheffield, UK S1 IWB c Dionex (UK) Ltd., Camberley, Surrey, UK Enviromental Research Centre, Chemistry Division, ShefSield Hallam Simultaneous determination of nitrate and nitrite in human serum has been accomplished using gradient ion chromatography and direct UV detection, with centrifugal ultrafiltration as the only pretreatment.The method has advantages over previous techniques in both ease of use and cost per assay. The sensitivity is adequate for use as a monitor of nitric oxide induced changes in serum nitrate and nitrite levels.A simple and reliable method for determination of nitrate and nitrite in human serum is needed because of the immense interest in the role of nitric oxide as a controller of vascular tone, and the implication of faulty NO balance in many diseases.' Determination of nitrate and nitrite in simple aqueous media may be carried out by any of several methods,2 but is most frequently performed by spectrophotometric assay of nitrite.In this method, a diazonium salt formed by reaction of nitrite with, for example, sulfanilamide is coupled with a chromophore such as N-( 1-napthy1)ethylenediamine to yield an azo dye of high molar absorptivity. The method has high sensitivity and works well in simple matrices such as water.For detection of nitrate, reduction to nitrite must be carried out before the spectrophoto- metric steps, and the result from a separate nitrite determination subtracted from the total. The reduction is most often performed with cadmium powder. In complex matrices such as biological extracts, however, the cadmium reduction step is often subject to interference. One standard method for wastewater3 advocates the use of hydrox- ylamine as an alternative reductant, but we have found this to be ineffective in serum. Several strategies have been proposed to overcome the difficulty of nitrate reduction in serum, either involving activation of the cadmium surface4,5 or use of an enzymic reduction step.6 Pretreatment to reduce the interference is also advocated.4 These approaches, however, add con- siderably to the cost and complexity of the determinations.The investigation of alternative methods for this matrix is therefore of considerable interest. Among the possible tech- niques, ion chromatography has gained considerable acceptance for the analysis of simple anions and cations. In relatively simple matrices, simultaneous determination of a range of common anions including nitrate and nitrite can be achieved with a run time of a few minutes.Determinations of nitrite and nitrate in serum and other biological media by ion chromatog- raphy have been reported,6-" but have been subject to several problems. The standard method of detection for ion chromatog- raphy is conductivity, which gives response to all eluted ions.Use of this in serum, however, requires pretreatment to reduce the chloride concentration which, at approximately 0.1 moll-1, is some four to five orders of magnitude higher than typical levels of nitrite. Nitrite elutes close to the chloride, and is often swamped by the chloride response if the latter is not removed by pretreatment.Detection by direct UV absorption, which will not record chloride, is an alternative to chloride removal, but the column packings commonly used for anion analysis are designed for high selectivity at the expense of capacity. The result is that, with sample loadings sufficient to give a reasonable response from nitrite, peak shapes and positions are distorted by the effects of column overload from the (invisible) chloride peak.We have developed a method which overcomes these problems by use of ( i ) direct UV detection to avoid recording of chloride; (ii) a chloride concentration gradient for elution to eliminate the distortion due to the sample chloride peak which is seen with other eluent ions; and (iii) a high capacity column to prevent overload.This combination of factors is novel and results in a determination for which the only pretreatment necessary is a simple centrifugal ultrafiltration to remove macromolecules. Experimental Fig. 1 shows the chromatogram obtained from a typical sample of human serum. The serum was centrifuged through a 3 kDa molecular mass cutoff filter (pre-washed Centricon-3; Amicon, Stonehouse, Gloucestershire, UK) before being loaded onto a Dionex Carbopac PA-100 guard column (Dionex UK, Camber- ley, Surrey), followed by a 4 X 250 mm analytical column of the same material.Peaks were identified by comparison of retention times with standards. Samples spiked with nitrite and nitrate showed superposition of the spike peak with the native peak, indicating that the peak identities were correct.The elution regime was a linear gradient of chloride concentration from 0.12 to 0.30 moll-' over 9 min, followed by I l l 2 4 6 8 Timelmin Fig. 1 direct UV absorption at 214 nm. Conditions as in text. Ion chromatogram of ultrafiltered human serum. Detection by52 Analytical Communications, February 1996, Vol33 4 min washing with 0.3 mol 1-1 chloride then re-equilibration with 0.12 mol 1-1 chloride for 6 min, provided by a Dionex 3000 quaternary gradient pump.The flow rate was 1.0 ml min-1. TRIS buffer (5 X 10-3 moll-1, pH 7.5) was present in all eluents. The water used was obtained by distillation then de-ionisation (Milli-Q, Millipore, Watford, UK), with a mini- mum resistance of 18 MSZ. HPLC grade sodium chloride (Merck, Poole, Dorset, UK) was used to produce the chloride solutions.All tubing was polyether ether ketone (PEEK) and metal contact was reduced to a minimum to avoid corrosion problems. The system was flushed with water at the end of each session. Ultraviolet detection was at 214 nm (Spectromonitor 111, Thermo-LDC, Warrington, Cheshire, UK), and chromatograms were recorded, processed and displayed by a Minichrom data station (VG Data Systems, Altrincham, Cheshire, UK).Peak areas were calculated from a baseline drawn to either side of the relevant peak. Results and Discussions Limits of detection for both nitrate and nitrite were approxi- mately 2 X 10-7 mol 1-1 (10 and 13 ppb, respectively). The calibration graph was linear up to 2 X 10-4 moll-' (the highest concentration used in the calibration set).The relative standard deviation for nitrite, as determined by replicate injections of serum samples, was 5% at the mean concentration (3.7 X 10-6 moll-1) found for a set of 200 random adult serum samples, and for nitrate was 4% at the mean adult level (4.0 X 10-5 moll-'). Spike recoveries, calculated from additions of equal masses of nitrite and nitrate, averaged 106%.In these, however, the nitrite value averaged 76% whereas the nitrate value averaged 137%, suggesting oxidation of nitrite to nitrate by a serum component; further investigations of this are under way. The average nitrite to nitrate molar ratio in the random adult serum set was 0.09. In contrast to previously described methods for determination of nitrite and nitrate in serum and similar complex matrices, this method requires minimal pretreatment, gives separate readings for each ion from one experiment, and is quick and inexpensive to perform.Sensitivity is adequate and selectivity is good. Work is in progress in applying the method to the investigation of nitric oxide function in various diseases. References 1 2 3 4 5 6 7 8 9 10 11 Knowles, R.G., and Moncada, S., Biochem. J., 1994,298,249. Sah, R. N., Commun. Soil Sci. Plant Anal., 1994, 25, 2841. Methods for the Examination of Waters and Associated Materials, HM Stationery Office, London, 1992, vol. 40, p. 36. Green, L. C., Wagner, D. A., Glogowski, J., Skipper, P. L., Wishnok, J. S., and Tannenbaum, S. R., Anal. Biochem., 1982, 126, 131. Gutman, S. I., and Hollywood, C. A., Clin. Chem. (Winston-Salem, N.C.), 1992, 38, 2152. Lyall, F., Young, A. and Greer, I. A., Am. J . Obstet. Gynecol., 1995, 173, 714. Lippsmeyer, B. C., Tracy, M. L., and Moller, G., J. Assoc. Of. Anal. Chem., 1990,73,457. Boermans, H. J., Am. J. Vet. Res., 1990, 51, 491. Marheni, Haddad, P. R., and Taggart, A. R., J. Chrom., 1991, 546, 221. Bianchi, E., Bruschi, R., Draisci, R., and Lucentini, L., Z. Lehensm.- Unters. Forsch., 1995, 200, 256. Everett, S., Dennis, M. F., Tozer, G. M., Prise, V. E., Wardman, P., and Stratford, M. R. L., J . Chrom., 1995, 706, 437. Paper 510791 3F Received December 5,1995 Accepted January 3,1996

