ISSN:0003-2654    

DOI:10.1039/AN98005FX001

出版商:RSC    

年代:1980

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98005BX003

出版商:RSC    

年代:1980

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98005BP007

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

Analyst, January, 1980, Vol. 105, p p . 11-17 11 Development and Construction of an Analyser for the Determination of the Hydrogen Peroxide Content of Natural Water Using a Chemiluminescent Reaction F. Shaw Interox Chemicals Limited, Product Research Department, Moorfeld Road, Widnes, Cheshire, WA8 0 J U The development and construction of a portable analyser for the determina- tion of hydrogen peroxide in natural water using the chemiluminescent reaction of luminol (3-aminophthalhydrazide) and hydrogen peroxide with potassium hexacyanoferrate(II1) as the catalyst is described. The method is simple and rapid and interfering organic compounds present in the water can be removed by dialysis. The apparatus can be powered by a 12-V car battery and is capable of determining hydrogen peroxide a t concentrations as low as 0.02 mg 1-l.Keywords : Hydrogen peroxide determination ; water analysis ; chemi- luminescence ; interference effects ; dialysis Hydrogen peroxide has been shown to occur naturally at trace levels in rivers and reservoirs in the USSR,1 as an intermediate product of the oxidation of organic matter. In view of the levels of hydrogen peroxide quoted in this study it was decided to investigate the levels of naturally occurring hydrogen peroxide in a series of rivers and reservoirs in the UK. Before such an investigation could be carried out it was necessary to develop a simple and sensitive method for determining hydrogen peroxide in water at very low levels. The method used in the USSR1 was based on the measurement of the chemiluminescent radiation produced by the uncatalysed oxidation of luminol by hydrogen peroxide in the sample.This type of static system, although extremely sensitive, may suffer from poor reproducibility due to mixing problems. Several other methods for the determination of hydrogen peroxide exist a t present but they suffer from several disadvantages. Titration of hydrogen peroxide with permanganate2 or iodide3 is not sufficiently sensitive. The phenolphthalein method, although sensitive, suffers from interference by other oxidants4 and the titanium method involving the forma- tion of the peroxytitanate complex, although fairly sensitive, could suffer from interference problems. A method has recently been reportede for the determination of residual concentrations of hydrogen peroxide based on the spectrophotometric measurement of the colour produced by the cobalt - hydrogen peroxide complex.This method also offers excellent sensitivity, but it was not used because coloured natural waters might contain substances that absorb ultraviolet radiation. A chemiluminescent method for the determination of glucose by the enzymatic conversion of glucose into gluconic acid and hydrogen peroxide has also been rep~rted.~eE The hydrogen peroxide formed is then determined by measurement of the intensity of the radiation emitted by its reaction with luminol in the presence of potassium hexacyanoferrate(II1) as catalyst. The system operates on a continuous basis and is capable of the determination of hydrogen peroxide at levels down to 0.0003 mg 1-l. It appeared likely, therefore, that a similar system might be suitable for the determination of trace levels of hydrogen peroxide in water provided that any potential interference problems could be overcome.Such a method could probably be more easily automated than existing methods. Consequently, the feasibility of using a chemiluminescent reaction for the determination of trace amounts of hydrogen peroxide in natural water was investi- gated.12 SHAW: CHEMILUMINESCENT ANALYSER FOR THE DETERMINATION Ana&.%!, VOl. 105 Experimental Apparatus Investigations were carried out using three different systems, one static and two on a continuous basis. Static system. A photomultiplier, Type 9502B, (EM1 Limited) connected to an amplifier manufactured under licence by the London Gas Council.The chemiluminescent cell was a 10-ml beaker mounted in optical contact with the photomultiplier. The photomultiplier and amplifier used in the static system were used to detect chemiluminescent activity on a continuous basis. The measurement cell was a 10-ml pipette mounted in optical contact with the photomultiplier. Solutions were passed through the measurement cell via a series of time-delay coils using a Technicon peristaltic pump. A Technicon thermostated dialyser unit was used in conjunction with this equipment for interference studies. Continuous-$ow system (1). The complete system is shown chagrammatically in Fig. 1. To waste I I water or sample L------J Fig. 1. Diagram of the continuous flow system for the determination of hydrogen peroxide. All tubing has an i.d.of 0.1 in. Portable continuous-flow system (2). The final system was constructed based on the results obtained from the initial investigation and was a portable analyser mounted in an aluminium box, using the equipment detailed in Table I. The system is shown diagrammatically in Fig. 2. The dialysers are connected in series, making one 90-cm path length dialyser. Both the mixing chamber and the measurement cell are constructed from glass and are cylindrical, measuring 2.5 cm in diameter and 1 cm deep. The measuring cell is mounted vertically in optical contact with the photomultiplier tube and isolated from any stray light. The distance between the mixing chamber and the measurement cell is 10 cm. In all instances a Servoscribe RE 541.20 chart recorder was used to record the chemi- luminescent signal.To waste 4 I 1 Potassium 1 hexacyanoferrate 4 Time delay mixino coils (in series) I (lO-*M) ro. I I . . . . . . . . . 1 I ! +~uminol I in .. a n - - 4 ..\ I I Glass measu, ment cell cx- purrs.. I I De-mineralised I De-mineralised I Mixing chamber I water / - I 1 I water or sample - - A(Sample/calibration I tube) To waste 2 Dialysers (in series) Peristaltic pump Fig. 2. Schematic diagram of portable hydrogen peroxide analyser.January, 1980 OF THE HYDROGEN PEROXIDE CONTENT OF NATURAL WATER 13 TABLE I COMPONENTS REQUIRED FOR HYDROGEN PEROXIDE ANALYSER Item Description Manufacturer Peristaltic pump . . . . 4-channel peristaltic pump, Model HP4 Gilson Radiation detection system .. Photomultiplier power supply, Type EM1 Ltd. Photomultiplier tube, Type 9502B EM1 Ltd. Voltage convertor . . . . Input 12 V d.c., output 230 V, 200 W Brunlec Appliances Ltd. Car battery . . .. . . Good quality 12-V car battery Numerous suppliers Recorder . . .. . . Potentiometric chart recorder, expandable E.g., Servoscribe, Smiths range 2-200 mV Industries Ltd. Dialysers . . .. . . Two dialysers: Technicon Instruments Ltd. (a) 30-cm assembly. Complete set P.M. 28A a t 50 Hz including dialyser plate, lower, Cat. No. 177-B010-00; dialyser plate, upper, Cat. No. including dialyser plate, lower, Cat. No. 157-B369-01; dialyser plate, upper, Cat. No. Pre-mounted membrane, Type “C,” 177-B078-00 (b) 60-cm assembly. Complete set 157-B370-01 Cat. No. 170-0472-02 Mixing coils, peristaltic- pump tubing and miscellaneous items .. 4 X mixing coils, Cat. No. 116-0104-08 Technicon Instruments Ltd. Standard manifold pump tubing, colour code purple and orange, Cat. No. Standard tubing, Cat. No. 116-0536-17 Various connectors and T-pieces 116-0532-1 7 Reagents Reagents of the highest available purity were used throughout and all solutions were prepared in de-mineralised water. Luminol solution. M) was prepared by dissolving 0.8 g of luminol (3-aminophthalhydrazide), 73 g of sodium hydroxide and 61.8 g of boric acid in 1 1 of water. M) was prepared by diluting 100 ml of stock luminol solution (4.5 x M) to 500 ml with water and adjusting the pH to 10.5 with hydrochloric acid (1 + 9). M) was prepared by dissolving 3.293 g of potassium hexacyanoferrate(II1) [K,Fe(CN),] in 1 1 of water, and adjusting the solution to pH 10.5 with sodium hydroxide solution (4 g 1-l). A stock solution (1 000 mg 1-l) was prepared by successive dilution of hydrogen peroxide (30% m/m) and for the most accurate work was standardised by titration with potassium permanganate solution on the day of use.A working solution of hydrogen peroxide (10 mg 1-l) was prepared by dilution of the stock solution (1 OOO mg 1-l). A working solution (0.1 M) was prepared by dissolving 3.15 g of barium hydroxide [Ba(OH),.8H20] in 100 ml of water, which had been boiled for 10 min and cooled under an atmosphere of nitrogen. A working solution (0.1 M) was prepared by dissolving 2.87 g of zinc sulphate (ZnS0,.7H20) in 100 ml of water.A stock solution (4.5 x A working solution of luminol (9 x Potassium hexacyanoferrate(III) solution. A working solution ( Hydrogen peroxide solution. Barium hydroxide solution. Zinc sulphate solution. Procedure For investigations involving the static system, viability tests were carried out using 1 ml of varying concentrations of hydrogen peroxide, added to a mixture of 5 ml of luminol solution (9 x M) and 5 ml of potassium hexacyanoferrate(II1) solution (10-2 M), and the chemiluminescent signals were recorded. Using the apparatus described under Continuous-flow System (1), each solution was passed through the appropriate channel. In the channel labelled de-mineralised water or sample (Fig. l), water was pumped and the base line set on zero on the recorder. The tube was14 Analyst, vol.105 placed in the hydrogen peroxide solution, the sample pumped for a period of 1 min and the signal recorded. The tube was then placed in water until the base line returned to zero. Interferences on the chemiluminescent signal were investigated by adding 1 ml each of barium hydroxide solution (0.1 M) and zinc sulphate solution (0.1 M) to 100 ml of reservoir and tap water containing added hydrogen peroxide and shaking and filtering before pumping a sample of the filtrate through the analyser. Further investigations of interference effects using dialyses were carried out by passing the hydrogen peroxide solution through the dialyser prior to entering the measurement cell. The utility of the portable analyser described under Portable Continuous-Flow System (2), but operated by mains voltage, was evaluated by the addition of a known volume of hydrogen peroxide to a known volume of water in a water main.The sample tube was then placed in the water and the analyser operated continuously. The hydrogen peroxide con- centration was monitored over a period of 24 h and displayed on a chart recorder. The calibration was checked manually several times during the operation. Further evaluation trials on the portable analyser involved a survey of the natural levels of hydrogen peroxide in a series of reservoir waters. The analyser was operated on-site from a car using a voltage converter and a 12-V car battery as the power source. SHAW : CHEMILUMINESCENT ANALYSER FOR THE DETERMINATION Results and Discussion Static System Fig.3 shows a typical signal obtained during the chemiluminescent reaction when 1 ml of hydrogen peroxide solution was added to a mixture of luminol and potassium hexacyano- ferrate(II1) solution. As can be seen the signal reaches a maximum approximately 2 s after the addition of the peroxide. The peak reduced in intensity immediately to about two thirds of its original value and remained at this level for approximately 2 min. After this time, a gradual reduction in intensity was observed although some activity still remained 3 h after the hydrogen peroxide addition. v) 50 CI .- C t 40 h $ 30 m z 20 10 i7j en 0 30 60 90 120’’ 45 90 135 180 Time/s Time/min Fig. 3. Change in intensity of chemiluminescent signal with The results obtained on adding varying concentrations of hydrogen peroxide to a mixture respect to time.of luminol and potassium hexacyanoferrate(II1) are shown in Table 11. TABLE II RESULTS OBTAINED USING STATIC SYSTEM Concentration of H,O,/ Signal, mg 1-l chart divisions 0.1 12 0.2 23 0.3 30 0.4 51 It was possible to obtain more sensitive measurements by increasing the sensitivity of the However, in this type of static system the hydrogen peroxide is not adequately amplifier.January, 1980 OF THE HYDROGEN PEROXIDE CONTENT OF NATURAL WATER 15 mixed with the luminol- hexacyanoferrate(II1) mixture and therefore the sensitivity of the amplifier required reduction to improve signal reproducibility. An additional problem associated with this static type of system is the fact that there is a significant background- emission signal from the luminol- hexacyanoferrate(II1) mixture.When a sample of de-mineralised water containing no added hydrogen peroxide is added to the luminol- hexacyanoferrate(II1) solution a reduction in the background intensity is observed owing to a dilution effect. When the hydrogen peroxide solution is added, therefore, the chemi- luminescent response increases owing to the hydrogen peroxide, but the effect of the dilution reduces the background simultaneously and therefore one process is competing against the other. This adversely affects the reproducibility. It was considered that these problems could be overcome by the use of a continuous-flow system. Continuous-flow System (1) The mixing coils (Fig. 1) serve the dual purpose of allowing the luminol and potassium hexacyanoferrate( 111) solutions to mix adequately and also allow the background chemi- luminescence to attain a steady value.A system employing four mixing coils produced the best signal to noise ratio. The effect of pumping hydrogen peroxide solutions of different concentrations through the system is shown in Table 111. TABLE I11 RESULTS OBTAINED USING CONTINUOUS-FLOW SYSTEM Concentration of H,O,/ Signal, 0.04 23, 24 0.08 46, 49 0.12 67, 67 0.16 95, 93 The results obtained indicate that the response is linear. mg 1-l chart divisions The detection limit expressed as the background noise plus 2a of ten determinations on a sample containing a concentration of 0.04mg1-1 of hydrogen peroxide was found to be 0.004 mgl-1 of hydrogen peroxide.The detection limit is adversely affected by the high level of background noise from the luminol- hexacyanoferrate(II1) solution, which necessitates a reduction in the sensitivity of the amplifier. The main cause of noise is due to the fact that a peristaltic pump is used to pass the solution through the system. This type of pump suffers from a high degree of surge and the detection limit quoted for this work is similar to that obtained by other workers using a peristaltic pump.g The use of an infusion pump, although offering the possibility of achieving lower detection limits, owing to lower surge characteristics, would be more difficult to operate on a continuous basis. The effect of pH and potassium hexacyanoferrate(II1) and luminol concentrations on the chemiluminescent response was investigated.The most intense emission signal occurs at pH 10.5 and this was used for all further investigations. The effect of increasing the luminol concentration was to increase the cherniluminescent activity, but this also decreased the signal to noise ratio. A concentra- tion of 9 x 1 0 - 4 ~ luminol appeared to offer a compromise and so this concentration was used throughout. A hexacyanoferrate(II1) concentration of M produced the most intense peak although there was little variation in the sensitivity over the concentration range investigated. A hexacyanoferrate(II1) concentration of M was used for further investigations. The completely optimised system was then applied to the determination of hydrogen peroxide in tap water.When tap water was passed through the sample channel, negative signals were obtained with respect to the background. When tap water was pumped continuously, the base line set on this signal and tap water containing increasing concentra- tions of added hydrogen peroxide pumped through the system, the results shown in Table IV were obtained. In addition to the negative behaviour, there is little difference between the signals for each concentration of hydrogen peroxide. Negative signals were also obtained by Bostick and Hercules' in the chemiluminescent determination of glucose using luminol and potassium hexacyanoferrate(II1). This was later attributed to the reducing effect of uric acid at the16 SHAW : CHEMILUMINESCENT ANALYSER FOR THE DETERMINATION AnaZyst, VoZ.105 TABLE IV EFFECT OF THE ADDITION OF HYDROGEN PEROXIDE TO TAP WATER Concentration of H,O,/ Signal, mg 1-1 chart divisions 0.04 10 0.08 12 0.12 13 0.16 16 basic pH required for luminol chemilumines~ence.~ The uric acid interference was removed by flocculation with barium hydroxide and zinc ~ u l p h a t e . ~ It appeared likely that a similar mechanism was operating for tap water, Le., tap water apparently contains a higher con- centration of reducing species than the de-mineralised water and a similar flocculation treatment might remove these species and eliminate the interference. When the filtrate from the flocculated tap water was pumped through the system the base line remained unaffected at the point where the tap water passed through the measurement cell, indicating that the interfering species had been removed.The effect of flocculation on the recovery of added hydrogen peroxide from tap water was investigated. The results obtained were compared with those obtained from de-mineralised water containing known concentrations of hydrogen peroxide and are shown in Table V. TABLE V EFFECT OF FLOCCULATION ON TAP WATER CONTAINING ADDED HYDROGEN PEROXIDE Hz?, Water type concentration/mg 1-1 . . .. .. 0 0.04 0.08 0.12 0.04 0.08 0.12 Tap De-mineralised . . .. 0 Flocculated Yes Yes Yes Yes No No No No Signal, chart divisions 0 23 45 63 0 24 46 68 There is good agreement between the signals from de-mineralised water with added hydrogen peroxide and those from tap water with added hydrogen peroxide after flocculation.Hydrogen peroxide recoveries are therefore unaffected by this treatment. Samples of reservoir water were treated similarly and in each instance no interference was observed. However, a flocculation procedure is difficult to automate and the utility of the chemi- luminescent technique could be extended if the complete system could be automated. A method for the automatic determination of sulphate in water has been reported,1° which involves the removal of interfering organic compounds in the water by dialysis. It appeared likely that this method may offer an alternative to flocculation for removal of interferent. The results obtained when applied to the chemiluminescent system are shown in Table VI. TABLE VI EFFECT OF DIALYSIS ON INTERFERENCE HZO, Signal, Water type concentration/mg 1-l chart divisions Tap .. .. .. .. 0 0 0.1 20 0.2 41 De-mineralised . . .. 0 0 0.1 20 0.2 40 The results indicate that the interfering substances in tap water are removed by dialysis. The recovery of hydrogen peroxide after dialysis, however, is approximately 50%. It is essential, therefore, that both samples and calibration solutions are passed through theJanuary, 1980 17 dialyser. The detection limit using dialysis is 0.01 mgl-1 of hydrogen peroxide and there- fore the method is sufficiently sensitive to render it analytically useful. Reservoir waters with added hydrogen peroxide were treated similarly and in each instance no interference was observed. OF THE HYDROGEN PEROXIDE CONTENT OF NATURAL WATER Portable Continuous-flow System (2) The results obtained when the analyser (Fig.2) was operated on a continuously monitoring basis are shown in Fig. 4. The results show the decomposition of the peroxide over the 24-h period. When used in conjunction with the voltage converter and car battery the stability decreased, resulting in a detection limit of 0.02 mg 1-1 of hydrogen peroxide. In the reservoir survey no hydrogen peroxide (less than 0.02 mg 1-l) was observed in any of the reservoirs tested. The analyser functioned satisfactorily, however, and the only maintenance required involved re-charging the battery after about 5 h of continuous use. No problems were encountered during the operation. 5 0 2 4 6 8 10 12 14 16 18 20 22 24 Time since peroxide additiodh Fig. 4. Results obtained on a continuous basis using portable analyser.Conclusions The method developed for the determination of hydrogen peroxide in natural water, based on the chemiluminescent reaction of luminol with hydrogen peroxide, is simple, sensitive and interference free when samples are dialysed before passing into the measurement cell. The analyser constructed is completely portable, being capable of operation from a car. The introduction of a simple automatic valve system connected to the sample - calibration tube would allow the analyser to be used automatically on a continuously monitoring basis and would therefore extend the analytical utility. The author thanks Interox Chemicals Limited for permission to publish this paper and Mr. C. Whalley of Laporte Industries Limited for his valuable advice and guidance in the project . 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. References Sinel’nikov, V. E., Gidrobiol. Zh., 1971, 7, 115. Vogel, A. I., “A Textbook of Quantitative Inorganic Analysis,” Longmans, London, 1966, p. 296. Vogel, A. I., “A Textbook of Quantitative Inorganic Analysis,” Longmans, London, 1964, p. 363. Lecoque. H., Bull. SOC. Cham. Belg., 1945, 54, 186. Internal communication, Laporte Industries Ltd. , Widnes. Masschelein, W., Denis, M., and Ledent, R., Wut. Sewage Wks, 1977, 124, 69. Bostick, D. T., and Hercules, D. M., Analyt. Chem., 1975, 47, 447. Williams, D. C., Huff, 6. F., and Seitz, W. R., Clin. Chem., 1976, 22, 372. Williams, D. C., Huff, G. F., and Seitz, W. R., Anulyt. Chem., 1976, 48, 1003. Ogner, G., and Haugen, A., Analyst, 1977, 102, 453. Received May 8th. 1979 Accepted August 14th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500011

