ISSN:0003-2654    

DOI:10.1039/AN98510FX001

出版商:RSC    

年代:1985

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98510BX003

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

ANALYST JANUARY 1985 VOL. 110 11 Effect of Surfactants on the Determination of Nitrate in Stream Waters by Using a Nitrate Ion-selective Electrode Hirokazu Hara Laboratory of Chemistry Faculty of Education Shiga University Otsu Shiga 520 Japan and Satoshi Okazaki Department of Chemistry Faculty of Science Kyoto University Kyoto 606 Japan Nitrate-nitrogen in stream waters containing anionic surfactants was determined with a nitrate ion-selective electrode. The results with 22 stream waters agreed well with those obtained by spectrophotometry. Nitrite, hydrogen carbonate and chloride interference were removed by the addition of sulphamic acid sulphuric acid and silver sulphate respectively. The interference from anionic surfactants was no problem provided its concentration was lower than 1 mg 1-1 as dodecylbenzenesulphonate (DBS).The error caused by DBS depends on its concentration and can be mostly eliminated by the addition of a cationic surfactant such as cetyltrimethylammonium bromide. An anomalous response behaviour in the presence of some cationic surfactants was observed and is discussed in relation to their extractability as nitrate salts. Keywords Nitrate ian-selective electrode; nitrate determination; stream waters; anionic surfactants; cationic surfactan ts One of the most attractive methods for the direct analysis of natural waters is a nitrate ion-selective electrode which can be used for the simple and rapid determination of nitrate in pIace of the conventional time-consuming spectrophstometric procedure. More than 20 papers have been published that deal with the application of nitrate ion-selective electrodes in water analysis.' Simeonov et a1.2 determined 1.1-2.9 mg 1-1 of nitrate-nitrogen in lake water by several potentiometric methods and compared the results.Tsuzuki and Hikosaka3 analysed river waters containing 0.8-4.7 mg 1-1 of nitrate-nitrogen after pre-treatment. In Shiga Prefecture is Lake Biwa which is the largest lake in Japan and into which over 100 streams flow. Some of the streams are polluted by domestic effluents e.g. anionic surfactants. Although many studies have been made of the determination of nitrate in environmental samples by using a nitrate ion-selective electrode few have dealt with the determination of nitrate in stream waters containing anionic surfactants.The development of a simple and reliable method for the determination of nitrate in polluted stream waters is essential because nitrate is one of the causes of eutrophication, and the prevention of the eutrophication of Lake Biwa is one of the most urgent environmental problems in Japan. The aim of this work was to evaluate a nitrate ion-selective electrode for stream water analysis. The interferences due to anionic surfactants and their elimination by the addition of cationic surfactants were examined. During the study an anomalous response behaviour was observed in the presence of some cationic surfactants when the concentration of nitrate was sufficiently high. Experimental Apparatus An Orion 93-07 nitrate ion-selective electrode and an Orion 90-02 double-junction reference electrode were used.The outer filling solution of the reference electrode was 0.4 M ammonium sulphate solution. The potentials at room tempera-ture (20 * 1 "C) were measured to 0.1 mV with an Orion Ionalyzer Model 701A and recorded with a pen recorder (Rikadenki Model R-10). Reagents Potassium nitrate dried for 2 h at 110 OC was used to prepare a standard solution. Analytical-reagent grade sodium dodecyl-benzenesulphonate (DBS) (Nakarai Chemicals) was used as an anionic surfactant. The cationic surfactants used were decyltrimeth ylammonium bromide dodecyltrimethylammo-nium bromide tetradecyltrimethylammonium bromide, stearyltrimethylammonium chloride cetylpyridinium bro-mide cetylbenzyldimethylammonium chloride (Tokyo Kasei Chemicals) cetyltrimethylammonium bromide (CTMA) (Nakarai Chemicals) and tetradecylbenzyldimethylammo-nium chloride (Dojin Chemicals) and they were used without further purification.Their solutions were prepared using distilled water. Analytical Procedure To 50.0 ml of sample solution were added 1 ml of 2 M ammonium sulphate solution and 0.5 ml of 0.2 N sulphuric acid that contained 2 g 1-1 of sulphamic acid. After allowing the mixture to stand for 1 h which was sufficient for the complete decomposition of nitrite by sulphamic acid 2 ml of 1.76 X 10-2 M silver sulphate solution were added4 and the electrode potential 1 min after immersion of the electrode pair was measured. This procedure brought the pH of the sample solution to approximately 3 which served to eliminate the interference from hydrogen carbonate and also those due to up to 50 mg 1-1 of chloride and 2.9 mg 1-1 of nitrite.A calibration graph was constructed using 0.1 0.2 0.3 0.5 1.0 and 10.0 mg 1-1 standard nitrate solution after the same pre-treatment as for the samples. For samples containing anionic surfactants. a cationic surfactant dissolved in the sulphuric acid containing sul-phamic acid was added to eliminate the interference. Nitrate was also determined by the spectrophotometric procedure in which nitrate was reduced to nitrite with zinc powder in hydrochloric acid and reacted with GR reagent (Griess - Romijn reagent a mixture of sulphanilamide and a-naphthylamine).5 Anionic surfactants in stream water were measured as methylene blue active substances (MBAS).6 Results and Discussion Stream Water Analysis Twenty-two stream waters flowing into L,ake Biwa and Seta River were analysed by the procedure described above.The calibration graph in the absence of cationic surfactants was linear (55.1 mV decade-') down to 0.5 mg 1-1 of nitrate-nitrogen but graphical evaluation was necessary below this concentration as shown in graph 1 of Fig. 1. A stable potential to within 0.1 mV was attained in less than 1 min unless th 12 ANALYST JANUARY 1985 VOL. 110 1 I I I 0.1 1 10 100 Concentration of nitrate-nitrogenimg 1-l Fig. 1. Calibration graphs for nitrate-nitrogen 1 in the absence o f cationic surfactants; 2 in the presence of CTMA; and 3 in the presence of tetradecylbeni-yldimethylammonium.Cationic surfactant added in an equimolar amount to 10 mg 1 - 1 o f Na DBS. Each graph is displaced by 6.25 rnV at 0.1 mg I- l sample was contaminated with anionic surfactants. '1 he average values and the concentration ranges of nitrate nitrite and chloride in the stream waters were 1.39 mg 1-1 (0.284-4.26 mg I - l ) 0.069 mg 1-1 (0.004-0.301 mg 1-1) and 12.3 mg 1-1 (4.34-58.4 mg 1-I) respectively. The agreement with the spectrophotometric results was fairly good as judged from the correlation coefficient of 0.990 and the correlation equation y = 0.082 + 0.945x where y is the value obtained by spectrophotometry and x that obtained by potentiometry. The average value and the concentration range of MBAS was 0.439 mg 1-1 (0.031-1.32 mg I-]).The selectivity coefficient of DBS calculated from measure-ment at 10 mg 1-1 was reported to be larger than 103 implying that the presence of several milligrams per litre of anionic surfactants made the determination impossible.7 However, our results implied that the interference from MBAS in such a concentration range caused an insignificant error as with inorganic interferents such as nitrite chloride and hydrogen carbonate. The reason why the interference of DBS was smaller than expected is that the selectivity coefficient of DBS was much smaller than the reported value when its concentra-tion was around 1 mg 1-1. For example the selectivity coefficient of DBS estimated at 1 mg 1-1 was about 5 (concentration of nitrate-nitrogen = 1 mg 1-1). Another reason is that the potential after 1 min by which time the interference was still not significant was used for the evaluation of nitrate.Subsequent work was concentrated on the interference of DBS and its elimination with cationic surfactants. Suppression of Interference from DBS by CTMA The interference of anionic surfactants in the determination of nitrate in environmental samples and its elimination were investigated by Takehara et al.7 They used a highly porous polymer bead packed column to remove anionic surfactants or added higher alkylamines such as laurylamine to precipitate anionic surfactants from the sample solution acidified to pH 3. Although the interference was reported to be eliminated effectively by these methods they seemed to be still incon-venient and time consuming because for example agitation of the sample solution for 10 min was recommended after the addition of laurylamine.We examined the effect of the addition of cationic surfactants to eliminate the interference of DBS. On the basis of the results reported later we recom-- T .- c S Q) 0 a 4-t 60 s 1 Time + Fig. 2. Potential - response graph for 1 mg I - I of nitrate-nitrogen 1. in the presence of CTMA; and 2 in the presence of both CI'MA and DBS. Concentration of DBS 10 mg I - ' ; CTMA added in an equimolar amount to 10 mg I-' of Na DBS 0 10 20 > E 2 30 . a 40 50 I I 5 1C 60' I 0 1 2 Concentration of DBS/mg 1-l Fig. 3. Effect of DBS on the response of nitrate ion-selective electrode 1 in the absence of CTMA; and 2 in the presence of CTMA added in an equimolar amount to 10 mg I * of Na DBS.Concentration of nitrate-nitrogen 1 mg 1 I mend the addition of CTMA. The calibration graph in the presence of CTMA is shown as graph 2 in Fig. 1. Little effect was observed with CTMA up to 10 mg 1-1 of nitrate-nitrogen. At 100 rng 1-1 of nitrate-nitrogen it took 3-5 min to obtain a stable potential reading and this equilibrium potential was several millivolts higher than that in the absence of CTMA. This anomalous behaviour of cationic surfactants has not previously been reported and will be discussed later. Fig. 2 shows the potential - response graph for standard nitrate-nitrogen and a sample solution containing DBS. The potential changed positively and rapidly just after the immer-sion of the electrode pair and about 10 s later it showed a gradual decrease to a more negative value.Great care should be taken with regard to the shift of the standard electrode potential of the nitrate ion-selective electrode after it has been immersed in a solution containing anionic surfactants. We stopped the measurement 1 min after the immersion and the potential difference AEbO shown in Fig. 2 was used as a measure of the supression effect by cationic surfactants. Fig. 3 demonstrates the suppression effect of CTMA on the interference of DBS. When the concentration of DBS was as low as 1 mg 1-1 its interference in the nitrate determination was fairly small but it showed extreme interference at 10 mg 1-1 as already pointed 0 ~ t .7 In contrast the interference of DBS was mostly removed by the addition of CTMA. However a small interfering effect of DBS remained even in the presence of a 10-fold excess of CTMA. The error still remained after the addition of CTMA was evaluated as shown in Table 1. The error was independent of the concentration of nitrate-nitrogen when it was higher than 1 mg 1-1 but became significant when it was as low as 0.1 mg 1 - 1 ANALYST JANUARY 1985. VOL. 110 13 Effect of Other Cationic Surfactants The effect of cationic surfactants can be divided into two parts, the suppression effect due to the interference of DBS and the interfering effect on the response of the nitrate ion-selective electrode especially at high concentrations of nitrate. In the latter instance no anionic surfactant was present in solution.The suppression effect of cationic surfactants is summarised in Table 2. In the alkyltrimethylammonium series the interference of DBS decreased with increasing carbon chain length. Although stearyltrimethylammonium was the most effective of the cations tested its low solubility in water prevented its convenient use. Other alkylpyridinium and alkyldimethylbenzylammonium cations were also as effective as CTMA as far as the suppression of the interference from DBS was concerned. This effect may be related to the association between DBS and the cationic surfactant in solution. Goto et al. 8 measured the association constant with a ferron (7-iodoquinolin-8-ol-5-sulphonic acid) - cationic surfac-tant complex system.Their association constants with the cationic surfactants listed in Table 2 correspond to the variation of A&() at least qualitatively the larger the logarithmic association constants the smaller is i.e. the more effective it becomes. The anomalous response of nitrate-nitrogen at 100 mg 1-1 that occurred with CTMA was also observed more signifi-cantly with alkylpyridinium and alkylbenzyldimethylammo-nium cations. In Fig. 1 the calibration graph in the presence of tetradecylbenzyldimethylammonium is shown as an example (see graph 3). The response of nitrate-nitrogen at 100 mg 1-1 with these cations was so slow that it took more than 5 min to reach a steady potential reading. With cetylbenzyldimethyl-ammonium a slow response was observed even in a solution of 10 mg 1-1 of nitrate-nitrogen.The sequence of the effect Table 1. Effect of CTMA on the suppression of interference from DBS. CTMA was added in an equimolar amount to 10 mg 1 - 1 of Na DBS Concentration of nitrate-nitrogen/mg 1- I Concentration of DBS/mg 1-1 Taken Found Error YO 2 0.1 0.112 22 5 0.1 0.149 49 10 0.1 0.220 120 2 1 .0 1.06 6 5 1 .o 1.12 12 10 1 .O 1.26 26 2 10 10.5 5 5 10 11.1 11 10 10 12.3 23 was cetylbenzyldimethylammonium > tetradecylbenzyl-dimethylammonium = cetylpyridinium > stearyltrimethylam-monium = cetyltrimethylammonium. Other alkyltrimethyl-ammonium ions tested showed no effect. Some studies have previously been made on the response behaviour of liquid membrane nitrate ion-selective electrodes in the presence of cationic surfactants.Takehara et al.7 stated that no effect on the electrode potential was observed with the addition of 1000 mg 1-1 of laurylamine acetate. Hulanicki et af.9 stated that cationic and non-ionic surfactants exhibited an effect of minor importance at higher concentrations or even showed no effect. Small positive interferences were observed with cetyltrimethylammonium p-toluenesulphonate by Campi et af.10 In this instance the interference may be due to the presence of a relatively large organic anion p-toluenesulpho-nate. Such an interfering effect of some cationic surfactants seems to deserve further investigation. Jyo et al. * I reported on the influence of a co-ion (cation) on the potential of nitrobenzene-based liquid membrane nitrate ion-selective electrode.They constructed a calibration graph using tetramethylammonium nitrate and obtained a positively curved graph at nitrate concentrations above 10-2 M when the concentration of the ion pair in a liquid membrane was as low as M. Their results were similar to our observations. They ascribed this interfering effect of the tetramethylammonium ion to the high extractability of the tetramethylammonium ion in comparison with the sodium ion and interpreted the results theoretically. The interfering effect of some cationic surfactants may be interpreted by taking their extractability into account. The sequence of the effect seems to correspond to the logarithmic extraction constants given by Goto et a1.8 and measured in a ferron - cationic surfactant system with dichloromethane as a solvent viz.cetylbenzyldimethylammonium (5.6) Bcetylpy-ridinium (5.2) = tetradecylbenzyldimethylammonium (5.2) > cetyltrimethylammonium (5.1) > tetradecyltrimethylammo-nium (5.0) > stearyltrimethylammonium (4.9) > dodecyl-trimethylammonium (4.6) > decyltrimethylammonium (3.6). In conclusion the higher the extractability of cation became, the more significant was the effect of the so-called "co-ion," because the nitrate ion was the only extractable anion with cationic surfactants in solution when its concentration was sufficiently high. Recovery Test A recovery test was carried out with stream waters containing none and a significant amount of anionic surfactants and the results are shown in Table 3.In these two instances a recovery of nearly 100% with a relative standard deviation of about 1% for five successive measurements were observed. Table 2. Suppression of interference from DBS in the determination of nitrate-nitrogen using cationic surfactants. Amount of N03-N determined 1 mg 1 - 1 . Amount of DBS 10 mg 1 - 1 . Concentration of cationic surfactant equimolar to 10 mg 1 - 1 of Na DBS Mean potential difference k s.d. Cationic surfactant after 1 min/mV None . . . . . . . . . . . . . . 53513 ( n = 4 ) Decyltrimethylammonium . . . . . . 4 0 2 6 ( n = 4 ) Dodecyltrimethylammonium . . . . . . 19 5 2 ( n = 3) Tetradecyltrimethylammonium . . . . 6.2 + 0.3 ( n = 3) Cetyltrimethylammonium . . . . . . 4.4 k 0.3 ( n = 10) Stearyltrimethylammonium .. . . . . 2.0 5 0.3 ( n = 3) Cetylpyridinium . . . . . . . . . . 3.9 2 0.2 ( n = 3) Tetradecylbenzyldimethylammonium . . 4.4 k 0.3 ( n = 3) Cetylbenzyldimethylammonium . . . . 3.5 20.6 ( n = 3) Logarithmic association constant with ferron * -2.0 2.6 3.4 3.9 4.3 3.9 3.5 3.8 * Data from reference 8 14 ANALYST JANUARY 1985 VOL. 110 Table 3. Recovery test with stream water Concentration of nitrate-nitrogen/ mg 1- * Recovery MBASi Stream Added Found* S.d. Yo mgl-1 Senjyot . . - 0.54 0.005 0.1> 1 .0 1.54 0.025 100 1.07 1.95 0.019 103 2.14 3.07 0.039 104 Sandai- . . - 0.824 0.008 1.3 * Mean values of five successive measurements. i Calibration graphs were constructed without (Senjyo) and with (Sanda) CTMA which was added in an equimolar amount to 10 mg 1-1 of Na DBS.Conclusion The method is simple and convenient for the determination of nitrate at concentrations above 0.1 mg 1-1 in stream waters. The interference of DBS below 1 mg 1-1 was found to be of no importance. It is possible to suppress effectively although not completely the interference of DBS in the concentration range 1-10 mg 1-1 by the addition of an amount of equimolar CTMA to 10 mg 1-1 of Na DBS. Some cationic surfactants caused the anomalous response? especially at high concentra-tions of nitrate. This behaviour may be interpreted by assuming that extraction of the ion pair of nitrate and the cationic surfactant occurs at higher concentrations of nitrate. This work was partially supported by a Grant-in-Aid for Scientific Research by the Ministry of Education Science and Culture of Japan (No. 58740255). 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. References Hulanicki A and Trojanowicz. M. Ion-sel. Electrode Rev 1979 1 207. Simeonov V. Andreev G and Stoianov A Fresenius Z. Anal. Chem. 1979 297 418. Tsuzuki T. and Hikosaka 0 Mizushori Cijyutsu 1981. 22, 431. “Instruction Manual Nitrate Ion Electrode Model 93-07,” Orion Research Cambridge MA 1981. Nishimura M . and Matsunaga K. Bunseki Kaguku 1969,18, 154. Nasu Y . and Tachibana H . in JSAC. Editors “Mizu no Bunseki (Water Analysis),” Third Edition Kagaku Dojin, Kyoto for Hokkaido Branch of the Japan Society for Analy-tical Chemistry 1981 p. 374. Takehara H. Hiratsuka Y and Harazono M. Bunseki Kagaku 1980 29 601. Goto K. Taguchi S . Miyabe K. and Haruyama K., Bunseki Kagaku 1983 32 678. Hulanicki A . Trojanowicz M. and Poboiy E. Anulyst, 1982 107 1356. Campi E. Saini G. and Castelli R. Ann. Chim. (Rome), 1982 72 471. Jyo A . Fukamachi K Koga. W. and Ishibashi N. Bull. Chem. SOC. Jpn. 1977 50 670. Paper A411 98 Received June 1 lth 1984 Accepted August 6th 198