ISSN:1359-7337    

DOI:10.1039/AC9963300051

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analytical Communications, February 1996, Vol33 (53-55) 53 Nitrate-selective Electrodes Containing an Amino Acid Betaine as Sensor Jim Bravena, Les Ebdona, Peter Suttona, Nicholas C. Framptonb and David Scholefield" 0 Plymouth Analytic-a1 Chemistry Research Unit, Department of Environmental Sciences, University l.f Plymouth, Drake Circus, Plymouth, UK, PL4 8AA UK, PL25 3JL Tlie West Company (UK) Ltd., Bucklers Lane, St.Austell, Cornw1al1, ICER, North Wykt., Okehampton, Devon, UK, EX20 2SB A novel type of nitrate-selective electrode is reported in which the sensor is glycine betaine hydrochloride. A membrane, hot-pressed from a mixture of glycine betaine hydrochloride (sensor 5 % m/m), dicumyl peroxide (initiator, 7.5% m/m), 2-nitrophenyloctyl ether (mediator, 41.5% m/m) and the acrylonitrile polymer Krynac (50.75) (46% m/m), gave a Nernstian response to nitrate from 0.1 to 1 X kr63-,CI-, of 9 x 10-4 for 0.1 mol dm-3 chloride. The figures of merit and the limit of detection, 1 X mol dm-3, were comparable to those for an established commercial electrode.Glycine betaine also functioned well as the sensor, provided that the membrane was treated with 5 mol dm-3 hydrochloric acid for 24 h.Possible correlations between nitrate selectivity and the role of betaines in marine plants are suggested. mol dm-3 with a selectivity coefficient, Experimental Analytical-reagent grade chemicals and AnalaR water [Merck (formerly BDH), Poole, Dorset, UK] were used unless stated otherwise. Tetrahydrofuran (THF) (Merck) was refluxed over potassium metal (Aldrich, Poole, Dorset, UK) and freshly distilled prior to use.The copolymer of acrylonitrile and butadiene, Krynac 50.75 (Bayer International SA, Newbury, Berkshire, UK) with 50% acrylonitrile content was purified by dissolving 15 g in 75 ml of THF and reprecipitating in 200 ml of cold methanol (Merck, SpectrosoL grade). The solvent mediator 2-nitrophenyloctyl ether (2-NPOE) (Fluka, Glossop, Derbyshire, UK) was used as received, as was dicumyl peroxide (DCP) (Merck, laboratory-reagent grade), glycine betaine (Aldrich) and glycine betaine hydrochloride (Aldrich).Triallyloctylammonium bromide was synthesized as pre- viously described? In two previous papers,IJ a range of nitrate selective mem- branes was described in terms of synthesis of the ion exchanger, fabrication of the polymeric membrane and performance of the corresponding nitrate-selective electrode.The membrane with the longest life and giving the best electrode performance was obtained using a membrane fabricated from an acrylonitrile- butadiene copolymer with a 50% acrylonitrile content and containing 6.5% m/m covalently bound triallyloctylammonium bromide as sensor ion exchanger.The electrode had a lifetime in excess of 500 d and the response to nitrate was Nemstian in the range 1 X 10-'-1 X mol dm-3 of nitrate. The LOD was 4.5 X 10-5 mol dm-3 and the selectivity coefficient for chloride over nitrate (kG3-,cl-) was 5.3 X 10-3. In addition to an on-going detailed mechanistic study aimed at improving the performance of the membrane described above, a limited screening programme was undertaken of qua ternary ammonium compounds known to occur naturally, with evaluation of their performance as sensor molecules.A simple entrapment of the sensor molecule within the membrane was used instead of the covalent attachment previously employed since it was most unlikely that the natural compounds would contain the ally1 groups required for covalent bonding to the polymer.Among the compounds selected were the amino acid betaines (CH3)3 RCHR COO-, which are found in plants, particularly in the halophytes,' and marine algae.4.5 The performance of electrodes fabricated using glycine betaine hydrochloride and glycine betaine as sensor molecules were compared with a commercially available nitrate ion-selective electrode and a slightly improved version of the electrode fabricated with triallyloctylammonium bromide as sensor as described above.Preparation of Membranes Membranes were fabricated as previously described' by dissolving the polymer in THF and adding the DCP and sensor compound with a small volume of methanol together with the solvent mediator. The solvent was removed under vacuum prior to hot pressing between Melanex film (ICI Films Division, Dumfries, UK) at an elevated temperature in a steel mould. A hydraulic press was utilized to press the membranes to a thickness of 0.3 mm at a temperature of 150 "C for 10 min to effect cross-linking of the polymer. The composition of the membranes prepared was as follows: ( i ) Glycine betaine membrane: 5% m/m glycine betaine; 7.5% m/m dicumyl peroxide; 41.5% m/m 2-nitrophenyloctyl- ether; 46% m/m Krynac (50.75).(ii) Glycine betaine hydrochloride membrane: 5% m/m glycine betaine hydrochloride; 7.5 % m/m dicumyl peroxide; 41 .S% m/m 2-nitrophenyloctyl ether; 46% m/m Krynac (iii) Triallyloctylammonium bromide membrane: 6.5% m/m triallyloctylammonium bromide; 7.2% m/m dicumyl peroxide; 39.7% m/m 2-nitrophenyloctyl ether; 46.6% m/m Krynac (50.75).(50.75). Electrode Evaluation Discs were punched from the pressed membrane and condi- tioned for 7 d in a 0.1 mol dm-3 potassium nitrate solution to effect replacement of Br- with NO'- anions and confer nitrate sensitivity. The membrane was assembled into the tip of a commercially available electrode body, IS 560 (Philips Analyt- ical, Cambridge, UK) and the internal filling solution was a54 Analytical Communications, February 1996, Vol33 mixture of 0.1 mol dm-3 potassium nitrate and potassium chloride solutions (1 + 1).The emf measurements were made by a digital voltmeter (Model PW 9409, Philips Analytical) and the electrochemical cell was completed by a double junction reference electrode (Model 90-02, Orion Research, Cambridge, MA, USA) with 0.04 mol dm-3 ammonium sulfate as the outer filling solution.Potassium nitrate standards were prepared using AnalaR water with the addition of 1 X 10-2 mol dm-3 potassium hydrogen phosphate solution as an ionic strength adjuster, held at 25 f 0.5 "C and stirred magnetically during the emf measurement. Selectivity coefficients (kg3-, cl-), calculated by the 18 mV method,6 were determined for a number of membranes using 1 X 10-1 rnol dm-3 potassium chloride.Comparisons were made with the performance of a com- mercially available nitrate selective electrode (Model IS 561-NO3, AT1 Unicam, Cambridge, UK). Results A comparison of the electrode responses for the membranes is given in Table 1.It is clear that the membranes containing glycine betaine hydrochloride and triallyloctyl ammonium bromide give ac- ceptable Nemstian responses and compare favourably with the commercial product in terms of LOD of NO3- ions and selectivity coefficient against C1- ions. The similarity in performance of the membranes is demon- strated more clearly in Figs. 1 and 2. The membrane containing betaine as the sensor gave a sub- Nemstian response.However, if this membrane was soaked in 5 mol dm-3 hydrochloric acid for 24 h a dramatic improvement in electrochemical performance was observed (Table 1 and Fig. 3). The electrode response times were approximately 5 s for solutions of 1 X 10-4 mol dm-3 and above; for more dilute solutions the response time was 1-2 min.Discussion Both the betaine hydrochloride- and hydrochloric acid-treated betaine membrane exhibited near-Nernstian response (-53 and -52 mV decade-', respectively) and a linear working range over four orders of magnitude. The LOD observed with the betaine hydrochloride-containing membrane was 1.1 X 10-5 mol dm-3 which is comparable to that obtained from the commercial electrode (0.8 X 10-5 rnol dm-3).More signifi- cantly, the selectivity towards chloride, probably the most important of all interferences in nitrate measurements, was similar to that of the established commercial electrode, see Fig. Table 1 Electrode response of hot-pressed membranes Slope/mV LOD/ Membrane decade-'" mol dm-3 k@,-, ~ 1 - Betaine Betaine hydrochloride Commercial membrane HC1-treated betaine membrane Triallyloct yl ammonium bromide -29 nd nd -53 1.1 x 10-5 ( 10-1 mol dm-3 Cl-) -58 0.8 x 10-5 (10-1 mol dm-3 Cl-) 9 x 10-4 8 x -52 6.5 x 10-5 nd 1.15 x 10-4 mol -56 0.95 x 10-5 dm-3 C1-) * Measured over the range 10-1 to 10-4 mol dm-3 nitrate.2. This highly promising performance has been obtained from a membrane where the sensor was not covalently bound to the polymer and cross-linking the sensor to the membrane may be expected to further enhance performance.Nitrate ion-selective electrodes in normal laboratory use generally have a quaternary ammonium salt containing four long n-alkyl chains as sensor ion-exchanger trapped in a PVC membrane. Similar performance electrodes have been produced by covalently bonding the quaternary ammonium ion-ex- changer to a suitable polymer.1v2 In all cases a significant amount of hydrophobicity arising from the long n-alkyl chains seems to be required.Glycine betaine hydrochloride, (CH3)3&CH$OZH C1-, is the simplest of all the amino acid betaines and the only one which we have investigated to date. It lacks any hydrophobic character. It is therefore somewhat unexpected that it should perform so well, and match the commercial and other electrodes in LODs, selectivity against chloride ions and response time.Clearly an ion-exchange mechanism is involved as betaine alone gave a 400 344 > 288 E h .- c 01 232 - s 176 10.~ 0.001 0.01 0.1 1 LOS"O31 Fig. 1 Variation of membrane potential with concentration of nitrate ions: triangle, commercial membrane; circle, triallyloctylammonium bromide; and square, betaine hydrochloride.400 344 > 288 E > .- .c) 01 232 - - s 176 10.~ 0.001 0.01 0.1 1 LOS"O,-l Fig. 2 Effect of 0.1 mol dm-3 C1- ions on the variation of membrane potential with concentration of nitrate ions: square, commercial membrane: circle, betaine hydrochloride; diamond, commercial membrane with 0.1 rnol dm-3 C1- ion interferent; and triangle, betaine hydrochloride membrane with 0.1 mol dm-3 C1- ion interferent.Analytical Communications, February 1996, Vol33 55 120 sub-Nerstian response which could be dramatically improved by treating the membrane with hydrochloric acid.Naturally occurring betaines are considered to have a role in plant osmoregulation in an environment high in salt content.3 In .- 1 ' 1 ! 1 1 ( 1 " ~ 1 " ' ~ I ~ ~ 1 1 * 1 1 , I "rn, I T l l l l l q r11w 400 344 , 288 E L .- - a, g 232 - 8 176 view of the performance of the electrode it is interesting to speculate whether their natural role may not also involve some selectivity mechanism in nitrate uptake in some plants. Further investigations on their electrode performance are presently being undertaken. Financial assistance for P.S. from the UK Biotechnology and the Biological Sciences Research Council is gratefully acknowledged. References 1 Ebdon, L., Braven, J., and Frampton, N. C., Analyst, 1990, 115, 189. 2 Ebdon, L., Braven, J., and Frampton, N. C . , Analyst, 1991, 116, 1005. 3 Harborne, J. B., in introduction to Ecological Biochemistry, Aca- demic Press, London, 4th edn., 1992, p. 22. 4 Blunden, G., Cripps, A. L., Gordon, S. M., Mason, T. G., and Turner, C. H., Bot. Mar., 1986, 29, 155. 5 Blunden, G., Smith, B. E., Irons, M. W., Yang, M. H., Roch, 0. G., and Patel, A. V., Biochem. Systematics Ecol., 1992, 20(4), 373. 6 Bailey, P. L., in Analysis with ion-selective Electrodes, Heyden, London, 2nd edn., 1980. Paper 5107376F Received November 9,1995 Accepted December 19, I995