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

18 Analyst, January, 1980, Vol. 105, pp. 18-24 Automated Colorimetric Determination of Acid Proteinase Activity in Fermentation Samples Using a Trinitrobenzenesulphonic Acid Reagent Kaj Andre Holm Analytical Laboratory. NO VO Research Institute, N O VO All&, DK 2850 Bagsvaerd, Denmark An automated method for the determination of acid proteinase activity has been developed. The enzyme is incubated with a haemoglobin substrate after which the reaction solution is dialysed. The concentration of the liberated amino acids and peptides is determined by reaction with a trinitro- benzenesulphonic acid reagent. The technical parameters of the process and the general analytical conditions have been optimised. The method offers great advantages in its flexibility of selecting substrate and buffer independently of the conditions for the colour reaction. The final result is an automated method with high sensitivity, precision and speed.The sensi- tivity of the colour reaction is 1 mg 1-1 of nitrogen from leucine. Keywords : Acid proteinase determination ; automated colorimetry ; haerno- globin substrate ; trinitrobenzeneswll5honic acid reagent The aim of the investigations described in this paper was to develop an automated analytical method for the determination of acid proteinases, which would provide great flexibility in the choice of substrate and buffer by the use of a specific and sensitive colour reagent. The background of the present work is a manual method described previously,l from which the incubation temperature of 40 "C, the incubation buffer of pH 4.70 and the haemo- globin substrate have been retained. Several possible substrates are mintioned in the literature, e.g., gelatin, bovine serum albumin, egg albumin, casein and haemoglobin.Normally casein or haemoglobin is preferred. At pH 4.70 the use of casein is complicated because this pH represents its isoelectric point, which makes it impossible to produce a sufficiently high casein concentration in the solution; the pH would have to be 5.50 or higher for this to be attained. Haemoglobin is therefore the only possibility left, and is also usually the substrate preferred. For some of the above-mentioned substrates the ionic strength may be a problem, but with haemoglobin the influence of the ionic strength is considerably reduced.2 Further, the haemoglobin molecule itself offers considerable buffer capacity, good solubility and good batch to batch reproducibility.For quantitative analysis of the peptides and amino acids liberated by the proteolysis, the following three methods have been used in most instances: (i) UV absorption of the aromatic amino acids.' (ii) The ninhydrin reaction, in which reaction takes place with the primary amino groups.3 This reaction is not particularly specific, as ammonia also reacts with ninhydrin and oligo- peptides react with it to only a limited extent. Ninhydrin is also very sensitive to oxygen. The trinitrobenzenesulphonic acid (TNBS) reaction, in which all primary amino groups in peptides and amino acids liberated by the proteolysis react with TNBS.4J The TNBS reagent was therefore chosen for use in the present method because of its sensitivity and specificity.The literature contains references to both manual methods6-* and automated methods utilising the TNBS rea~tion.~-l~ The kinetics of the TNBS reaction are rather complicated; an explanation has been offered by Satake and co-~orkers*~~ and G01dfarb.l~ The product of the reaction is generally considered to be a nucleophilic aromatic substitution adduct. (iii) Experimental Principle The acid proteinase enzyme hydrolyses haemoglobin to liberate peptides and amino acids. The reaction mixture is dialysed, after which the concentration of liberated primary aminoHOLM 19 groups in the recipient stream from the dialyser is determined by the TNBS reaction.The sensitivity of the method is 1 mg 1-1 of nitrogen from leucine. Unit Definition This method gives a relative measurement by comparison with an acid proteinase enzyme preparation standardised by a manual meth0d.l One haemoglobin unit tyrosine (HUT) is here defined as that amount of enzyme which in 1 min under the given standard conditions produces a hydrolysate, the absorption of which at 275 nm is equal to 1.10 pg ml-1 of tyrosine. Standard Conditions for the AutoAnalyzer Substrate . . .. .. . . . . Haemoglobin, 2 g 1-1. pH . . .. .. .. . . . . 4.70. B u f e r .. . . . . . . , . Acetate, 0.1 M. Incubation tem$erature . . . . . . 4 0 ° C . Incubation time . . . . . . . . 7.2 min. Dialysis time. . .. . . . . . . 0.1 min. Colour reaction time at 40 "C . . . . 3.6min.Apparatus The Technicon AutoAnalyzer 2 system consisted of the following components (the number of each component used is indicated in parentheses) : sampler 4; proportioning pump; DIY manifold cassettes including heating bath, 40 "C, 10.6 ml (2); heating bath, 40 "C, 5.37 ml (2) ; dialyser, 6 in (2) , Cuprophan membranes; colorimeter type SCIC (2) ; and recorder (2 pens). Reagents All chemicals used were of analytical-reagent grade. Acetate buffer, p H 4.70, 0.1 M. Substrate buffer, pH 4.70, haemoglobin 2 g l-l, concentration of acetate ion 0.1 M. The solution is prepared by dissolving 10.8g of sodium hydroxide and 28.6 ml of glacial acetic acid in 5 1 of de-mineralised water. A 100-ml volume of de-mineralised water is added to 2.00 g of dialysed, freeze-dried haemoglobin with stirring for 10 min. Stirring is continued while 5.7 ml of glacial acetic acid and approxi- mately 5.5 ml of 1 M hydrochloric acid are added until the pH reaches 1.70.The stirring is continued for 20 min. Finally, approximately 67 ml of 1 M sodium hydroxide solution are added until the pH reaches 4.70. De-mineralised water is then added to give a total volme of 11. It is stable at 4 "C for up to 3 days. A borate stock solution (0.1 M) is prepared by dissolving 38.1 g of sodium borate (Na,B,O,.lOH,O) in 1 1 of de-mineralised water. The final reagent is prepared by dissolving 0.50g of sodium sulphite in 1 1 of borate stock solution. This buffer should be used on the day of preparation. This is prepared by dissolving 0.100 g of trinitrobenzenesulphonic acid in 100 ml of de-mineralised water. The reagent is prepared freshly every day and must be protected against light. L-Leucine (93.7 mg), 0.05 ml of 1 M hydrochloric acid and 0.05 ml of a 1% m/V solution of thymol in ethanol are dissolved in 200 ml of de-mineralised water. Acetate buffer of pH 4.70 is used for the preparation of the standards.They may be kept frozen a t -15 "C for up to 1 month. A special enzyme, isolated at NOVO from an Aspergillus niger var. strain, was used as the proteinase. Brij 35. Procedure The sample is intro- duced into both the sample and the blank channels. These two physically independent channels are completely identical in the manifold construction, the only difference between The solution must be clear before use and may be filtered if necessary.Borate - sulphite buffer, pH 9.30. Trinitrobenzenesulpphonic acid reagent (TNBS-R) , 1 g 1-l. Leucine standard, 50 mg 1-1 of nitrogen from NH,. The solution is stable for up to 1 month at 4 "C. Acid proteinase standards, 10, 25, 50 and 75 HUTml-l. Used as the detergent. The flow diagram (Fig. 1) shows the construction of the manifold.Sampler I V Fig. 1. Flow diagram of the automated acid proteinase method. The sampling rate is 60 samples h-l with a sample to wash ratio of 6 : 1. The wash The colorimeter consists of two interference filters (420 nm) and solution used is de-mineralised water. two flow cells of length 15 mm and internal diameter 1.5 mm. The flow-rates (ml min-l) are given in parentheses. Part numbers refer to the Technicon catalogue.January, 19SO PROTEINASE ACTIVITY I N FERMENTATION SAMPLES 21 them being that on the sample side substrate and buffer (pH 4.7) are introduced, whereas on the blank side only buffer (pH 4.7) is used.On the sample side, sample and substrate are mixed and the incubation takes place at 40 “C. The enzyme solution is then dialysed. Borate - sulphite buffer and TNBS-R are mixed and introduced as the recipient stream in the dialyser unit. Reaction between the primary amino groups and the TNBS-R takes place from this point, through a heating bath at 40 “C, and a ten-turn mixing coil, before the final colorimetric measurement is made a t 420 nm. Registration on the recorder appears about 15 min after the sample has been taken up by the sampler. Twenty drops of Brij 35 per litre are used in all reagents, except for the TNRS-R.After operation, the AutoAnalyzer system is cleaned with Deconex (2% V / V ) for 5 min, de-mineralised water for 2 min, hydrochloric acid (0.2 M) for 5 min and then de-mineralised water for 20-30 min. Calculation A 50 mgl-1 leucine standard is used for calibration of the blank channel against the sample channel and is introduced at the start of each analysis series. The blank channel is adjusted to give the same response as the sample channel to the 50 mg 1-1 leucine standard by use of the “standard calibration.” The drift-corrected blank molar absorptivities are deducted from the corresponding drift-corrected sample molar absorptivities. The differ- ence obtained in this way is converted into HUT ml-l via the enzyme standard graph for the sample channel.The final result may then be corrected for any dilution of the sample that has taken place prior to the AutoAnalyzer procedure. Results After optimising the technical parameters for the process (Fig. 1), the analytical conditions were optimised using the manifold as mentioned above. Selection of the Haemoglobin Concentration Various concentrations of haemoglobin were tested, under routine analytical conditions, with various concentrations of the acid proteinase enzyme (Fig. 2). Haemoglobin con- centrations of more than 5gl-1 could not be used as they caused changes in the dialysis conditions (even clotting). The optimum concentration, which gave a rectilinear relation- ship up to an absorbance of 0.700, was 2 g 1-l.This concentration was used in this work (Michaelis constant, K , = 0.2 g 1-I.) 70 8 30 2 m -9, 20 10 0 10 20 30 40 50 60 Concentration of acid proteinase enzyme/HUT nil-’ Fig. 2. Haemoglobin substrate assay using various concentrations of acid proteinase enzyme. A, 1.0; x , 2.0; 0, 3.0; and 0, 4.0 g 1-1 of haemoglobin.22 HOLM : AUTOMATED COLORIMETRIC DETERMINATION OF ACID Analyst, VoZ. 105 Choice of pH for the TNBS Reaction pH values between 8 and 10.4 have been used. Higher sensitivity is obtained at a higher pH, but the formation of breakdown products, such as picric acid, is increased, resulting in an increased reagent blank. The optimum pH for this reaction has not been established in the literature. A pH of 9.30 was chosen as a reasonable compromise.Selection of the Trinitrobenzenesulphonic Acid Concentration Addition of TNBS-R gave the same result whether the reagent was added to the recipient stream before or after the dialysis procedure, ie., the TNBS molecule does not pass through the dialysis membrane under the analytical conditions used. For practical reasons the TNBS-R was added before the stream reached the dialyser. Using leucine standards of various concentrations the TNBS concentration was varied from 0.5 to 2 g 1-1 (Fig. 3). Only a t 1 g 1-1 was a satisfactory rectilinear result obtained and therefore this concentration was used in this work. 0 10 25 50 Leucine concentration/mg I-' Fig. 3. TNBS assay using various concentrations of 0, 0.5; x , 1.0; and 0, 2.0g 1-1 of TNBS.leucine. Optimisation of the Sodium Sulphite Concentration If an amount of sodium sulphite corresponding to the TNBS concentration on a molar basis should be present, this would correspond to approximately 55 mg 1-1 of sodium sulphite. The experimental results show that this is very nearly the case (Fig. 4). With different leucine concentrations, the sodium sulphite concentration was varied from 0.025 to 18.9 g 1-1. A level of maximum sensitivity was obtained with between 0.1 and 0.5 g1-1 of sodium sulphite. For routine use, a concentration of 0.5 g 1-l of sodium sulphite is used, which gives a rectilinear result. As can be seen from Fig. 4, the addition of sodium sulphite, at this concentration, increases the sensitivity of the method by about 300%. A slightly lower sensitivity may be obtained by measuring the absorbance at 340nm without addition of sodium sulphite. TNBS Reaction Temperature 25 to 50 "C (Fig.5). sensitivity. With different leucine standards, the temperature of the TNBS reaction was varied from Only at 40 "C was the desired rectilinear result obtained with sufficient This work was therefore carried out a t 40 "C. Specificity of the TNBS Reaction peptides and amino acids. ammonia. The light-sensitive TNBS reagent reacts with primary amino groups, including those in No reaction occurs with uric acid, proline, hydroxyproline or Copper(I1) ions cause a reduction in sensitivity, whereas ethanethiol causes an23 increase in sensitivity. However, these two agents do not normally occur in the samples to be analysed.TNBS also reacts with free SH groups in peptides and amino acids, but the complexes formed are so unstable that no SH group interference in the TNBS reaction with the free NH, groups will occur. The possibility of interference has been taken into con- sideration in the method described by including a sample blank with each analysis. January, 1980 PROTEINASE ACTIVITY I N FERMENTATION SAMPLES 0 10 25 50 Leucine concentration/mg I-’ Fig. 4. of leucine. 0, 18.9 g 1-’ of sodium sulphite. Sodium sulphite assay using various concentrations 0, zero; 0, 0.025; x , 0.10; 0, 0.50; A, 6.3; and Precision An acid proteinase sample containing 50 HUT ml-l was analysed; 20 determinations were made for which the mean sample absorbance was 0.4065 (after correction for the blank), with a standard deviation of 0.0015 absorbance unit.Measurement of the Percentage of the Steady-state Value A standard is sampled until a constant (“steady-state”) deflection of the recorder pen is obtained. The absorbance of the same standard is then measured at the sample frequency used for analysis to obtain a peak signal. The concentration values corresponding to both the steady state and the peak signals are obtained from the calibration graph, and the 0 10 25 50 Leucine concentration/mg I-‘ Fig. 5. TNBS assay using various concentrations of leucine at different temperatures. A, 25; x, 40; and 0, 50 “C.24 HOLM percentage of the steady state (yoSS) is calculated from the ratio of these two values from the equation where Cpeak and C,, are the concentration values corresponding to the peak and the steady- state signals, respectively.The value of yoSS obtained in this work was 0.99. Measurement of the Percentage Interaction To determine the percentage interaction (or “carry-over”) , the absorbances of a low standard, a high standard and then a low standard again are measured at the frequency used for analysis. The concentration values corresponding to these absorbances are obtained from the calibration graph and the percentage interaction (%I) is calculated from the equation where Ca, Cb and C, are the concentration values of the first low, the high and the second low standard, respectively. The value of %I obtained in this work was 0.55. Discussion The present method is distinguished by the fact that the substrate and buffer for the incubation can be chosen independently of the TNBS colour reaction, This means that the method may in principle be used, with the same TNBS colour reaction, for the determination of different proteolytic enzymes, regardless of the incubation pH and substrate used.Some experiments have been performed using casein (without sodium sulphite addition at the incubation step) at pH 5.50 and 8.0; in the latter instance the results obtained were com- parable to those obtained with the method described in reference 10. The method described here is also prone to fewer possible interferences than the TNBS colour reaction because of the dialysis procedure used. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. References Method 22-62, American Association of Cereal Chemists, St. Paul, Minn., 1968. Lanoe, J., and Dunnigan, J., Analyt. Biochem., 1978, 89, 461. Moore, S., and Stein, W. H., J . Biol. Chem., 1954, 211, 907. Satake, K., Okuyama, T., Ohashi, M., and Shinoda, T., J . Biochem., 1960, 47, 654. Satake, K., Take, T., Matsuo, A., Tazaki, K., and Hiraya, Y., J . Biochem., 1966, 60, 12. Mokrasch, L. C., Analyt. Biochem., 1967, 18, 64. Brown, H. H., Clin. Chem., 1968, 14, 967. Snyder, S. L., and Sobocinski, P. Z., Analyt. Biochem., 1975, 64, 284. Palmer, D. W., and Peters, T., Clin. Chem., 1969, 15, 891. Dunn, E., and Brotherton, R., Analyst, 1971, 96, 159. Preston, K. R., Cereal Chem., 1975, 52, 451. Adams, C. A., Robberts, T. C., and Butler, K. C., Analyt. Biochem., 1976, 70, 181. Goldfarb, A. R., Biochemistry, N.Y., 1966, 5 , 2570. Received July 2nd, 1979 Accepted July 30th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500018