ISSN:0003-2654    

DOI:10.1039/AN9851000011

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

ANALYST JANUARY 1985. VOL. 110 I 5 Novel Ion-selective Electrode System for the Simultaneous Determination of Fluoride and Calcium in Acid Solution John E. Tyler Medical Research Council Dental Unit The Dental School Lower Maudlin Street Pristo! BS12LY 1lK and John E. A. Comer Orion Research AG Fahnlibrunnenstrasse 5. 8700 Kiisnacht Switzerland A method is described that utilises two independent electrode systems; a new ultra-sensitive fluoride - glass pH electrode differential cell and a calcium electrode monitor for the direct and simultaneous measurement of both fluoride and calcium in unbuffered acid solution. Fluoride Nernstian response mav be extended by at least one decade towards the lower limit of detection with an appreciable reduction in electrode response time at low concentrations when compared with the conventional use of fluoride electrodes using TISAB-type buffers.The system has been successfully applied as a biopsy technique for the determination of fluoride concentration profiles within the outer 100 vm of human tooth enamel. Keywords Fluoride de te rm in a tion; calcium determination; acid media ; ion -selective electrodes; tooth enamel The role of fluoride in the reduction of dental caries is not fully understood and considerable research has been directed towards the evaluation and distribution of fluoride in dental enamel associated tissues and integuments. One current technique used for the determination of fluoride to calcium ratios in mineralised tissues depends upon the serial acid etching of small isolated areas for example on a tooth enamel surface followed by the analysis of individual etch solutions for fluoride and calcium using fluoride ion-selective elec-trodes' and atomic-absorption spectrometry.The calcium assay is accepted as a measure of the mass of dissolved matrix and depth of etch.2 In a patent described by Diggens and ROSS,~ the electrochemical analysis of fluoride may be conducted at pH values below a pK of 3.2 for hydrofluoric acid. Under these conditions HF will be the predominant fluoride-con taining species. A com bination glass pH electrode and a single fluoride electrode formed a differential cell the potential difference between the electrodes being a logarith-mic function of the total fluoride concentration. Fluoride species present with changing pH were defined as follows: .. [H+l [F-I [HFI H F e H+ + F- and K = ~ Response of fluoride electrode: Response of pH electrode: Substituting for [F-] in equation (2): E F = E O F - kln[F-] . . . . . . ( 2 ) EH= pH+ kln[H+] . . . . . . (3) = E"F - klnK - kln[HF] + kln[H+] . . . . (4) Subtracting equation (3) from (4): EF - E H = ETF = EOTF - kln[HF] . . . . ( 5 ) where E is the observed e.m.f.; E" the standard potential; and k = RT/zF the Nernstian slope factor R being the gas constant T the absolute temperature F Faradays constant and z the ionic charge including the sign. The behaviour of the combination glass pH - fluoride system conforms to a full Nernstian response where the potential of the cell ( E F ) is proportional to In[HF] or total fluoride (TF) in strong acid conditions.The patent applies to other systems of weak acids; given an electrode sensitive to small concentrations of weak acid anions and acidification below the plc of the protmarcd species the total anion concentration may be determined using a combination pH - anion-sensitive electrode cell. Although atomic-absorption spectrometry is the favoured technique for calcium determination it may be inconvenient and disadvantageous compared with a non-destructive elec-trochemical method with the possibility of using other electrode sensors without loss of sample. Therefore this study describes a development of a multi-electrode system for the direct and simultaneous determination of flyoride and calcium in acid solution with a specific appiicatior to apiititic biological tissues.Experimental Apparatus Two specific ion meters were used an Orion Model 001 microprocessor digital lonalyzer (Orion Research Incorpor-ated Cambridge MA) for the direct reading of total flucride concentration and an Orion 401 meter in the divalent cation mode for the direct readout of calcium. Both the fluoride electrode and the pH glass electrode have an extremely high electrical resistance. In order to measure the potmid difference between these two electrodes. it is necessary to use an instrument such as the Orion 901 which has two high impedance inputs. In the 901 the potential of the fl~xiride electrode and the pH electrode are measured reiative to the Ag - AgCl reference electrode which in this instance i s buiit into the combination pH electrode.Two separate Orion electrode systems were applied; a single fluoride e!ectrode (94-09) in conjunction with a flat-surfaced gei-filled pi 4 combination electrode (91-35) for fluoride deterniiiiaticli;h and a calcium electrode (93-20) together with a sir;p,ie-iuncticn Ag - AgCl 4~ KCl reference electrode (90-!)1). M y -propylene calibrated flasks beakers micropipette tips and a shallow microtitration cell (35 x 35 x 8 mmj machined frc:m H 45 mm diameter 20 mm thick disc of polypropylene. \ ' t i t employed throughout the experiments. A sinall P T F E - L ~ ~ ~ L ~ magnetic stirring bar (8 x 4 x 4 mm). located in ii 2 inin drcp recess within the centre of the titration cell pmnittcd vigorous agitation of 2-ml aliquots of HCI containing C11111iiIil-tive additions of fluoride and calcium ions.It was essw!iel ti) immerse the electrode membranes completely irlta th:. 7cL.t solutions and use a magnetic stirrer with a11 isolateif e!e;t!-i 16 ANALYST JANUARY 1985 VOL. 110 motor to avoid changes in Nernstian response due to temperature variation. To test the efficacy of the calibrated fluoride and calcium electrode systems in acid media a millimetre diameter exposed surface of crystal fluorapatite was immersed for 5-min periods in the acid titration cell and the fluoride and calcium d i sso 1 u t i o n product s s i mu 1 tan e o us 1 y m o n i t o re d e 1 e c t r o c h e m -ically. Fluorapatite of known stoicheiometric composition, was an ideal matrix with which to test the multi-electrode system before application to apatitic biological specimens.Reagents De-ionised distilled water was used throughout. Hydrochloric acid. Dilutions were prepared from Aristar grade concentrated acid from BDH Chemicals. Fluoride solution 4 pg ml-1 in 0.1 M hydrochloric acid. Prepared from analytical-reagent grade sodium fluoride. Calcium solution 1 mg ml-1 in 0.1 M hydrochloric acid. Prepared from analytical-reagent grade calcium carbonate or hydroxyapatite. Mineral fluorapatite Durango Mexico Ca5(PO3)3F. Cal-cium 39.1% (theoretical 39.7"/0). Synthetic hydroxyapatite Cai( P 0 4 ) 3 0 H . Calcium 38.4% (the ore t ical 39.9 % ) . Electrode calibration solution Fluoride. A concentration range within 0.001 to 0.1 pg ml-I was obtained by the cumulative addition of 1 to 50 p1 of standard fluoride solution to 2-ml aliquots of standard HCl.Calcium. Cumulative additions of microlitre aliquots of standard or dilute standard calcium solutions to 2 ml of 10-1 M HCl provided a concentration range within 0.025 to 250 pg ml-1. Synthetic hydroxyapatite of known composition was used as a standard for calcium electrode calibration for the analysis of fluorapatite or tooth enamel hydroxyapatite is an appropriate standard in relation to the apatitic composition of both fluorapatite and inorganic tooth matrix. Results The response of the fluoride electrode in combination with a glass pH electrode in aqueous solutions of varying pH and fluoride concentration is shown in Fig. 1 which demonstrates the Nernstian response of the system.Note that the results become increasingly independent of pH below pH 2. Linear deviation for a fluoride concentration of less than 0.005 pg ml- 1 may in part be due to nanogram amounts of fluoride in the de-ionised water and hydrochloric acid used in the experiments which is most pronounced at pH 0.1. Using TISAB I1 at pH 5.5 Fig. 1 demonstrates a negligible response of the differential cell below the 0.1 pg ml-1 concentration of fluoride. The mean Nernstian slope factor for fluoride concentrations of 0.001-0.1 pg ml-1. over a pH range of 1-4. was observed to be +55.8 (s.d. 1.6) mV pF-* at 20 "C ( n = 4), with an average correlation coefficient r = 0.999 (where r is the correlation coefficient calculated from regression coeffi-cients).At a specific pH value of 1.0 the electrode slope was observed to be +53.4 (s.d. 0.56) mV pF-1 at 20 OC average r = 0.999 ( n = 6) for the fluoride concentration range as above. The time response of the differential cell is shown in Fig. 2. The mV recorder reading showed a 95% response within 30 s for a 0.005 pg fiuoride addition to a vigorously stirred 1-ml sample of 0.1 M HCI complete equilibration being attained within 2 min. The Orion calcium electrode a neutral carrier PVC sensor and a single-junction reference electrode containing 4 M KCI operating at an acid pH of 1.0 provided a convenient cell for the direct monitoring of calcium ion concentrations released by acid etching of mineralised tissues. The constant back-5c 0 > I c: E E ui - 50 -100 pH 0.1 pH 1.0 pH 2.0 pH 3.0 aH 4.0 0.001 0.005 0.01 0.05 0.1 Concentration of fluorideipg ml-1 Fig.1. Uncorrectcd millivolt response of the fluoride differential cell with variation of acid concentration (HCI). Extrapolation of calibration lines computed a blank value of 0.09 pg ml- I of fluoride for Aristar concentrated hydrochloric acid and 0.001 pg ml- 1 for distilled water I 1 0 1 2 Time/m in Fig. 2. Response time of the fluoride - glass pH differential cell. Using a chart speed of 20 mrn min I . a 9X'% mV response was observed within 60 s for a 5-ng addition of fluoride to a 1-ml aliquot of 1 O - l M HCI ground ionic strength (10-1 M HCI) that was maintained throughout all experiments ensured a constant calcium activity coefficient and therefore calcium activity was directly proportional to concentration.Calcium calibration under the above specific conditions showed a Nernstian response over a range of 0.025-250 pg ml-1 of calcium using hydroxyapatite as a standard average electrode slope i-26.6 (s.d. 0.50) mV pCa-i at 19 "C average r = 0.999 ( n = 9). Although the calcium electrode was being used at a low pH no problems with respect to instability drift or response were identified over a two-year period of constant use. Using the same calcium electrode module and a calcium carbonate standard following recommended direct measurement procedures at pH 6 slightly greater electrode slopes were obtained average slope +27.9 (s.d. 0.41) mV pCa-1 at 19 "C average r = 0.999 ( n = 4).A Student t-test showed a significant difference between electrode calibration using an acid solution of hydroxyapatite and calcium carbonate at pH 6 (p < 0.001) ANALYST JANUARY 1985 VOL. 110 As described in this paper the application of the fluoride differential cell and the calcium electrode system to a titration cell containing 2-ml aliquots of acid provided a technique for the simultaneous monitoring of fluoride and calcium ions. The intermittent introduction of a small window of crystal fluora-patite for serial acid etching into the titration cell produced fluoride and calcium analyses as shown in Fig. 3. An average fluoride to calcium ratio of 0.099 (s.d. 0.001) was determined for fluorapatite Y = 0.999 ( n = 12) compared with the theoretical value of 0.095.The observed fluoride content was 3.86% (s.d. 0.05) as compared with 3.77% the theoretical value. Assuming the density of fluorapatite (3. lo), the calcium content and area of window exposed the acid etch depth and mass of fluorapatite dissolved for each 5-min etch may be calculated Fig. 3. Similarly applying this technique to windowed 1 mm exposed tooth enamel surfaces fluoride to calcium depth profiles may be obtained for the outer enamel. Discussion Investigation of the differential fluoride cell which functions at a low pH showed a Nernstian-type response extended by one concentration decade towards the lower limit of detection compared with the TISAB method for fluoride determination. This cell responds to total fluoride in the form of undissociated hydrogen fluoride which is the predominant fluoride-containing species in acid solutions below the pK value of this acid.Diggens and Ross3 recommended pH values to be adjusted to below 2.5 for fluoride analyses. However as shown in Fig. 1 a Nernstian response may be attained for pH values between 3 and 4 but no response of this system was observed using TISAB at pH 5.5 for fluoride concentrations of less than 0.1 pg ml-1. Extrapolation of the pH 1 to pH 4 fluoride calibration lines from Fig. 1 provided a calculated value of 0.09 pg mlLl for the fluoride content of BDH Aristar HCl. This is a value for the reagent blank necessary in order to correct for the slight curvature of response within the 0.001-0.005 pg ml-1 fluoride concentration range.However this value does not account for a corrected electrode slope response of 39.7 mV pF-1 at pH 0.1 (Y = 0.993) for a fluoride concentration range 0.01-0.1 pg ml-1. At such pH values so few fluoride ions are present in solution that the fluoride electrode kinetics would be impaired. Although it is well known that hydrogen fluoride readily reacts with glass no 0 5 10 15 Concentration of calcium/pg ml-1 Fig. 3. Sequential analyses of mineral fluorapatite CaS(P0&F by serial acid etching a 1 mm diameter exposed crystal surface with direct electrochemical analysis of each acid etch for fluoride and calcium. Observed fluoride to calcium ratio calculated from regression 0.099, compared with a theoretical value for fluorapatite of 0.095.Calculated average values for the depth and mass of mineral dissolved for 5-min individual etches were 2.6 pm and 6.4 pg respectively 17 deterioration or loss in response of the glass membrane of the pH electrode was observed during 12 months of continuous use. The fast response of the fluoride differential cell and the low concentrations of fluoride analysed are not conducive to fluoride - glass reactions. The Orion calcium sensor adequately functioned at pH 1 with an unexpected life span in excess of 24 months. Adverse effects produced by calcium activity junction potential variation or interference due to the high mobility of hydrogen ions were reduced or avoided by using a constant background acid (ionic) strength for specimen and calibration measurements.Although complexing ligands reduce the level of free calcium ions in solution at low pH there will be a reduction in complexation and precipitation of insoluble calcium phosph-ates carbonates hydrogen carbonates and sulphates com-pared with the recommended pH 5-6 procedures for normal use of the calcium electrode. The small but significant effect of 0.1 M acid hydroxyapatite solutions on chlcium electrode response was less than 5% of the slope factor when related to phosphate-free calcium solutions analysed at pH 6. With calculated correlation coefficients approaching unity for the electrode calibration the Orion calcium electrode was shown to be quantitative over a wide range of calcium concentrations under controlled acid conditions.4 Serial acid etching of a fluorapatite crystal of known X-ray diffraction and chemical composition followed by analysis of the individual etches for fluoride and calcium confirmed the viability of the multi-electrode system.The fluoride to calcium ratio and fluoride content for fluorapatite were observed to be greater than 95% of the theoretical values. Following these evaluations the method was successfully applied to the analysis of mineralised tissues where the inorganic matrix is predominantly hydroxyapatite. Electrochemical analysis of microgram amounts of surface tooth enamel as shown in Fig. 4 produced enamel fluoride profiles to a depth of 80 pm within 18 min of the analysis time. This method provides a rapid technique for the analysis of fluoride and calcium in biopsy specimens with characterisa-tion of the enamel fluoride profiles to a 10-pm depth within 4 min.Compared with the microlitre drop method for fluoride analysis,5 the present extension of fluoride Nernstian response facilitates the use of larger analyte volumes which not only reduces evaporation problems but also permits the simul-taneous use of other electrode sensors without the loss of 20 40 60 80 Depth from enamel surface/ym Fig. 4. Analysis of serial acid etches from 1 mm diameter enamel “windows” provided fluoride concentration profiles for a child treated with fluoride tablets since birth. A Permanent premolar after 12 years; B deciduous canine after 7 years. The profiles demonstrate the elevation of surface enamel fluoride which is partially attributed to fluoride tablet therapy during permanent tooth developmen ANALYST JANUARY 1985 VOL.110 fluoride sensitivity. Nevertheless it may be possible to apply the fluoride - glass pH differential cell to a micro-drop system, in which instance nanogram amounts of mineralised tissues could be analysed for fluoride. It is interesting to speculate as to why the fluoride electrode shows a lower limit to the Nernstian response under our experimental conditions of low pH compared with the usually observed limit of 10-6 M fluoride in solution above pH 4. It is generally held that the lower limit of Nernstian response is determined by the finite solubility of the membrane used in an ion-selective electrode; within this limit the fluoride electrode responds to fluoride released by its own dissolution.Lingane6 suggested an enhanced solubility of lanthanum fluoride in acid media. Ikrami et al. 7 showed that this solubility increased with decreasing pH and proposed a linear relationship between log (solubility) and pH. Increased solubility of the lanthanum fluoride membrane would indicate a higher not lower limit for Nernstian response at low pH values. However the reaction, H+ + F- HF leads to a reduction of fluoride ion content in solutions below pH 4. Accordingly a solution of 10-6 M fluoride (at pH 5.5) would show a fluoride ion activity of only 10-9 M at pH 1 and we have obtained consistent reliable results with good recoveries of fluoride in many solutions with fluoride contents of less than 10-6 M.We suggest that the lower limit of Nernstian response of the fluoride electrode is determined not only by the solubility of the membrane but also by other factors.8 One such factor might be the reduced interference from the hydroxide ion in acid solution because at pH 5.5 the hydroxide concentration of 10-8.5 M is significantly reduced to 10-13 M at pH 1. Although the hydroxide ion is a well known interferent of the fluoride electrode in low-level fluoride measurement and in alkaline solutions the above explanation is speculative and further research will be necessary to elucidate the mechanism of fluoride electrode response at low pH. Nevertheless analysis of tooth enamel for fluoride and calcium contents as described in this study yield a fluoride concentration profile related to depth of enamel. Using a constant volume of acid etchant for each profile the stepwise cumulative observations become an analytical advantage when small fluoride and calcium increments are to be evaluated. The fast response of the fluoride differential cell together with the quantitative performance of the calcium electrode operating in an acid mode provides a routine electrochemical technique for the simultaneous determination of fluoride and calcium in mineralised tissues. 1. 2. 3. 4. 5 . 6. 7. 8. References Moody G. J. and Thomas J . D. R. Zon-Sel. Electrode Rev., 1979 1 187. Berndt A. F. and Stearns R. I. Editors “Dental Fluoride Chemistry,” Charles G. Thomas Springfield IL 1978 pp. Diggens A. A. and Ross J . W. UK Patent Application GB 2064 131 A 1981. Thomas J. D. R. Anal. Proc. 1981 18 350. Weatherell J.A. Hallsworth A. S. and Robinson C. Arch. Oral Biol. 1973 18 1175. Lingane J . J . Anal. Chem. 1968 40 935. Ikrami D. D. Komilove G. and Khaitova M. Zzv. Akad. Nauk Tadzh. SSR Otd. Fiz.-Mat. Geo1.-khim. Nauk 1973,3, 65. Midgley D. Anal. Proc. 1980 17 306. 25-26. Paper A4193 Received March 7th 1984 Accepted August 29th 198

