ISSN:0003-2654    

DOI:10.1039/AN98005FX045

出版商:RSC    

年代:1980

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98005BX047

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

December, 1980 SUMMARIES OF PAPERS I N THIS ISSUESimultaneous Determination of the Noble Metals in GeologicalMaterial by Radiochemical Neutron- activation AnalysisA simple procedure is described for determining the noble metals in geologicalsamples rich in chromium and copper. Powdered rock samples (0.1-0.3 g)are irradiated in an epithermal neutron flux, under a cadmium filter, toreduce interference from 51Cr and 64Cu. After digestion with a peroxidefusion, the noble metals are separated on Srafion NMKR ion-exchange resin.Activity from chromium(VI), retained on the resin, is eliminated by reductionto chromium(II1) with iron(I1) sulphate before separation. The gamma-rayactivity is measured with a lithium-drifted germanium detector for palladium(IOgPd), platinum (lg9Au), iridium (Ig41r) and gold (IB8Au).The separationyield (iridium 40%, palladium SO%, gold 90%) is determined for eachsample by irradiating the resin to activate a stable noble metal carrier.Multi-element noble metal standards (0.1-8 pg) are prepared from dilutesolutions (100 pl) evaporated on resin (1 ml). The practical detection limitsare 2.5 ng of palladium and platinum, 0.1 ng of iridium and 0.01 ng of gold.The reliability of the procedure was confirmed by analysing sulphidestandards (PTM, PTC) and standard rocks (PCC-1, DTS- 1). The methodhas been applied to the analysis of chrome-spinels.Keywords : Noble metal determination ; epithermal neutvon-activation analysis ;ion-exchange separation ; vefere'lzce vocks ; chotlze-spilzelSUSAN J.PARRYUniversity of London Reactor Centre, Silwood Park, Sunninghill, Ascot, Berltshire,SL5 7PY.Analyst, 1980, 105, 1157-1162.Determination of Aluminium in Wood Pulp Liquors Using GraphiteFurnace Atomic- absorption SpectrometryA graphite furnace atomic-absorption spectrometric method for the cleter-mination of 0.5-100 p g ml-l of aluminium in matrices containing chlorineand high concentrations of sulphur is described. The precision of the over-allprocedure was 4%. Graphite tube conditions and the composition of thematrix were found to change the sensitivity for aluminium.Keywords : Aluminium determination ; graphite fzivnace atomic-absovptioiispectrometvy ; szzlphuv intevferevzce ; chlovine ilztevfevence ; gvapliite tubeconditionsJ.-A. PERSSON, W.FRECH, G. POHL and K. LUNDGRENDepartment of Analytical Chemistry, University of lime%, S-901 87 Cme%, Sweden.Analyst, 1980, 105, 1163-1170.Determination of 2-Aminobutane in Potatoes usingHigh-performance Liquid ChromatographyDevelopment work on the use of 2-aminobutane to control certain potato-tuber diseases necessitated the development of a convenient and sensitiveanalytical method for residue determination. The amine was distilled frompotatoes, dansylated and determined using reversed-phase high-performanceliquid chromatography with fluorescence detection. Recovery of 2-amino-butane by distillation was about 95% from standard solutions and 92:d) fromtreated potatoes. The lower limit of detection is below 0.2 p g kg-l.Keywords : 2-A minobutane determination ; potatoes ; vesidues ; high-pevfovmnnceliquid c?womnntogvapJ$y ; flzzovescence detectionK. A. SCUDAMOREMinistry of Agriculture, Fisheries and Food, Slough Laboratory, London Road,Slough, Berkshire, SL3 7H J .Analyst, 1980, 105, 1171-1175