ISSN:1359-7337    

DOI:10.1039/AC9963300053

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analytical Communications, February 1996, Vol33 (57-59) 57 Application of a Novel lminodiacetate Chelating Material to Automated Matrix Separation for Inductively Coupled Plasma Mass Spectrometry Gillian M. Greenwaya, Simon M. Nelmp, and Dagmar Kollerb a School of Chemistry, University of Hull, Cottingham Road, Hull, North Humberside, UK HU6 7RX Cheshire, UK CW7 3BX Fisons Instruments Elemental Analysis, Ion Path, Road Three, Winsford, A novel iminodiacetate chelatftlg reagextt immobilized onto a controlled pore glass st~ppwt tias Men tested as a reagent for trace element matrix separation, priot to analysis by ICP-MS. A commercial autohated systetn was employed in the study.The reageht has been shown to be effective for the analysis of effluent samples at a sampling frequency of 10 h-1.Calibrations from pure and saline solutions showed good linearity and compared well, showing that retention of the selected analytes was unaffected by the matrix. In order to remove residual matrix elements from the column after sample loading a short buffer wash was found to be necessary. The effectiveness of the matrix separation process was illustrated in terms of the 63Cu : 65Cu ratio.This is abnormally high if sodium is present in the plasma, owing to formation of the 40Ar23Na+ interference, and is shown in this study to remain unaltered for saline samples after matrix separation. Initial capacity and recovery data for copper are also shown for the material. The requirement for separation of trace elements from complex matrices prior to analysis by ICP-MS is well established.Traditionally, the methods which have been developed to achieve the separation process have been based on the use of chelating reagents covalently immobilized onto either a poly- meric or silica-controlled pore glass support. In recent years, the trend for ICP-MS has moved from batch separations to on-line procedures. The latter have generally been shown to be faster and less susceptible to contamination.Polymer based resins, particularly of the iminodiacetate (IDA) type, have been extensively used.1.2 These have the disadvantage that they tend to swell and contract under changing solution conditions, such as pH and ionic composition. This can lead to back pressure problems in on-line systems and often necessitates lengthy conditioning of the column material between samples, which reduces sample throughput.For this reason, some workers have assessed the use of controlled pore glass as the chelate supp0rt.39~ This material is dimensionally stable under changing flow conditions, has a reactive surface which can be easily functionalized and is chemically stable over a wide pH range.The active surface of this material also leads to rapid chelation of trace elements, which is a particularly desirable feature in on- line analysis, as analyte residence times in such procedures can be of the order of just a few seconds. An automated sample preparation system was used to illustrate how effective matrix separation could be performed on real samples within an acceptable time scale.The use of a commercial controlled pore glass immobilized IDA (CPG-IDA) material for trace element matrix separation, which combines the benefits of the effective IDA chelating agent with the robust CPG support, will be described in this paper. Experimental Reagents The CPG-IDA material, named as Prosep IDA (Bioprocessing, Consett, Durham, UK) was used as supplied.The preparation and structure of this material are commercially sensitive so cannot be described herein. Calibration solutions were prepared from elemental stock solutions (1000 pg ml-l: SpectrosoL, Merck, Poole, Dorset, UK). Ammonium acetate buffer (Sigma, Poole, Dorset, UK) was prepared from the solid. Adjustment to the buffer pH was made using glacial acetic acid or aqueous ammonia as appropriate.Measurement of pH was made using a portable pH probe (Piccolo 2, Hanna Instruments, Kings Langley, Hertfordshire, UK). Artificial sea-water (Instant Ocean, Aquarium Systems, Mentor, OH, USA) was prepared by dissolving the powder (approximately 330 g) in water (10 1). High purity de-ionized water (18 MQ cm resistivity) (Elgastat UHQ PS, Elga, High Wycombe, Buckinghamshire, UK) was used throughout.Instrumentation ICP-MS measurements were made using a Fisons Instruments PlasmaQuad 2 Plus (Cheshire, UK). The instrument was calibrated before use, using a tune solution containing the elements Be, Co, Y, La, Eu and Bi at 10 ng ml-1 in 5% nitric acid. The transient analyte peaks were acquired in peak jump mode, by first setting an uptake delay to allow for the injected sample to reach the plasma, then collecting data across the peak.The instrument operating parameters are given in Table 1. The Table 1 Operating parameters for the ICP-MS Instrument Fisons Instruments PQ2 Plus Forward power/W Gas flow rated min- 1- Outer gas Intermediate gas Aerosol gas Nebulizer Spray chamber Mode of data acquisition Dwell time/ms Uptake delay/s Acquisition time/s Selected isotopes 1350 13.0 0.1 0.939 De Galan type Glass, water-cooled, 10 "C Peak jumping, 1 point per peak 10.24 18 35 4*Ti, 51V, 55Mn, T o , 6*Ni, h3C~, 64Zn, Y u , *07Ag, 114Cd, '40Ce, 208Pb58 Analytical Communications, February I996, Vol33 matrix separation process was automated using a Fisons Instruments PrepLab system.Matrix Separation Procedure The manifold used in this study is illustrated in Fig.1. It incorporated a glass mini-column (2 cm X 3 mm id) (Omnifit, Cambridge, UK), packed with the Prosep IDA material located in a 3 ml volume loop across the Teflon dual 6 port valve (D- 6-V) on the PrepLab. The PrepLab was used to mix the sample on-line with buffer before loading it into the sample loop, using a peristaltic pump (P2).The contents of this loop were then automatically injected onto the column, allowing the matrix and buffer to pass to waste. A post sample buffer wash was passed through the column via valve 2 (V2) to ensure that any residual matrix was removed, using a second peristaltic pump (Pl). This was followed by a water wash via V2 for 1 min, to ensure that residual buffer was eluted from the system.Finally, the retained elements were eluted into the ICP-MS via valve 1 (Vl) in the opposite direction to the matrix separation flow, thereby yielding sharp elution peaks. With this manifold design, a preconcentration factor of approximately 3 was obtained although the capacity of the material is sufficient to allow a much larger preconcentration.This would, however, lead to an increase in analysis time. Results and Discussion Operating Conditions for Matrix Separation The conditions used in this preliminary study are given in Table 2. At this stage the system has not been fully optimized. An ammonium acetate buffer at pH 5.5 was selected as this has been shown to be effective for the quantitative retention of a range of elements on IDA-based resins.5 The manifold design led to a matrix separation flow rate double that of elution.For ICP-MS, uptake rates between 0.7 and 1.5 ml min-1 are typical. This dictated the choice of elution flow rate and, hence, that of matrix separation. Evaluation of the Matrix Separation Efficiency The influence of residual matrix elements on the column following matrix separation was evaluated in terms of the 63Cu : 65Cu ratio as described previously.6 Residual sodium in the eluted sample led to the formation of 40Ar23Na+ in the plasma which overlapped with 63Cu, giving an erroneously high result for this isotope.Gradual deposition of the residual salt on the sampler and skimmer cones also led to a continuous upward drift in the G3Cu response, even when FI sample introduction was used.Table 3 illustrates that residual matrix problems did not arise in this study. The accepted 63Cu : 65Cu ratio is slightly greater than those measured in this study because of mass discrimination effects and mass bias, owing to sequential scanning of the transient peak. Capacity and Recovery Evaluation for Copper The capacity of Prosep IDA under batch conditions has only been evaluated for Cu at this stage using the compromise matrix separation pH and buffer conditions described in Table 2.A result of 0.5 mmol g-I for Cu was obtained which is significantly greater than has been reported for the 8-hydroxy- quinoline analogue of this material.7 This capacity is suffi- Acid Water Water diently high to ailow high preconcentration factors o i over 100, even with the mini-column arrangement. For Cu, a recovery of 83% was obtained under the compromise pH conditions used.A more comprehensive recovery and capacity study is currently being undertaken. Calibration and Analysis of Real Samples Column U + I Multi-element calibration solutions were prepared across the range 0-10 ng ml-1 in both pure water and artificial sea-water Waste matrices.Both sets of solutions were passed through the matrix separation manifold. Five repeat analyses were made for each solution. The results for both matrices compare well, as illustrated in Table 4, thereby facilitating the use of simple water external calibration for saline water analysis. To validate this preliminary study, two independently analysed real effluent samples were analysed and the results compared. Analyses of saline CRMs are currently being undertaken to further validate the procedure.The independent analysis has been performed by Buffer (1 mol I-') Rear of D-6-v (3mO Waste Fig. 1 The automated matrix separation manifold. Table 2 Operating parameters for matrix separation Flow ratedm1 min-r' Sample volume Buffer conditiom- On-line buffering Post sample buffering Column wash Column eluent Total analysis time Sampling frequencyh- I Matrix separation, 3; elution, 1.5 Sample loop, 3 ml Ammonium acetate, 2.0 mol I-*, pH 5.5, 1.5 ml min-1 Ammonium acetate, 1.0 mol I-l, pH 5.5, 0.5 ml Deionized water, Elga UHQ quality (18 MS2 cm) Nitric acid (Aristar), 1.0 mol 1-1 6 min sample-' 10 ~~~~~ Table 3 Measured isotope ratio of 63Cu : 65Cu for pure and saline water samples using the IDA column Sample description 63Cu/65Cu ratio* Natural 63Cu:65Cu ratio, accepted value Cu in sea-water injected/lO ng ml-1 Cu in 2 mol 1-l HN03 injected/50 ng ml-I Cu matrix separated from spiked pure water Cu matrix separated from spiked sea water Estuarine river water ( = 1.5% salinity), matrix Industrial process sample, matrix separated fidence limit, n = 5 ) , except natural ratio.For all real samples, n = 2. 2.24 5.31 -f; 0.79 2.19 -f; 0.09 ( 5 ng ml-1) 2.09 f 0.10 ( 5 ng ml-1) 2.17 f 0.12 separated 2.10 f 0.04 2.11 f 0.14 * Values quoted with uncertainty (2 standard deviations, 95% con-Analytical Communications, February 1996, VoE33 59 ___ dilution of the samples with direct ICP-MS analysis.The dilution methodology degraded the detection limits for some of the analytes of interest and rendered analysis of 63Cu im- practical. Table 4 Comparison between saline and pure water calibration data for the IDA column Calibrations Mn Spiked pure water- s, (%) at 5 ng ml-' Correlation Sensitivity (counts Detection limit (3s)/ (n = 5) 2.5 coefficient, I' 0.9990 ng-1 mlllO5) 0.02 ng ml-I 0.32 Spiked synthetic sea-water- s, (9%) at 5 ng ml-' Correlation Sensitivity (counts ng-1 Detection limit (3s)/ (n = 5 ) 1.6 coefficient, I' 0.9970 m 1/ 1 05) 0.02 ng ml-' 0.30 c o 2.6 0.9998 0.18 0.002 1.1 0.9994 0.17 0.03 c u 3.9 0.9982 0.09 0.20 1.6 0.9889 0.11 0.64 Ce 1.8 0.9983 0.16 0.05 2.2 0.998 1 0.16 0.18 Cd 4.5 0.9965 0.03 0.1 1 2.7 0.998 1 0.03 0.08 Table 5 Comparison of data obtained for two real effluent samples Sample I* Sample 2" Element A B A B V 38 29 316 290 c o 117 110 624 62 1 Ni 20 <20 223 178 c u 183 180 142 141 Mn 1399 1400 482 46 1 * Results in ng ml-1.For this study, n = 5 . Furthermore, prolonged exposure of the sampling cones to even a diluted sample of this type would result in salting up, ionization suppression and signal drift.Conclusions The Prosep IDA immobilized chelate was shown to be effective for the separation of several trace elements from saline matrix samples. Comparable linear calibrations were obtained from both pure water and sea-water matrices, illustrating that simple water matrix external calibration could be used for quantifying analytes in saline samples.A sampling rate of 10 h-l was achieved. Efficient matrix separation was demonstrated by the unaltered 63Cu : 65Cu ratio measured for saline samples after separation. Trace analytes measured in two effluent samples compared well with the results of an independent analysis, illustrating the practical application of the system. Additional capacity, recovery and saline CRM analyses are currently being carried out to further validate the use of the material for on-line matrix separation.The authors wish to thank Fisons Instruments Elemental Analysis and the EPSRC for their funding of this project and also Dr. P. Clarke (Bioprocessing, Consett, Durham, UK) for supplying the Prosep IDA material. References Ebdon, L., Fisher, A., Handley, H., and Jones, P., J . Anal. At. Spectrom., 1993, 8, 979. Hirata, S., Honda, K.. and Kumamaru, T., Anal. Chim. Acta, 1989, 221, 65. Akatsuka, K., McLaren, J. W., Lam, J. W., and Berman, S. S., .I. Anal. At. Spectrom., 1992, 7, 889. Elmahadi, H. A. M., and Greenway, G. M., J. Anal. At. Spectrom., 1993, 8, 101 1. Heithmar, E. M., Hinners, T. A., Rowan, J. T., and Riviello, J. M., Anal. Chem., 1990, 62, 857. Nelms. S. M., Greenway, G. M., and Hutton, R. C., J . Anal. At. Spectrom., 1995, 10, 929. Sturgeon, R. E., Berman, S. S., Willie, S. N., and Desaulniers, J. A. H., Anal. Chem., 1981,53, 2337. Paper 51078361 Received December 1, 1995 Accepted January 2,1996