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

Analyst, January, 1980, Vol. 105, pp. 25-29 25 Polarographic Behaviour and Analysis of 2-Nitro- 5,6-Dimethylindane-l,3=dione in Raw Materials and Pharmaceutical Formulations J. T. Browne and W. Franklin Smyth" Beecham Laboratories, Betchworth, Dorking, Surrey Chemistry Department, Chelsea College, University of London, Manresa Road, London, S W3 6 L X 2-Nitro-5,6-dimethylindane-lI 3-dione is a new orally active anti-asthmatic drug and is administered in the form of the monohydrate sodium salt. It possesses three electro-reducible groups, namely the two carbonyl and the nitro group, the reduction of which are pH dependent. The best defined waves for analytical purposes are those produced by the 4e reduction of the nitro group in acidic media, which, when combined with the fact that the known breakdown products and impurities do not contain this electro-active group, endows a polarographic method for the assay of this drug in raw materials and formulations with enhanced selectivity.Keywords Polarography ; 2-nitro-5,6-divnethylindane- 1,3-dione ; raw materials ; pharmaceutical formulations The polarographic behaviour of 2-nitro-5,6-dimethylindane-l,3-dione (BRL 10833), I, was investigated using both d.c. and differential-pulse polarography over a wide range of pH and in a concentration range suitable for its determination in raw materials and formulations. This led to a postulation of a mechanism of reduction and establishment of optimum analytical conditions for the assay of I in the presence of its known breakdown products and impurities, i.e., 5,6-dimethylindanetrione-2-oxime (11), 5,6-dimethylindane-l,3-dione (111) and 43- dimethylphthalic acid (IV), the first two being electro-reducible in their own right, but at different potentials at the chosen pH for the determination of I, Polarography has been applied to the assay of I in capsule and tablet formulations both for initial and for stored samples.CH3 cH3m o2 0 I CH3 cH3m 0 Ill CH3 cH3@N0H 0 II CHncooH c H3 COOH IV Experimental A 2 x M stock solution of the sodium salt (monohydrate) of compound I in aqueous solution was prepared. Aliquots of 1 ml of this stock solution were diluted with 100ml of Britton - Robinson (BR) buffer solutions to give a series of solutions that differed by pH 1 * Present address: Department of Chemistry, State University of New York at Bnffalo, Buffalo, N.Y.14214, USA.26 BROWNE AND SMYTH : POLAROGRAPHIC BEHAVIOUR AND Analyst, VoZ. 105 and covered the pH range 3-1 1. Solutions covering the lower and upper ends of the pH scale were prepared by diluting 1-ml aliquots of the stock solution with 100 ml of 1 N hydrochloric acid and 1 N sodium hydroxide solution, respectively. The solutions were de-oxygenated with nitrogen and subjected to polarography in d.c. and differential-pulse polarographic modes using the PAR, Model 174A, Polarographic Analyser equipped with a drop timer and suitable recorder. A three-electrode system involving a dropping-mercury electrode as the indicator electrode, a platinum counter electrode and a saturated calomel reference electrode were used.The scan rate was 5mVs-l, modulation amplitude 100mV, current range 20 PA, drop time 1 s, low pass filter 0.3 s and the initial potential was commonly -0.3 V. A calibration graph was constructed as follows: 0.08-, 0.10-, 0.12- and 0.14-g amounts of the sodium salt (monohydrate) reference standard of compound I were accurately weighed and dissolved in 200 ml of distilled water in separate 200-ml calibrated flasks. Aliquots of 1 ml of these solutions were diluted to 100 ml with pH 5 BR buffer solutions and subjected to differential-pulse polarography using the above conditions. A graph of the differential- pulse polarographic peak height (in millimetres) versus concentration of I (in milligrams per litre) was then constructed for the range 3.94-7.07 mg 1-l.The precision of the method was examined for five solutions of similar concentration and the standard deviation of the ratio of peak height to sample mass was found to be satisfactory (0.02). The raw material was assayed for content of compound I by accurately weighing approxi- mately 0.1 g of homogenised sample and dissolving it in 200 ml of distilled water. A 1-ml aliquot of this solution was diluted to 100 ml with pH 5 BR buffer, subjected to differential- pulse polarography and the height of the resulting peak related to the concentration of I in the raw material by reference to the calibration graph. The contents of 10 capsules or the crushed powder from 10 tablets were thoroughly mixed. Sufficient powder was weighed so as to contain approximately 0.05 g of I, dissolved in 100 ml of water in a calibrated flask and shaken mechanically for 30 min.After allowing the contents to settle, a 1-ml aliquot of the clear supernatant liquor was pipetted into a 100-ml calibrated flask and diluted to the mark with pH 5 BR buffer. Approximately 10 ml of this solution were trans- ferred into a clean polarographic cell and de-aerated with nitrogen for 5 min. Using the instrumental conditions described earlier, this solution was subjected to polarography. The height of the first peak, corresponding to the reduction of the nitro group (i’), was measured. By reference of the peak height of each sample to the calibration graph, the concentration of I in each final sample solution ( A ) could be calculated in milligrams per millilitre.The content of com- pound I in capsules or tablet formulations (C) was then calculated from the equation C = 100APM1/M2 mg per capsule or tablet, where P% is the potency of the reference standard of compound I, M,g the tablet mass or average capsule fill mass and M 2 g the initial sample mass. Capsule or tablet formulations were assayed by the following method. This procedure was repeated for all other sample solutions. Results and Discussion Polarographic Behaviour of I in pH Range 0-14 at a Concentration of 2 x Compound I gave rise to one predominant 4e d.c. polarographic wave in the pH range 0-14 (i’), corresponding to reduction of the nitro group. E, and E, values became increasingly negative with increasing pH, e.g., at pH 0 E, and E, = -0.22 V (versus a saturated calomel electrode) and at pH 14 E, and E, = -1.0 V (versus a saturated calomel electrode).The height of the d.c. polarographic wave remained unaltered in the pH range 3-1 1 whereas the peak height of the diff erential-pulse polarographic peak remained constant in the pH range 3-7 and then decreased markedly with increasing pH so that by pH 11 the peak height had decreased to 33% of its original value. The optimum peak for differential- pulse polarographic analysis of I was therefore chosen as pH 5 with the appropriate BR buffer as the supporting electrolyte. Two further d.c. polarographic waves (i” and P’) were observed at more negative potentials and at pH 0-3; these were equal in height and corresponded to two 2e processes. In the intermediate pH range 5-9 both of these waves decayed and eventually disappeared giving a wave (i””) corresponding to a 2e process at more negative potentials. This last wave was unaltered in the value of its limiting current over the pH range 9-14.MJanzcary, 1980 ANALYSIS OF %NITR0-5,6-DIMETHYLINDANE-1,3-DIONE 27 Although aliphatic hydroxylamines are generally not reducible in the available potential range,l the anion of compound I is pseudo-aromatic in character owing to delocalisation processes that can occur in the five-membered ring in addition to its being a yellow solid. It is therefore believed that wave it’ corresponds to reduction of the protonated hydroxyl- amine, formed on reduction of the nitro group at more positive potentials. Wave 2”’ occurs at potentials corresponding to the reduction of similarly conjugated carbonyl compounds and, as it corresponds in height to a 2e process, it must involve only one of the carbonyl groups in compound I.The behaviour of this wave is similar to that observed by Stradins et aL2 for a similar molecule, 2-phenylindane-l,3-dione. Waves i” and i”’ decay in neutral and alkaline media as, respectively, hydroxylamines are not reduced unless they are proton- atedl and carbonyl compounds can be reduced by alternative reaction mechanisms when there is inadequate proton availability, e.g., involving metal cations such as Naf. The proposed mechanism of reduction is shown in Fig. 1. 0 ;‘ 4e (pH 0-14) I 0 Fig. 1. Proposed mechanism of reduction. *As the ultraviolet - visible spectrum of I does not change in the pH range 0-14, it is assumed that this species (or its tautomers) are the predominant ones in solution.It is, in fact, yellow in solution. Analysis of I in a Capsule Formulation A capsule formulation had been prepared containing a nominal 59 mg of the sodium salt (monohydrate) of compound I in a total capsule fill mass of approximately 150 mg. Samples of this formulation were analysed in order to examine the applicability of this method to the analysis of I in pharmaceutical formulations. The contents of ten capsules were mixed thoroughly and five samples of approximately 65 mg were weighed into separate 50-ml calibrated flasks. A 40-ml volume of distilled water was added and the mixture shaken for 30 min. It was found that all the solid dissolved except for a trace amount of an insoluble excipient.This material settled when the mixture was allpwed to stand for 5 min. A 1-ml aliquot of the clear supernatant liquor was diluted to 100 ml with pH 5 BR buffer and subjected to polarography as described under Experimental. By reference to the calibration graph the concentration of the sodium salt (monohydrate) of compound I in each sample was calculated. The experimental details and results are shown in Table I.28 BROWNE AND SMYTH : POLAROGRAPHIC BEHAVIOUR AND Analyst, Vol. 105 TABLE I ASSAY OF CAPSULE FORMULATIONS CONTAINING I Sample mass/ Peak height/ Content of compound Sample No. mg m m per capsulelmg 1 61.0 67.5 58.3 2 71.1 79.8 58.0 3 65.5 74.3 59.4 4 69.8 77.0 57.4 5 69.7 77.7 57.8 This gives a mean capsule content of the sodium salt (monohydrate) of compound I of 58.2 mg, which is well within the specifications set for this formulation, The relative standard deviation is 1.30%, which is *2.6y0 for 95% confidence and equal to *1.5 mg per capsule.This shows that this assay method may be applied to capsule formulations of this type with high precision. Effect of Impurities and Breakdown Products on Polarographic Behaviour of I The differential-pulse polarographic behaviour of these potential interferences, 11, I11 and IV, was investigated in pH 5 BR buffer at a concentration of 5 x mg ml-1. The oxime (11) gave two peaks at -0.75 and -1.03 V ('ueysus a saturated calomel electrode) and 5,6- dimethylindane-l,3-dione gave one peak at -0.9 V ('ueyszbs a saturated calomel electrode) compared with an E, of -0.58 V (veysus a saturated calomel electrode) for I under the same conditions.4,5-Dimethylphthalic acid (IV) was not reducible in the available potential range. To evaluate the effect these compounds would have on the differential-pulse polaro- graphic behaviour of I, 10% of each was added separately and then together to a solution containing 5 x 10-3mgml-1 of the parent compound (I). No change in the peak height corresponding to i' was observed. Thus, the analytical method is selective for compound I in the presence of its breakdown products. Stability Testing three types of samples were assayed as described below. In order to examine the application of the method to the analysis of I in storage trials Aqueous solution (1.5%) of I stored at 80 "C in a sealed vial for 6 months When the solution was diluted to 5 X 10-3mgml-1 in BR buffer solution and subjected to polarography it was noted that the main polarographic peak (i') for compound I was absent.This indicated that compound I in the sample had been completely degraded and further confirmed the usefulness of this method for the assay of stability-test samples. Cafiule formulation samples stored for 6 months at 5 "C in a sealed canister and at 37 "C exposed to 75% relative humidity These capsules contained a nominal 48 mg of I as the free nitro compound in a total capsule fill mass of 250 mg. The contents of ten capsules from each storage condition were emptied from the capsule shells and thoroughly mixed and treated as described under Experimental. A duplicate sample was taken in all instances.The sample data and results obtained are shown in Table 11. The concentration of the free nitro compound I was calculated from the equation A x purity of standard (%) x average capsule fill mass (mg) Sample mass (mg) x 100 - A x 84.0 x 250.9 x 10000 mg per capsule Content of I = - M x 100 where A = concentration of sample solution found by reference to the calibration graph.January, 1980 ANALYSIS OF 2-NITRO-5,6-DIMETHYLINDANE-1,3-DIONE 29 TABLE I1 ASSAY OF CAPSULE FORMULATIONS AFTER STORAGE Sample mass/ Peak height/ Content of compound I Storage conditions mg mm per capsulelmg 5 "C sealed .. . . 252.1 75.5 47.8 5 "C sealed . . . . 248.4 75.5 48.3 Mean: 48.1 37 "C, 75% RH* . . . . 253.8 74.0 45.7 37 "C, 75% RH* . .. . 251.0 74.0 47.0 Mean: 46.4 * RH = relative humidity. . It would be expected that the samples stored at 5 "C in a sealed container would remain the same and suffer very little or no degradation of the active component, compound I. However, samples stored under severe conditions such as 37 "C and openly exposed to 75% relative humidity would be expected to have suffered some decomposition. The above results show that this method can determine I in the presence of its breakdown products. Tablet formulation samples stored at 5 "C in a sealed canister and at 30 "C openly exposed to 75% relative humidity As with the previous analysis of capsule formulations this method has been applied to the analysis of tablet formulations containing I at a nominal level of 50mg in a total tablet mass of approximately 260 mg.As this analysis was carried out simultaneously with the previous capsule analysis the same calibration graph was applicable. Following powdering of the tablets, samples containing approximately 5 x mg ml-l of I in pH 5 BR buffer solution were prepared and subjected to polarography. Sample data and results are shown in Table 111. TABLE I11 ANALYSIS OF TABLETS CONTAINING I AFTER STORAGE FOR 6 MONTHS Sample mass/ Peak height/ Content of compound I Storage conditions mg mm per tabletlmg 5 "C sealed .. . . 246.7 73.1 49.2 5 "C sealed .. . . 245.8 72.9 49.2 Mean: 49.2 30 "C, 75% RH* . . . . 250.6 71.8 47.7 30 "C, 75% RH* . . . . 249.1 72.0 48.0 Mean: 47.9 * RH = relative humidity. These results again show that the method is sensitive to breakdown of I as it can detect the approximate 2.6% drop in potency of the sample stored under more severe conditions. These results also show good agreement between duplicate assays for each sample. This work has shown that this polarographic assay method is suitable for the determina- tion of I in capsule and tablet formulations. Also, it can be used to measure the stability of I in the raw material state and in capsule and tablet formulations. References 1. 2. Zuman, P., and Perrin, C. L., Editors, "Organic Polarography," Interscience, New York, 1969. Stradins, J. P., Tutane, I. K., and Vanag, G. J., Zh. Analit. Khim., 1965, 20, 1239. Received April 27th, 1979 Accepted July 9th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500025