ISSN:0003-2654    

DOI:10.1039/AN98005FP141

出版商:RSC    

年代:1980

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98005BP145

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

DECEMBER 1980 Vol. 105 No. 1257 The Analyst Generation of Covalent Hydrides in Atomic-absorption Spectroscopy A Review R. G. Godden and D. R. Thomerson Bawd-Atowiac L t d , Warnev Draw, Spyangwood Industraal Estate, Braantree, Essex, CM7 7 Y L Summary of Contents Introduction Historical Methods of reduction Design of reaction vessels/collection of the hydride Liquid nitrogen trap Silver nitrate solution Balloon method Other collection devices Direct introduction Automated systems Methods of atomisation Flames and tubes Air - acetylene flame Argon - hydrogen flame Silica tubes Electrically heated tubes Flame-heated tubes Graphite tube furnaces Various Nolecular emission cavity analysis (RIECA) Atomic fluorescence Plasmas D.c. arc discharge Light sources Interferences Background absorption Valency Acid concentration Organic binding Inter-element effects Detector temperature Detection limits and sensitivities Applications Waters and effluents Metallurgical Foods Clinical Agriculture Geological Tobacco Environmental Petroleum products Future trends References Keywords : Review; atomic-absorption spectroscopy; covalent hydride genera- tion 11371138 GODDEN AND THOMERSON : GENERATION OF COVALENT AWd&.St, Vd.105 Introduction Covalent hydrides are a series of compounds whose elements are of the C, N and 0 groups, where the number of valency electrons is equal to, or greater than, the number of orbitals. These elements include As, Al, B, Bi, Ge, P, Pb, Se, Sb, Si, Sn, Te and Ti. Of these thirteen elements, eight have been induced to form covalent hydrides in sufficient amounts to be of practical analytical use; these are As, Bi, Ge, Pb, Se, Sb, Sn and Te.The basic design of a hydride generation system, with subsequent atomic absorption, may conveniently be considered as four steps. First, generation of the h:ydride ; second, collection of the hydride (if necessary); third, transfer of the hydride to the atomiser; and fourth, decomposition of the hydride to the gas-phase metal atoms within the optical axis of the atomic-absorption spectrophotometer. The measured atomic-absorption signal is directly proportional to the number of free metal atoms per unit optical axis cross-section at a given instant. In order to increase this signal it is necessary to generate the hydride quickly, or collect it and then transfer it as quickly as possible to the atorniser.This reduces dilution of the hydride by the carrier gas. Historical Arsenic is both extremely toxic and an indicator of precious metals, even a t low concentra- tions. The poor sensitivity of this metal makes the application of conventional atomic absorption rather difficult. The arsenic resonance lines lie in the far ultraviolet region of the spectrum where flame absorption can produce an unfavourable signal to noise ratio (Fig. 1). With the application of the argon - hydrogen - entrained air flame to atomic- absorption spectroscopy in 1967 a much improved signal to noise ratio was obtained, This flame, usually referred to as the argon - hydrogen flame, when used with conventional hollow- cathode lamp sources, absorbs about 15% of the available light compared with 62y0 for the air - acetylene flame at 193.7 nm.l Interferences are encountered when the argon - hydrogen flame is used to determine microgram amounts of arsenic in real samples.The argon - hydrogen flame is cooler than the air - acetylene flame, and interferences due to incomplete salt dissociation and molecular absorption are common. In order to overcome this problem, modifications of the Gutzlet method, utilising the generation of arsine, have been applied to atomic-absorption spectroscopy. Holak2 collected the arsine in a liquid nitrogen trap, which was then warmed, and the evolved arsine was swept into the flame with a flow of nitrogen. By removing the arsenic from the sample matrix, interferences in the flame are minimised.Madsen3 collected the arsine in dilute silver nitrate solution anld aspirated the resulting solution directly into an argon - hydrogen flame of the type dexribed by Kahn and Schallis.1 A spectrophotometric method, which involves reaction of arsine with silver diethyldithio- carbamate to form a red complex, has also been proposed for the determination of arsenic. Hundley and Underwood4 applied this method to the determination of arsenic in a wide I Nodamping Damping int 5-s I egration A Fig. 1. Arsenic trace at 193.7 nm using the air - acetylene flame.December, 1980 HYDRIDES IN ATOMIC-ABSORPTIOS SPECTROSCOPY 1139 variety of food products. Spectrophotometric methods have the disadvantage of requiring large amounts of sample in order to obtain sufficient sensitivity.Dalton and Nalanoskij were the first to report the direct aspiration of arsine into a