ISSN:1359-7337    

DOI:10.1039/AC9963300057

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analytical Communications, February I996, Vol33 (61-64) 61 Quantitative Analysis of Chlorine in Air by Gas Chromatography A. L. Hunt and J. F. Alder Department of Instrumentation and Analytical Science, University of Manchester Institute of Science and Technology, P.O. Box 88, Manchester, UK M60 lQD A GC method was developed for the separation of gaseous chlorine in air using a dual column switching system eluting to a thermal conductivity detector.The limit of detection was found to be 3 ppm; the linear range up to 300 ppm. Initial separation was achieved with a PoraPLOT Q (25 m X 0.53 mm id), column with the separation of the permanent gases being achieved with a Molecular sieve 5A PLOT column (25 m x 0.53 mm id). The resolution of oxygen, nitrogen and carbon dioxide in gaseous samples provides an indication of chromatographic performance.Many methods have been employed for the detection of chlorine. Colorimetric methods, such as that used in Draeger tubes are c0mmon.1~ With Draeger tubes the quantitative measures are subject to error, owing to the lack of precision in reading the distance advanced by the colour change. The colorimetric indicator is o;tolidine adsorbed onto silica gel.The limit of detection is dependent on the o-tolidine analogue and the A,,,, of absorbance. The most sensitive tubes detect chlorine at concentrations as low as 0.001 ppm.' As well as colour reactions, the UV absorption of free and combined chlorine may be employed. Molecular chlorine dissolved in carbon tetrachlo- ride has an absorption maximum at 331 nm and a molar absorption coefficient of 95.7 dm3 mol-I cm-1.This method, however, is relatively insensitive and is consequently not popular for chlorine determination.' Electrochemical methods have been employed for ch1orine;Gg potentiometric determina- tion is based on interfacial electrochemical potentials and mainly involves measurement of potential differences that exist across membranes that separate the test solutions.The potential difference results from transport of the species through the membrane or the binding of charged species to it and interferences result if selective membranes are not employed. The presence of other chlorine containing compounds can result in erroneous results. A continuous chlorine analyser was produced commer- cially,l" based on the production of iodine from PbIz by the chlorine.The iodine was subsequently determined using an iodide electrode. Only one publication has directly tackled the problem of GC analysis of chlorine." The method was used for the determina- tion of both free chlorine and other chlorine containing organic products such as chlorinated hydrocarbons.Known aliquots of chlorine were mixed with acetone and decane (the internal standard). The resultant monochloroacetone was then analysed by GC-flame ionization detection (FID) or GC-MS. Problems were encountered with this method due to production of the dichloroacetone in varying quantities. In the present study, a method was developed for the direct determination of chlorine in a gaseous atmosphere comprising mainly air.The aim was to provide a reference method for a chlorine gas blending apparatus used to test chemical sensors, and then develop a general method for separation and quantification of chlorine and other permanent gases in air. Experimental Gas Chromatography was carried out using a Chrompack 9001 gas chromatograph (Chrompack) equipped with a flame ionization detector (Chrompack) and thermal conductivity detector (TCD) (Chrompack).Injection onto each of the two channels was via two six-port valves (Valco) equipped with a 0.25 ml sample loop. The two configurations used are shown in Figs. 1 and 2. Data acquisition was achieved using a 386DX Opus PC equipped with a SUMMIT (Comus, Kingston-upon-Hull, UK) chromatography control package.All gases were from BOC Special Gases (BOC, London, UK). Helium was used for the carrier and make-up gases. A 1% chlorine in helium mixture was diluted with air in a blending manifold for calibration purposes. Air for the switching valves and FID was cleaned with an RS 724-330 gas regulator and filter system (RS Components, Stockport, UK). All instrument gases were purified using SAMPLE hi OUT CARRIER DETECTOR (A or B) T SAMPLE LOOP ( A or B) Fig.1 Schematic diagram of single column configuration. Column A, PoraPLOT Q; column B, Molecular sieve 5A; detector A, FID: and. detector B, TCD. SAMPLE SAMPLE OUT SiWPLE LOOP THREE WAY UNION Fig. 2 Schematic diagram of dual column switching system.62 Analytical Communications, February 1996, Vol33 Air Purichem towers (Chrompack).Gas flows were measured with bobbin rotameters (Platon, London, UK) calibrated against a bubble flow meter. The blending apparatus produced samples in the range 10000-1 ppm gaseous chlorine in air. 1:10.4 2:00.3 Chlorine 1 General Procedure Initial studies were carried out using the configuration in Fig. 1. The usefulness on the single column with each detector was investigated.Conditions for all single column applications are shown in Table 1. Dual column applications and valve switching timed events are shown in Tables 2 and 3, respectively, and the system set-up in Fig. 2. J i . - Results and Discussion Single Column Studies PoraPLOT Q Using the conditions outlined in Table 1, gaseous chlorine and air were satisfactorily separated. The use of the FID was \ Table 1 GC operating conditions for single column studies Molecular sieve 5A PoraPLOT Q GC conditions TCD FID TCD FID Oven temperature limit/"C Initial oven temperature/"C Final temperature limit/"C Temperature risePC min-1 Detector temperaturePC Injector temperaturePC Sampling valve temperaturer Sample volume/cm3 Carrier gas flow/cm3 min-* Make-up flow/cm3 min-1 Air flow/cm3 min-1 Hydrogen flow/cm3 min-1 180 60 60 0 250 100 'C 200 0.25 17.6 29.15 N/A NIA 280 60 120 30 250 1 00 200 0.25 11.2 8.8 NIA N/A 280 60 60 0 250 100 200 0.25 13.1 31.2 255.2 33.1 Table 2 GC operating conditions for dual column studies GC conditions Level Oven temperature limit/"C Initial oven temperaturePC Final oven temperaturePC Temperature risePC Detector temperaturePC Injector temperature/"C Sample valve ternperature/"C Sample volume/cm3 *Total flow/cm3 min-1 *Total flow less make-up/cm3 min-1 180 60 100 39.9 250 100 200 0.25 20.2 11.2 * These flows relate to the system in by-pass configuration. Table 3 GC timed events control Time/min 0.00 0.00 0.10 0.10 0.11 1.60 5 .oo 5.01 Description Valve 1 to sample Valve 2 to series Valve 1 to inject Start computer integration Reset computer integration Valve 2 to by-pass Stop computer integration Reset computer integration rejected for further study due to its inability to detect the permanent gases.For chlorine the FID gave good peak shapes with a limit of detection of less than 2 ppm chlorine in air. A linear range was achieved up to approximately 1500 ppm.The results were reproducible with respect to retention time and peak area, but the lower range was reduced with extended column use. Fig. 3 shows the resulting chromatogram. Similar studies were attempted on a similar 10 m PoraPLOT Q (Chrompack) column. While the limit of detection was reduced to less than 1 ppm the retention time was only 12 s and the rise in signal caused by the valve firing tended to interfere with the chlorine peak itself.The 25 m column was attached to the TCD, resulting in the separation shown in Fig. 4. The retention time was increased due to the much reduced flow rate required to increase detector sensitivity, A limit of detection of 3 ppm was achieved with a linear range up to 300 ppm. The limit of detection was calculated as three times the standard deviation of the baseline noise.The detectors easily became contaminated with samples containing more than 100 ppm chlorine in air resulting in a 10-fold increase in the limit of detection. This was largely avoided by regularly dismantling the detectors and thoroughly cleaning. When not in use the GC sampled dry clean air and the detectors held at 280 "C.v) Valve firing 2 .- a, x a Chlorine 0:41.5 I 1 Time/min Fig. 3 column (FID). Separation of chlorine in air on a PoraPLOT Q (25 m X 0.53 id) c v) > E 2 0 a, r x a, .- n 2 Separation of 100 ppm chlorine in air on a PoraPLOT Q (25 m X Time/min Fig. 4 0.53 id) column (TCD).Analytical Communications, February 1996, Vol33 63 I c I > E 2 c)) Q) r Y 0, .- n 1 Timelmin Nitrogen Fig.5 0.53 1D (TCD). Separation of air on a Molecular sieve 5A, 25 m PLOT column, The single column results led to the development of the system shown in Fig. 2. The installed system was configured to run in a serial by-pass mode with both columns continuing to elute to the TCD. Tables 2 and 3 show the condition and timed events control respectively for the dual column system.Molecular sieve 5A Fig. 5 shows the separation of the air constituents achieved by the Molecular sieve 5A (Chrompack) column eluting to the TCD. This separation could be implemented by passing the flow eluting from the PoraPLOT Q column to the Molecular sieve 5A for further separation. The molecular sieve column could not be brought into contact with the gaseous chlorine especially in humid air.The heat of the gas sampling valve caused the reaction of chlorine with water as shown in the following reaction scheme: Clz + H20 + HCl + HOCl The products are strongly retained by the Molecular sieve 5A; while this does not harm the column, its presence changes the chromatography and the retention time decreases due to site blocking. Reconditioning of the column restored the original separation.This was achieved by heating the column at 180 "C for 10 h then cooling (10 "C min-1) to 100 "C for 5 h. Drying of the sample before entrance into the GC eliminates the problem, as does separation of chlorine from the test mix prior to the air passing to the molecular sieve. The switching time from serial to by-pass mode was such as to avoid passage of chlorine to the second column.Dual Column Studies The final system employed (Table 2, Fig. 2) resulted in the separation shown in Fig. 6. Valve 1 is the sampling valve. At the beginning of the analysis, a 0.25 cm3 sample is injected onto the PoraPLOT Q column which is connected to valve 2, and the carrier gas supply to it controlled by the channel A flow controls.At the start, valve 1 connects the PoraPLOT Q directly to the Molecular sieve 5A. The air constituents and other permanent gases are poorly contained by the PoraPLOT Q and so rapidly pass to the molecular sieve. Valve 2 switches the flow from the first column directly to the TCD via a three-way union, Oxygen 2:20.5 1 2 Time/min 3 Fig. 6 system. Separation of 15 ppm chlorine in air using dual column switching at the same time the sampling valve switching the channel flow control to column B.In this way the chlorine passes directly from the PoraPLOT Q column to the detector as does the air, which is now separated into its constituents by the Molecular sieve 5A. The flow rates were carefully controlled to avoid coelution of chlorine and oxygen. The air separation served to act as an internal standard.Since the air constituents in the blending apparatus are standard in each sample, nitrogen will be easily seen if there are any changes in the relative retention times of oxygen. Similarly, a peak area ratio of oxy- gen:chlorine will help to show any degradation of system sensitivity. A reduction in the excellent reproducibility can be used as an indicator of system degradation. This needed to be carefully monitored due to the highly corrosive effects of chlorine and its hydrolysis products.The limit of detection of chlorine was unaffected by the use of the Molecular sieve 5A in conjunction with the PoraPLOT Q. The cut time for switching from serial to by-pass mode did affect the limit of detection as well as overall retention times.Careful optimization was required if coelution was to be avoided . Conclusion The dual column system developed proved to be a reliable way of monitoring chlorine levels in air. The method was reprodu- cible over any 24 h period even at low ppm levels, although the single column studies showed that there was a reduction in the linear range with time. The use of the FID was rejected due to its inability to detect the permanent gases.References 1 Leggett, D. J. , Chen, N. H., and Mahadevappa, D. S., Fresenius' J . Anal. Chem., 1994,315,47. 2 Belcher, R., Nutten, A. J., and Stephen, W. I., Anal. Chem., 1954,26, 772.64 Analytical Communications, February 1996, Vol33 Gabbay, J., Davidson, N., and Donagi, A. E., Analyst, 1976, 101, 128. Nicholson, N. J., Analyst, 1965,90, 187. Feniol, M., and Gazet, J., Anal. Chim. Acta, 1988, 209, 321. Matszewski, W., and Trojanowicz, M., Anal. Chim. Acta, 1988,207, 59. Mari, C. M., and Terzaghi, G., Sens. Acfuators, 1989, 15, 569. Liu, J., and Weppner, W., Sens. Actuators, 1992, B6, 270. Galdikas, A., Martunas, Z., and Setkus, A., Sens. Actuators, 1992, B7, 633. 10 Willard, H. H., Menitt, L. L., Dean, J. A., and Settle, F. A., Instrumental Methods of Analysis, van Nostrand, Reinhold, 6th edn., 1981, p. 916. Batlin, F., Dzierzynski, M., Corne, C . M., and Barronet, F., Analusis, 1991, 19, 36. 11 Paper 51065561 Recerived October 5 , 1995 Accepted January 6, I996