ISSN:0003-2654    

DOI:10.1039/AN9851000015

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

ANALYST JANUARY 1985 VOL. 110 19 Evaluation of the L’vov Platform and Matrix Modification for the Determination of Aluminium in Serum Maurizio Bettinelli and Umberto Baroni Central La bo ra tory EN E L - DCO Piacenza ltal y and Franco Fontana and Piergiorgio Poisetti Ila Medical Division Section of Nephrolog y and Dialysis Civil Hospital Piacenza Italy The method previously described by Casetta eta/. for the determination of aluminium in dilute (1 + 1) human serum using matrix modification and a stabilised temperature platform furnace has been considered. The combination of the platform integrated absorbance new coated tubes and oxygen addition to the charring step provided better precision and smaller variation during the life of the tube. Good results were achieved by standardising the procedure against a calibration graph if integrated absorbance signals were used for quantitation.The calibration was linear up to at least 150 pg I-’ of aluminium; the within-run and between-run precision was 5.5 and 6.5% respectively (at 14.3 pg 1-1 of aluminium); and the recovery of aluminium added to pooled serum ranged between 97 and 102%. Furnace lifetimes in excess of 200-250 firings using oxygen ash i ng were routinely achieved. Keywords Aluminium determination; L’vov platform; matrix modification; serum Several methods are used to determine the aluminium concentration in serum including neutron activation X-ray fluorescence flame atomic-absorption and -emission spectro-photometry inductively coupled plasma emission spectropho-tometry and graphite furnace atomic-absorption spectropho-tometry.1 In their present state neither X-ray fluorescence nor flame absorptionlemission methods are sensitive enough to measure trace levels of aluminium in biological samples. Neutron activation analysis produces excellent results but the methods developed are time consuming and the facilities are not always readily available. Graphite furnace atomic-absorption spec-trophotometry is the most frequently used technique. Literature concerned with the determination of aluminium in a furnace is extensive,’ but sometimes unclear. Some of the more recent developments in the atomic-absorption analysis of biological samples use various matrix modifiers,”-s the addition of oxygen during the ashing step to facilitate removal of carbonaceous residues,6.7 the introduction of improved pyrolytically coated graphite tubes8 and the stabilised temper-ature furnace with the L’vov platform.9-12 In deciding which analytical approach (wet ashing extrac-tion procedure direct injection etc.) to adopt for routine use in a clinical laboratory we considered three fundamental criteria.These were that the method should be simple rapid and that sample pre-treatment should be minimal to reduce the source of contamination. In this work the approach based on the direct injection of dilute (1 + 1) serum with accurate control of instrumental conditions proposed by Casetta et al. ,13 has been evaluated as it appeared to meet the above criteria. In this technique dilution with a matrix modifier is necessary to reduce the viscosity of the samples; an injection volume of 10 pl is suitable to prevent improper deposition of sample on to the walls of the tube; the sampling capillary tip must be aligned perfectly (adjusting the tip distance from the platform and the tip dipping in the cup) with the sample introduction hole in the graphite tube; and the serum in the cup must be shaken well before the sample is placed on the platform; a programme step must be added between char and atomisation (at the same char temperature but at a “gas stop” condition) to minimise the temperature difference between the char and atomisation steps.Experimental Apparatus A Perkin-Elmer Model 5000 atomic-absorption spectropho-tometer equipped with an HGA-500 graphite furnace AS-40 autosampler and a Model 056 strip-chart recorder was used for the absorbance measurements.A deuterium arc lamp was used to correct for non-specific absorption. New pyro:ytically coated graphite tubes with a solid pyrolytic graphite platform were used. The purge gas was argon except in the charring step where oxygen was employed. Reagents All of the water used to wash laboratoryware and prepare solutions and standards was prepared by a Milli-Q system (Millipore Ltd.). All glassware plastic tubes and stoppers, disposable pipette tips polypropylene cups for the AS-40 sampler etc. were pre-cleaned by soaking them overnight in a saturated solution (about 0.5 mol 1-1) of disodium ethylene-diaminetetracetate (BDH Chemicals) rinsed with copious amounts of de-ionised water and dried in a clean atmosphere.Serum samples were diluted with an equal volume of a solution containing 2.0 g 1-1 of Mg(N03)* (Merck Suprapure grade); no Triton X-100 solution was added to the sample. A stock solution of 100 mg 1-1 of A1 (aluminium nitrate, BDH standard solution for AAS) was used to prepare the standards. From this solution a 1 mg 1-1 A1 intermediate stock standard was prepared daily; 2.5 5.0 10.0 and 15.0 ml of this solution were pipetted into separate polypropylene containers and made up to 100 ml with de-ionised water. These standardscorrespond to25.0,50.0,100.0and 150.0pgl-1 of aluminium respectively and were used for the standard additions method. - . Sample Collection Blood samples obtained by a plastic cannula with stainless-steel needles were transferred into 10-ml polypropylene sample tubes centrifuged at 2000 rev min-1 for 10 min and the serum was transferred into 3-ml plastic containers.In the determination of reference values samples were taken from “normal” persons through a plastic cannula that had been washed with blood. For each sample a double 10-ml portion of blood was then collected centrifuged as above and stored at 4 “C. Sera determined to have a low aluminium content by the automated method of standard additions were pooled and used as the matrix to prepare calibration graphs. All serum samples were analysed within four weeks to date of sampling and to minimise contaminations all operations were per-formed under dust covers 20 ANALYST JANUARY 1985 VOL.110 Results and Discussion The instrumental operating conditions for the HGA-500 (Table 1) were those reported by Casetta et al. l3 In this study, the linearity of the calibration graph the sensitivity the within-run and between-run precision the detection limit the analytical recovery and the tube life have been verified. Graphite furnace analyses were occasionally subject to a random error resulting from a system malfunction. To detect these we repeated the determination on each sample at least three times. A random result clearly variant from the others was rejected and the remainder were averaged. As evident from the calibration graph (Fig. 1) linearity up to at least 150 pg 1-1 of A1 was obtained. Using 10 pl of dilute (1 + 1) serum the sensitivity of the platform system was 1.05 pg 1-1 of A1 for 0.0044 absorbance seconds.In order to study the need for background correction 50 and 100 pg 1-1 of A1 were added to aliquots of pooled serum and the analyses were performed with and without using background correction. The difference between the absorbances were in both the tests within the analytical variation of the method which indicated that background correction was unnecessary at this wavelength. Precision data are presented in Table 2. Within-run precision was estimated by analysing a serum pool that was spiked, divided into aliquots and kept at 4 "C until absorbance measurements were made. Between-run precision was obtained from the analysis of two serum pools performed on separate days.Recovery studies were performed by adding two known amounts of aluminium to aliquots of three serum samples. The original serum and the two spiked aliquots were processed and the calculated recovery data are given in Table 3. The detection limit (20) measured with a 20 pg 1-1 A1 standard was 2.0 pg of Al. In one test a serum sample (containing about 100 pg 1-1 Al) was fired repetitively in a pyrolytically coated graphite tube until the tube failed. Figs. 2 and 3 show the chart recorder response of this serum sample after ten and 150 firings, respectively. After 150-180 firings there was no systematic change in the aluminium peak area signal but a remarkable change in the peak height signal. A good example of the advantage of using area integration is reproduced in Fig.4, which shows the change in peak shape recorded by increasing the firings. Table 1. Instrument and furnace conditions for the determination of aluminium in serum Model 5000 spectrophotometer; hollow-cathode lamp current 20 mA; wavelength 309.3 nm; slit width 0.7 low; background corrector yes; signal peak area; integration time 6 s; chart recorder, 20 mm min-1; and span 10 mV full scale. HGA-500 graphite furnace; pyrolytically coated graphite tubes with a stabilised temperature platform furnace; sample volume 10 PI; alternative volume 10 PI HGA programme Step 1 2 3 4 5 6 Temperature/"C . . . . 80 130 500 1500 1500 1500 . . . . . . . . 4 Holdis 4 25 55 25 30 Flow-rate of internal gas Ar/mlmin-l 300 300 300 300 0 Flow-rate of alternative gas 02/mlmin-l .. . . . . 50 Recorder . . . . . . Read . . . . . . . . Base line . . . . . . . . Ramp/s . . . . . . . . 1 30 30 1 15 1 . . . . . . 7 8 9 2400 2600 20 0 1 1 6 6 20 0 300 300 - 10 X X 0. 0 50 100 150 Aluminium concentration/pg I-' Fig. 1. Calibration graph for aluminium in dilute (1 + 1) serum using the pyrolytically coated graphite tube and the L'vov platform. Standard additions method results A from pooled sera at low aluminium content (five determinations); a from 50 serum samples analysed on different days with five different tubes. The natural level of aluminium has been subtracted from all readings. 0 Results from aqueous solutions Table 2. Precision data for the determination of aluminium Relative standard deviation 7'0 Sample Pool/yg 1-1 Within-run* Between-runt SerumA .. . . 14.3 5.5 6.5 +50.0 3.6 4.1 + 100.0 1.5 2.7 SerumC . . . . 39.2 3.1 4.6 +50.0 1.1 3.1 + 100.0 0.7 2.6 * Results of five determinations. t Calculated on five separate days. Table 3. Recovery data of aluminium added to pooled sera Pool/ Added/ Sample pg1-1 Pg 1-SerumA . . 4.3 50.0 100.0 SerumB . . 17.0 50.0 100.0 SerumC . . 39.2 50.0 100.0 Found/ 55 103 66 119 86 136 Pgl-' Recovery, O/O 102 99 99 102 97 9 ANALYST JANUARY 1985 VOL. 110 21 J I 4- Time Fig. 2. Chart recorder response (10 X) of 100 pg 1 - I of A1 in serum after ten firings. Peak area signal reproducibility (coefficient of variation) 2.20% t Time Fig. 3. Chart recorder response (10 X) of 100 pg I-’ of A1 in serum after about 150 firings.Peak area signal reproducibility (coefficient of variation) 2.50% Fig. 5 shows the repeatability of the aluminium peak area signal at intervals of ten firings. Each point of the graph represented an average of five firings. This particular tube failed after about 400 firings of diluted serum. Fig.’5 also shows the progressive loss in sensitivity of tube (1) (after about 200 firings) which might be attributed to the erosion of the pyrolytic inner surface. Using the same “old” platform and a new pyrolytic tube the normalised peak area signal agreed very well with the earlier sets of data. No ash build up was found after about 300 firings of dilute serum samples. In another test we used five pyrolytically coated tubes and five graphite platforms to analyse serum samples.Any one of these tubes showed a precision after about 150 firings of less than 5.9% at 100 yg 1-1 of Al. Furnace lifetimes in excess of 200-250 firings using oxygen ashing were routinely achieved. In this study 198 serum samples were analysed by the method of standard additions and on pooling the results a line with slope = 38.8 x 10-4 y-intercept = 9.0 x 10-4 and r = 70 60 .g 50 c 3 ?-F .= 40 e 2 0 .-a 30 -c Y m a 20 1 I I 1 20 1 00 150 180 Number of firings Fig. 4. Change in peak shape with increased number of firings. Signal of 1 ng of A1 in serum from a platform in a pyrolytically coated tube. Values on peaks are signal areas in absorbance seconds I 1 I 1 1 1 1 I 0 50 100 150 200 250 300 60 Number of firings Fig.5. Signal repeatability of a dilute (1 + 1) serum sample during the life of the pyrolytically coated graphite tube with the L’vov platform. Average of five determinations (+la) is plotted every ten firings. The tube failed after 400 firings. (a) Tube No. 1; ( b ) tube No. 2 with the “old” platform 0.998 was produced. This line (Fig. 1) was not significantly different from the calibration graphs prepared initally with pooled sera with low aluminium content (slope = 38.0 x 10-4; y-intercept = 7.5 x 10-4; and r = 0.988) and with aqueous solutions (slope = 38.6 x lo-4;y-intercept = 1.2 X 10-4; and r = 0.999). Results obtained for 169 serum samples by the method of standard additions minus those from the calibration graph did not differ significantly [mean difference = +2.05 k 3.99 (1 s.d.) yg 1-1 of All.The concentration of aluminium in sera ranged from 4 to 108 yg 1-1 of Al. From these results it was established that for an accurate determination of aluminium in serum the method of standard additions is not necessary and a single calibration procedure utilising aqueous solutions is adequate. The calibration method permits the direct calculation by microprocessor of the aluminium concentration from a single calibration point using the specific function available on the Model 5000 spectrophotometer. To obtain reliable results it i 22 important particularly as the tube ages to check that the calibration has not changed (recalibration function).The aluminium concentration in serum from healthy subjects (n = 40) was determined to be 17.3 f 6.1 (range 2.C36.0 pg 1-1 of Al). The main criterion used for including samples was that the donor was not taking aluminium-containing antacids. Serum aluminium concentration in 68 samples (haemodialysis patients and patients on continuous ambulatory peritoneal dialysis) was 29 k 16 pg 1-1 A1 (range 8-106 pg 1-1 Al). A significant statistical difference between these two groups of data was present as the t-test for the mean produced a value of p < 0.001. Further suitable considerations such as sex and age of the subjects type and length of treatment etc. will be presented in a later paper. Many workers have suggested that sample contamination during blood collection can contribute to the higher alumin-ium values in normal subjects.In order to verify that all manipulations were performed without the introduction of random contaminations we collected and analysed separ-ately two portions of serum for each subject (87 samples) and four aliquots of serum for one normal subject. The aluminium concentrations determined in the four serum specimens were respectively 16 17 16 and 18 pg 1-1 of A1 (mean 16.7 k 1.0) and the results obtained for 84 samples (16 normal subjects and 68 patients) did not differ significantly (t-test for paired observations produced a value of p > 0.01). In this test only three samples were found to give anomalous results and were rejected. (a) 64 and 35 pg 1-1 of Al; (b) 25 and 44 pg 1-1 of Al; and (c) 24 and 44 pg 1-1 of Al.Samples (a) and (c) were “normal” subjects and (b) was a haemodialysis patient. Conclusions The results obtained show that the determination of serum aluminium concentration using the stabilised temperature furnace with chemical matrix modification is acceptable. Reliable results can be achieved by standardising the pro-cedure against a calibration graph if integrated absorbance signals are used for quantitation. The combination of the ANALYST JANUARY 1985 VOL. 110 platform matrix modification integrated absorbance oxygen addition to the charring step and new coated tubes provided better precision and smaller variation during the life of the tube. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. References Ward M.K . and Parkinson I. S . “Analytical Difficulties in Plasma and Serum Aluminium Determination,” paper presen-ted at an International Workshop on The Role of Biological Monitoring in the Prevention of Aluminium Toxicity in Man, Luxembourg 5-7th July 1982. Fell G. S . “Electrothermal Atomic Spectrophotometric Analysis of Aluminium in Blood Serum Plasma,” paper presented at an International Workshop on the Role of Biological Monitoring in the Prevention of Aluminium Toxicity in Man Luxembourg 5-7th July 1982. Slavin W. Carnrick G. R . and Manning D . C. Anal. Chem. 1982 54 621. Slavin W. Manning D. C. Carnrick G. R. and Pruszkow-ska E . At. Spectrosc. 1981 2 69. Leung F. Y. and Henderson A. H. Clin. Chon. 1982 28, 2139. Delves H. T. and Woodward. J. At. Spectrosc. 1981 2 65. Eaton D. K. and Holcombe J. A Anal. Chem. 1983 55, 946. Slavin W Manning D. C. and Carnrick G. R . AGal. Chem. 1981 53 1504. L’vov B. V. Spectrochim. Acta Part B 1978 33 153. Manning D . C . Slavin W. and Carnrick G. R . Spectrochim. Acra Part B 1982 37 331. Leung F. Y. and Henderson A. R. At. Spectrosc. 1983,4,1. Slavin W. Manning D. C. and Carnrick G. R. At. Spectrosc. 1981 2 137. Casetta B. Nardini R. and Plazzotta M. “Proceedings of a Workshop on the Study of Serum Aluminium Determination,” USL 7 Udinese Udine 22nd June 1983 in the press. Paper A411 78 Received May 14th 1984 Accepted August 8th) 198