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

30 Analyst, Janztary, 1980, VoZ. 105,pp. 30-36 Determination of Trace Amounts of Copper in Palm Oil by Differential-pulse Anodic-stripping Voltammetry K. H. Wong, Y. S. Fung and K. W. Fung Department of Chemistry, University of Hong Kong, Hong Kong A method is described for the determination of copper in palm oil by differential-pulse anodic-stripping voltammetry after dry ashing. Only a few grams of sample are needed and the relative standard deviation is about 3% for oils with copper contents less than 100 p.p.b. (parts per 108). Keywords : Copper determination ; palm oil ; diflerential-pulsa anodic-stripping voltammetry The detrimental effects of copper on the storage life and bleachability of lipids are well kn0wn.1-~ Its pro-oxidative power has been demonstrated even at concentration levels as low as 30 p.p.b.(parts per The current specification for the upper concentration limit of copper in crude palm oil in the food industry is 80 p.p.b. and it is proposed to decrease this limit to 50 p.p.b. in the near f u t ~ r e . ~ Such a low concentration level, together with the complications in triglyceride fractions of palm oil, place serious limitations on the choice of a method for the determination of copper. introduced a %yo solution of soya bean oil in isobutyl methyl ketone directly into the flame for atomic-absorption spectro- photometric measurement. However, a concentration of 250pg of copper per gram of sample (250 p.p.m.) was needed in order to achieve 1% absorption. Palm oil, which is more viscous than soya bean oil, requires a higher solvent dilution ratio to prevent clogging of the aspiration capillary, and this method therefore cannot be applied to the determination of trace amounts of copper in palm oil.Atomic-absorption spectrophotometry with electro- thermal atomisation has a lower detection limit, and can be used for direct determinations even after dilution with a suitable solvent. However, the reproducibility of this method is poor.' Also, it is difficult to remove the non-atomic absorption interferences in untreated oil samples.* Kundu and Prevotg used an oxygen-rich atmosphere during the drying and ashing stages in order to minimise smoke formation during the subsequent atomisation. The relative standard deviation was only 3.5% at the 100 p.p.b. level but increased to about 743% at the 25 p.p.b.level for the direct determination of copper in oil. X-ray fluorescence spectroscopy and neutron-activation analysis are less subject to interference by the lipid matrix and have sufficient sensitivity. Unfortunately, the instruments are expensive and the precision is relatively 10w.l~ Hence the direct determination of copper in palm oil without sample pre-treatment is not promising. Wet ashing, acid extraction and dry ashing are the common methods of sample pre- treatment in order to destroy or to separate a metallic ion from the organic matrix in a biological sample for trace-metal analysis. Palm oil, like other lipids, is relatively resistant towards oxidation and a large volume of a perchloric acid - sulphuric acid mixture is needed in wet ashing. Because of the associated safety hazards11s12 and the lengthy procedure, wet ashing is not recommended.Jacob and Klevay13 and Deck and Kaiser14 used a constant-boiling hydrochloric acid - EDTA mixture to extract copper from oil samples prior to colorimetric and atomic-absorption spectrophotometric determination, respectively. Colorimetry and conventional flame atomic- absorption spectrophotometry have relatively high detection limits of about 0.1 mg 1-1 for copper in s01ution.l~~~~ A large volume of palm oil is therefore needed and the preparation of the sample for analysis requires a considerable amount of time. Fung and Fungl' proposed a two-stage dry-ashing method to decompose the organic matter, the residue being dissolved in acid, the resulting solution mixed with a buffer solution and the copper content determined by a potentiometric method. The method has a low detection limit of about List etWONG, FUNG AND FUNG 31 10 p.p.b.At such a low level the response time is very long, however, and careful control of the measuring medium and conditions are needed in order to minimise interferences. In this work, a more sensitive and reproducible method, differential-pulse anodic-stripping voltammetry (DPASV) was used for the determination of copper in the resulting solution after pre-treatment. The feasibility of using acid extraction and dry ashing as the pre- treatment method was investigated. The effect of the nature of the crucible (porcelain, silica, Pyrex, platinum) in dry ashing on the results of the subsequent analysis was also studied.Experiment a1 Apparatus A PAR 174A Polarographic Analyser was used for DPASV studies and the voltammograms were recorded with an Esterline Angus, Model 575, X - Y recorder. A conventional three- electrode system was used. The reference electrode was a saturated Ag- AgCl type separated from the bulk electrolyte with a 1 M potassium nitrate salt bridge. A PAR 9223 hanging mercury drop electrode was used as the working electrode and a platinum wire inside a frit was used as the counter electrode. A Varian Techtron, Model 1200, atomic-absorption spectrophotometer with an air - acetylene atomiser or a Model 63 carbon rod atomiser was used for atomic-absorption spectrophotometric analysis. A Griffin electric furnace with a built-in Eurotherm thermo- static control was used in dry ashing.A 5-pl Excalibar pipette was used to introduce the sample solution into the carbon tube furnace and a 25-pl Oxford pipette for introducing the potassium thiocyanate solution or standard copper(I1) solution into the voltammetric cell. All glassware and crucibles were first cleaned by immersion in a 1 + 1 sulphuric acid - nitric acid bath overnight, followed by ultrasonic cleaning in a water-bath for 1 h. Reagents unless otherwise specified. pared by triple distillation of the acids in all-glass apparatus. a Nanopure purification system (Rarnstead). removed by scrubbing through vanadium(I1) chloride solution. was prepared by dissolving freshly cleaned copper foil (>99.9%) in dilute nitric acid.standard solutions were prepared daily. All chemicals were of analytical-reagent grade and were used without further treatment Constant-boiling hydrochloric acid and nitric acidl8 were pre- The water was purified with Residual oxygen in commercial nitrogen was A stock solution of copper(I1) Fresh Procedure The palm oil samples were melted on a hot-plate a t 60 "C. A graduated pipette was used for random sampling of the homogenised oil. In acid extraction about l o g of sample were pipetted into a 250-ml round-bottomed flask fitted with a reflux condenser. A 50-ml volume of the appropriate acid- EDTA mixture and acid-washed anti-bumping granules were then added. The mixture was refluxed for 3 h at 250 "C in a silicone oil-bath. After cooling, 25 ml of the aqueous solution were removed by pipette for DPASV or atomic-absorption spectrophotometric analysis.The covered crucible was heated in the furnace at 350 "C for the first 2 h, then at 480 "C for the next 1-2 h for complete oxidation. A 2.5-ml volume of triply distilled nitric acid was then added to dissolve the ash. The resulting solution was made up with water to 25ml for atomic-absorption spectrophotometry with electrothermal atomisation. In DPASV, 20 ml of the resulting solution were pipetted into the cell, followed by 25 pl of 1 M potassium thiocyanate solution. The solution was deaerated by purging with purified nitrogen for 5 min. Copper was deposited on the mercury electrode at -1.0 V (similar results were obtained by setting the potential at -0.7 V) with respect to the reference electrode for 2 min, after which the magnetic stirrer was switched off and the solution was allowed to stand for 30 s. The copper was then stripped off a t a scan rate of 5 mVs-l, modulation amplitude 50 mV and pulse interval 0.5 s.Both the standard-additions method and the In dry ashing, about 4 g of sample were pipetted into a pre-weighed crucible.32 WONG et al. : DETERMINATION OF TRACE AMOUNTS OF COPPER IN PALM Analyst, Vol. 105 calibration graph method were used to calculate the copper concentration. Similar sampling and ashing procedures were used in flame atomic-absorption spectrophotometric studies, except that 60 g or more of sample were used. The operating conditions were as follows: wavelength, 324.7 nm; lamp current, 5.0 mA; spectral band pass, 0.5 nm; flame, oxidising air - acetylene; and settings on Model 63, drying 7.5 (250 "C) for 20 s, ashing 6.3 (650 "C) for 20 s and atomisation 7 (2300 "C) for 3 s.Results and Discussion A high efficiency for the extraction of copper from oil with hydrochloric acid in the presence of EDTA was reported by Deck and Kaiser14 and Jacob and K1e~ay.l~ The possibility of using the same method coupled with DPASV for the determination of copper in palm oil was investigated, and the effects of the concentrations of EDTA and hydrochloric acid, the acid to oil ratio and the time of reflux on the extraction efficiency were studied. The efficiency was calculated with reference to the result obtained by dry ashing, which has been shown to give a 100% recovery (see later discussion).DPASV was used for the determination of copper in the acid layer after separation. Tables I-IV show the effects on extraction efficiency of various experimental parameters. TABLE I EFFECT OF EDTA CONCENTRATION ON EXTRACTION EFFICIENCY EDTA concentration, yo Extraction,* yo 0 40 0.005 40 0.007 5 69 0.01 93 * 10 g of crude oil (0.42 p.p.m. of copper) refluxed t O.Olyo was about the maximum solubility. with 50 ml of 20.2% HCI. No bumping was observed if the volume ratio of acid to oil was higher than 5 : 1 in the presence of anti-bumping granules. The extraction efficiency increased as the concentration of EDTA increased. The role of EDTA in the extraction process can be explained by the following mechanism. In highly acidic solution, EDTA should exist mainly in the form of H4Y, which would be soluble in oil.The copper in oil exists mainly as chelated compounds with hydroperoxides and secondary oxidation products.lg An equilibrium between H4Y and the copper complexes in oil to form CuY2-, which is soluble in water, would be expected. TABLE I1 EFFECT OF HYDROCHLORIC ACID CONCENTRATION ON EXTRACTION EFFICIENCY Acid concentration, yo Extraction,* yo 10.0 25 15.0 36 20.2t 93 * log of crude oil (0.42 p.p.m. of copper) with t Constant-boiling HCl. 50 ml of 0.01% EDTA. The EDTA complex would then be converted into the stable chloro-complex in hydrochloric acid solution. Thus, copper could be extracted continuously from the oil layer into the acid layer in the presence of a small amount of EDTA. The higher extraction efficiency with increasing hydrochloric acid concentration as shown in Table I1 supports the proposed mechanism.The conditions chosen