flame without the use of a collection device. This work utilised an argon - hydrogen flame, and the liberated arsine was carried to the nebuliser by the released hydrogen. A detection limit of 0.1 pg of arsenic was obtained by this method. The sensitivity was further improved by Fernandez and Manning,6 who collected the arsine in a balloon reservoir before introducing it into an argon - hydrogen flame. The determination of selenium as hydrogen selenide (SeH,) was also described. A detection limit of about 0.02 pg of arsenic was obtained by this method. This apparatus was further improved by Ilanning,’ and this design became the first commercially available attachment for atomic-absorption spectrophotometers.A total reaction time of 4-5 min using tin(I1) chloride -hydrochloric acid - potassium iodide - zinc is normal for this method. It was suggested by Manning that the technique might also be applied to the generation of the gaseous hydrides of antimony and bismuth. The successful generation of stibine was accomplished by Pollock and West,8 but they found that the commonly used tin(I1) chloride - hydrochloric acid - potassium iodide - zinc reaction was not suitable. They obtained good results with titanium(II1) chloride - hydrochloric acid - magnesium. The first reference to the use of a tube detector, in this instance electrically heated, as an alternative to the argon - hydrogen flame appears in the work reported by Chu et ~ 1 .~ in 1972. A tin(I1) chloride - potassium iodide - zinc reaction in an acidic matrix was employed together with a balloon collection device. A reaction time of 10 min was allowed before the arsine was swept to the heated tube by a flow of argon, and a sensitivity of about 5 ng was obtained for arsenic. A further interesting aspect of this work is that the reaction time necessary to complete the reduction of the sample is about 8 min and not 4-5 min as allowed in previous reports. Also in 1972 the use of sodium tetrahydroborate(II1) to produce arsine was reported by Braman et al.,1° who also reported an emission in a d.c. arc discharge method. In 1973 the use of sodium tetrahydroborate(II1) was extended to the determination of Se, Sb and Bi by Schmidt and Royerll and to the determination of Ge by Pollock and West.12 Fernandez13 expanded this range to the determination of Sn and Te.All the above methods continued to use the argon - hydrogen flame. In 1974 Thompson and Thomersonl4 published a report on the use of sodium tetrahydro- borate(II1) as a reducing agent, coupled with a flame-heated silica tube as a detection device. Fig. 2 shows the sensitivity obtained for arsenic (cf. Fig. 1). This paper also included a report on the generation of plumbane (PbH,) for analytical purposes. At this stage the various combinations of reducing agents, collection devices and detection systems had been established. No detection limit was quoted. r 1 ml 0.01 pg rnl-’ arsenic (in 1.5 M HC I) Time ---+ Fig.2. Typical arsenic trace (2 ml of 1% NaBH, solution in the generator cell).1140 GODDEN AND THOMERSON: GEXERATION OF COVALENT AIzalyst, VOL. 105 Methods of Reduction Various reducing agents and sources of nascent hydrogen have been suggested in order to convert the element of interest into its hydride. The Marsh reaction, using zinc metal and dilute hydrochloric or sulphuric acid, was the most frequently used in the early methods. A major drawback of this method is that it can only be used for arsenic, and the time taken for the completion of the reaction may be as long as 10 min. I t is therefore necessary to store the evolved gas in some form of reservoir. A mixture of dilute hydrochloric acid, 40% m/V potassium iodide solution, 40% mjv tin(I1) chloride solution (using the dilute hydrochloric acid as a solvent) and granular zinc metal has been d e ~ c r i b e d .~ ~ ~ This appears to give a faster reaction time as no collection vessel, or carrier gas, was found to be necessary. This mixture also appears to be usable for the determination of selenium. Pollock and West* concluded that the above mixture was not suitable for the generation of stibine (SbH,) from solutions containing antimony(II1). However, they found that a magnesium - hydrochloric acid - titanium(II1) chloride medium was entirely satisfactory. The advantage of this reduction mixture is that it can also be used for the determination of Bi and Te. The use of sodium tetrahydroborate(III), reported by Schmidt and Royer,ll is a land- mark in the development of the technique.It has been applied successfully to the genera- tion of the hydrides of As, Bi, Ge, Sb, Se, Sn and Te,13 and subsequently Pb.l4 Collection times were reduced to 30 s for Bi and Te and 2-3 min for As, Ge, Sb, Se and Sn.13 However, the system described by Thompson and Thomerson14 requires no collection vessel, and complete reduction of most metals was obtained within 30 s. The sodium tetrahydroborate(II1) method offers