ISSN:0003-2654    

DOI:10.1039/AN9851000019

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

ANALYST JANUARY 1985 VOL. 110 23 Determination of Electroactive and Non-electroactive Gases Using a Membrane Polarographic Detector in a Flow System Andrew Mills and Carl Lawrence Department of Chemistry University College of Swansea Swansea SA2 8PP UK A simple method has been devised for analysing toxic corrosive gases such as C12 NO2 and SO2 using an oxygen membrane polarographic detector (02-MPD) coupled to a flow system and an inert carrier gas such as N2. The flow system minimises the degree of corrosion and poisoning of the MPD so often found when analysing these gases. It was shown that by using an electroactive carrier gas such as air (instead of N2) the same 02-MPD - flow system could also be used to detect non-electroactive gases such as N2 Ar He CH4 and C02. The peak current observed following injection of an electroactive or non-electroactive gas was found in all instances to be directly related to the amount injected.The application of MPD‘s in gas chromatography is discussed in the light of this work. Keywords Electroactive gas determination; non-electroactive gas determination; flow system; membrane electrode; Clark cell Following the development of the oxygen membrane polaro-graphic detector (02-MPD) by Clark,’ many groups2-10 have tried with varying degrees of success to develop MPD’s for gases other than 02. Indeed quite early on Sawyer et a1.2 showed that C12 Br2 NO2 and SO2 could be detected using an 02-MPD whereas H2S H2 and N2 could not. Over the last decade a great deal of progress has been made in this area.For example Ben-Yaakov,3 using a porous membrane construc-ted an MPD for the determination of Br2 C12 I2 and the acids HOCl and HOBr; Albery and co-workers developed MPD’s for N204 and C025 and more recently we developed ones for H26 and D2.7 In addition to these developments with conventional “Clark-type” MPD’s Bergman and co-workers8-10 have shown that the “metallised membrane electrode” (which does away with the electrolyte layer and all the complications it affords) can be used to detect 0 2 Hz CO and NO2. Subsequent work by Pletcher and Gibbs” con-firmed these findings and showed that this type of electrode could be also used to detect SO2 and NO. A detailed report by Bergman on the polarographic curves for many oxidising and reducing toxic gases (such as C12 NO2 SO2 NO and H2S), recorded using a metallised membrane electrode has been published recently.12 However as has been noted by both Sawyer et a1.2 and Bergman,Q detection of large amounts ( i e .partial pressures ca. > 0.5 atm) of some of these toxic gases can prove difficult. For example both these workers noted some degree of corrosion of their MPD’s whilst monitoring high levels of C12 NO2 or SO2. Thus high partial pressures of these gases can lead to the permanent damage of an MPD as well as erratic and irreproducible readings. Indeed to avoid problems of this nature Bergman used only dilute forms of these gases (usually 1%) in his polarographic study. 12 In addition to corrosion of the MPD erratic and irrepro-ducible results can also occur if the working electrode is poisoned by the electroactive gas.Sawyer has reported2 this to be so for SO2 and H2S with the latter gas appearing particularly effective as a poison. Pletcher et a1.,13 using a metallised membrane electrode have reported that CO will also poison the working electrode (in this instance Au or Pt supported on a PTFE membrane) of an MPD. As with the problem of corrosion dilution of the electroactive gas with an inert gas (such as N2) appears to offer a partial solution to the problem as it reduces the rate of deactivation of the electrode surface. In the first part of this paper we describe how a variety of toxic electroactive gases can be detected and how the problems of poisoning and corrosion can be reduced substan-tially by using an 02-MPD coupled to a flow system with an inert carrier gas (such as N2).The second part of the paper is concerned with what at first sight appears to be a contradiction in terms i. e. the detection of non-electroactive gases using an 02-MPD. In it we describe how this rather unusual situation can be achieved by using the same 02-MPD coupled to a flow system as mentioned in the previous paragraph but this time with an electroactive carrier gas (such as air). Experimental Apparatus and Reagents The main components of the flow system used are shown in Fig. 1. A fine control of carrier gas flow-rate was achieved using an Edwards needle valve (Model LB1 B) attached at the head of the cylinder (1). Flow-rates were measured using a calibrated flow meter (2) (Glass Precision Engineering Ltd., Model R S X ) incorporated in the line (see Fig.1). The carrier gas flowed from the cylinder (1) into the mixing chamber (3) [a 125-cm3 Pyrex Dreschel bottle modified to receive a rubber septum (4)] and then on to the 02-MPD ( 5 ) . It is worth noting that as the 0,-MPD measures the partial pressure of the electroactive gas it is important to keep the flow-rate (and therefore the pressure of the gas reaching the detector) stable. A detailed description of the 02-MPD ( 5 ) (designed and constructed in this department and now available commercially from Rank Bros. Cambridge) has been given in a previous paper.14 The gases C12 NO2 and SO2 were supplied by BDH Chemicals Ltd. all others were obtained from BOC.2 \ 4 I 1 3 Fig. 1. Schematic diagram of the flow system coupled to the 0,-MPD. The components are 1 carrier gas cylinder; 2 flow meter; 3 mixing chamber; 4 rubber septum and 5 0,-MP 24 ANALYST JANUARY 198.5 VOL. 110 Procedure Nitrogen (white spot BOC) was used as the carrier gas in the work on electroactive gases whereas air (BOC) was used in the work on non-electroactive gases. In either instance small volumes (usually 0.1 5 Vi I 30 ml) of the gas under study (electroactive or non-electroactive) were injected via the rubber septum (4) (see Fig. l) into the mixing chamber (3), and then swept out by the appropriate carrier gas to the O2-MPD for detection. In all of this work the Pt working electrode of the 02-MPD was polarised at -0.7 V verms the Ag - AgCl counter - reference electrode using either a Metrohm potentiostat (Model E611) or a potentiostat desig-ned by Mills and Pavlou and now available commercially from Rank Bros.(Cambridge). Both potentiostats had a compensa-tion current facility which was essential for the work on non-electroactive gases. The output of the 0,-MPD was recorded on a Servogore 210 x - t chart recorder. Results and Discussion Detection of Toxic Electroactive Gases The electroactive gases selected for study (i.e. C12 NO2 and SO2) are good oxidising agents and therefore will not only be readily reduced at the working electrode of an 02-MPD but will also corrode and damage other parts of it (e.g. the silicone rubber retaining ring2). In addition as discussed in the introduction SO2 can poison the surface of the working electrode (particularly Pt) of an 02-MPD.2 The extent of corrosion and/or poisoning of an 0,-MPD by any one of these electroactive gases will amongst other things depend upon: (a) the partial pressure of the gas and ( b ) the duration of its exposure to the detector.Thus when using an MPD to detect such gases it is important to minimise both (a) and ( b ) and this we can do by coupling the detector to a flow system see 'Fig. 1. The partial pressure of the electroactive gas (Peg) immediately after ( t = 0) injection into the mixing chamber, assuming perfect mixing would be given by vi P' v * * (1) Peg = - . . . . . . where V; = volume of electroactive gas injected (cm3); V = volume of the mixing chamber (cm3); and P' = the total pressure in the mixing chamber (atm).This would then be expected to decay exponentially as the electroactive gas is swept out of the mixing chamber by the inert carrier gas to the MPD.14 The subsequent response of the MPD would be given by where id ( t ) = diffusion-controlled current at time t after injection (PA); ieg = diffusion-controlled current when the electroactive gas instead of N2 is flowed through the system (PA); andf = flow-rate of the carrier gas (cm3 min-1). Equations (1) and (2) have been verified recently using O2 as the electroactive gas.14 They imply that the degree of poisoning and/or corrosion of an 02-MPD by a toxic electroactive gas such as C12 or SO2 can be minimised by incorporating the MPD into a flow system using high flow-rates 0 low injection volumes (Vi) and a large volume mixing chamber (VJ.This is because under these conditions, the partial pressure and exposure time of the electroactive gas to the Oz-MPD would be relatively small. A large variety of volumes (Vi = 0.2-5 ml) of each dlectroactive gas (i.e. air NO2 C12 and SO2) were injected into the mixing chamber of the flow system (see Fig. 1). In order to minimise the contact time between the 02-MPD and the gas under study a high flow-rate cf = 310 cm3 min-1) was employed. Fig. 2(a) shows a typical current versus time output from the 02-MPD on injection of 5 ml of one of these E $ 4 0 c 7 . c ? 3 0 2 1 1 1 1 1 0 2 4 6 8 1 0 1 1 1 1 1 1 0 2 4 6 8 1 0 1 2 Ti me/m i n Fig. 2.Typical output from the 0,-MPD following an injection of 5 ml of (a) NO2 and ( b ) SO2 (third injection) into the mixing chamber of the flow system using N2 as the carrier gas cf = 310 cm3 min-1). The time origin is arbitrary Table 1. Data on non-electroactive gases Property Air No c12 so2 No.ofpoints . . 8 7 8 7 Least-squares analysis* GradienthA ml-1 13.7 k 0.1 12.2 k 0.1 47.5 k 0.4 14.3 k 0.4 Correlation coefficient . . 0.9998 0.9999 0.9998 0.9986 Peak half-life (f4)?/s . . 20 30 30 361 PTFEatNTP§/ . . 0.87 1.39 - 1.2.5 Permeability in * For a plot of ip0 versus V over the range 0.2 5 V 5 5 ml;f = 310 t t+ = time taken for id(f) = ipO/2. $ This is f + for the first injection r4 increases thereafter with each subsequent injection [see Fig. 2(b)].§ From reference 11; units lo-' cm-3 per cm atm s. cm3 min-l. electroactive gases (in this instance NOz). Ideally the peak current (ipo) should be given by (3) i.e. ipo should be directly related to the volume of the electroactive gas injected (Vi). This was in fact found to be so and the results of a simple least-squares analysis of the data for an ipo versus Vi plot for each gas are given in Table 1. From Table 1 we can see that the flow system provides an excellent method for the quantitative analysis of gaseous samples containing large amounts of an electroactive component such as 02 C12 and NO2 and in all this work none of the problems2 due to MPD corrosion were encountered. However although corrosion of the MPD appeared minimal some poisoning of the Pt working electrode by SO2 was observed particularly when large (Vi > 3-5 ml) injections were made.Unlike the other electroactive gases used large injections of SO2 led to current versus t profiles that changed in shape after each injection. Fig. 2(b) shows a typical output from the 0,-MPD after a third injection of 5 ml of SO,; the first injection appearing very similar to that of NO2 [see Fig. 2(a)]. The poisoning of the Pt working electrode by SO2 appeared to manifest itself in the form of peak broadening and an increase in the residual or base line current after each injection, however the peak current (ip") (as measured with reference to the pre-injection base line) appeared less affected. The effects of poisoning decreased with decreasing volume of SO2 injected becoming unnoticeable when Vi < 1 ml ANALYST JANUARY 1985 VOL.110 In a previous paper14 we showed that the time taken for id(t) to reach i,o/2 i.e. the half-life of a peak (f$) should be V” f ti=-ln2 . . . . . . . (4) for an “ideal” MPD i.e. one with a zero response time. As in all our work V = 156 ml and in this study f = 310 ml min-1, then using equation (4) we can calculate that ti = 21 s. In Table 1 are listed typical half-life values observed for a 5-ml injection of each gas. In agreement with our previous work,14 equation (4) appears to hold when 0 is the electroactive gas. However, a slightly more sluggish response by the 02-MPD was observed for C12 SO2 and NO2 and is probably due to the interaction of these gases with the electrolyte layer to form acids (e.g.HOC1 H2SO3 and HN02) which can subsequently diffuse to and then react at the Pt working electrode. This effect could be avoided if a metallised membrane electrode, rather than a Clark-type electrode is used and this indicates an important advantage of the Bergman-type electrode. 12 The gradients from the i,O versus Vi plot for each gas given in Table 1 correspond to the average peak currents per ml of gas injected. The peak current is determined by amongst other things (a) the degree of mixing with the carrier gas ( b ) the permeability of the electroactive gas toward the mem-brane material (in this instance PTFE) covering the sensing electrodes and (c) the degree of catalytic activity shown by the Pt working electrode towards the reduction of the gas.Similar peak shapes and t+ values indicate that at these high flow-rates (f = 310 cm3 min-I) the degree of mixing of the electroactive gas with the carrier gas is similar for all gases studied. As the permeabilities of these gases towards PTFE are also similar (see bottom of Table l ) it appears likely that the large differences in peak current between O2 and C12 on the one hand and NO2 and SO2 on the other are due to the catalytic factor outlined in (c). So far we have concerned ourselves with only the detection of toxic oxidising gases however it is obvious that any gas that can be made to react at the working electrode of an MPD (these include 02 Cl, Br, I, NO, NO N20 CO C02, SO2 H2S H2 and D2) could be analysed using a flow system coupled to an MPD.As mixtures of such gases could be separated by chromatography they too could be analysed using such a system. However by its very nature the MPD is limited to the detection of electroactive gases and in the next section we describe how this major problem can be overcome simply with the result that non-electroactive gases can be detected using the same 0,-MPD coupled to a flow system as described here. Detection of Non-electroactive Gases If air instead of N2 is used as the carrier gas in the flow system (see Fig. l ) then it should be possible to “detect” non-electroactive gases using the 0,-MPD as such gases will cause a transitory lowering of the partial pressure of O2 in the mixing chamber on injection which can be subsequently monitored, down the line by the O2-MPD.Assuming perfect mixing the current versus time output of the 0,-MPD following injection of V; cm3 of a non-electroactive gas into the mixing chamber would be id(t) = I,ir [v - vi exp(-ft/~,>l . . . . ( 5 ) where iair = diffusion-controlled current when only air is flowing through the system. It follows from equation ( 5 ) that the maximum drop in current which we shall call the peak current (i,’) and which should be observed immediately ( t = 0) following injection, can be written as v o 25 Fig. 3(a) shows a typical current versus time output from the 0,-MPD immediately following injection of 5 ml of N2 into the flow system (f = 180 cm3 min-1). The observed peak current of 64 nA is in good agreement with that predicted by equation (6) where i,’ = 70 nA (as Vi = 5 mi V = 156 ml and i,, = 2.2 PA).In addition as predicted by equation ( 5 ) the decay of the peak current i,’ is exponential and a plot of ln[id(t) - iair] versus t yields a reasonable straight line ( r = 0.9976) over two half-lives with a gradient (m) = 1.21 f 0.04 min-1. This value for the gradient agrees well with that predicted by equation (9 i.e. rn = 1.15 (= flV,). However, injection of other non-electroactive gases (i.e. gases that cannot be reduced at the Pt working electrode of the 02-MPD) such as He Ar CH4 and C02 did not lead to similar “ideal” responses from the 02-MPD. From Table 2 we can see that the typical peak currents (i,’) recorded for a 5-ml injection of each gas. are dissimilar.Leaving aside CH4 and C 0 2 it is interesting to note that the peak currents are in the order He > N2 > Ar although the peak areas are approxi-mately ( 2 5 % ) equal. As the densities of these three gases are in the order Ar > N2 > He it may be that the trend in i,’ reflects different degrees of mixing of the injected inert gases with the carrier gas possibly due to the low flow-rate (f= 180 cm3 min-1). For CH4 however the peak current is greater 220 218 216 214 a : 5) ?? 5 220 m I . + 0 216 21 2 208 0 1 2 3 4 5 Tirne/mi n Fig. 3. Typical output from the 02-MPD following an injection of 5 ml of (a) N and ( b ) C02 into the mixing chamber of the flow system using air as the carrier gas (j = 180 cmj min-I). The time origin is arbitrar 26 ANALYST JANUARY 1985 VOL.110 Table 2. Data on non-electroactive gases Property N Ar He CH4 C02 Typicalpeakcurrent*/nA 64 47 78 86 154 TypicalpeakareahAmin 67 61 64 58 -Result of least squares analysis? No.ofpoints . . . . 12 12 - 16 13 Correlation coefficient 0.9993 0.9972 - 0.9988 0.9960 * For a 5-ml injection of gas; f = 180 cm3 min-1. -t For a plot of i,’ versus V over the range V = 30-0.2 ml; f = 260 cm3 min-1. than that of He even though its density is lower. It is probable that this is due to the adsorption of CH4 on to the surface of the Pt working electrode and subsequent inhibition of the reduction of 0 2 and possible oxidation of the adsorbed CH4. All the gases discussed so far gave similar peak shapes i.e., an initial high peak current following injection which then decayed exponentially but with C 0 2 this was not found to be so and Fig.3(b) shows a typical response of the 02-MPD following injection of 5 ml of C 0 2 into the mixing chamber. Possible causes for this anomolous behaviour include altera-tion of the pH of the electrolyte surrounding the Pt working electrode due to carbonic acid formation and adsorption of C02 on to the electrode surface. Despite this unusual peak shape the peak current (ip’) was found to be directly related to the amount of C 0 2 injected. Indeed this was found to be so for all the gases used in this study (i.e. CH4 C02 N2 He and Ar) over the range 0.2 ml 5 Vi I 30 ml f = 260 cm3 min-1. At the bottom of Table 2 are listed the correlation coefficients, along with the number of points used in its calculation from a plot of i,’ versus Vi for each gas.In conclusion by using an electroactive carrier gas such as air it has proved possible to detect quantitatively non-electroactive gases (i.e. gases that cannot be reduced at the Pt working electrode of the 02-MPD) with an 02-MPD coupled to a flow system (see Fig. 1). Quantitative analysis of a mixture of non-electroactive gases (or electroactive gases for that matter) using an 02-MPD would require separation (by gas chromatography for instance) of the mixture prior to detec-tion due to the limited selectivity of the 0,-MPD. This idea is not original however as Bergman et a1.10 have already shown that mixtures of CO and H2 could be separated by gas chromatography and subsequently detected using a metal-lised membrane electrode.In addition Blurton and Stetterls have used a PTFE-bonded diffusion electrode as the detector in the gas chromatography of many electroactive gases, including H2S NO CO SO2 and NO2. So far the use of MPD’s in gas chromatography has been limited to the detection of electroactive gases but from the work described in the latter part of this paper we can now see that non-electroactive gases may also be detected using an MPD, after separation by gas chromatography provided an elec-troactive carrier gas is used. Indeed just recently we have shown that the major components of petroleum spirit (b.p. 40-60 “C) (including 2-methylbutane pentane 2,2-dimethylbutane 2,3-dimethylbutane 2-methylpentane, 3-methylpentane and hexane) can be detected after separation by gas chromatography using an H2-MPD and H2 as the carrier gas.A detailed account of this work will appear at a later date but the principles upon which it is based are those described in the latter part of this paper. Interestingly the idea of indirect detection is not altogether new to chromato-graphy; indeed its use in the form of indirect UV detection is becoming increasingly popular in ion chromatography. 16 We thank the SERC for financial support of this work. In addition we gratefully acknowledge Mr. R . Enos and Miss S. L. Giddings for their technical assistance. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. References Clark L.C. Trans. Am. SOC. Artq. Intern. Organs 1956 2, 41. Sawyer D. T. Raymond S. G. and Rhodes R. C. Anal. Chem. 1959 31 2. Ben-Yaakov S . J. Electroanal. Chem. 1979 98 15. Albery W. J. Brooks W. N. Gibson S. P. Heslop M. W., and Hahn C. E. W. Electrochim. Acta 1979 24 107. Albery W. M. and Baron P. J. Electroanal. Chem. 1982. 138 79. Mills A . Harriman A . and Porter G. Anal. Chem. 1981, 53 1254. Mills A . Analyst 1984 109 95. Bergman I. and Windle D. A . Ann. Occup. Hyg. 1972 15, 329. Bergman I. Proc. Conf. Environ. Sensors Appl. 1974,29,67, Institution of Electronic and Radio Engineers London. Bergman I. Coleman J. E. and Evans D. Chromato-graphia 1975 8 581. Pletcher D. and Gibbs T. K. Electrochim. Acta 1980 25, 1105. Bergman I. J. Electroanal. Chem. 1983 157 59. Pletcher D. McCallum C . and Gibbs T. K. Efectrochim. Acta 1977 22 525. Mills A . and Lawrence C . Analyst 1984 109 1549. Blurton K. F. and Stetter J. R. J. Chromatogr. 1978 155, 35. Small H. and Miller T. E. Anal. Chem. 1982 54 462. Paper A 412 4 1 Received July 18th 1984 Accepted August 9th 198