for acid extraction were 1 part of oil mixed with 5 parts of 20.2% hydrochloric acid - O.Olyo EDTA, the mixture being refluxed for 3 h. UnderJanuary, 1980 OIL BY DIFFERENTIAL-PULSE ANODIC-STRIPPING VOLTAMMETRY 33 these conditions a high extraction efficiency was obtained with both crude and hydrogenated palm oils. Unfortunately, the potential of the copper stripping peak in the resulting acidic solution was even closer to the mercury dissolution potential than that in pure hydrochloric acid solution. Hence the drawing of the base line is arbitrary in dilute copper solution. The precision of the peak current measurement would be seriously affected.Even though the copper stripping peak was well separated from the mercury dissolution peak in sulphuric acid, the two were close to each other in the resulting acidic solution after extraction. The stripping voltammogram was affected by the acid-soluble species in the oil sample. In view of this, the acid extraction method is not recommended as the pre-treatment method for trace copper determination in palm oil by DPASV. TABLE I11 EFFECT OF EXTRACTION TIME ON EXTRACTION EFFICIENCY Reflux time/h Extraction,* yo 1 24 2 63 3 93 * l o g of crude oil (0.42 p.p.m. of copper) with 50ml of 20.2% HCl + 0.01% EDTA. Dry ashing has the unique advantages of requiring minimum operator's attention and low reagent contamination. However, retention on and leaching from the container and the volatilisation loss would be a serious problem in trace metal analysis.Volatilisation loss would not be a problem for copper if the ashing temperature was lower than 550 oC.20 A low and variable recovery was reported by Deck and Kaiserf4 with the use of an open flame to remove part of the organic matter in oils before ashing in a Vycor or platinum crucible. Evans et used a lengthy char-ashing method and obtained a very high recovery. Quantitative recovery of copper has also been achieved by a two-stage dry ashing.17 The last method was used in this study. TABLE IV EFFECT OF ACID TO OIL RATIO ON EXTRACTION EFFICIENCY Acid to oil ratio ( V / V ) Extraction,* % 3: 1 86 5 : 1 93 * 10 g of crude oil (0.42 p.p.m. of copper) with 20.2% HCl + 0.01% EDTA for 3 h.More discrepancies were centred on the choice of the crucible material used for dry In a non-reducing medium, platinum was to be preferred because of its inert properties.20 B ~ w e n ~ ~ claimed that a borosilicate beaker could be used for biological samples at 390 "C with an alkali metal nitrate as an ashing aid. Fung and Fung17 found that a porcelain crucible could be used to determine copper in palm oil. Evans et aZ.21 reported that Vycor was better than porcelain as the container in dry ashing. In order to obtain more information on this problem we further studied the effect of the nature of the crucible on the results of the determination of copper in palm oil. The copper content of the palm oil samples was determined by the recommended flame atomic-absorption spectrophotometric method.26 In order to minimise the effect of retention and leaching, a 60-g sample was ashed in a large platinum crucible.The copper content found by flame atomic-absorption spectrophotometry was used as a reference in the subse- quent comparison. Results with less than a 3% error were obtained by using a 4-g sample in a platinum crucible with the proposed method, as shown in Table V. The results obtained by using porcelain crucibles depend considerably on the history of the crucible and the34 Analyst, vo,?. 105 cleaning procedure. The results are close to or higher than the expected values, as shown in Table V. The amount of copper leached from the crucible was found to be less than 0.4pg. Thus, the error resulting from using a porcelain crucible as the container in dry ashing would be less significant if the copper content in the resulting solution was higher than 1Opg.It was found that a 1 + 1 mixture of sulphuric and nitric acids was more effective than nitric acid alone for cleaning the crucible. On the other hand, the retention losses were significant if a silica or Pyrex crucible was used as the container. The amount of copper retained depends on the degree of roughness of the surface. Hence a platinum crucible should be used as the container for samples containing low levels of copper, and a porcelain crucible could be used for samples containing high levels of copper or for larger amounts of sample. Silica or Pyrex would not be suitable as the container in the absence of an ashing aid.WONG et al. : DETERMINATION OF TRACE AMOUNTS O F COPPER IN PALM TABLE V EFFECT OF CRUCIBLE MATERIAL ON THE RECOVERY OF COPPER I N DRY ASHING OF PALM OIL Crucible material Sample Platinum . . . . Crude palm oil Porcelain?. . . . Crude palm oil Silica . . . . Crude palm oil Platinum . . . . Bleached, neutralised, hydrogenated palm oil neutralised, hydrogenated palm oil Borosilicate beaker Bleached, Sample size: 3-4 g. Copper concentration found r Relative standard No. of Range, p.p.b. Mean, p.p.b. deviation, % measurements 90-96 93 2.9 6 A 60: (94)* 90-228 48-68 56 4 35-37 36 2 (37) * 3-34 14 6 * Result obtained by flame atomic-absorption spectrophotometry with 60-g sample. t Cleaned with 1 + 1 nitric acid - sulphuric acid mixture.$ With 30 different crucibles. The advantage of adding thiocyanate to the electrolyte in the determination of copper in a chloride medium by DPASV has been discussed by hlann and De~tscher.~' After the addition of 25 pmol of thiocyanate to 20 ml of the electrolyte (dilute nitric acid), the stripping peak was slightly broadened and the peak current was slightly decreased. However, it was shifted cathodically and was well separated from the mercury dissolution peak, as shown in Fig. 1. A relatively flat base line was obtained, and the accuracy on the peak current measurement was therefore increased significantly. The addition of thiocyanate is essential in trace copper determinations by DPASV with a mercury electrode. A linear calibration graph with a correlation coefficient of 0.9996 was obtained at copper concentrations in the electrolyte in the range 1-31 p.p.b.The results obtained by the calibration graph method and the standard-additions method showed no statistically significant differences. No interferences from iron(II), iron(II1) and nickel(II), which are the common major metallic ions in oils, were observed at concentrations up to 1000 p.p.m. The relative standard deviation was about 3% for a 4-g sample containing 90 p.p.b. of copper. Atomic-absorption spectrophotometry with electrothermal atomisation is also a popular technique for trace metal analysis. A parallel study was made of the simultaneous deter- mination of copper in the resulting solution by both DPASV and atomic-absorption spectro- photometry.The reproducibility of the latter technique was found to be greater than loyo, which was much inferior to that of DPASV. This poor reproducibility might be the result of the unpredictable change in the pyrolytic surface of the carbon tube after each atomisation, as the atomisation temperature was high for the determination of copper. The times required for the two techniques were about the same. Together with its lowJaizztary, 1980 OIL BY DIFFERENTIAL-PULSE ANODIC-STRIPPING VOLTAMMETRY 35 J ! , , , , -0.6 -0.4 -0.2 0.0 0.2 E/V versus Ag - AgCl Fig. 1. Differential-pulse anodic-stripping voltam- mograms of 10 p.p.b. of copper(I1) in the resulting acid solution. A, in the absence of potassium thiocyanate; and B, in the presence of 25 pniol of potassium thiocyanate.running costs and easy operation, DPASV is preferred to atomic-absorption spectrophoto- metry with electrothermal atomisation in this instance. Flame atomic-absorption spectro- photometry was not sensitive enough for such low levels of copper. The proposed method could be applied to the determination of copper in other oils and fats. Conclusion Dry ashing followed by diff erential-pulse anodic-stripping voltammetry is a sensitive and reproducible method for the determination of trace amounts of copper in palm oil. The two-stage dry-ashing procedure is simple and relatively rapid. A platinum crucible is recommended as the container in the ashing process for copper contents less than 1Opg and a porcelain crucible can be used for the samples with higher copper contents.We thank Lam Soon 9il and Soap Manufacturing Sdn. Bhd., Malaysia, and Harrisons and Crosfield (Malaysia) Sdn. Bhd. for supplying the palm oil samples. This research was supported by the Higher Degrees and Research Grants Committee of the University of Hong Kong. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. References Marcuse, R., “Metal Catalysed Lipid Oxidation,” Swedish Institute of Food Preservation Research Evans, C. D., Schwab, A. W., Mosu, H. A., Hawley, J. E., and Melvin, R. H., J . Am. Oil Chem. SOC., Hampson, G. C., and Hudson, B. J . F., “Physical and Chemical Properties of the Constituents of Malaysian Palm Oil Technical Bulletin, Malaysian Palm Oil Producer’s Association, Kuala Lumpur, Johansson, G., Chemy Ind., 1975, 2, 902.List, G. R., Evans, C. D., and Kwolek, W. F., J . Am. Oil Chem. SOC., 1971, 48, 438. Black, L. T., J . Am. Oil Chem. SOC., 1975, 52, 89. Olejko T. J., J . Am. Oil Chem. SOC., 1976, 53, 480. Kundu, M. K., and Prevot, A., Analyt. Chem., 1974, 46, 1591. List, G. R., Evans, C. D., and Kwolek, W. F., J . Am. Oil Chem. SOC., 1971, 48, 438. Analytical Methods Committee, Analyst, 1960, 85, 643. Analytical Methods Committee, Analyst, 1973, 98, 458. Jacob, R. A., and Klevay, L. M., Analjt. Chem., 1975, 47, 741. Deck, R. E., and Kaiser, K. K., J . Am. Oil Chem. SOC., 1969, 47, 126. Analytical Methods Committee, Analyst, 1963, 88, 253. Analytical Methods Committee, Analyst, 1971, 96, 741. (SIK), Goteborg, Sweden, 1968. 1951, 28, 68. Edible Oils and Fats,” Pergamon Press, Oxford, 1961, pp. 13-16. 1973, No. 1, p. 14.36 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. WONG, FUNG AND FUNG Fung, Y. S., and Fung, K. W., Analyst, 1978, 103, 149. Vogel, A. I . , “A Textbook of Quantitative Inorganic Analysis,” Third Edition, Longmans, London, Cooney, Y. M., Evans, C . I)., Schwab, A. W., and Cowan, J. C., Analyt. Chem., 1958, 35, 152. Gorsuch, T. T., Analyst, 1959, 84, 135. Evans, C. D., List, G. R., and Black, L. T., J . Am. Oil Chem. SOC., 1971, 48, 840. Piccolo, B., and O’Connor, R. T., J . Am. Oil Chem. SOC., 1968, 45, 789. Burrows, J . A., Heerdt, J. C., and Willis, J. B., Analyt. Chem., 1965, 37, 579. Persmark, U., and Toregard, B., J . Am. Oil Chem. SOC., 1975, 48, 650. Bowen, H. J. M., Analyt. Chem., 1968, 40, 969. Horwitz, W., Editor, “Official Methods of Analysis of the Association of Official Analytical Chemists,” Mann, A. W., and Deutscher, R. L., Analyst, 1976, 101, 652. 1961. Twelfth Edition, Association of Official Analytical Chemists, Washington, D.C., pp. 432-436. Received May 17th, 1979 Accepted July loth, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500030