ISSN:1359-7337    

DOI:10.1039/AC9963300061

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analytical Communications, February 1996, Vol33 (65-68) 65 Chromatographic Resolution of Enantiomers on a Chiral Stationary Phase Physically Anchored to Porous Graphitic Carbon Lotfi I. Monser and Gillian M. Greenway School of Chemistry, University of Hull, Hull, UK HU6 7RX A chrysene-tartramide derived chiral stationary phase (CSP) was prepared and adsorbed onto the surface of porous graphitic carbon (Hypercarb).The stereoselective properties of the carbon-based chiral stationary phase was evaluated in a pre-packed high-performance liquid chromatography column. The carbon-based CSP was found to be capable of separating the enantiomers of a range of compounds including l,l’-bi(2-naphthol), 2,2-diaminobinaphthalene, 1,5-hexadiene-3,4-diol, benzoin, 1,4-di-0- benzyl threitol, 1,l’- binaphthalene-2,2’- diylhydrogenphosphate, propranolol, nadolol and labetolol using a non-aqueous mobile phase.The addition of ammonium acetate to the mobile phase was found to improve the resolution of enantiomers for l,l’-binaphthalene-2,2’-diylhydrogenphosphate, propranolol, nadolol and labetolol. The amount of chrysene-tartramide adsorbed on porous graphitic carbon was calculated from the cycling process and the value corresponded to 2.6% by mass of porous graphitic carbon.It has been reported that a chiral stationary phase (CSP) in which an (R,R)-N,N’-dialkyltartramide is linked to the surface of silica gel by derivatization is capable of chiral recognition of broad categories of enantiomers.1 This CSP recognizes the molecular chirality of enantiomers through its dual hydrogen bond association, but the presence of silanol groups on the silica surface can be problematical because they may interact with the solute enantiomers by forming hydrogen bonds in competition with the chiral moiety of CSP if care is not taken to avoid the problem.Porous graphitic carbon (PGC) is an alternative non-polar stationary phase for liquid chromatography. As well as having excellent flexibility in terms of the pH at which it can be used (0-14) it also has an almost homogeneous surface and shows good peak symmetry for a variety of solutes.2J The structure of the surface.however, means that large planar molecules are strongly retained on the surface. Although this would normally be a disadvantage Wilkins et al.4 exploited this strong retention by preparing a tartaric acid derivative bound to anthracene.The planar part of the molecule was adsorbed on the carbon surface, thus providing a semi-permanent coating on the column, which was used for the enantiomeric separation of anti-inflammatory agents and tropic acids by supercritical fluid chromatography. Chiral separations have been achieved on PGC in HPLC, but only when a chiral additive is added to the mobile phase.5-15 By employing tartaric acid derivatives as chiral complexing agents, Heldin and co-workers 1 6 y 1 7 separated enantiomeric acids, esters and amino alcohols of moderate hydrophobicity using dynam- ically coated PGC.In comparing different solid supports, they found that high separation factors were observed when using a support with a minimum number of polar groups, namely PGC.The aim of this work was to design a novel chiral selector with a polyaromatic hydrocarbon anchor that would be strongly adsorbed to the graphite surface, thus producing a new chiral stationary phase that would be stable under the majority of mobile phase conditions. The novel compound consisted of 2-aminochrysene linked to an @,I?)-N-alkyltartramide deriva- tive.Experimental Instrumentation A gradient modular HPLC system (Gilson Medical Electronics Middleton, WI, USA) was used for the work with a variable- wavelength UV detector. The detector was set at 225 and 254 nm, with an attenuation of 0.05 AUFS. Samples were introduced via an auto-sampler with a 5 p1 loop (ISS 100 Perkin- Elmer Corporation, Norwalk, CT, USA).The column was packed with ‘Hypercarb’ porous graphitic carbon and was 100 cm long X 4.6 mm id, with 7 pm particle size (Shandon Scientific, Runcorn, Cheshire, UK), and a flow rate of 1 ml min- was used. Reagents HPLC grade solvents were obtained from Fisons (Loughbor- ough, Leicestershire, UK) and Rathburn Chemicals (Walk- erburn, Peebleshire, Scotland).The high purity water used was produced by reverse osmosis and ion exchange (Elgastat UHQ PSI1 system Elga, High Wycombe, Buckinghamshire, UK). The 2-aminochrysene was obtained from Lancaster (Lancashire, UK), N-methylmorphine, ethylchloroformate and isopropyla- mine were from Fluka (Gillingham, Dorset, UK), while (R,R)- tartaric acid and acetic anhydride were from BDH Laboratory Supplies (Merck, Poole, Dorset, UK).The enantiomeric compounds used in this study are shown in Table 1. All other chemicals were of analytical grade and were used without further purification. Synthesis of CSP In order to anchor the chiral selector to PGC, chrysene was selected to attach the tatramide moiety directly. The carbon- Table 1 Enantiomeric compounds No.1 2 3 4 5 6 7 8 9 Name (R,S)-1 ,l’-Bi(2-naphthol) (R ,S)-2,2-Diaminobinaphthalene (&)- 1,5-Hexadiene-3,4-diol (*)-Benzoin (k)- 1,4-Di-O-benzylthreitol (R,S)- 1,l ’-Binaphthalene-2,2’-diylhydrogenphosphate ma-Propranolol N a d o 1 o 1 Labetolol Source Fluka Fluka Sigma Fluka Fluka Aldrich Sigma Zeneca Zeneca66 Analytical Communications, February 1996, Vol33 based CSP (V) used in this work was synthesized as shown in Fig.1. Procedure for Column Coating A solution containing 0.5 mg ml-1 of the synthesized compound [CSP (V)] in tetrahydrofuran was cycled through the column with a flow rate of 1 ml min-l. The amount adsorbed was monitored off-line by UV spectrophotometry until almost no further uptake appeared to occur. The amount of CSP (V) in the reservoir solution before passing it through the column was 50 mg and the amount remaining in the solution after cycling through the column was 24 mg.Chromatographic Parameters Enantiomeric separation on the carbon-based CSP column was investigated by measuring the capacity factor, k = (tR - to)/to where tR is the retention time of the solute peak and to is that of an unretained peak.The selectivity was calculated from the capacity factors of enantiomers 1 and 2; a = kz/kl. The resolution was calculated according to the following equation Rs = 1.18 [ ( ~ R Z - tRi)/(whI 4- fl'h2) where whl and wh2 are the peak widths at half height of the first and second eluted isomers. Results and Discussion A range of aromatic hydrocarbons were investigated to identify a suitable molecule to anchor the chiral moiety.Chrysene was selected as being the most appropriate because it was stable on the carbon surface under a wide range of polar and non-polar solvents. This meant that the molecule would provide good stability for a range of solvents but could be removed if required from the carbon surface. The surface coverage is an important parameter for