ISSN:0003-2654    

DOI:10.1039/AN9851000023

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

ANALYST JANUARY 1985 VOL. 110 27 Preparation of Coated Plastic or Glass Rod Epoxy-based Voltammetric Electrodes Using a Multi-layer Coating and Vapour Hardening Technique Hilbert P. Henriques and Arnold G. Fogg Chemistry Department Lo ugh borough University of Techno log y Loug h borough Leicestershire LE I I 3TU UK ~ ~ ~ ~ ~~ ~~ A method is recommended for producing graphite-loaded epoxy-based voltammetric electrodes by coating plastic or glass rods. A multi-layer coating technique is used each coating being hardened by absorbed hydrofluoric acid vapour. Electrodes of sufficiently low electrical resistance are produced by rubbing graphite into several of the lower coatings immediately after they are hardened. Unlike coated-wire electrodes complete coverage of the surface being coated is not essential.Keywords Graphite-loaded epoxy-based electrodes; hydrofluoric acid vapour hardening; voltammetry; coated rod electrodes Previously l a new multi-layer coating and hardening technique was reported for the preparation of highly satisfactory graphite-loaded epoxy-based voltammetric electrodes and the technique was applied to the preparation of disc and coated-wire electrodes. The surface that was to support the electrode was coated thinly and smoothly with a graphite-loaded epoxy base containing no hardener. The surface was then placed in a vapour chamber over 40% mlm hydrofluoric acid for a few minutes before being removed from the chamber and the epoxy coating was hardened at 50-60 "C by the hydrofluoric acid it had absorbed.The electrode surface was built up by repeating the process. The coated wire electrodes were prepared from new copper wire. This was treated first with 5 M nitric acid solution and then with acetone and allowed to dry without washing it with water at any stage. A small amount of graphite-loaded epoxy was placed on the wire and smoothed with a tissue and then a thin flexible polythene or Cellophane sheet. The very thin and smooth layer that was produced was then hardened using the vapour hardening technique. Further hardened layers were built up by repeating the process. To produce perfectly satisfactory graphite-coated wire electrodes it was found to be essential to cover the copper surface completely as otherwise the copper surface affected the response.A particularly vulnerable place on the surface is the tip and it was recommended that whenever feasible this should be isolated from later contact with the electrolyte. An unloaded epoxy containing 40% rnlm hydrofluoric acid as hardener was used to isolate the tip and also to seal the coated-wire electrode into a glass tube holder. In the work described here the multi-layer coating and vapour hardening technique has been extended to the production of coated plastic and glass rod electrodes. An advantage of using plastic or glass instead of wire is that it is not as critical to coat the surface as thoroughly because the plastic or glass is non-conducting. A disadvantage arises from this non-conductivity however in that coated plastic rod or coated glass rod electrodes produced in a similar way to the coated wire electrodes' have too high an electrical resistance.This difficulty has been overcome here by enriching the first few epoxy layers with graphite powder immediately after they have been hardened. Electrodes produced in this way have a sufficiently low resistance. Experimental Graphite-loaded Epoxy Base The graphite-loaded epoxy was prepared by mixing 0.44 g of epoxy base ( i e . the contents of the adhesive tube of a two-tube Araldite pack) (Ciba-Geigy) with 0.56 g of Specpure graphite pelletable grade 1 (Johnson Matthey Chemicals) on a suitable plastic or glass surface. This mixture can be stored indefinitely in a suitable glass or plastic container. Gaseous Hardener The plastic rod or glass rod to be coated was treated with a thin layer of the graphite-loaded epoxy base as discussed later.A plastic screw-cap container was used as a vapour chamber. A small amount (2 ml) of 40% mlm hydrofluoric acid was placed in a plastic beaker which in turn was placed in the container. Previously 1 articles to be hardened were either suspended inside the container in a nylon net or suspended through small holes drilled in the cap. In this work an alternative method was also used. A holder for a plastic beaker was constructed from copper wire such that the holder could be hooked over the side of the container. Articles to be hardened were placed in the beaker in the container for 5 min to absorb hydrofluoric acid. The articles were then removed from the container and heated at 50-60 "C over a hot-plate.Caution-Extreme precautions should be taken when using hydrofluoric acid. Work should be carried out in a fume-cupboard using the relevant safety equipment. Care must also be taken not to handle directly surfaces that may still retain some hydrofluoric acid. Graphite-loaded Epoxy Resin This was prepared immediately before use by mixing 1 drop (0.1 ml) of 40% mlm hydrofluoric acid with 0.5 g of graphite-loaded epoxy base prepared as above. The mixture hardened in about 40 min at room temperature and was used to form electrically conducting bonds between conducting materials. Isolating Epoxy Resin This was prepared immediately before use by mixing 1 drop (0.1 ml) of 40% mlm hydrofluoric acid with 0.5 g of unloaded epoxy base.The mixture hardened in about 40 min at room temperature. Preparation of Graphite-loaded Epoxy Plastic Rod or Glass Rod Electrodes Plastic and glass rods (2-3 mm diameter) were used in 7-cm lengths. The plastic rods used were in fact knitting needles made of polystyrene. The surface of the rods was roughened using an abrasive paper (fine-grade abrasive Acton and Borman); the ends of plastic rods were flattened using this abrasive paper. The surfaces were then wiped with a tissue 28 ANALYST JANUARY 1985 VOL. 110 A B ( b ) 9raphite-loaded epoxy resin C Isolating epoxy resin Graphite-loaded epoxy resin Metal foil Fig. 1. (a) Types of coated rod electrodes A end surface active (disc); B end cylinder and C central cylinder. ( b ) Methods of making electrical contact A with a graphite-loaded epoxy-resin coating; B with metal foil; and C with a metal wire bonded in place with graphite-loaded epoxy resin washed with ethanol and dried.A very small amount of the graphite-loaded epoxy base was placed on the cylindrical surface and thoroughly smoothed over most of the length using a thin flexible Cellophane sheet. It was essential that only a thin film was left in order to obtain effective hardening. The coated rod was suspended in the vapour chamber over 40% mlm hydrofluoric acid for 5 min. The rods were then removed and hardened at 50-60 "C over a hot-plate for a further 5 min. For some types of electrode the end of the rod was also coated. A small portion of Specpure graphite was placed on the hardened surface and smoothed into the layer using a tissue.The graphitised layer was then rubbed intensively with a tissue in order to remove the excess of graphite powder. The surface now had a shiny mirror-like finish owing to the entrapment of graphite in the surface. The process of coating the surface with graphite-loaded epoxy and of hardening and graphitising the coating was repeated twice. The coating and hardening process was then repeated twice more without the graphitisation process. The combination of three graphitised layers and two normal layers produces electrodes of low resistance (<150 ohm for a 6-cm length) with a sufficiently robust surface. Three types of plastic and glass rod electrodes were prepared as illustrated in Fig.l(a). In type A both the cylindrical and end surfaces were coated with graphite but the cylindrical surface was isolated subsequently from later contact with the electrolyte by coating it with the isolating epoxy resin. This electrode is thus a coated disc electrode. In types B and C the end surface was left uncoated. The type B electrode is essentially an end cylinder the remainder of the cylindrical surface being treated with isolating epoxy resin; the end surface that is uncoated plastic was cleaned finally with an abrasive to remove any graphite particles. The type C electrode is of the central cylinder type. In this work the coated glass rod electrodes that were prepared were of types B and C. Electrical contact was made at the end of the rod away from the electrode surface.Several methods of making contact were used three of which are illustrated in Fig. l(6). In type A the graphite coating at the end of the electrode was covered with graphite-loaded epoxy resin which was allowed to harden. Electrical contact was then made very conveniently by means of a crocodile clip. In type B metal foil (e.g. aluminium foil) was wrapped round the coated surface and was held firmly in position by some means; electrical contact was made to this again possibly with a crocodile clip. In type c a copper wire was fixed to the coated surface by means of graphite-loaded epoxy. Testing the Electrodes The electrodes were tested in the static mode using slow (5-25 mV s-1) linear sweep voltammetry and differential-pulse voltammetry.The three-electrode system consisted of the test electrode a calomel reference electrode and a platinum counter electrode. Voltammetry was carried out by means of a Metrohm Polarecord 626 which incorporates its own recorder. The background currents associated with the electrodes in a range of buffer solutions were investigated and the oxidation of iodide food colouring matters and dopamine was studied as examples. Results As an extension of previous work' on the production of graphite-loaded epoxy-coated wire electrodes attempts were made to coat plastic rods in a similar manner. The resistance of a 6-cm length of plastic rod coated in the same manner even with eight to ten coatings was typically 1000 ohm or more. Using the present graphitisation procedure the resistance of similar electrodes with only five coatings was <150 ohm.Coated plastic and coated glass rod electrodes have an advantage over coated wire electrodes in that the material of the rod does not interfere with the response of the electrode. Thus highly satisfactory electrodes can be made without the same degree of care to ensure that the original surface is covered adequately. During the development stage in this work numerous plastic rod electrodes of the three types shown in Fig. l(a) were prepared. All were readily produced and behaved similarly. Most of the later tests were carried out using the end-cylinder type and the results presented are for this type. Coated-glass rod electrodes behaved similarly. The coating and vapour hardening procedure worked particularly well on slightly roughened glass presumably owing to the etching action of the hydrofluoric acid.The electrodes behave very similarly to the graphite-loaded epoxy-coated electrodes described previously' and to well behaved glassy carbon electrodes. Three slow linear sweep scans from 0 to +1.2 V were used to condition the electrodes as previously. 1 Background currents associated with a typical conditioned electrode in pH 2 Britton - Robinson buffer 0.18 M sulphuric acid pH 3 ammonium citrate buffer and pH 7.5 Britton - Robinson buffer are shown in Fig. 2. In all these buffers without deoxygenation the background current between -0.3 and + 1.1 V was less than 0.2 PA. Typical linear sweep voltammograms for the oxidation of 10-3 M potassiu ANALYST JANUARY 1985 VOL.110 r 29 I -0.8 0 +1.0 Potent i a l/V Fig. 2. Background currents (linear sweep voltammetry) for a conditioned graphite-loaded coated plastic rod electrode Lend-cylinder type). No deoxygenation. Three successive scans in A pH 2 Britton - Robinson buffer; B 0.18 M sulphuric acid; C pH 3 ammonium citrate buffer; and D pH 7.5 Britton - Robinson buffer. Scan rate = 10 mV s-1 + 0.3 +0.6 0 +0.5 Potent i a l/V Fig. 3. Typical voltammograms at a graphite-loaded coated plastic rod electrode (end-cylinder type). (a) Four successive linear sweep scans (10 mV s-1) without cleaning in a 10-3 M solution of potassium iodide in 10-l M potassium chloride solution. ( b ) Three successive scans (differential-pulse voltammetry 5 mV s-l 50 mV pulse height) of a 40 pg ml-1 solution of dopamine in pH 6.2 Britton - Robinson buffer.Cleaning between scans effected by placing electrode in 2.5 M sodium sulphite solution for 5 min at + 1.5 V iodide in 10-1 M potassium chloride solution are shown in Fig. 3(a); these show the excellent reproducibility of the signal without cleaning the electrode between scans. At the levels of determinand used in static systems, contamination of the electrodes with the product of electrode reactions is a problem in many instances as also occurs with other solid electrodes including glassy carbon electrodes. Chemical and solvent cleaning of electrodes coated with graphite-loaded epoxy was discussed previously. 1 The same methods were used with the present electrodes.Glassy carbon electrodes are slightly easier to clean because harder physical pressure can be applied during the cleaning procedure. The effectiveness of a particular cleaning process depends on the nature of the adsorbed product. In this work an effective way of cleaning electrodes that had been used with dopamine was to place them in 0.1 M sodium hydroxide solution or 1.5 M sodium sulphite solution and hold them at a potential of +1.5 V for 5 min. The differential-pulse voltammograms for dopamine shown in Fig. 3(b) illustrate this point. Linear sweep voltammograms obtained with food colouring matters were similar to those obtained previously.1 Discussion A method has been described for producing coated plastic and coated-glass rod voltammetric electrodes of low resistance and excellent performance.Their main advantage over coated metal wire electrodes is that complete coverage of the material of the rod is not essential. Indeed in one type of electrode described here the plastic or glass end of the rod is left uncovered. The electrodes described are easy to construct and may be regarded as disposable. Coated wire electrodes in general whether ion-selective or these new voltammetric electrodes have the advantage of being simple to make and there is the possibility of making them very small. The graphite-loaded epoxy disc electrodes have so far been tested only briefly in flow systems but good results have been obtained and no difficulties in this use are envisaged. As with glassy carbon electrodes contamination problems are expec-ted to become minimal at the low determinand concentrations that are measured in flow systems. A logical extension of the present work is the production of coated plastic and glass electrodes with other geometries, Indeed the coating and graphitisation procedures described here should allow the production in situ of electrode surfaces on a wide range of electrically non-conducting surfaces of virtually any shape. A preliminary study has been made of the production of a thin-film voltammetric cell constructed from coated plastic plates but further work is required to assess this. Work on the coating of carbon fibres and indeed of plastic fibres and glass fibres for possible use in vivo pharmacological applications is also being investigated. H.P.H. thanks the University of ViGosa Brazil for leave of absence and CAPES (Brazil) for financial support. Reference 1. Henriques H. P. and Fogg A. G. Analyst 1984 109 1195. Paper A41150 Received April 12th 1984 Accepted May 21st 198