several advantages o\rer both the zinc - tin(I1) chloride and magnesium - titanium(II1) chloride procedures. Although sodium tetrahydroborate(II1) was initially available only in powder form, the use of 0.25-g pellets soon became established. These are sufficiently uniform to eliminate the need for weighing, are easy to handle and are relatively inexpensive.Although some problems were experienced with reagent purity in early determinations,14J5 the quality of the sodium tetrahydroborate(II1) supplied by the major companies in recent years has improved considerably. The hydride generator described by Fernandez13 requires a pellet of sodium tetrahydro- borate(II1) to be used for each measurement, whereas the system described by Thompson and T h ~ m e r s o n ~ ~ requires one or two 0.25-g pellets to be dissolved in 25 ml of water and 1- or 2-ml aliquots of this solution are used. The latter system is obviously more economical with regard to consumption of reagent. Several workers16-l* have reported that sodium tetrahydroborate(II1) solution is relatively unstable, and periods of several hours to 1 d have been quoted as the useful life.However, Knechtel and liraserlS reported the use of a 2% m/V solution of sodium tetrahydroborate(II1) containing one pellet of potassium hydroxide per 100 ml of solution. The solution was filtered through a 0.45-pm membrane filter, and it was reported that this solution remained usable for 3 weeks. A different reducing mixture, used by Goulden and Brooksbank,*O consists of a mixture of tin(I1) chloride, potassium iodide and aluminium powder. h heated packed stripping column was used to strip the selenium hydride generated from the reduced specimen so that atoms, and not molecules, would enter the optical path. A further method, but one which has not gained wide usage, is the zinc column to which an acidic solution of the analyte is added.14v21 This method is applicable only to the genera- tion of arsine and stibine.Design of Reaction Vessels/Collection of the Hydride The speed of reaction of many of the reducing agents used is such that it can take several minutes for the reaction to reach completion. As peak-height measurements are normally made it is obviously advantageous to collect the liberated hydride and then sweep it to the detection system in the shortest possible time. Several methods of collection have been proposed and are outlined below. It should be noted that when sodium tetrahydroborate(II1) is used as a reducing agent the need to collect the liberated hydride is often minimised.December, 1980 HYDRIDES I N ATOMIC-ABSORPTION SPECTROSCOPY 1141 Liquid Nitrogen Trap This method was first proposed by Holak2 and consists of an arsine generator connected to a U-tube via a calcium chloride drying tube.The U-tube was packed with glass beads and immersed in a Dewar flask containing liquid nitrogen. LYhen all of the arsenic had been converted into arsine the generator outlet valve was closed. The rubber connecting hoses on each arm of the U-tube were sealed with pinch-clips and the U-tube was then removed from the liquid nitrogen and allowed to come to room temperature. Nitrogen a t a flow-rate of about 50 ml min-l was connected to one arm of the U-tube and the other arm was con- nected to the nebuliser capillary of an atomic-absorption spectrophotometer. The pinch- clarnps were removed and the arsine was allowed to flow into the mixing chamber.Several worker^^^-^^ have reported the use of a liquid nitrogen trap, but it is a specialised technique and seems to be slow and complex for routine use. However, Braman and Foreback25 have shown that it is a useful method for the collection, and subsequent selective volatilisation, of methylated hydrides of arsenic generated from environmental samples. Silver Nitrate Solution This method was proposed by Madsen3 and consists of an arsine generator constructed using a 125-ml conical flask, 6-mm glass tubing and a 10-ml graduated measuring cylinder. The end of the delivery tube was drawn to a fine tip and was arranged so that the tip of the delivery tube was placed against the bottom of the graduated cylinder, which contained 5 ml of 0.01 M silver nitrate solution. The evolution of the hydride was allowed to continue for 15 min and the delivery tube was then removed from the absorbing solution.This solution was aspirated into the argon - hydrogen flame of an atomic-absorption spectrophoto- meter. The method is time consuming and later methods have found more widespread use. Balloon Method Probably the most popular collection device is the balloon attached to the side of the reaction vessel. This form of collection device was first described by Fernandez and Manning.6 The apparatus was modified by Manning,' who incorporated a dosing stop-cock so that zinc could be introduced into a closed system (Fig. 3). A four-way valve permitted the system to be flushed with an inert gas, and the balloon