character- izing the CSP.Complete monolayer coverage of the carbon with chrysene molecules would give the best separation. The surface coverage found in this study shows that the chiral selector is COOH l I COOH I OH 0 strongly adsorbed by graphite from tetrahydrofuran, giving rise to a surface concentration comparable to that of the chemically bonded phases. The amount of uptake of the chiral selector by PGC was determined as described in the column coating procedure and was found to be 26 mg (2.6% by mass of PGC).If the moleplar surface area of the chrysene molecule is taken to be 77 A, this result then suggests a 50.3% monolayer coverage (PGC surface area is 104.8 m2 g-1). Once prepared, the modified column was used to separate the enantiomers of a range of different compounds by HPLC (Table 2).The separation factor observed here typically ranged from 1.1 to 1.3. These degrees of chiral recognition are quite sufficient to ensure satisfactory resolution, as is evident from the R, values ranging from 0.8 to 2.1. An example is given in Fig. 2, which shows the successful separation of the enantio- mers of benzoin and propranolol.In this present system, dual hydrogen bonding between solute enantiomers and the chiral moiety of the CSP is assumed to be essential for the chiral recognition. The primary results obtained with this system encourage further study of chiral separation on PGC. Stability and Reproducibility of the Chromatographic System The stability of the carbon-based CSP was studied by passing a large volume of eluent through the column [approximately 4000 ml of solvents ( - 800 injections)].The capacity factors (Fig. 3) of the enantiomers were stable up to an elution volume of approximately 1500 ml. Above this volume a 9-12% decrease in retention was observed, especially for highly retained enantiomer. Similarly, a decrease in separation factors (Fig.4) was observed for most of the enantiomers. At first, the decrease in retention and separation factors was believed to be due to leaching of the chiral selector from the column. It was, however, found that the original retention and separation factor could be recovered by flushing the column with 250 ml of mobile phase made up of 15% tetrahydrofuran in hexane. These results suggest that leaching of the CSP is not the problem and that, in fact, it is the masking of the attractive interaction sites of the CSP with impurities of highly retained substances during the continuous use of the column that causes the problem. The OAC I f; I I1 V IV Fig.1 methanol. Synthesis of chrysene-@,I?)-tartramide CSP: a, acetic anhydride, sulfuric acid; b, isopropylamine; c, 2-aminochrysene; d, 0.6 moll-' ammonia inAnalytical Communications, February 1996, Vol33 67 - flushing of the column with the mobile phase, described above, therefore overcomes the problem by removing the highly retained substances to leave the chiral selectors active again.The stability of the CSP was also further tested with polar solvents (methanol and methanol-water mixtures) and at pH values of 3, 7 and 12.The results obtained were very similar to those obtained for Figs. 3 and 4, thus showing these factors did not affect the stability of the column. I0 -~ -r 5 Column Regeneration In an attempt to restore the original surface character of the PGC, the CSP-coated column was subjected to elution with tetrahydrofuran. After the adsorbed molecules had been washed out, the column was tested with substituted benzene solutes under the same chromatographic conditions before coating.A complete recovery of the column was seen with no change in ~ -- ~~~~~ ~ Table 2 Chromatographic resolution of enantiomers on the PGC-chrysene based CSP. Mobile phase: A, 95 : 5 hexane-propan-2-01; B, 98 : 2 hexane- dichloromethane; C, 1 mmol 1-1 ammonium acetate in methanol; D, 0.1 moll-' ammonium acetate in methanol; E, 10 mmol l-l ammonium acetate in methanol No.Solute Mobile k'l k'2 (x R, phase 1.22 (S) 1.40 (R) 1.16 0.86 A 4.01 (R) 4.39 (S) 1.10 1.05 B ZNH2 N H 2 0.86 OH @ 2.69 o&v(-) OH 10.1 1.05 2.98 11.2 1.19 (S) 1.4 (R) OH Bo*NH< 11.98 15.05 OH -< 2.4 2.9 1 Hd bH H2NOC HO @ 5.74 7.9 CH3 1.22 0.8 A 1.11 1.18 B 1.12 1.05 A 1.18 1.02 C 1.26 2.1 D 1.2 1.2 E 1.31 1.66 D retention time for a range of test solutes before coating and after removal of the coating.Conclusion The aromatic base (chrysene) of the CSP showed a strong adsorption characteristic as its planar configuration matched with that of the graphite surface and it proved to be successful in anchoring the chiral selector to PGC, yielding a successful I I I I 2 . 0 0 4,00 6.00 8.00 1 .I I I I I I I I 10.00 20.00 30.00 Fig. 2 Enantiomeric resolution of a, benzoin in 98 : 2 hexane-dichloro- methane and b, propranolol in 0.1 moll-1 ammonium acetate in methanol. Column: PGC 100 x 4.6 mm id coated with 26 mg CSP(V); mobile phase flow rate, 1 ml min-*. li/ 0 0 1000 2000 3000 4000 Volume of mobile phasehl Fig.3 Stability of chromatographic system: change in capacity factor with volume of mobile phase passed through column. +, Binaphthol; U, diaminobinaphthalene; A , benzoin; x, binaphthalene-2,2'-diylhy- drogenphosphate; x, propranolol; @. labetolol.68 Analytical Communications, February 1996, Vol33 1.4 I tc 1.2 - 0 1000 2000 3000 4000 Volume of mobile phase/ml Fig. 4 with volume of mobile phase passed through column.For key see Fig. 3. Stability of chromatographic system: change in separation factor carbon-based CSP. This method offers great flexibility, because it is possible to strip off the adsorbed material to give a fresh graphite surface. References 1 2 Dobashi, Y., and Hara, S., J . Org. Chem., 1987, 52, 2490. Gilbert, M. T., Knox, J. H., and Kaur, B., Chromatographia, 1982, 16, 138.3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Knox, J. H., Kaur, B., and Millward, G. R., J. Chromatogr., 1986, 352, 3. Wilkins, S. M., Taylor, D. R., and Smith, R. J., J . Chromatogr. A, 1995,697,5874. Huynh, A., Karlsson, A., and Pettersson, C., J . Chromatogr. A, 1995, 705, 275. Chm, W. C., Micklewright, R., and Barrett, D. A., J. Chromatogr. A , 1995,697,213. Josefesson, M., Carlsson, B., and Norlander, B., J . Chromatogr. A , 1995, 684, 23. Knox, J. H., and Wan, Q. H., Chromatographia, 1995, 40, 9. Gorog, S., and Gazdag, M., J . Chromatogr. B: Biomed Appl., 1994, 659, 51. Josefesson, M., Carlsson, B., and Norlander, B., Chromatographia, 1993, 37, 129. Karlsson, A., Luthan, K., Pettersson, C . , and Hacksell, U., Acta Chem. Scand., 1993,47,469. Stefansson, M., and Hoffmann, K., Chirality, 1992, 4, 509. Karlsson, A., and Pettersson, C., Chirality, 1992, 4, 323. Karlsson, A., and Pettersson, C., J . Chromatogr., 1991, 543, 287. Clark, B. J., and Mama, J. E., J. Pharm. Biomed. Anal., 1989, 7, 1883. Heldin, E., Lindner, K., Petersson, C., Lindner, W., and Rao, R., Chromatographia, 1991,32,407. Heldin, E., Huynh, N., and Petersson, C., J . Chromatogr., 1991,585, 35. Paper 51076850 Received November 24,1995 Accepted January 12,1996