ISSN:0003-2654    

DOI:10.1039/AN9851000027

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

ANALYST JANUARY 1985 VOL. 110 31 Polarographic Determination of Nitroxazepine Hydrochloride in Tablets" Arvind K. Mishra and Kamalakar D. Godet Pharmaceutical Chemistry Laboratory Department of Pharmacology Institute of Medical Sciences, Banaras Hindu University Varanasi-22 1005 India A polarographic method has been developed for the quantitative determination of nitroxazepine hydrochloride in the pure form and in tablets by the aid of different buffer systems. The substance is extracted from the sample with water the appropriate buffer of selected pH is added to an aliquot and the solution then polarographed at a dropping mercury electrode versus a standard calomel electrode. The resultant single reduction wave is well developed and permits a precise quantitative determination.The method of standard additions is used. Keywords Nitroxazepine hydrochloride determination; d.c. polarograph y; dosage forms Nitroxazepine hydrochloride 10-[3-(dimethylamino)propyl]-2-nitrodibenz[b,fl[ 1,4]oxazepine-l1 (10-H)-one hydrochloride (I) is an efficient tricyclic antidepressant drug endowed with a high therapeutic ratio that is used in the treatment of patients with depressive mood disorders. A review of the literature reveals that no attempt has been made to study the polarographic behaviour of nitroxazepine hydrochloride. However several reports pertaining to the polarographic assay of different tricyclic antidepressants with phenothiazine rings'-3 are available. The purpose of this work was to establish the experimental conditions that permit the study of the polarographic behav-iour of nitroxazepine hydrochloride and its determination in tablets.Most of the work was performed under purely aqueous conditions but as the addition of an aqueous solution shifts the half-wave potential of an electroactive ion to either a more positive or negative potential,46 various non-aqueous sol-vents such as methanol ethanol and propanol were used in order to study the effect of mixed-solvent composition on the half-wave potential ( E J of nitroxazepine hydrochloride. Experimental Apparatus and Conditions for Polarographic Analysis A manual polarograph S (Adept Laboratory Poona) in conjunction with a spot galvanometer was used for the current -voltage measurements. A two-electrode combination was used consisting of a saturated calomel electrode (S.C.E.) and a dropping mercury electrode (D.M.E.).All the measure-ments were performed at 25 & 0.2 "C. The D.M.E. had the following characteristics (in distilled water at 0.0 V open-circuit potential) m2W6 = 1.98 mg2'3 s-4 at a mercury column height ( h ) of 50 cm and an applied potential range of o.cb1.20 v. * Presented at the 34th Indian Pharmaceutical Congress held in i- To whom correspondence should be addressed. Varanasi 20t h-23rd December 1982. Controlled-potential Electrolysis A modified H-type Lingane cell,' with mercury pool cathode, platinum wire gauze electrode and spot galvanometer was used for the controlled-potential electrolysis. Suitable volumes ranging from 25 to 40 ml of buffer of selected pH (at which a well defined wave was obtained) were placed in each of the two compartments.The solution in the electrode compartment was deaerated with a stream of 02-free nitrogen for 10 min. An appropriate volume of purified mercury was placed at the bottom of the working electrode compartment and the applied potential was set, usually at 1.00 V for each of the samples and the corresponding current decay was noted using a galvanometer. The galvanometer was then disconnected from the cell and a known volume (25 ml) of nitroxazepine hydrochloride stock solution was added to the working electrode compartment and deaerated. The galvanometer was turned on again and electrolysis was allowed to proceed virtually up to completion (ca. 4 h). The corresponding current decay was plotted against time.Reagents and Solutions All of the chemicals used were of either AnalaR grade from BDH Chemicals or general-reagent grade from E. Merck. Four different buffer systems namely acetate (pH 3.60-5.60),8 McIlvaine (pH 2.20-8.00),9 borate (pH 7.80-10.00)10 and Britton - Robinson (pH 2.00-12.00),11 were prepared in distilled water. A stock solution (10-3 M) of nitroxazepine hydrochloride was also prepared in distilled water. A 0.2% aqueous solution of Triton X-100 was used12 to eliminate the polarographic maxima encountered throughout the polaro-gram. General Procedure for Studying Polarographic Behaviour Nitroxazepine hydrochloride (99.5%) was obtained from Hindustan Ciba-Geigy Ltd. Bombay (Sintamil) and was used without further purification.A 1.0-ml volume of the stock solution of nitroxazepine hydrochloride was taken in a polarographic cell 0.1 ml of Triton X-100 and 8.9 ml of the appropriate buffer of selected pH were added and the solution was purged with 02-free nitrogen for 10 min prior to each run. The stream of nitrogen was allowed to flow gently on the surface of the solution during the electrode reaction. The selected pHs were as follows acetate 4.00; McIlvaine 4.60; borate 8.00; and Britton - Robinson buffer 5.0 32 pH Dependence Studies The polarograms of nitroxazepine hydrochloride were obtained in each of the four buffer systems taken over the entire pH range and the optimum pH range which gave a well defined wave for each sample was also found.Effect of Mixed-solvent Composition To study the effect of mixed-solvent composition on the half-wave potentials of nitroxazepine hydrochloride varying percentages of methanol ethanol and propanol were added separately to the polarographic test solution and the polaro-grams were recorded after deaeration for 15 min. Analysis of Tablets Twenty tablets were weighed and the average mass per tablet was determined. A portion of the finely ground sample, containing 2540 mg of nitroxazepine hydrochloride was accurately weighed and transferred into a 100-ml calibrated flask containing 75 ml of distilled water. The contents of the flask were shaken for at least 20 min on a magnetic stirrer and then diluted to the mark with water. The solution was next filtered through a fine-pore filter-paper discarding the first 20 ml of the filtrate.A 5-ml aliquot of the clear filtrate was pipetted into a 50-ml calibrated flask 0.5 ml of Triton X-100 was added and the solution again diluted to the mark with the respective buffer of selected pH as previously described. A 10-ml volume of this solution was injected into a polarographic cell and polarograms were recorded for a 0.G1.20 V applied potential at a D.M.E. versus S.C.E. After obtaining the polarograms 1.0 ml of the standard solution (0.5 mg ml-1) of nitroxazepine hydrochloride was added to the cell deaerated for 2 min and again polarograms were recorded under the same conditions. The wave heights H and h were measured and the mass of nitroxazepine hydroch-loride per tablet was calculated using the following equationl3: Mass of nitroxazepine hydrochloride per tablet (mg) = ahb x 1000 where a is the mass of nitroxazepine hydrochloride reference standard in 100 ml of standard solution (mg); b is the average mass of a tablet (g); Wis the mass of sample (mg) taken for the polarographic determination; h is the wave height of nitrox-azepine hydrochloride before standard additions; H is the wave height of nitroxazepine hydrochloride after standard additions; and 1.10 is the dilution factor.* * (1) (1.10H-h)W * * * * Recovery Experiments In order to establish the reliability and suitability of the proposed method known amounts of the pure drug were added to various pre-analysed formulations of nitroxazepine hydrochloride and the mixtures were analysed by the pro-posed method.ANALYST JANUARY 1985 VOL. 110 Interference Studies The polarograms of the drug and suitable amounts of pharmaceutical adjuvents used in the tablet formulations i . c . , starch microcrystalline cellulose lactose talc and mag-nesium stearate were also recorded in order to study the possible interference of excipients on the nature of the wave. Results and Discussion The determination and study of the electrochemical behaviour of most of the tricyclic antidepressants containing a phenothiazine ring reported so far either involved an indirect determination3,14 or utilised the oxidative property of the phenothiazine ring at a dropping mercury or rotating platinum electrode2.15 whereas in this work nitroxazepine hydroch-loride (an oxazepine derivative) gave a single four-electron, reduction wave which may be assigned to the facile reduction of the nitro group at position 2 giving a characteristic E+ in all the buffer systems namely acetate (pH 3.60-5.60) McIlvaine (pH 2.2@8.00) borate (pH 7.80-10.00) and Britton - Robin-son (pH 2.00-12.00) over the entire pH range.A well defined wave was observed in each of these buffers in certain pH ranges (Table 1). The nature of the wave was found to be diffusion controlled in the buffer systems taken as shown by the linear depen-dence of limiting current on %‘h,,, and [depolariser] con-stancy of the wave height in the pH range studied and the fact that dildT had a very low temperature coefficient. The irreversible nature of the wave was confirmed by logarithmic plots.16 The slope value of the plot of Ed versus (did - i), which appreciably exceeded 59.2/n mV and the numerical value of E+ - E+ of the polarographic wave (appreciably exceeding 56.4ln mV) confirm the irreversible17 nature of the wave.The E4 of nitroxazepine hydrochloride was dependent on pH and shifted towards more negative potentials with increase in pH of the buffer systems. Fig. l(a) and ( b ) show the graphs of E+ versus pH for all the buffer systems taken and a straight line is observed in each example. The value of an (Table 1) was determined by the method of Oldham and Perry,lg where an is the product of the transfer coefficient and number of electrons per molecule of the reactant involved in the rate-determining step of the electrode process.The value of P (number of protons involvi?d per molecule of the reactant in the rate-determining step) was also determined19 (Table 1). An attempt to estimate n, the number of electrons involved in the rate-determining step gave a value of 2 because for totally irreversible systems as in this instance a should be less”) than 0.5. However according to Meites” only a single electron can be transferred at a time during the course of the electrode reaction and a value of n exceeding 1 should merely mean that successive steps are too close together to be distinguished on the time scale implicit in the polarographic measurements. However the total number of electrons involved in the reduction process of nitroxazepine hydro-chloride was found to be four as determined by the controlled-potential electrolysis.Table 1. Polarographic characteristics of nitroxazepine hydrochloride in various buffer systems. c = 10 -I M Optimum pH Selected Buffer range* PHt E,IV idlpA dE;ldpH at” P Acetate . . . . . . 4.00-5.00 4.00 0.38 0.374 0.077 0.96 1.26 McIlvaine . . . . . . 4.50-5.50 4.60 0.39 0.395 0.074 0.95 1.20 Britton - Robinson . . 4.0&5.00 5.00 0.43 0.384 0.074 0.95 1.20 Borate . . . . . . 7.80-8.60 8.00 0.59 0.291 0.076 0.90 1.16 * The pH range giving a well defined reduction wave. tThe pH at which the effects of various parameters were studied ANALYST JANUARY 1985 VOL. 110 33 2.0 3.0 4.0 5.0 6.0 7.0 8.0 0.40 ' I I I I I I I 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 PH Fig.1. Effect of pH on the half-wave potential of the nitroxazepine hydrochloride reduction wave in the buffer systems taken at c = 1.0 x 10-4 M . A . Acetate buffer; B McIlvaine buffer; C borate buffer; and D. Britton - Robinson buffer After establishing the stoicheiometry of the rate-determining step i.e. n = 2 and H+ (number of protons taking part) = 1 the following mechanism can be suggested for the polarographic reduction of nitroxazepine hydroch-loride which corresponds to the usual reduction mechanism22 for the nitro group. The total four-electron reduction process represents the reduction of nitroxazepine hydrochloride to the corresponding phenylhydroxylamine derivative. Table 2. Effect of increasing the percentage of non-aqueous solvents on E of nitroxazepine hydrochloride in McIlvaine buffer.c = M; pH 4.60 Ethanol Methanol, 5% E4/V idIpA % EdV idlpA 20 0.40 0.367 20 0.41 0.357 30 0.42 0.336 30 0.43 0.326 40 0.44 0.306 40 0.46 0.295 50 0.45 0.285 50 0.48 0.275 60 0.45 0.275 60 0.48 0.265 2e- "+ (rate-deter-mining step) I ,CH3 CH2CHZCH2N I 0 'CH3 V IV An increase in the percentage of methanol ethanol and propanol in the polarographic test solution shifted the half-wave potentials towards a more negative potential with simultaneous decrease in diffusion current (Table 2). It should be noted that with propanol it was not possible to study the effect at concentrations above 20% of alcohol as the wave became distorted. However at lower concentrations of propanol a significant negative shift in comparison with methanol and ethanol from -0.38 to -0.43 V versus S.C.E.in McIlvaine (pH 4.60) buffer was observed. An increase in the organic solvent content resulted in a rise in pH23324 and an increase in the dissociation constant of the protonated species.25 Both of these factors lower the rate of protonation and consequently lead to a shift in E+ of the reduction wave towards a more negative potential in all such situations where protonation proceeds electron transfer. The decrease in diffusion current may be partly due to an increase in the viscosity of the medium and partly to an ion-pair factor.26 The ion-pair factor must be considered because a continuous decrease in the diffusion current was observed. It also appeared that the above factors are not the only ones responsible for the observed shift in Ea; for nitroxazepine hydrochloride the observed shift is greater than it should be owing to the change in pH and dissociation constant.This additional shift in EQ may be ascribed to a decrease in adsorbability and hence surface concentration of the depolariser with an increase in the percentage of non-aqueous solvent in the aqueous - organic mixture.27 A decrease in surface concentration would retard the electrode process resulting in a decrease in EJ and id. Table 3. Assay of nitrovazepine hydrochloride tablets by d.c. polarography Buffer Acetate (pH 4.00) . . . . . . McIlvaine (pH 4.60) . . . . Borate (pH 8.00) . . . . . . Britton - Robinson (pH 5.00) Amount found Labelled by proposed amounthg method*/mg 25 75 25 75 25 75 25 75 24.7 74.9 23.9 74.9 24.6 74.5 24.9 74.9 Recovery by proposed method % 99.13 99.7 101.2 99.9 99.3 99.2 99.8 99.7 Standard deviation 0.01 58 0.0186 0.0 1 26 0.0 1 34 0.0164 0.0206 0.0158 0.0 164 Coefficient of variation % 0.0633 0.0248 0.0504 0.0178 0.0658 0.0274 0.0632 0.0218 * Each value is the average of five determinations 34 0.60 0.50 0.40 f .c 5 0.30 0 L 3 0.20 0.10 0 0.2 0.4 0.6 0.8 1.0 1.2 Ed.e. N VS. S.C.E. Fig. 2. Polarographic waves of nitroxazepine hydrochloride in McIlvaine buffer ( H 4.60) A before and B after the addition of the standard solution fn.5 mg ml-I) I Table 3 gives the results of the assay of the pharmaceutical dosage forms in all of the four buffer systems at their selected pH.The best results are observed with McIlvaine (pH 4.60) and Britton - Robinson (pH 5.00) buffers for both 25- and 75-mg tablets. In all four buffer systems analysis of the dosage forms is best performed in a less acidic pH range (pH > 3.60). With a more acidic medium it is better to use a smaller percentage of non-aqueous solvents. However in these analyses small percentages of only ethanol and methanol gave good results. Fig. 2 shows the polarograms of the extracted drug before and after addition of the standard solution of pure nitroxaze-pine hydrochloride. The method of standard additions is preferred because it is more rapid than a concentration -diffusion current plot method.Care should be taken with the size of the standard additions because it influences the relative error of the result. However there is an upper limit of concentration,28.29 which may vary considerably from one substance to another because above this the wave height of a substance is no longer proportional to its concentration. Moreover the best result is observed if the standard addition is large.30 None of the excipients commonly employed in the tablet dosage form of nitroxazepine hydrochloride were found to interfere with the assay of the drug. Apparent variations of idle can be produced by the potential impurities (if present) and react with the electroactive substance actually responsible for the wave. These are only apparent because it is actually c that is affected in each instance while the diffusion current may be accurately proportional to the concentration of the electro-active substance that remains.3’ bay, 1 2 3.4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21, 22. 23. 24. 25. 26. 27. 28, 29. 30. 31. ANALYST JANUARY 1985 VOL. 110 Financial assistance from Indian Drugs and Pharmaceutical Ltd. New Delhi is gratefully acknowledged. The authors also acknowledge with thanks Hindustan Ciba-Geigy Ltd. Bom-€or providing pure nitroxazepine hydrochloride. References Ellaithy M. M Zndiun J . Phurm. Sci. 1980 42 41. Chuen N. and Riedel B. E . . Curl. Pharm. J . Sci. 1961. 94, 51. Bhatt S . K. Arora K . K .Chakrabarty S. P . and Gode. K. D. Zndian J . Hosp. Pharm. 1979. 6 182. Parkar A. J. Q. Rev. Chem. Soc. 1962 16 163. Takahashi R . Tulanta 1965 12 1211. Kolthoff I. M and Hills G . J Editors “Polarography,” Volume 1 Macmillan London 1966. p. 1. Lingane J. J . Swain C. G. and Fields M. J . Am. Chem. SOC. 1943 65 1348. Walpole G . S . J. Chem. Soc. 1514 105. 2501. McIlvaine T. C. J. Biol. Chem. 1921 49 183. Clark W. M. and Lubs H. A . J. Bucteriol. 1921 2 1. Britton H. T. S . Editor “Hydrogen Ions,” Volume 1 Van Nostrand New York 1956. Meites L. Editor “Polarographic Techniques,” Interscience, New York 1965 p. 321. Marjan S. J . Pharm. Sci. 1976 65 736. Dumortier A. G. and Patriarche C. J . Fresenius 2. A n d . Chem. 1973 264 153. Kabasakalian P. and McGlotten J .Anal. Chem 1959. 31, 431. Meites L. Editor “Polarographic Techniques,” Interscience, New York 1965 p. 219. Meites L. Editor “Polarographic Techniques,” Interscience, New York 1965 p. 289. Oldham. K. B. and Perry E. P. Anal. Chem. 1965 40 65. Meites L. Editor “Polarographic Techniques,” Interscience, New York 1965 p. 248. Goto R . and Tachi I. “Sbornik Mezinarod. Polarograf. Sjezdu Prazu 1st Congress Part 1,” 1951. p. 69; Chem. Abstr. 1952 46 6967c. Meites L. Editor “Polarographic Techniques,” Interscience, New York 1965 p. 245. Chandra K. Shakya S. S . and Singh. M. Zndian J. Chem., 1982 21 254. Schwabe K. “Advances in Polarography.” Volume 3 Per-gamon Press Oxford 1960 p. 911. Schwabe K. “Advances in Polarography,” Volume 1 Per-gamon Press Oxford 1962 p. 333. Mairanovskii S. G. and Gul’tyai V. P Elektrokhimiya 1 , 460; Chem. Abstr. 1965. 63 977%. Srivastava 0. N. and Gupta C. M. Analyst 1972 97 204. Mairanovskii S . Talanta 1965 12 1209. Reinmuth W. H. Anal. Chem. 1956 28 1356. Eckschlager K . Collect. Czech. Chem. Commun. 1962 27, 1521. Meites L. Anal. Chem. 1956 28 139. Meites L. Ediror “Polarographic Techniques,” Interscience, New York 1965 p. 127. Paper A411 91 Received May 30th 1984 Accepted August 14th 198