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

Analyst, January, 1980, Vol. 105, p p . 37-42 Analytical Applications of Emulsions: Determination of Lead in Gasoline by Atomic-absorption Spectrophotometry 37 L. Polo-Diez, J. HernBndez-Mhdez and F. Pedraz-Penalva Department of Analytical Chemistry, Faculty of Sciences, University of Salamanca, Salamanca, Spain A rapid and simple atomic-absorption spectrophotometric method for the determination of lead in commercial gasolines, based on the formation of a stable oil - water emulsion using a suitable emulsifier and further aspiration into the flame, has been developed. The effect of different matrices and of lead alkyl compounds on sensitivity is not significant, several calibration pro- cedures being possible. The accuracy was confirmed by statistical com- parison with the ASTM titrimetric method.Keywords : Lead determination ; emulsions ; atomic-absorption spectrophoto- metry ; gasolines It is well known that lead is present in commercial gasolines mainly as lead alkyl anti-knock compounds. The amount of lead in these samples generally falls within the range 0.1- 1.0 g 1-l. There are several methods reported in the literature for the determination of the total lead content of motor gasolines. In the ASTM method, the total lead content is determined titrimetrically, after converting the lead alkyl compounds into lead chloride by refluxing with concentrated hydrochloric acid.l*3 In another method, the lead is titrated with EDTA after treatment with iodine mon~chloride.~ All these methods are time con- suming. X-ray methods are also used for this p ~ r p o s e .~ Atomic-absorption spectrophoto- metric (AAS) methods are based on dilution of the gasoline sample in a suitable solvent before aspirating the sample into the flame. Robinson5 used isooctane as the diluent and standards of tetraethyllead (TEL) in isooctane. Other solvents have also been employed for this purpose.6 The main problem involved, besides the large volumes of the organic solvents used, seems to be the variation in the response of the different lead alkyl compounds. Several methods have been proposed to overcome this inconvenience, based on the addition of an excess of iodine in isobutyl methyl ketone (MIBK)697 and suitable liquid anion exchangers to stabilise the iodine - lead compounds.8 Although it seems that some of the problems involved in these determinations have been overcome, others such as noise, fluctuations of the sensitivity and the characteristic problems of the flame are still present and it would be useful to provide a method in which these disadvantages are absent.Basically, it involves the formation of an emulsion from a small amount of gasoline and a larger amount of water in the presence of a suitable emulsifier. The oil - water emulsion obtained is a strictly heterogeneous system, but it is homogeneous enough to be aspirated directly into the flame, being essentially aqueous, and has the subsequent advantages that this involves. We have applied this principle successfully to the determination of lead in lubricant In addition, we have reported preliminary results for the method that we propose here.l0 Apparatus mental parameters of the instrument were adjusted to obtain maximum sensitivity.Kerry Pulsatron 250 ultrasonic device was used for emulsification. Reagents were used. solvents. In this paper, we propose a method based on the use of emulsions and AAS. Experimental A Pye Unicam SP 90 Series-2 atomic-absorption spectrophotometer was used. Experi- A For the application of the ASTM titrimetric method, the reagents specified elsewhere1s2 Benzene, hexane and isobutyl methyl ketone (analytical-reagent grade) were used as Caution-Benzene is highly toxic and a known carcinogen and extreme care should be taken.38 Analyst, vol. 105 A non-leaded gasoline (G) from ENPETROL (Empresa Nacional del Petrbleo, Spain) and three other non-leaded gasolines, G,, G, and G,, were used.These gasoline samples were prepared by mixing three solvents, supplied by CAMPSA (Compaiiia Arrendataria de Petr6leo Sociedad Anhima, Spain). The three solvents S, (boiling range 150-200 "C), S, (boiling range 50-150 "C) and S, (diethyl ether) were mixed in different proportions to give the gasoline samples: G, (1 + 1 + l), G, (1 + 1 + 4) and G, (4 + 2 + 1). The emulsifiers used were : non-ionic Brij 30 [polyoxyethylene-4-lauryl ether, Br015, hydrophilic - lipophilic balance (HLB) 9.71 and Tween 80 (polyoxyethylene-20-sorbitan monooleate, TW080, HLB 15.0), both supplied by Scharlau-Ferosa. An emulsifier with an HLB of 13.5 was prepared by mixing Brij 30 and Tween 80. Standards A gasoline standard containing 0.59g1-1 of lead alkyl [as a mixture of TEL and tetra- methyllead (TML) in a ratio of 1.0: 1.01 was obtained from ENPETROL.Ethyl Corpora- tion TEL orange, supplied by CEPSA (Compaiiia Espaiiola del Petr6leo Sociedad Anhima, Spain) was used to prepare standards of TEL by dissolving 0.50 g of TEL in benzene or in the G,, G, and G3 non-leaded gasolines. Gasoline samples and TML in unstated ratios. Procedure Transfer 1 ml of the gasoline sample into a 50-ml calibrated flask containing about 20 ml of water and 5 drops of the emulsifier with an HLB of 13.5 (about 0.7% m/m); make the volume up to 50 ml, shake vigorously for a few seconds and put the flask inside the ultra- sonic device for 10 min. The calibration graph was obtained by applying the above procedure to increasing volumes of a gasoline standard or to standards of TEL in benzene.Alternatively, the standard-additions method may be used by adding increasing volumes of a gasoline standard or TEL in benzene to 1 ml of the sample. POLO-DfEZ zt al. : EMULSIONS I N THE DETERMINATION OF The commercial motor gasolines were rated a t 85, 90, 96 and 98 octane, containing TEL This emulsion is then introduced directly into the flame. Results and Discussion The emulsions obtained using the above procedure produce instrumental signals at the resonance line of lead (283.3 nm), which increases when larger amounts of gasoline sample as well as of lead nitrate solution are emulsified, although with the latter the sensitivity is lower. Thus, we may conclude that the instrumental signals obtained are due to the lead present in the emulsion.Spectral interferences from other elements present in the gasoline are not expected.ll The conditions required in order to obtain and stabilise these emulsions for analytical purposes were then optimised. Choice of Emulsifier It is known that the formation of these emulsions requires the presence of a suitable emulsifier.12J3 Each oil - water system has its optimum emulsifier, which is characterised, particularly for the non-ionic ones, by its hydrophilic - lipophilic balance. Although general rules may give the rough range within which the suitable HLB must lie, the optimum emulsifier must be chosen experimentally for each particular instance because commercial surfactants contain impurities. In addition, the non-foaming properties, necessary to make up the volume easily and to avoid heterogeneous distributions of the gasoline droplets in the bulk of the emulsion, must be taken into account.Because of this we preferred non-ionic emulsifiers. The literature recommends HLB values in the range 11-16 to emulsify derivatives of petroleum corn pound^.^^^^^ We have obtained stable emulsions (stable for at least 1 h) by using emulsifiers prepared from Brij 30 and Tween 80 having HLB values in the range 13- 15. However, an HLB of 13.5 was considered optimum because for higher HLB values, emulsions are less stable, fluctuations of the signal are observed and measurements must be made just after their preparation. However, for lower HLR values the sensitivity decreases.The blank test for non-leaded gasoline was negligible. As has been mentioned, our interest lies in obtaining oil - water emulsions.39 January, 1980 The influence of the percentage of emulsifier present on the analytical response is shown in Fig, 1. As can be seen, a certain amount of emulsifier is necessary in order to obtain maximum sensitivity, the signals being stable for levels below 2% m/m (the highest per- centage studied). These results have led to the use of about 1% m/m of emulsifier. LEAD I N GASOLINE BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 0 0.1 0.3 0.5, 0.7 0.9 Amount of emulsifier, % Fig. 1. Effect of the amount of emulsifier on the absorbance using 1 ml of the standard gasoline sample; emulsification time 10 min; total volume of the emulsion 50 ml; emulsifier with an HLB of 13.5.a~ 0.40 5 0.30 t m 2 Q 0.20 3 5 7 9 1 1 1 3 Time/min Fig. 2. Effect of emulsification time on the absorbance using five drops of the emulsifier with an HLB of 13.5 (about 0.7%). Other conditions are those specified in Fig. 1. Effect of Emulsification Time As is known, the formation of an emulsion requires the supply of a certain amount of energy, mechanical methods being most often used for this purpose. We employed an ultrasonic device for emulsification purposes because it allows easy handling of the sample and, in addition, the reproducibility is better than that obtained with mechanical systems. Fig. 2 shows the effect of the emulsification time; the absorption signals increase with time, being roughly constant above 9 min.We used 10 min for the optimum emulsification time, when the signals are stabilised, minimising small effects from the geometric differences of the flasks. Determination of Lead in Gasoline Samples Calibration graphs, obtained by applying the above procedure to increasing volumes of a gasoline standard, are linear below 6 p.p.m. of lead in the emulsion, the sensitivity (slope of a graph of the absorbance versus parts per million of lead) being 14.4 times greater than that of the aqueous lead solutions in the presence of the same amount of emulsifier. The calibration graph obtained using the gasoline standard that contains TEL and TML was employed to determine the content of lead in four commercial gasoline samples, which also contain these two alkyl compounds in unknown ratios; the results are shown in Table I.TABLE I COMPARISON OF ASTM TITRIMETRIC AND EMULSION METHODS Emulsion method r A , ASTM method: Octane rating Lead content*/ Standard Coefficient of lead content*/ of sample g 1-1 deviation/g 1-l variation, % g I-' 85 90 96 98 0.32 0.29 0.35 0.39 0.006 1.88 0.004 1.38 0.003 0.86 0.010 2.56 0.32 0.29 0.35 0.40 * Each result is the average of 10 determinations. The standard deviation for ten replicate determinations is within the range 0.003-0.010 p.p.m. of lead, for a sample containing about 0.35 p.p.m. of lead. The accuracy was determined by comparing the results obtained by the emulsion method with those supplied by the ASTM titrimetric method (Table I). The Student &test1* showed that differences are not signifi- cant at the 95% level of probability.Thus, we may conclude that the proposed method allows the accurate determination of lead in gasoline samples.40 Analyst, VoZ. 105 These results seem to show that neither the matrix nor the type of alkyl compound that the gasoline contains affects the sensitivity significantly; other calibrations and tests were carried out. Constant volumes of standard gasoline solutions, prepared by diluting the gasoline standard sample that contains TEL and TML with hexane, MIBK, benzene or non-leaded gasoline, were used to prepare the calibration graphs. The slopes of the straight lines obtained, adjusted by the least-squares method, were compared with each other by the Student t- test (Table 11). The results show that these slopes are coincident at a 95% level of probability.However, the results obtained in the analysis of the gasoline samples, using these graphs, do not show significant differences, at a 95% level of probability, from the results supplied by the ASTM titrimetric method. POLO-D~EZ et al. : EMULSIONS IN THE DETERMINATION OF TABLE I1 COMPARISON OF THE SLOPES OF THE CALIBRATION GRAPHS OBTAINED BY DILUTING THE GASOLINE STANDARD WITH DIFFERENT SOLVENTS Student’s &values* for samples in- I A I Non-leaded Diluent Hexane MIBK Benzene gasoline Hexane .. - 0.15 1.50 1.50 MIBK .. .. . . 0.75 - 0.18 0.18 Benzene . . .. . . 0.75 0.75 - 1.50 Non-leaded gasoline . . 1.50 1.50 3.0 - * Five standard samples were used to prepare each calibration graph. The tabulated value of t is 3.18 a t a 95% level of probability.In another series of experiments, calibration graphs were prepared using constant volumes of solutions prepared by diluting the TEL standards in GI, G, and G, with their respective solvents. These solvents were prepared (as described previously) from those usually employed by CAMPSA to obtain commercial gasolines. The percentages of the basic components used to prepare the solvents were chosen to obtain gasolines with very different boiling-points. The slopes of the calibration graphs were compared with each other and it was found again that the differences are not significant at a 95% level of probability (Table 111). The lead contents in the four commercial gasoline samples, obtained from these graphs, were also compared with those derived from the ASTM titrimetric method (Table IV).These results do not show significant differences at a 95% level of probability. TABLE I11 COMPARISON OF SLOPES OF CALIBRATION GRAPHS OBTAINED FROM TEL I N DIFFERENT NON-LEADED GASOLINES Student’s t-values* for standards in- Diluent of the A I TEL standards G,t G2 t Gat - 2.52 1.03 - 2.57 3.05 0.87 1.80 - Gl G2 G3 * Calibrations were obtained from five standard samples. The t GI, G2 and G3 are non-leaded gasolines. tabulated value of t is 3.18 a t a 95% level of probability. Finally, the standard-additions method was applied by using standards of TEL in benzene with two different procedures. In method A, 1-6 ml of the standard solution of TEL in benzene were added to 1 ml of the sample; in method B, 1 ml of the benzene solutions containing increasing amounts of TEL was also added to l-ml of the gasoline samples.The slopes obtained from these two methods were compared (Table V). Again, according to the Student t-test the slopes show no significant differences at a 95% level of probability. However, these graphs also permit accurate determinations of the lead contents in the gasoline samples studied. Similar results were found when gasoline standards (including that which contains TEL and TML) instead of a standard of TEL in benzene was used.January, 1980 LEAD IN GASOLINE BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 41 TABLE IV COMPARISON OF RESULTS OBTAINED BY USING CALIBRATION GRAPHS PREPARED FROM TEL I N DIFFERENT NON-LEADED GASOLINES Concentration of TEL*/g 1-1 A I \ Developed method (emulsion) of sample Glt Gzt G3t7 (ASTM) Octane rating I A Check method 85 0.52 0.49 0.49 0.50 90 0.45 0.45 0.45 0.46 96 0.53 0.47 0.51 0.54 98 0.67 0.64 0.62 0.64 * Each result is the average of five determinations.t GI, G, and G3 are non-leaded gasolines used as diluents of the TEL to prepare the standards for calibrations. Conclusions The results referred to above show the applicability of the oil - water emulsions in con- junction with atomic-absorption spectrophotometry for the determination of lead in commercial gasolines. For practical purposes water is used as the diluent to obtain the emulsion, thus avoiding the use of large amounts of organic solvents. The HLB of the emulsifier, necessary in order to obtain stable emulsions, is not critical but is within the range 13-15.The sensitivity of the absorption signal is about 15 times higher than that obtained from aqueous solutions. The signals from several matrices are significantly alike, as is shown by the results found from the analysis of different gasoline samples, containing TEL and TML, by employing calibration graphs obtained from the gasoline standard, which TABLE V COMPARISON OF SLOPES OF GRAPHS OBTAINED BY TWO DIFFERENT STANDARD-ADDITIONS METHODS Octane rating of sample 98 96 90 98 96 90 Student’s t-values* Method A Method B - - 98 96 90 98 96 90 - 0.71 1.77 1.06 2.32 3.25 0.47 - 1.65 0.75 0.15 2.71 1.18 1.65 - 3.05 2.72 2.12 1.41 2.51 1.75 - 2.36 3.30 2.50 0.15 3.21 1.18 - 0.47 3.10 2.58 4.24 3.30 0.94 - 7 A > * Calibration graphs were obtained from four commercial gasoline samples. The tabulated value of t is 4.30 a t a 95% level of probability.also contains TEL and TML, and from solutions of TEL in benzene or in non-leaded gaso- lines. Consequently, it would appear that differences in the analytical signals from different alkyl compounds are minimised by this method and different calibration procedures are possible. The accuracy, determined by comparing the results with those supplied by the ASTM titrimetric method, shows the valihty of the proposed method.42 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. POLO-D~EZ, HERNANDEZ-M~NDEZ AND PEDRAZ-PENALVA References “IP Standards for Petroleum and its Products,” Part 1, Sections 1 and 2, Institute of Petroleum, “Annual Book of ASTM Standards. Petroleum Products and Lubricants,” Moss, R., and Campbell, K., .I. Inst. Petrol., 1967, 53, 89. Blears, D. G., Williams, J. I., and Melloy, D. G., Publication No. G6/19/26, The Associated Octel Robinson, J. W., Analytica Chzm. Acta, 1961, 24, 451. Kashiki, M., Yamazoe, S., and Oshima, S . , Analytica Chinz. Acta, 1971, 53, 95. Robbins, W. K., Analytica Chzm. Acta, 1973, 65, 285. Rusell, T. J., and Campbell, K., The Associated Octel Company Limited, January 1977 Hernindez-Mkndez. J ., Polo-Diez, L., and Bernal-Melchor, A., Analytica Chim. A cta, 1979, 108, 39. Polo-Diez, L., Hernandez-Mkndez, J ., and Pedraz-Penalva, F., paper presented a t Euroanalysis I11 Dagnall, R. M., and West, T. S., Talanta, 1964, 11, 1553. Lissant, K. J ., Editor, “Emulsions and Emulsions Technology,” Marcel Dekker, New York, 1974, Becher, P., “Emulsions, Theory and Practice,” Blume Editorial, Madrid, 1972, pp. 217 et seq. Peters, D. G., Hayes, J. M., and Hieftje, G. M., “Chemical Separations and Measurements,” Saunders, Received May 22nd, 1979 Accepted August 22nd, 1979 London, 1976. American Society for Testing and Materials, Philadelphia, 1975. Parts 23, 24 and 25. Company Limited, London, 1969. Conference, Dublin, 1978. pp. 397 and 749. Philadelphia, 1974, pp. 29 et seq.