ISSN:1359-7337    

DOI:10.1039/AC9963300065

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analytical Communications, February 1996, Vo133 (69-70) 69 Optical Characterization of Copper Oxides Formed by Oxygen Plasma N. Bellakhala, K. Draoua, J. L. Brisseta and M. Lengletb a Laboratoire d' Electrochimie Interfaciale et de Chimie Analytique, UFR des Sciences, F-76821 Mont Saint Aignan, Cedex, France h Laboratoire d'rlnalyse Spectroscopique et de Traitement de Surface des Mate'riaux, UFR des Sciences, F-76821 Mont Saint Aignan, Cedex, France Copper foils are oxidized by an inductively coupled low pressure oxygen plasma.The resulting oxides formed are identified by their FTIR and photoluminescence spectra. The nature of the oxidized layer and its thickness depend on the time of exposure to the plasma and on the operating conditions. During the past few years, semiconductor oxides have been investigated extensively for their unique physical and chemical properties in connection with microelectronics, photochemistry, heterogeneous catalysis, and the production of low-temperature superconductors, etc.Copper oxides are the subject of increased interest due to their use as catalysts in methanol synthesis or as components in solar cells.1.2 They are usually prepared by thermal oxidation of copper,3 electrodeposition or reactive sputtering for Cu20,4,5 sputtering or thermal chemical vapour deposition (CVD) for CuO.6 The chemical properties of a plasma are investigated in the case of an oxygen rf plasma in the range 100-200 Pa, which is suitable for studying the oxidizing properties.The literature reports on the oxidizing properties of oxygen plasma, and for example, the oxidizing properties of the oxygen atoms are applied in the plasma treatment of YBaCuO thin films.7 In this case the particles inductively created in this non-equilibrium luminescent electric discharge (i.e., excited molecules, atomic oxygen, ozone, molecular ions and atomic ions), react at the copper surface and form copper oxides.After treatment, the copper oxides yielded can be characterized by optical methods, such as FTIR and photoluminescence spectroscopy and the film thickness measured by the interferometric method. Experimental The plasma device was composed of a tubular pyrex reactor. An oxygen plasma was produced by means of an rf generator with a frequency of 13.56 MHz which was inductively coupled to the reactor.The rf power could be varied in the range 100-2200 W. The vacuum (200 Pa) was sustained by a primary pump. The sample holder with a stainless-steel quenching head was immersed in the vertical pyrex reactor. The influence of some parameters of the plasma oxidation process were shown to be important for the growth rate of the copper oxide layers: the distance between the copper target and the rf coil was maintained at 10 cm, in order to avoid rf heating effects, the transmitted power was maintained at 200 W and the flow rate of oxygen at 0.25 1 min-1.The specular IR spectra were recorded on a Nicolet (Madison, WI, USA) FTIR 7 10 spectrophotometer (analysed range: 5000-225 cm-1, angle of incidence 80 degrees: RS 80) and the photoluminescence spectra on a Jobin-Yvon (Long- jumeau, France) 3C spectrophotometer (analysed range: 300-1200 nm).The copper samples (Cu, 0.3 mm thick with 1.1 cm2 surface area) were mechanically polished with different grinding papers (400,800 and 1200 grade) washed in absolute ethanol and dried before treatment by the oxygen plasma. Results and Discussion Photoluminescence Spectrosocopy Czanderna and co-workers have obtained a composition C U O ~ .~ ~ (or Cu302) by low-temperature oxidation (120-160 "C) for polycrystalline (industrial grade) copper.* The most sig- nificant aspect of the X-ray and electron diffraction measure- ments is that C U O ~ , ~ ~ yields the same diffraction lines as those of Cu20 and only those lines. This composition is a gross defect structure of Cu20 corresponding to one copper vacancy per cell unit.The copper(1) oxides (Cu20, Cu302) are the only forms of copper that can be identified by photoluminescence spectro- scopy. Lenglet and co-workers confirmed the differences between the optical properties of these oxides.9 Emission photoluminescence spectroscopy spectra were realized with an excitation wavelength of 530 nm.Fig. 1 shows the spectra recorded for samples treated for 15 and 20 min by the oxygen plasma. The spectrum of the sample treated for 15 min exhibits two emission bands: an intense luminescence emission at 760 nm and a second one at 820 nm, due respectively to the presence of Cu302 and Cu20 in the oxide film. For the samples treated for 20 min, the band relevant to Cu302 disappears and only the band at 820 nm can be observed (Cu20).The disappearance of the emission at 760 nm shows that Cu302 was initially present in the oxide layers and later oxidized into CuO. The photo- luminescence spectra reveal that the formation of CuO can be obtained by oxidation of both Cu20 and Cu302. t - J cu,o 600 700 800 900 Unm Fig. 1 Photoluminescence spectra (A,,, = 530 nm) of copper(1) formed after oxygen plasma treatment of samples copper (transmitted power, 200 W; flow rate of oxygen, 0.25 1 min-1, pressure, 200 Pa).(a) 15; (b) 20 min.70 Analytical Communications, February 1996, Vo133 Infrared Reflectance Spectra Fig. 2 shows the spectra of samples treated for 15 min or more, which provided IR reflectance spectra which were recorded under an incidence angle of 80 degrees.The spectra of samples treated for 15 rnin show an intense band at 650 cm-1 which can be attributed to the longitudinal optical vibrations (LO) of Cu20 and Cu302 (Cu302 has similar vibrations to those of Cu20). The intensity of this band decreases with the treatment duration when the thickness (d) of copper oxide layers increases and the band position moves to the high energy.For layers thicker than 0.25 vm, which correspond to layers treated for more than 25 min, a new band can be observed at 610 cm-I. This band is attributed to the transverse optical vibrations (TO) to the oxide 40 min (d = 0.98 prn) 35 min (d = 0.77 pm) 30 min (d = 0.50 pm) 25 rnin (d = 0.35 pm) 20 rnin (d = 0.22 pm) 15 rnin (d = 0.15 pm) 700 600 500 400 v/cm-’ Fig.2 Infrared reflectance spectra of copper oxidized in oxygen plasma (transmitted power, 200 W; flow rate of oxygen, 0.25 1 min-1, pressure, 200 Pa), at 80 degrees incidence angle. CuzO alone. The presence of CuO can be identified by low intensity bands at 605, 530, 470 cm-l (TO) present on the spectra of samples treated for at least 30 min. After oxidation the film thickness was measured by an interferometric method.Most of the spectra maxima and minima come from the interference on the thin copper oxide layers. The measured thickness of oxide layers as a function of the treatment time are shown in Fig. 2. The thickness of the layer increases significantly with the treatment time. Conclusions Optical methods have proved to be of great assistance in the characterization of the copper oxides formed on a copper surface after plasma treatments of various durations.Cu302 and CuzO can only be identified by photoluminescence spectros- copy. The oxides formed and the film thickness depend strongly on the working conditions of the plasma. The mechanism of the oxidation remains to be determined, since we have only identified the oxides formed as reported and the gaseous species present in the plasma by means of their emission lines: atomic oxygen was characterized by the 777.4 nm 0 (5P) triplet and the 844.6 nm 0 (3P) singlet.References Chinchen, G. C., Denny, P. D., Jennings, J. R. Spencer, M. S., and Waugh, K. C., Appl. Catal., 1988, 1, 36. Horntrom, S. E., Kadsson, S. E., Roos, A., Westerstrandh, B., and Kauf, A., Solar Energy Mater., 1984,9, 367. Lenglet, M., Kartouni, K., and Delahaye, D. J., Appl. Electrochem., 1991, 21, 697. Drobny, V. P., and Pulvrey, D. L., Thin Solid Films, 1979, 61, 89. Beensh-Marchwicka, G., Krol-Stepniewska, L., and Slaby, M., Thin Solid Films, 1982, 88, 33. Holzschuh, H., and Suhr, H., Appl. Phys. A , 1990, 51, 486. Yoshida, A., Tamura, H., Morohashi, S., and Hasuo, S., Appl. Phys. Lett., 1989, 55, 2354. Wieder, H., and Czanderna, A. W., J. Phys. Chem., 1962,66, 816. Beucher, E., Lefez, B., and Lenglet, M., Phys. Status Solidi A , 1993, 136, 139. Paper 510681 5K Received October 16, I995 Accepted January 2,1996