ISSN:0003-2654    

DOI:10.1039/AN9851000031

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

ANALYST JANUARY 1985 VOL. 110 35 Determination of Lead Mercury and Cadmium by Liquid Chromatography Using On-column Derivatisation With Dithiocarbamates Roger M. Smith Arif M. Butt and Arun Thakur Department of Chemistry University of Technology Loughborough Leicestershire L E I I 3TU UK The combination of HPLC with direct injection and on-column derivatisation provides a method for the multi-element identification of metals in pollution studies. However problems have occurred when this technique has been applied to the determination of heavy metal ions. In this work the reasons for these difficulties has been studied and a modification of the procedure enabling lead cadmium and mercury(l1) ions to be determined is suggested. Keywords Heavy metal determination; liquid chromatography; dithiocarbamates; on-column derivatisation In recent years there has been considerable interest in the use of normal- and reversed-phase liquid chromatography for the determination of metal ions as their chelates.1.2 A range of different chelating groups have been used including dithizone, acetylacetone and the dithiocarbamates.In almost all of the studies the reagent was added in a pre-column reaction and the complex was extracted from the excess of reagent before analysis. It has been shown that for some metals a simpler procedure is to use a reversed-phase column and incorporate the reagent in the mobile phase.3.4 The sample ions can be injected as an aqueous solution and the chelates are then formed by an on-column reaction before separation.The chelates can be selectively detected by either UV - visible3 or electrochemical detectors4 without interference from the excess of reagent in the mobile phase. This combination of HPLC with on-column derivatisation and direct injection provides a potentially quick and easy method for the multi-element identification and determina-tion of metals in pollution studies or in trade wastes as an alternative to inductively coupled plasma spectroscopy or repeated atomic-absorption measurements. The technique has been successfully applied to the determi-nation of copper nickel iron(II1) and cobalt ions using sodium diethyldithiocarbamate (0.05%) as the reagent but lead and mercury ions gave poor peak shapes and low reproducibility of peak areas.3 However the pre-formed complexes of the heavy metals are apparently stable on reversed-phase chromatography as a number of workers have reported the separation of the dithiocarbamates of lead,5-11 merc~ry,6~8-13 and cadmium.5-8 11,13 This paper describes a study of the reasons for the problems encountered in the determination of heavy metals and suggests a modification of the original on-column reaction technique by changing the mobile phase that enables lead, cadmium and mercury(I1) ions to be determined following their direct injection as aqueous solutions.Experimental Reagents and Solutions Metal salts. Reagent-grade lead nitrate cadmium nitrate and mercury(I1) chloride were obtained from Fisons Scientific Apparatus Loughborough. Dithiocarbamate. Sodium diethyldithiocarbamate was obtained from BDH Chemicals Ltd.Poole Dorset. Methanol and chloroform. HPLC grade from Fisons Scien-tific Apparatus Loughborough. Apparatus Liquid chromatography was carried out using a Pye Unicam XPS pump connected to a column (10 cm X 5 mm) packed with ODs-Hypersil (Shandon Southern Runcorn) and a Pye Unicam PU 4020 variable-wavelength detector operating at 350 nm. Samples (10 PI) were injected using a Rheodyne 7010 valve into the mobile phase of methanol - water - chloroform (70 + 20 + 10) containing 0.05% m/V sodium diethyldithio-carbarnate which had a flow-rate of 1 ml min-1. Some of the studies with cadmium used methanol - water - chloroform (55 + 40 + 5) containing 0.05% m/V sodium diethyldithiocarba-mate. Results and Discussion In previous studies good linearity and repeatability were obtained from the determination of aqueous solutions of copper and other ions but not for lead or mercury by injection into methanol - water (80 + 20 V/V) containing 0.05% m/V sodium diethyldithiocarbamate.3 The first aim of this study was therefore to locate the cause of this poor reproducibility in peak heights and shapes for the heavy metal complexes.In earlier studies of pre-formed dithiocarbamate chelates by other workers some problems had also been encountered with this group of metals. Hutchins et al. suggested that the peak shapes were poor because of exchange reactions of the chelates with the nickel in the stainless-steel columns.11 They tried to suppress this effect by the addition of EDTA to the mobile phase and the use of plastic-walled columns.However, the EDTA totally displaced cadmium and lead ions from their Table 1. Separation of metal ions by reversed-phase HPLC by injection into a mobile phase containing dithiocarbamates Metalion t,l$min k' Pb(II)* . . . . 4.8 3.80 Cd(II)* . . . 3.7 2.70 Hg(II)* . . . . 7.2 6.20 Hg(I)* . . . . 5.9 4.90 Cd(I1)t . . . . 4.2 3.2 * Mobile phase methanol - water - chloroform (70 + 20 + 10) + t Mobile phase methanol - water - chloroform (55 + 40 + 5 ) + $ Retention time. 0.05% mlV sodium diethyldithiocarbamate. 0.05% mlV sodium diethyldithiocarbamate 36 ANALYST. JANUARY 1985. VOL. 110 Table 2. Calibration graphs for peak areas against concentration of Pb(I1) ions injected into a mobile phase containing sodium diethyldithiocarbamate.Mobile phase as in Table 1. Detection at 350 nm. Mean peak area/mm* Concentration Area of Pb(II) p.p.m. (attenuation) 0.1 93 (0.005) 0.2 198 (0.005) 0.4 349 (0.005) 0.6 286 (0.01) 1 .0 488 (0.01) 5.0 513 (0.04) 10.0 567 (0.08) Adjusted to 0.08 6 12 22 36 61 256 567 Correlation I . . . 0.9988 Slope . . . . . . 55.7 Intercept . . . . -0.6 * Results based on five replicate injections. Coefficient of variation * YV 14.4 3.1 1.6 4.6 2.0 1.6 2.1 Table 3. Calibration graphs for peak areas against concentration of Hg(I1) ions injected into mobile phase containing sodium diethyldithiocarbamate solution. Mobile phase as in Table 1. Detec-tion at 350 mn Mean peak arealmmz Concentration Area of Hg(II),p.p.m.(attenuation) 0.5 127 (0.005) Q.8 241 (0.005) 1 .O 312 (0.005) 5.0 207 (0.04) 10.0 404 (0.04) Correlation . . . . Slope . . . . . . Intercept . . . . * Results based on five injections. Adjusted Coefficient 8 12.3 15 6.6 19 2.7 103 3.1 202 2.4 to 0.08 of variation,* "/o 0.9998 20.4 -1.0 t i Pb +-PI .- J 6 7 6 5 4 3 2 1 0 Timeimin Fig. 1. Separation on an ODS-Hypersil column of chelates formed by the injection of 10 1 of a mixture containing 1 p.p.m. of lead, mercury(II) mercurp(1Y and cadmium ions into methanol - water -chloroform (70 + 20 f 10) containing 0.05% mlV sodium diethyldi-thiocarbamate as eluent. Detection at 350 nm and 0.01 a.u.f.s. chelates and they were unretained.Other workers have used EDTA to prevent exchange reactions during the separation of mercury dithiocarbamates. 12 To test if exchange was occur-ring in this study samples of lead ions were left in the injector loop or the syringe needle for varying times before injection into the mobile phase of methanol - water (80 + 20) containing 0.05% mlV sodium diethyldithiocarbamate. In other studies different ff ow-rates were used to change the residence time of the chelate in the column. In each instance the results were effectively the same with no evidence of an interaction changing the peak shapes or reproducibility. In a separate study the column was removed and the flow passed directly into the detector so that the system was effectively used in a flow injection mode.With this system the peak heights of the chelates formed from ten injections of 1000 p.p.m. lead nitrate solutions showed poorer reproducibility [coefficient of variation (c.v.) 8.1Y03 than azobenzene (c.v. 1.7%) which was used as a neutral non-reactive test sample. It appeared therefore that it was the derivatisation step rather than the column that was causing the problems. Either the reaction was not continuing to completion or the product was being partially precipitated causing peak broadening and poor reproducibility. Drasch and co-workers found that injection of pre-formed lead and cadmium diethyldithiocarbamates on to an octykilyl bonded silica column gave severely tailing and misshapen peaks with methanol - water (70 + 30) as the mobile phase.5.h As these chelates are readily soluble in organohalogen solvents they examined methanol - water - chloroform (50 + 25 + 25) as a possible mobile phase and good peak shapes and separations on the octylsilyl bonded silica column were obtained.They also successfully used methanol - water -chloroform (60 + 32 + 8) with an octadecylsilyl bonded silica column. Other laboratories reporting the separation of pre-formed chelates have also used unusual mixtures of solvents for the mobile phase including acetonitrile - water -ethyl acetate - 0.05 M sodium dithiocarbamate (60 + 24 + 5 + 5),7 acetonitrile - water - 0.01 M ammonium tetramethylenedi-thiocarbamate (69 + 31 + 0.15)s and methanol - water -diethyl ether - pH 7 buffer - ammonium hexamethylenedithio-carbamate (82 + 9 + 3 + 3 -t- 3).13 As it seemed that the low solubility of the the heavy metal chelates in methanol - water mixtures might be the source of the poor results the effect of using different eluents, containing less polar organic modifiers was studied.Acetonit-rile - water (70 + 30) plus sodium diethyldithiocarbamate and an ODs-Hypersil column gave good results for 100 p.p.m. solutions of lead (c.v. 2-3%) but much poorer results with 10 p.p.m. solutions. When methanol - water - chloroform (70 + 20 + 10) containing 0.05% mlV sodium diethyldithiocarba-mate was studied the injection of lead ions gave a sharp peak for the chelate (k' = 3.8) (Table 1). The peak areas were much more reproducible than earlier (1 p.p.m. lead ions C.V. 2.0%) and gave a linear calibration graph from 0.1 to 10 pp.m.with good reproducibility (Table 2). The lower limit was set by a deterioration in reproducibility. Detection was carried out at 350 nm to reduce interference from the excess of reagent in the eluent. As the samples were injected in aqueous solution rather than the mobile phase there was a major disturbance o ANALYST JANUARY 1985 VOL. 110 37 the base line from k’ = 0-2 because of solvent depletion effects. Because of this it was not possible to reduce the sample retention times significantly by using a stronger eluent. As this mobile phase gave satisfactory results for lead ions, mercury(I1) ions were also examined. The injection of mercury(I1) chloride gave a sharp peak at k’ = 6.2 and repeated injections gave reproducible retentions and peak areas.The calibration graph was linear from 0.5 to 10 p.p.m. of mercury(I1) ions (Table 3). It was of interest to determine if the system could distinguish mercury(1) from mercury(I1). The injection of a saturated solution of mercury(1) sulphate gave a peak at k’ = 4.9 well resolved from the mercury(I1) peak. However the low solubility of mercury(1) compounds made the preparation of a calibration graph difficult and no further studies were carried out. As expected from the spectra of the chelates the extension of the technique to cadmium ions was much less successful because the chromophore of the cadmium chelate is very similar to that of the dithiocarbamate reagent. There is only a small change in absorption at 350 nm on the formation of the chelate resulting in a much lower sensitivity than for the other metals studied.The retention time of the cadmium chelate was very short (k‘ = 2.70) and suffered interference from the base-line disturbance caused by the sample solvent. A longer retention time (k’ = 3.2) and better peak shapes with freedom from interference could be obtained by using methanol -water - chloroform (55 + 40 + 5 ) plus dithiocarbamate as the mobile phase. However the response was still very poor and using the maximum sensitivity range of the detector a 5 p.p.m. cadmium solution gave a similar peak area to a 0.03 p.p.m. lead solution. On increasing the cadmium ion concentration from 5 to 100 p.p.m. there was a marked increase in peak width from 0.2 to 1.2 min but only a small change in peak height although the peak shapes were symmetrical.The relationship of the resulting peak areas with cadmium concentration was non-linear. This may be caused by a kinetic or slow mixing effect rather than sample precipitation but was not studied further. A mixture of 1 p.p.m. of the four heavy metal ions could be readily resolved (Fig. 1) and the elution order corresponded to that found for the pre-formed complexes. 10 The sensitivity for lead and mercury(I1) is similar to that of flame AAS but cadmium is probably too poor for effective use. In order to test if the method could be suitable for multi-element analysis a sample of trade waste was examined. It was found to contain 1.7 p.p.m. of lead and <1 p.p.m. of cadmium by HPLC compared with the atomic-absorption spectroscopic analysis of 2.1 p.p.m.of lead and 0.19 p.p.m. of cadmium. Conclusion The simple and rapid on-column derivatisation method with the modified eluent can therefore be used for the direct separation and determination of lead and mercury ions at the parts per million level but reagent interference and low sensitivity mean that it would not be suitable for cadmium ions. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. References Schwedt G. “Chromatographic Methods in Inorganic Analy-sis,” Hiithig Heidelberg 1981. Veening H. and Willeford B. R. Adv. Chrornatogr. 1983, 22 117, Smith R. M. and Yankey L. E. Analyst 1982 107 744. Bond A. M . and Wallace G. G. Anal. Chem. 1981 53, 1209. Drasch G. von Meyer L. and Kauert G. Fresenius Z. Anal. Chem. 1982,311 695. Drasch G. Kauert G. and von Meyer L. Trace Element Anal. Chem. Med. Biol. 1983 2. 1109. Yamazaki M. Ichinoki S . and Igarashi R. Bunseki Kagaku, 1981 30,40. Ichinoki S . and Yamazaki M. Bunseki Kagaku 1982 31, E319. Schwedt G. Chrornatographia 1978 11 145. Schwedt G. Chromatographia 1979 12 289. Hutchins S. R. Haddad P. R. and Dilli S. J. Chromatogr., 1982 252 185. Inoue S. Hoshi S . and Sasaki M. Bunseki Kagaku 1982, 31 E243. Ichinoki S . Morita T. and Yamazaki M. J. Liq. Chromat-ogr. 1983 6 2079. Paper A41230 Received July 6th 1984 Accepted August 15th 198