ISSN:0003-2654    

DOI:10.1039/AN9800500037

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

Analyst, January, 1980, Vol. 105, pp. 43-47 Determination of Lead in Drinking Water by Atomic- absorption Spectrophotometry Using an Electrically Heated Graphite Furnace and an Ammonium Tetramethylenedithiocarbamate Extraction Technique 43 R. P. Mitcham Essex Water Company, Langford, Maldon, Essex, CM9 6QA The determination of lead in drinking water using electrothermal atoniisers has been shown to suffer from serious suppression effects. This paper deals with a semi-micro technique that chelates and separates the lead into an organic extract prior to analysis. The extract is then injected directly into the atomiser without the need to separate the two phases. The method is shown to conform to the now generally accepted standard adopted by the Water Industry for trace metal analysis.Keywoyds Lead determination ; water analysis ; atomic-absorption spectro- photometry ; electrothermal atomiser Within the past few years there has been increasing interest in the amount of lead consumed by people in their drinking water. There have been several local surveys and a national survey. The national survey was conducted by the Water Authorities on behalf of the Department of the Environment, and the requirements of accuracy for any analytical method were established. The bias of any method must not exceed 0.005 mg 1-1 or 10% of the lead concentration (whichever is the greater). The standard deviation must not exceed 0.0025 mg 1-1 or 5% of the lead concentration (whichever is the greater). In addition, the limit of detection should be of the order of 0.005 mg 1-l.These requirements are now generally accepted throughout the industry as the standard for any method used to deter- mine lead in water.l The determination of lead by atomic-absorption spectrophotometry using an electrically heated graphite furnace has at face value many attractions. The method is rapid, very sensitive and requires almost no pre-treatment of the sample. Unfortunately, the applica- tion of the technique has been found to be limited by the inorganic constituents in river and treated waters that interfere to cause marked suppression. All of the waters within the Essex Water Company’s area of supply have high dissolved solids contents and, because of the interference effects, direct analysis using an electrothermal atomiser has been found to give results that do not comply with the now recognised standard.Various methods have been proposed to overcome these effects. These include the use of EDTA,2 ascorbic acid3 and orthophosphoric acid4 as releasing agents. Our investigations of the use of these reagents have not provided a general method applicable to all of the different types of water that we encounter. With ascorbic acid and EDTA, it was found that waters with high sodium contents still exhibited high suppression effects. For example, in some samples containing 300 mg 1-1 of sodium the suppression effects have still been of the order of 40%. These results are in agreement with a review by the Water Research Centre,5 which recommends the standard-additions technique as the only satisfactory method of determining lead in water by electrothermal atomisation techniques.The orthophosphoric acid method relies on the removal of interfering anions as their volatile acids, leaving the cations as their phosphates. The problem found with this method was that during atomisa- tion considerable background absorption was experienced, which was too great for efficient compensation by continuous background correction. The method of sample and standard additions, although the inost accurate method tested to date, is time consuming and much of the advantage of the furnace is lost. There is also the possibility that after a large number of samples, the accumulated residues within the tube may reduce the sensitivity.44 MITCHAM: DETERMINATION OF LEAD I N DRINKING Analyst, Vol. 105 A more satisfactory method was sought that would preferably separate the lead from the interferences prior to analysis.The obvious course was to use some form of extraction technique and the tetramethylenedithiocarbamate - 4-methylpentan-2-one system used for flame atomic-absorption spectrophotometry seemed the most suitable. The method currently used in the water industry6 is very time consuming and involves extracting200 ml of sample with 10 ml of 4-methylpentan-2-one and separation of the organic phase prior to analysis. A more rapid method has been developed using a semi-micro separation, based on the technique recommended by Barnard and Fishman.' In this instance an aliquot of the organic phase is transferred directly by pipette from the extraction tube into the furnace.The method appears to be preferable to other reported methods and is described in this paper. Experimental Reagents Concentrated nitric acid. Atomic-spectroscopy grade. Ammonium tetramethylenedithiocarbamate. Laboratory-reagent grade. This is used as a 1% m/V solution in distilled water and is filtered prior to use. 4-Methylpentan-2-one. Analytical-reagent grade. Apparatus The atomic-absorption spectrophotometer used was a Perkin-Elmer 360 fitted with a deuterium arc background corrector. This instrument is fitted with a special band pass mode for carbon furnace work that consists of a slit of reduced height. The carbon furnace was an HGA 74, and nitrogen (with a low oxygen content) was used as the purge gas. This particular furnace can operate in three modes of gas flow during atomisation but for this work the miniflow mode was used. In this mode the purge gas flow is reduced from the normal rate of 4 ml s-l to 0.5 ml s-l for 8 s of atomisation.This results in an enhancement of sensitivity. All furnace injections were made with Oxford micropipettes utilising disposable plastic tips. The extraction vessels were glass specimen tubes, 25 ml in capacity, with plastic snap-on caps. A Perkin-Elmer 56 flat-bed recorder was used to trace the absorbance readings at a chart speed of 0.10 cm s-l. Optimisation of Instrument Operation For the operating conditions for the atomic-absorption spectrophotometer the manu- facturer's instructions were followed. The resonance line used was 283.3 nm and the band pass was set at 0.7 nm (carbon furnace mode).The operating conditions of the furnace had to be established as the stability of the lead chelate was unknown. This tempera- ture was selected as it is slightly lower than the boiling-point of 4-methylpentan-2-one and would avoid spluttering within the tube during drying, and the time was sufficient to evaporate off the solvent completely. The atomisation temperature was set at 2000 "C. The time for atomisation was set at 10 s, which is 2 s longer than the programme for mini- flow operation. The ashing temperature was established by studying the effect of various ashing temperatures on the atomisation signal of a 0.1 mg 1-1 lead organic extract. Parallel with this a distilled water standard, acidified with nitric acid, was also studied.The period of ashing was fixed at 15s and the experiment was conducted twice on both organic and aqueous samples, in two different carbon furnace tubes. From Fig. 1 it appears that the ideal ashing temperature for the organic extract is 700 "C. This is slighly lower than the aqueous lead nitrate standard. These findings are similar to those reported by H o d g e ~ , ~ who studied the effect of orthophosphoric acid on lead deter- minations. The Environmental Protection Agency* also recommends this asliing tempera- ture setting for lead determinations using electrothermal atomisation methods. Procedure previously acidified to pH 2.5 (k0.3) with nitric acid. pH 2.5 (k0.3) upon receipt. and distilled water prior to use. The initial drying stage was set at 100 "C for 30 s.Standards of 0.010, 0.025, 0.050 and 0.100 mg 1-1 of lead are prepared in distilled water Water samples are also acidified to All glassware and plastics are washed with dilute nitric acidJanuary, 1980 WATER BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 45 \ 1 0 200 400 600 800 1000 1200 Ashing temperature/OC Fig. 1. Effect of ashing temperature on absorbance signal for 0.1 mg 1-1 of lead atomised at 2000 "C. A, Organic ex- tract; and B, lead nitrate solution. Results are the mean of 5 readings. A 10-ml aliquot of each drinking water sample is transferred by pipette into a specimen tube, then 1 ml of ammonium tetramethylenedithiocarbamate solution is added, followed by 5 ml of 4-methylpentan-2-one. The mixture is shaken vigorously for 1 min and allowed to stand for 3 min to allow the two phases to separate.A 5O-pl volume of the organic phase is transferred by pipette into the carbon furnace and the sample atomised using the miniflow programme given in Table I. TABLE I MINIFLOW PROGRAMME Drying Ashing Atomisation Temperature/"C . . . . 100 700 2 000 Time/s . . .. . . 30 15 10 Results and Discussion The method and operating conditions of the equipment having been established, the linear The results of this investigation are given in Table 11. working range was then determined. TABLE I1 LINEAR WORKING RANGE OF THE PROPOSED METHOD Pb concentration in Absorbance of organic Relative standard mg 1-1 determinations) (10 determinations) aqueous sample/ phase (mean of 10 deviation, 0.010 0.055 5 4.0 0.025 0.1490 1.4 0.050 0.3160 3.4 0.100 0.569 5 1.9 It appears from Table I1 that the method has a linear working range up to approximately 0.075 mg 1-1 of lead using this equipment.The recoveries of a lead standard added to potable waters of different compositions were also studied. Three types of water were selected, which, from previous analyses, were known to cause severe interference effects. Two were from deep well sources and the third was a lime - soda softened river water. The waters were acidified to pH 2.5 (k0.3) with nitric acid and 0.1 mg of lead was added to46 MITCHAM: DETERMINATION OF LEAD I N DRINKING Analyst, vol. 105 1 1 of each. These aqueous samples were first analysed by direct injection into the furnace, and subsequently using the proposed procedure, by extraction with ammonium tetra- methylenedithiocarbamate ; the volume of 4-methylpentan-2-one used for this extraction was increased to 10 ml to avoid the concentration effect that occurs in the proposed method.To establish that the improved recovery was not simply a function of added ammonium tetramethylenedithiocarbamate, the aqueous solutions were also analysed for lead after the addition of the complexing agent. The compositions of the water samples are indicated in Table I11 and results for the lead determinations given in Table IV. TABLE I11 MINERAL ANALYSIS OF THE WATERS USED Electrical Contentlmg 1-1 conductivity/ A > Sample $3 cm-l ;a2+ Mg2+ Naf C1- 50, P 0 4 a - 6.1 0.20 Potable water . . 786 40 13 112 93 171 Well .. I . . . 1618 102 57 225 299 188 16.0 <0.01 Well . . . . . . 1600 19 15 356 336 74 10.9 0.16 The results in Table IV illustrate clearly how the response of 0.10 mg 1-1 of lead is reduced in water containing high levels of dissolved salts. Sample 2 is a particularly good example of suppression effects as absorption is reduced by almost 50% in comparison with the distilled water standard. The addition of the complexing agent has further reduced the absorbance signals, in two instances by almost 40% of their original values. The results for the organic extracts, however, agree well with those for the extracted standard. The lowest result represents a 94% recovery, which is adequate for practical purposes. From the results of a recent inter-laboratory analytical quality control exercise for lead, conducted by the Anglian Water Authority, it was possible to compare bias and standard deviation results for two independent standards analysed by the proposed method and the solvent extraction - flame method.Table V gives the results of this exercise. The results for the extraction - flame technique are the composite results of ten laboratories. The standard deviation has been calculated by the method of pooling estimates of the same popula- tion standard deviation. I t is apparent from these results that the method is comparable to the recognised solvent extraction - flame technique and is also capable of obtaining the standards for bias and standard deviation required by the water industry. To establish that the method was satisfactory over a wide composition of waters, six samples were collected for sample and standard-additions analysis.Four were taken from a consumer distribution network, one was an estuary water having a conductivity of 2500 pS cm-l and the sixth was a sample of estuary water having a conductivity of greater than 28000 pS cm-l. The spike increased the concentration of lead in the samples by 0.05 mg 1-1 and the results of this exercise are given in Table VI. The high residual lead levels for the sea- estuary water sample cannot be completely explained, but this does not detract from the exercise, which was to establish the recovery of a spike standard, which for this as with all the samples was acceptable. TABLE IV ABSORBANCE OF 0.10 mg 1-1 OF LEAD IN DIFFERENT WATERS The results are the means of 10 determinations and are corrected fcr the blank.Aqueous + complexing Aqueous sample agent Organic extract No. Type of water Absorbance yo Absorbance % Absorbance yo - Acidified distilled water 0.330 7.7 0.253 0.2 0.290 1.5 1 Acidified potable water 0.282 5.1 0.265 4.6 0.302 1.8 2 Acidified well water 0.183 15.7 0.107 4.4 0.273 1.8 3 Acidified well water 0.215 11.9 0.140 5.1 0.310 5.3 * R.S.D. = relative standard deviation. Sample R . S . D . I r * r *January, 1980 WATER BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 47 TABLE V RESULTS OF AN INTER-LABORATORY CONTROL EXERCISE Technique Micro-extraction, electro- thermal atomisation Solvent extraction, flame atomisation . . .. Micro-extraction, electro- thermal atomisation Solvent extraction, flame atomisation .. .. Mean concentration Standard Relative True value/ obtained/ deviation/ standard Bias a t 95% mg 1-l mg 1-1 mgl-l deviation, yo confidence, yo . . 0.0959 0.095 1 0.002 2 2.3 1.9 . . 0.0959 0.096 7 0.002 6 2.7 2.9 . . 0.0501 0.050 9 0.001 5 2.9 3.0 . . 0.0501 0.050 0 0.002 1 4.2 3.2 Conclusion The method appears to overcome all of the problems associated with the analysis of lead using electrothermal atomisers. In this work, no electronic scale expansion of instrumenta- tion has been used and the concentration step during extraction is minimal. The method may therefore have wider use than that is at present required by the water industry. TABLE VI RECOVERY OF LEAD FROM SPIKED SAMPLES The results are the means of five determinations.Hard water: total hardness >250 mg 1-1 CaCO,. Medium-hard water: total hardness 150-250 mg 1-1 CaCO,. Pb in sample after spiking Pb in sample/ Sample mg 1-’ Valueslmg 1-1 Hard water . . .. . . 0.0092 0.0565 Hard water . . .. . . 0.0032 0.0503 Medium-hard water. . . . 0.0037 0.053 Hard water . . .. . . 0.0035 0.050 7 Eastuary water . . . . 0.0034 0.053 3 Sea - estuary water . . . . 0.0083 0.055 1 Relative standard deviation, yo Recovery, 3.9 94.6 2.1 94.2 1.4 100.2 1.9 94.4 3.5 99.8 2.5 93.6 % I thank Mr. J. G. Slack and Dr. J. M. Ottaway for their helpful advice, and the Anglian Water Authority, in particular Dr. B. T. Croll and Mr. C R. Whitfield, for permission to use the statistical analysis data. References 1. 2. 3. 4. 5. Ransom, L., and Wilson, A., “Determination of Lead in Tap Water: Interlaboratory Harmoniza- DolinSek, F., and Stupar, J., Analyst, 1973, 98, 841. Regan, J . G., and Warren, J.. Analyst, 1976, 101, 220. Hodges, D. J., Analyst, 1977, 102, 66. Ransom, L., and Orpwood, B., “An Evaluation of an Electrothermal Device for the Determination of Lead and Cadmium in Potable Water,” Water Research Centre Technical Report No. 49, July 1977. 6. “Lead in Potable Waters by Atomic Absorption Spectrophotometry,” HM Stationery Office, London, 1976. 7. 8. “The Determination of Antimony, Arsenic, Beryllium, Cadmium, Lead, Selenium, Silver and Tellurium in Environmental Water Samples by Flameless Atomic Absorption,” Environmental Protection Agency, Chicago, PB 269 902, January 1977. tion,” Water Research Centre Technical Report No. 28, July 1976. Barnard, W. M., and Fishman, N. J., Atom. Absorption Newsl., 1973, 69, 118. Received June 7th, 1979 Accepted July 25th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500043