ISSN:1359-7337    

DOI:10.1039/AC9963300069

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analytical Communications, February 1996, Vol33 (71-73) 71 Analysis of Nanogram Spots of Ink by Fourier Transform Infrared Microscopy and Spectral Stripping David Crowther, Markus Best and Claudia Wohlfarth Environmental Research Centre and Materials Research Institute, Shefield Hallam University, Shefield, UK SI I WB Dot matrix printing inks for industrial printers consist of solvent, binder, dye and various minor additives.We have used FTIR microspectroscopy to examine the composition of these inks from the analysis of single dots of less than 1 mm diameter. Good spectra were obtained by reflection-absorption on metal surfaces and by ATR on plastic surfaces. Paper substrates have proved more troublesome. Analysis of the printed inks was performed by an iterative process of component identification through library searching and spectral subtraction. Some bandshifts were observed which complicated this process, but two or three successive identification-subtraction cycles could often be made.A comparison of real sample spectra with synthetic spectra produced by addition of pure component spectra was also useful in highlighting spectral differences, indicating component interactions in the formulated inks.Long-throw dot-matrix printers are used extensively for low resolution printing on high speed production lines. Examples can be found in any supermarket in the form of date codes on merchandise. Each dot is formed from approximately 1.5 nl of ink, of which about 70% is volatile solvent, 15% binder, 10% dye and 5% other components. The printed spot, of perhaps 0.5 mm diameter, contains about 500 ng of dry material.' Our aim was to see how much information concerning the ink composition can be extracted from the non-destructive FTIR analysis of a single spot, preferably in situ, on the variety of substrates (e.g., glass, metal, paper and various plastics) that are commonly encountered.This information could be useful in forensic applications or in process developments.Previous reports on the FTIR analysis of inks have concerned other types of ink.24 The IR spectrum of part of a spot was recorded and the principal component (the binder) identified by library search- ing. A reference spectrum of this component was subtracted from the spectrum of the ink and a second identification- subtraction cycle performed for the next most abundant component (the dye).A final search was used to attempt identification of the minor components, e-g., plasticizers. Experimental The sampling technique used was chosen to suit the nature of each substrate. For inks which could easily be scraped off their substrate (e.g., glass), spectra were recorded by transmission following thinning and flattening of the sample in a diamond window compression cell (Spectra Tech, Warrington, UK).For ink spots printed on metal surfaces the reflection-absorption technique was used in situ. An ATI-Mattson Galaxy 4020 FTIR spectrometer (ATI-Unicam Ltd., Cambridge, UK) linked to a Spectra Tech IR-Plan advanced analytical IR microscope was used to record spectra with these sampling techniques, and apertures were used to limit the sampling area to less than 100 pm in the largest dimension.For inks on paper (results not shown here) the apertures were removed to allow collection of the maximum amount of reflected radiation (principally diffusely reflected). For inks on plastic substrates, spectra were recorded on a Nicolet Magna FTIR spectrometer fitted with a Nic-Plan microscope having a ZnSe attenuated total reflectance (ATR) objective.Results and Discussion Transmission Spectra of Dried Ink Flakes A commercial black ink of known composition was dried on to a glass slide, then flakes of dried ink were scraped off and transferred to the compression cell. After squashing to yield a suitable area and thickness of sample, spectra were recorded in transmission mode using circular apertures allowing a sampling area of 100 pm diameter.The spectrum of the complete black ink (Fig. 1 a) indicated a major contribution from the nitrocellulose binder, a reference spectrum of which (Fig. 1 b) was subtracted to yield the difference spectrum in Fig. 1 c. This indicated the presence of the black dye, a reference spectrum of which (Fig.1 d) was subtracted to yield the spectrum in Fig. 1 e. This final spectrum yielded a reasonable match to the plasticizer (reference spectrum Fig. 1 0. Transmission is the most reliable mode in which to obtain infrared spectra, and so these data, with good sample area and thickness, provide a marker for what can be achieved by microspectroscopy and spectral stripping.Reflection-Absorption Spectra From a Dot on the Base of a Beer Can The impact of a travelling drop of ink results in a spot of dried ink which is thin in the centre, with a thicker periphery. The thickness of the ink layer was less than that of the scraped flakes and absorbances were lower, despite the reflection-absorption sampling in which the effective pathlength is double the sample thickness.Fig. 2 shows spectra obtained from a section of a single dot from the date code printed on the base of an aluminium can. The best results were obtained by masking a small rectangular area at the edge of the dot, where the ink was thickest. Despite this small sampling area successive subtrac- tions (Fig. 2 a-f), performed as for the dried ink flakes of Fig.1, yielded spectra in which the principal features of each component could still be seen. The mass of dried ink sampled, based on an estimate of the percentage area sampled, was only72 Analytical Communications, February 1996, Vol33 about 10 ng. This was in contrast to the dried flakes, where a greater area of a thicker sample was examined. a C -F w 2 3500 3000 2500 2000 1 500 1 000 Wavenumber Fig.1 Spectra of black ink recorded by transmission in a diamond window compression cell; a, ink; b, binder reference spectrum; c, ink minus binder; d, dye; e, ink minus binder minus dye; f, plasticizer. 3500 3000 2500 2000 1500 1 000 Wavenumber Fig. 2 base of a can. For key, see legend to Fig. 1. Spectra of black ink recorded by reflection-absorption from the Micro-ATR Spectra From a Dot on a Polystyrene Sheet Inks on plastic substrates are usually difficult to remove mechanically without scraping off some of the substrate as well.Solvents may sometimes be used to remove an ink spot, but there is the danger of substrate coating, bulk polymer or plasticizer being removed as well. These substrates are, however, ideally suited to micro-ATR sampling because the substrate surface deforms just enough to allow good, even contact across the sampling area (usually between about 10 and 50 pm in diameter). A spectrum of the same black ink was obtained with an ATR objective on a different instrument to that used for the reference spectra (Fig.3a) and a mathematical correction was applied to the spectrum to compensate for the variation in sampling depth with frequency.Despite this, and the fact that the reference spectra were obtained in transmission mode, reasonable results were obtained. Successive identifications of binder and dye were made and, after subtraction of reference spectra of each, the final difference spectrum (Fig. 3 f) still showed the major peaks of the plasticizer (Fig.3 g), although the relative peak intensities are not retained as well as in the previous spectra. 'Synthetic' Spectra The carbonyl band of the plasticizer, which occurs at 1730cm-1 in pure plasticizer is shifted when in the ink to 1716 cm-1, but the complexity of the spectrum makes it difficult to see whether any other band shifts are present. Addition of the spectra of individual components (binder plus dye plus plasticizer) by computer and overlay of the result with 3500 3000 2500 2000 1500 1000 Waven urn ber Fig.3 Spectra of black ink on a sheet of polystyrene recorded by ATR; a, ATR spectrum of ink; b, ink spectrum after ATR correction; c, binder reference spectrum; d ink minus binder; e, dye; f, ink minus binder minus dye; g, plasticizer.Analytical Communications, February 1996, Vol33 73 the ‘natural’ ink spectrum showed that the carbonyl was the only band undergoing a major shift (Fig.4). This technique of comparison of ‘real’ with ‘synthetic’ spectra of mixtures is a useful tool in establishing inter-component interactions. Ink Dots on Paper Spectra were recorded from ink dots on papers by both diffuse reflectance and ATR techniques (data not shown).In both cases, however, spectral contributions from the substrate showed through. Because the spatial heterogeneity in a paper occurs at only slightly smaller dimensions than the sampling areas, it is impossible to obtain a good background spectrum for subtrac- tion without averaging spectra obtained from several different areas of sample.This is very time consuming, and the results were poor when subtractions were attempted. With the ATR technique there is also the problem of obtaining even contact over the whole sample area; the ink tends to collect in small depressions on the surface whereas the fibres of cellulose are more likely to contact the ATR crystal. Harris2 reported the successful subtraction of paper components from spectra of ball-point pen ink on paper, using a specialized diffuse s I ti 1800 1600 1400 1200 1000 Wavenumber Fig.4 Spectrum of black ink by transmission (thin line) and ‘synthetic’ sDectrum made bv addition of sDectra of binder. dve and Dlasticizer (thick reflectance accessory, where the sampling area may have been greater than used here. Bartick et al.4 have published spectra of toner dots on paper using ATR microscopy, but thermoplastic toners form a lump on the paper surface, unlike the solvent based inks used here. Conclusions FTIR microscopy on nanogram samples can yield spectra that are good enough for analysis by successive identification and subtraction of principal components if the substrate is reason- ably flat and an appropriate sampling technique is used.The results described here are for inks, but the technique can be applied to other organic or inorganic mixtures, either as coatings or discrete samples. Despite initial reservations about detector linearity, these results show that if the original spectra do not show saturation at peaks then reasonable subtraction results can be obtained, allowing identification of a mixture in situ from nanogram amounts of sample. We thank R. Burr for providing the samples and initiating the project. Thanks are also due to M. Sweeney of Spectra Tech UK Ltd. for recording the ATR spectra snd for helpful discussions. C. W. and M. B., students from the Fachhochscule Aalen, Germany, would like to acknowledge financial assistance from the EU Erasmus programme. References 1 Hudd, A., in The Printing Ink Manual, ed. Leach, R. H., and Pierce, R. J., Chapman and Hall, London, 5th edn., 1993, ch. 12, pp. 678-698. 2 Hams, J., Can. Soc. Forens. Sci. J . , 1991, 24, 5. 3 Bartick, E. G., and Merrill, R. A., J. Forens. Sci., 1992, 37, 528. 4 Bartick, E. G., Tango], M. W., and Reffner, J. A., Anal. Chim. Acta, 1994, 288, 35. Paper 5/08028B Received December 1 I , 1995 Accepted January 9,1996 I , line).

ISSN:1359-7337    

DOI:10.1039/AC9963300071

出版商:RSC    

年代:1996

数据来源: RSC