ISSN:0003-2654    

DOI:10.1039/AN9851000035

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

ANALYST JANUARY 1985 VOL. 110 39 Determination of the Herbicides Frenock and Dalapon in Soil and River Water by Mass Fragmentography Tadashi Tsukioka and Shigenori Shimizu Nagano Research Institute for Health and Pollution 1978 Komemura Amori Nagano-shi Nagano Japan and Tetsuro Murakami Department of Chemical Engineering Kogakuin University 1-24-2 Nishishinju ku Shinjuku-ku Tokyo, Japan A method is described for determining frenock and dalapon in environmental samples such as river waters and soils based on the reaction of I-benzyl-3-p-tolyltriazene with an extract of frenock or dalapon from strongly acidified sample solutions to form benzylated species which are subsequently analysed by mass fragmentography (with soil samples steam distillation is applied prior to the extraction).Apiezon L grease - phosphoric acid serves as the stationary phase in GC and the molecular ions with mlz = 236 and 232 are used to monitor frenock and dalapon respectively. The detection limits for frenock and dalapon are 0.05 and 0.5 ng respectively with 5 pl of sample injected. Recovery experiments using actual river water and soil samples showed recoveries of more than 92% with a coefficient of variation of less than 5% ( n = 7 ) . The method is capable of the simultaneous determination of frenock and dalapon with sufficient sensitivity and selectivity to be applicable to environmental samples. Keywords Frenock and dalapon determination; mass fragmentography; herbicide residues; soil; water Sodium 2,2,3,3-tetrafluoropropionate (frenock) a contact-type herbicide is widely used in forestry and sodium 2,2-dichloropropionate (dalapon) a permeation-spreading herbicide is employed in orchards and fields.These herbi-cides have been used in large amounts and in wide ranges of sites and it is therefore desirable from the standpoint of environmental pollution control to establish a simple precise method for the trace determination of herbicides. A literature search revealed no reference methods for frenock but some chromatographic methods have been reported'-6 for dalapon. Ermolaeva et al.1 used a gas chromatograph with a flame-ionisation detector (FID-GC); Chmil'2 and Frank and Deminti applied a gas chromatograph with an electron-capture detector (ECD-GC) to the methyl ester derivative Cotterill4 to the 1-butyl ester derivative and Van der Poll and De V O S ~ to the 3-phenylpropyl ester derivative; and Chalaya and Gorbonos6 applied thin-layer chromatography separation (TLC) with bromophenol blue as the colour reagent.Application of these methods to frenock was tried but without success as frenock when converted into its acid form is highly volatile in comparison with dalapon. Subsequently as described here a method was developed in which both frenock and dalapon were extracted from strongly acidic solutions and converted into benzylated species with a carboxylated esterifying agent and then subjected to a determination by mass fragmentography (MF). The proposed method has the advantages over the conven-tional methods for dalapon that frenock and dalapon can be determined simultaneously the analytical procedure is easy to perform and the selectivity is so high that coexisting sub-stances hardly influence the analytical results thus making it possible to obtain accurate results even for environmental samples containing various foreign matter.Experimental Chemicals Kogyo Co. Japan. Co. Japan. , Sodium 2,2,3,3-tetrafluoropropionate (frenock) . Daikin Sodium 2,2-dichloropropionate (dalapon) . Tokyo Kasei 1-Benzyl-3-p-tolyltriazene. Tokyo Kasei Co. Japan. Diethyl ether. Suitable for detection of pesticide residues. Anhydrous sodium sulphate. As for diethyl ether. All other chemicals used were of guaranteed grade. Instrumental The gas chromatograph - mass spectrometer was a Model JMS-D300 from Japan Electron Optics Laboratory Co.(JEOL). A glass column of 2 m x 2 mm i.d. packed with Apiezon L grease - H3P04 ( 5 + 2) on Chromosorb W (AW-DMCS) (60-80 mesh) was used. The column temperature was programmed from 150 to 200 "C at 10 "C min-1 the injection port temperature was 220 "C the enricher temperature 220 "C and the ion source temperature 250 "C. The ionisation voltage was 30 eV and the ion multiplier voltage 2.5 kV. The carrier gas was helium at a flow-rate of 40 ml min-1. The ions monitored were at mlz = 236 for frenock and mlz = 232 for dalapon. Analytical Procedure For aqueous samples A 100-ml portion of the sample is placed in a 200-ml separating funnel to which 35 g of NaC1 6 ml of 9 M H2S04 and 20 ml of diethyl ether are added. The funnel is shaken for 10 min and the aqueous layer is discarded.The ether layer is washed with 10 ml of saturated NaCl solution dried with anhydrous Na2S04 and transferred into a 100-ml flask. To this flask is added 1 ml of a 2% miV solution of l-benzyl-3-p-tolyltriazene in diethyl ether. The flask is held in a water-bath at 60 "C for 1 h to effect benzylation under refluxing conditions. After cooling the solution is transferred with 10-20 ml of diethyl ether into a 100-ml separating funnel and washed with 10 ml of 1.2 M HCl and then twice with 10 ml of saturated NaCl solution. The ether layer is dried with anhydrous Na2S04 and concentrated in a Kuderna - Danish apparatus to 2 ml of which 5 1.11 is injected into the GC - MS system to be determined by MF. For soil samples About 20 g of soil are weighed into a distillation flask to which 40 ml of water 20 g of NaCl and 10 ml of 9 M H2S04 are added 40 80 8 2 60 ANALYST JANUARY 1985 VOL.110 -1 I I I I The mixture is subjected to steam distillation until 180 ml of distillate have been obtained. Water is added to this distillate to give a volume of 200 ml of which 100 ml are taken followed by treatment as described above for aqueous samples. It should be noted that a dry-mass correction is applied in such a way that a portion of sample is weighed in a stoppered weighing bottle and dried at 105-110 "C to constant mass in order to obtain the water content. Results and Discussion Selection of Extraction Conditions for Aqueous Samples As pointed out by Frank and Demint,3 both frenock and dalapon are extracted with difficulty unless the medium is strongly acidic and the NaCl concentration is high.The relationship between their recovery and the concentrations of both H2S04 and NaCl were therefore studied. First to choose the H2S04 concentration 20 pg each of frenock and dalapon 30 g of NaCl and different amounts of concentrated H2S04 were added to 100 ml of distilled water to produce a final solution with an H2S04 concentration of 0-1 M. Each solution was extracted with 20 ml of diethyl ether. A maximum and constant recovery was attained at 0.25 M H2SO4 for dalapon and at 0.5 M H2SO4 for frenock (see Fig. 1). Next to establish a suitable NaCl concentration a similar experiment was carried out using &30% m/V NaCl in 0.5 M H2S04.It was found that the recovery increased slightly at higher NaCl concentrations (see Fig. 2). Hence a solution containing 0.5 M H2S04 and saturated with NaCl is the optimum for the ether extraction. Selection of Steam Distillation Conditions for Soil Samples As in the extraction of aqueous samples the yields of the steam distillation of both frenock and dalapon are affected by the concentrations of H2SO4 and NaCl. The following three experiments were therefore performed. First to decide the optimum amount of H2S04 20 g of soil, 50 ml of distilled water 20 pg each of frenock and dalapon, 30 g of NaCl and different amounts of concentrated H2SO4 in the range 0-10 ml were placed in a distillation flask. The mixture was subjected to steam distillation until 200 ml of distillate had been obtained.As shown in Fig. 3 the recovery of frenock and dalapon increased with increase in the amount of H2S04 added until it reached a constant value at 3 ml. Next to establish a suitable amount of NaCl a similar series of NaCl solutions containing amounts in the range 5-50 g were prepared while the amount of H2S04 was kept at 5 ml. The results (Fig. 4) demonstrated that constant recoveries of both frenock and dalapon were obtained when the amount of NaCl added was more than 15 g. Another experiment was conducted to establish the opti-mum amount of distillate in which 1 mg each of frenock and dalapon were added the amounts of H2SO4 and NaCl were kept at 5 ml and 20 g respectively and the recovery was measured for 40-ml increments of the subsequent distillate.It was found that 160 ml of distillate gave more than a 95% recovery of dalapon which has a lower distillation rate than frenock. Hence the optimum steam distillation procedure was concluded to be as follows to 20 g of soil sample were added 10 mi of 9 M H2S04 20 g of NaCl and 40 ml of distilled water, and the mixture was subjected to steam distillation until 180 ml of distillate had been collected. Study of Conditions for Benzylation The conditions for benzylation were examined with a view to reducing the polarity of the compounds in order to prevent tailing during chromatography after selection of a high-boiling derivative of frenock and to increase the selectivity of I I I I 0.25 0.50 0.75 1 .o 50 ' 0 H2S04 concentrationh Fig.1. extraction from aqueous samples. A Frenock; and B dalapon Relationship between recovery and H2S04 concentration for B 100 I I I I 1 10 20 30 NaCl concentration % m/V Fig. 2. extraction from aqueous samples. A Frenock; and B dalapon Relationship between recovery and NaCl concentration for I I I I I 0 2 4 6 8 10 60 ' H2S04 added/ml Fig. 3. for steam distillation. A Frenock; and B dalapon Relationship between recovery and amount of H,SO added 'ooT===- 50 0 10 20 30 40 50 NaCl added/g Fig. 4. for steam distillation. A Frenock; and B dalapon Relationship between recovery and amount of NaCl adde ANALYST JANUARY 1985 VOL. 110 > v) +- .-C k 500 c -41 i 4-- 20 *E Mt .-a) c 236 -al - 10 ; - al the monitored mass by increasing the relative molecular mass of the monitored species.An experiment was carried out to determine the optimum volume of diethyl ether to be used in the benzylation 20 yg each of frenock and dalapon were transferred into a 100-ml flask and 1 ml of 2% mlV 1-benzyl-3-p-tolyltriazene solution in diethyl ether and different amounts of diethyl ether in the range 5-40 ml were made to react by placing the flask in a water-bath at 60 "C for 1 h under refluxing conditions. As shown in Fig. 5 dalapon gave a constant yield when 10-40 ml of diethyl ether were added whereas the yield of frenock was constant when 10-30 ml of diethyl ether were added and decreased with volumes over 30 ml. Next to decide a suitable reaction time a similar experi-ment was conducted by keeping the amount of diethyl ether at 20 ml and varying the reaction time in the range 1&100 min.Fig. 6 indicates that the yield increased with increasing reaction time until a constant yield was obtained at 40 min. It was noted that the yield of frenock decreased if the reaction time was longer than 80 min. Further to investigate the relationship between the amount of benzylating agent added and the yield 0.5 mg each of frenock and dalapon were taken an amount of a 2"/0 ethereal solution of 1-benzyl-3-p-tolyltriazene in the range 0.1-2 ml was added and the reaction was allowed to proceed under the above-specified conditions. A constant yield was obtained when a 2% ethereal solution of benzylating agent was used in amounts of more than 0.25 ml.The procedure finally adopted was as follows 20 ml of diethyl ether and 1 ml of 2% mlV 1-benzyl-3-p-tolyltriazene solution in diethyl ether were used and benzylation was effected by heating the system in a water-bath at 60 "C for 1 h. This procedure gave quantitative results when 2 mg each of frenock and dalapon were used. 0 ' Study of Conditions for GC - MS - MF Several types of GC column packings were tested. The best separation was obtained with Apiezon L grease - H,P04; other packings such as OV-17 SE-30 and DEGS can also be used. To select monitor ions for the MF determination benzyl-ated frenock and dalapon were subjected to EI mass spectral measurement and the results obtained are shown in Figs. 7 and 8. Both benzylated frenock and dalapon gave their major fragment ion at mlz = 91 which corresponded to the benzyl group.However this fragment ion was unsuitable because it is common to many other compounds found in environmental samples. For this reason the derivative molecular ions (Mf) at mlz = 236 and 232 were selected as the monitor ions for frenock and dalapon respectively. To select the ionisation voltage capable of producing Mf in the largest amount voltages in the range 10-70 eV were tested. The most Mt was produced at 30 eV for an ion source temperature of 250 "C. The conditions selected were therefore Apiezon L grease -H3P04 as the GC column packing molecular ions M t with mlz = 236 and 232 as the monitor ions and an ionisation voltage of 30 eV. LT . o Effect of Similar Substances To detect any interferences in the determination benzylations were tried on mono- di- and trichloroacetic propionic, butyric and valeric acid.In GC dichloroacetic acid had almost the same retention time as dalapon but gave no interference in the MF determination. l o o t 8 8o t I I 1 I I 0 10 20 30 40 Amount of ether/ml Relationship between relative yield and amount of ether for Fig. 5. benzylation. A Frenock; and B dalapon 80 /-- 60 I I 1 1 I I 1 0 20 40 60 80 100 Reaction time/min Fig. 6. benzylation. A Frenock; and B dalapon Relationship between relative yield and reaction time for 1 0 0 0 ~ 9' 130 rnlz Fig. 7. EI mass spectrum for benzylated frenock. Column 2 m, Apiezon L grease - H3P04 ( 5 + 2); column temperature programmed from 120 to 200 "C at 10 "C min-1; injection port and enricher temperatures 220 "C; ion source temperaure 250 "C; ionisation voltage 70 eV; and carrier gas He at a flow-rate of 40 ml min-I 91 50 I 1000 Calibration Graph and Recovery Experiments To construct calibration graphs standard frenock and dalapon in 100 ml of distilled water containing 0.5 1 2 3 4 and 5 pg of mJz Fig.8. Fig. 7 EI mass spectrum for benzylated dalapon. Conditions as i 42 ANALYST JANUARY 1985 VOL. 110 Table 1. Recovery of dalapon and frenock from river water and soil Coefficient of Average variation % Sample Compound Addedlpg recovery Yo (n = 7 ) Riverwater . . . . Dalapon 5.0 96.8 2.5 Frenock 1 .0 97.0 3.1 Frenock 2.0 92.7 3.6 Soil . . . . . . .. Dalapon 10.0 93.5 4.7 B 1 1 1 1 1 I 1 1 2 3 4 5 6 7 8 9 10 Timelmin Fig. 9. Mass fragmentogram for a river water. Conditions as in Fig. 7 except column temperature programmed from 150 to 200 "C at 10 "C min-1 and injection port and enricher ionisation voltage 30 eV. A mlz 236 (frenock); and B miz 232 (dalapon) A T /c 1 2 3 4 5 6 7 8 9 10 Tirnelmin Fig. 10. Fig. 9 Mass fragmentogram for a soil. Conditions and traces as in each compound were subjected to the procedure described for aqueous samples. The MF calibration graphs for both frenock and dalapon were linear over the range examined. The detection limits were 0.05 ng for frenock and 0.5 ng for dalapon. Recovery experiments were conducted by adding constant amounts of frenock and dalapon to 100 ml of river water or 20 g of soil without frenock and dalapon and carrying out the procedure as described for aqueous samples or soil samples, respectively.Based on the recoveries shown in Table 1 the proposed procedure was considered to be satisfactory with recoveries of more than 92% and a coefficient of variation of less than 5 % . Application of the Procedure to Environmental Samples The proposed method was applied to real samples of river water and soil. Fig. 9 shows the detection of frenock at 0.008 pg ml-1 in a river water sampled near a site where frenock had been released 1 week before. Fig. 10 shows the detection of dalapon at 0.23 pg g-1 in a field soil sample. Conclusion The procedure for the determination of frenock and dalapon can be summarised as follows.For extraction of aqueous samples the sample is saturated with NaCl and contains H2S04 at a concentration of 0.5 M before being subjected to diethyl ether extraction. For steam distillation of soil samples 20 g of soil are mixed with 20 g of NaCl 10 ml of 9 M H2S04 and 40 ml of water and the mixture is subjected to steam distillation until 180 ml of distillate have been obtained. For benzylation 20 ml of diethyl ether and 1 ml of 2% mlV 1-benzyl-3-p-tolyltriazene in diethyl ether are added to the sample in a flask and the mixture is refluxed for 1 h. For MF determination molecular ions M-f with m/z = 236 for benzylated frenock and 232 for benzylated dalapon are used as monitor ions. The detection limits are 0.05 ng for frenock and 0.5 ng for dalapon. Recoveries from river waters and soils are more than 92% for both frenock and dalapon with coefficients of variation less than 5%. The proposed procedure is considered to be applicable to many other low relative molecular mass carboxylic acids, including halogenated compounds such as trichloroacetic acid. We thank T. Okuhara G. Yamaura and F. D. Chow for many helpful suggestions. References 1. 2. 3. 4. 5. 6. Ermolaeva L. P. Ogloblina I. P. and Il'icheva I. A . Zh. Anal. Khim. 1977 32 2429. Chmil' V. D. Zh. Anal. Khim. 1978 33 2232. Frank P. A. and Demint R. J . Environ. Sci. Technol. 1963, 3 69. Cotterill E. G. J . Chromatogr. 1975 106 409. Van der Poll J . M. and De Vos R. H. J . Chromutogr. 1980, 187 244. Chalaya Z. I. and Gorbonos T. V. Zh. Anal. Khim. 1980, 35 1352. Paper A4/115 Received March 2211d 198.1 Accepted August 6th 198

ISSN:0003-2654    

DOI:10.1039/AN9851000039

出版商:RSC    

年代:1985

数据来源: RSC