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

48 Analyst, January, 1980, Vol. 105, pp. 48-51 Determination of Trace Concentrations of Mercury in Biological Materials After Digestion Under Pressure in Nitric Acid Catalysed by Vanadium Pentoxide Vlasta Korunova and Jifi Dedina Institute of Physiology, Czechoslovak Academy of Sciences, Videhkd 1083, 142 20 Prague 4, Czechoslovakia The determination of trace amounts of mercury in biological materials is described. Vanadium pentoxide-catalysed digestion in nitric acid under pressure in a PTFE autoclave was used for destruction of the samples. Satisfactory destruction was achieved in the absence of sulphuric acid. Mercury was determined in the digest solution by atomic-absorption spectro- photometry using a cold vapour circulating in a closed system The detection limit was optimally 1 ng g-l.No losses of mercury occurred. The accuracy of the determination was confirmed by the close agreement of certified and determined concentrations of mercury in reference materials. Keywords : Mercury determination ; vanadium pentoxide-catalysed nitric acid autoclave digestion ; cold-vapour atomic-absorption spectrophotometry ; biological standard reference materials Conventional wet ashing, preventing mercury losses during digestion, is very laborious. Alternatively, digestion of biological materials in PTFE autoclaves can be ~ s e d , l - ~ usually with nitric acid alone1s3 and sometimes in combination with sulphuric acid2** or perchloric acid.586 The main advantage of autoclave digestion is its speed and virtually complete elimination of 10sses.~~~ In addition, the risk of contamination is also lower due to the small surface area in contact with the sample during digestion.However, increased speed of digestion should not be achieved at the expense of complete sample destruction. This is especially important in the atomic-absorption spectrophotometric determination of mercury by the cold-vapour technique, when all of the mercury is to be in an inorganic form. After incomplete destruction some mercury may be left in a biologically bound form. Another source of error may be the non-specific absorption of volatile fractions of residual material that remains after incomplete destruction and is entrained into the absorption cuvette.7J3 Our preliminary experiments showed that in pure nitric acid complete destruction of materials such as plants, animal tissues or dried milk is not achieved even after 2 h a t an autoclave temperature of 140 "C.Nitric acid - sulphuric acid mixtures are unsuitable as the resulting nitrogen oxides, which interfere in the determinati~n,~~lO are difficult to remove from the digest ~olutions.~ As shown earlier, vanadium pentoxide in the presence of sulphuric acid is an efficient catalyst for the digestion of biological materials at atmospheric pressure.ll A mixture of nitric and sulphuric acids with vanadium pentoxide ensures the satisfactory destruction of different types of biological materials at a temperature of 140 OC.l2 We found that in an autoclave, in which a temperature of 140 "C can be attained even in the absence of sulphuric acid, digestion proceeds satisfactorily in a mixture of nitric acid and vanadium pentoxide.Experimental Reagents Working mercury standard solutions (0.1 and 1 pg ml-l) were obtained by diluting a stock standard solution (1000 pg ml-l) with 8% m/V nitric acid containing 0.01% of potassium bichrornate. The reduction solution was 20% m/V tin(I1) chloride in 2 M sulphuric acid. All solutions were prepared from de-ionised water. Laboratory implements were washed with chromic acid (glassware) or concentrated nitric acid (PTFE autoclave vessels) and dilute potassium permanganate solution and then repeatedly rinsed in de-ionised water.KORUNOVA AND DEIDINA 49 Apparatus Mercury absorption was measured a t 253.7 nm (slit width 0.7 nm) on a Perkin-Elmer 503 atomic-absorption spectrophotometer with a Perkin-Elmer mercury hollow-cathode lamp.The determinations were carried out using the Perkin-Elmer Mercury Analysis System (Fig. 1) with a modified reaction vessel of a more suitable shape and smaller volume. Digestion was conducted in a Perkin-Elmer Autoclave 11, consisting of a 120-ml PTFE vessel and an outer coating; the autoclave can be heated to a maximum temperature of 160 "C. All glassware was made of Simax glass (Pyrex-type glass). Fig. 1. Experimental arrange- ment of the mercury analysis sys- tem: 1, air pump: 2, reaction vessel (volume of lower part 20 ml); 3, drying tube with Mg(C10,),.2H20; 4, absorption cuvette (length 14.5 cm); and A, connector. Sample Digestion A maximum of 1 g of a dry sample was heated in the autoclave for 30 min at 100 "C and 60 min at 150 "C with 5 ml of concentrated nitric acid and 10-20 mg of vanadium pentoxide.After cooling, the autoclave was opened and nitrogen oxides dissolved in the digest solution were allowed to escape. After several minutes the digest solution was supplemented with an excess of 5% potassium bichromate solution (2-7 ml). The contents of the PTFE vessel were then transferred quantitatively into the reaction vessel together with the wash liquid (10 ml of 0.025% potassium permanganate in 0.5 M sulphuric acid), and the volume was made up to 20ml with 0.5 M sulphuric acid. All blank solutions were prepared by the same procedure. The reaction vessel was then closed. Determination The method with a cold vapour circulating in a closed system was used.13 A maximum of 24 h was allowed to elapse between the digestion and the determination.A digest or blank solution in the reaction vessel was supplemented with 2 ml of the reduction solution, the vessel was immediately fitted with a ground-glass joint (Fig. 1) and the air pump was switched on. When the recorder trace of the signal reached a maximum, the tubing between the absorption cuvette and the pump was disconnected (point A) and the mercury vapour was shunted into the exhaust. The complete disappearance of mercury vapour from the sample was indicated by the decrease of the absorbance to the base line. The dependence of the signal height on the reaction vessel matrix and volume was compensated for by adding standards in small volumes (maximum lOOp1) to each measured solution.The amount of mercury in the digest or blank solution was calculated from the heights of the first signal (mercury in the digest or blank) and from those of the other signals (mercury from added standards) measured in the same solution.50 KORUNOVA AND D ~ D I N A : DETERMINATION OF TRACE Analyst, Vol. 105 Results and Discussion The calibration graph was linear throughout the measured range (up to 100 ng of mercury in the reaction vessel). The sensitivity of the determination (the slope of the calibration graph) depended on the volume and composition of the fluid in the reaction vessel. The dependence of the sensitivity on volume was slight, a 10% increase or decrease in the volume in the reaction vessel (20ml) causing only a 3% increase or decrease in sensitivity.The composition of the solution in the reaction vessel affected the sensitivity more markedly; the slope of the calibration graph (S) expressed in absorbance units per nanogram of mercury changed from 1.1 x in some digests. The standard deviation of the noise (sN) was about 5 x absorbance unit; thus, in measure- ments of this type, the detection limit (2s,/S) was usually 1 ng of mercury (this corresponds to a relative detection limit of 0.05 ng ml-l). The relative standard deviation of the signal heights for 50 ng of mercury measured in the same solution was 1-1.5%. in pure 0.5 M sulphuric acid to 0.5 x TABLE I RECOVERY OF MERCURY(II) CHLORIDE ADDED TO SAMPLES BEFORE DIGESTION Hg added/ Hg content in digest Material ng solution*/ng Recovery, % Dried milkt .... . . 0.0 7.7 f 1.7 NBS orchard leaves: . . . . 0.0 26.7 f 1.3 NBS orchard leavest . . . . 20.0 46.2 f 1.0 98 f 8 Dried milkt .. .. . . 50.0 57.7 f 1.5 99 f 4 * Arithmetic mean f standard deviation from seven samples. t 0.2-g samples. ; Standard reference materials from the National Bureau of Standards; 0.15-g samples. Nitrogen oxides cause non-specific absorption. Their residues in digests were therefore removed by oxidation with potassium bichromate; the excess of this reagent also served for stabilisation of the digest. The accuracy could also be affected by mercury losses due to volatilisation or sorption throughout the procedure; as can be seen from Table I, these losses are negligible. Obviously, no losses occur during autoclave digestion (which is in agreement with earlier studies3), during the transfer of the digest into the reaction vessel or during its stay in this vessel.In a digest solution containing nitric acid and potassium bichromate kept in the reaction vessel, mercury is stable for at least 20 d (Table 11). Also, in agree- ment with earlier data,14 standards are stable in 5% nitric acid containing potassium TABLE I1 RELATIVE CONCENTRATIONS OF Hg2+ IN VARIOUS SOLUTIONS Each value is the mean of concentrations deter- mined independently in all solutions of a given type. Type of solution* I A \ Solution age/d A, % B, % C, % 0 100 100 100 2 96 96 102 10 84 104 105 20 82 101 102 * A: 1 ng ml-l of Hg in 5% nitric acid (in glass calibrated flask), a total of three solutions.B: 2 ng ml-1 of Hg in 5% nitric acid and 0.01% potassium bichromate (kept in a glass calibrated flask), a total of eight solutions. C: 3.7-7.1 ng ml-1 of Hg in digest solution from 0.5 g of lyophilised rat kidney (kept in reaction vessels), a total of six solutions .January, 1980 CONCENTRATIONS OF MERCURY IN BIOLOGICAL MATERIALS 51 bichromate whereas Hg2+ is unstable in pure nitric acid. The precision and accuracy of the determination, however, can best be assessed from repeated mercury determinations in reference materials (cf., Table 111). The detection limit (c,) depends on the content of mercury in blanks, i.e., on the contami- nation of reagent and glassware and on contamination from the laboratory environment. The detection limit is defined by the equation cL = %bl/m, where sbl is the standard devi- ation of the amount of mercury in the blank and m is the sample mass.The content of mercury in the blanks fluctuated, depending on the contamination of reagents, in the range 2-10 ng, with a relative standard deviation of about 20%. For l-g samples, the optimum detection limit was therefore about 1 ng of mercury. The size of the sample depends on the type of niineralised material. With dried milks and the bovine liver standard reference material from the National Bureau of Standards (NBS), a sample of 1 g can be digested. TABLE I11 DETERMINED AND CERTIFIED MERCURY CONTENTS I N REFERENCE MATERIALS Determined/ Certified/ Number of Material ng g-l ng g-‘ samples Bovine liver*, t . . .. 16 f 5 16 2 6 Orchard leaves*, $ .. . . 153 & 8 155 15 8 Bowen’s kale:, 5 . . . . 166 & 4 1605 3 * NBS standard reference material. t 1.0-g samples. 0.15-g samples. Reference material described in reference 15. With some other materials, digestion is accompanied by a higher pressure rise and the maximum applicable sample sizes are therefore lower. For instance, the maximum digestible amounts were 0.6 g for lyophilised rat kidneys and for samples of laboratory diet and cbout 0.2 g for the NBS orchard leaves. This method of digestion is highly efficient (e..g., in dried milks digestion leaves no perceptible residues), yet some materials (lyophilised kidneys and liver, NBS orchard leaves, laboratory diets) leave remnants of undecomposed fat after digestion. The agreement between the determined and certified mercury concentra- tions in reference materials (Table 111) indicates that this fact has no effect on the accuracy of the determination.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. References Holak, W., Krinitz, B., and Williams, J. C., J . Ass. Off. Analyt. Chem., 1972, 55, 741. Paus, P. E., Atom. Absorption Newsl., 1972, 11, 129. Kotz, L., Kaiser, G., Tschopel, P., and Tolg, G.. 2. Analyt. Chem., 1972, 260, 207. Von Auslitz, H. J., Arch. Lebensmittelhyg., 1976, 27, 68. Van Eenbergen, A., and Brunin, X. E., Analytica Chim. Acta, 1978, 98, 405. Luyten, S., Smeyers-Verbeke, J., and Massart, D. I., Atom. Absorption Newsl., 1973, 12, 131. Poluektov, N. S., Vitkun, R. A., and Zeljukova, J. V., Zh. Analit. Khim., 1964, 19, 937. Lindstedt, G., Analyst, 1970, 95, 264. Woidich, H., and Pfannhauser, W., 2. Lebensmittelunters. u. -Forsch., 1974, 155, 271. Adrian, W. J., Atom. Absorption Newsl., 1971, 10, 96. Malaiyandi, M., and Barrette, J. P., Arch. Envir. Contam. Toxacol., 1974, 2, 315. Deitz, F. D., Sell, J. L., and Bristol, D., J . Ass. Off. Analyt. Chem., 1973, 56, 378. Hatch, W. R., and Ott, W. L., Analyt. Chem., 1968, 40. 2085. Feldman, C., Analyt. Chem., 1974, 46, 99. Bowen, H. S. M., in Lenihan, J . M. A4., and Thomson, S . J., Editors, “Advances in Activation Analysis,” Academic Press, London, 1969, Volume 1, p. 101. Received May 22nd, 1979 Accepted July 18th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500048

出版商:RSC    

年代:1980

数据来源: RSC