was attached to a hose nipple, which it was not necessary to remove and replace for each sample.The system was connected to the burner in the same way as the previously described apparatus. Various modifications of the balloon - reaction flask system have been p r o p ~ s e d ~ ~ * ~ ~ ~ in order to simplify the system. However, few, if any, commercially available systems continue to use a balloon as a collection device; the use of sodium tetrahydroborate(II1) and silica detection tubes has almost eliminated the need to collect large amounts of hydride. A further drawback of balloons is the frequency of replacement; Chu et aL9 mention 15-20 and Manning7 50 determinations per balloon. This was an improvement over the previous device. Other Collection Devices tainers.450-ml plastic jar. As an alternative to balloons several workers have advocated rigid or semi-rigid con- Aggett and Aspe112' used a 250-ml glass bottle and Pollock and \Vest12 employed a Direct Introduction systems now allow direct introduction of the hydride to the atomiser. reported systems that do not require collection of the hydride.5~10~14~16~2*-~ I t is not always necessary to collect the liberated hydride, and most commercially available Several workers have1142 GODDEN AND THOMERSOK : GENERATION OF COVALENT Aqzalyst, Vol. 105 Argon by-pass Auxiliary view apparatus rotated 90" Fig. 3. System devised by Fernandez and IhIanning. Automated Systems Several workers have reported the design and application of automated and semi-automated system~.~1-3~ The systems reported to date fall into two categories; the first uses a peristaltic pump and the second a pressurised reagent pumping system.The peristaltic pump system is the more automated of the two in that it requires less manipulation by the operator. However, the pressurised reagent pumping system appears to allow greater control over the reaction conditions. A typical peristaltic system33 employs a multi-channel pump to transfer sample and reagent, in suitable proportions [e.g., 7 . 5 ml of hydrochloric acid and 7 . 5 ml of sodium tetrahydroborate(II1) per 50 ml of sample]. The hydrochloric aci.d is mixed with the sample via a mid-feed coil. This acidified stream is allowed to react with a stream of sodium tetra- hydroborate(II1) solution in an expanded T-connector.The m.ixture of gaseous hydride and liquid then flows to a gas separator, where the gaseous portion is purged from the liquid phase with argon. This mixture of gaseous hydride and argon is intsoduced into the burner chamber via the supplementary oxidant inlet. The semi-automated system described by Fiorino et aZ.38 consists of a single-stage gas regulator with a single nitrogen inlet and two outlets. One outlet is connected to a pressurised reagent bottle containing sodium tetrahydroborate(II1) solution. The other outlet is connected to a pressurised reagent bottle containing distilled water. The outlet from each bottle passes through separate solenoid valves, which are controlled by a single four-cam motor-driven timer (90 s ) . The reagents are connected to PTFE metering valves contained within the reaction cell head.Prior to the actual determination the duration and flow-rates of the sodium tetrahydro- borate(II1) solution and water wash, and the duration of the rest -period, are adjusted to ensure complete reaction of the analyte and to maximise the absorption signal. An argon - hydrogen flame is employed.December, 1980 HYDRIDES IN ATOMIC-ABSORPTION SPECTROSCOPY 1143 An aliquot of the sample is transferred to the 8-in test-tube reaction cell and diluted to 20 ml with acidic diluent. When necessary sodium iodide solution is added (0.5 ml of 10% m/V solution) to reduce As and Sb to the trivalent state. The test-tube is then con- nected to the cell head and the timer is actuated. Gaseous reaction products from the generator are delivered to the flame via a glass adaptor, which replaces the nebuliser.Methods of Atomisation Flames and Tubes Air - acetylene jiame Holak2 used a conventional air - acetylene flame supported on a three-slot Boling burner, Although the air - acetylene flame was used by many workers in the early stage of develop- ment of the technique it was soon superseded by the argon - hydrogen - entrained air flame (usually called the argon - hydrogen flame). Argon - hydrogen flame The use of this type of flame for the determination of arsenic in aqueous solution was first described by Kahn and Scha1lis.l They discovered that it absorbed about 15% of the light output from the hollow-cathode lamp, a t 193.7nm, compared with 62% absorbed by the air - acetylene flame.This resulted in a much better signal to noise ratio. The cooler flame also gives much better sensitivity, but it is more prone to interferences. Dalton and Malanoskis introduced arsine directly into the argon - hydrogen flame, and Fernandez and Manning6 collected arsine in a balloon before releasing the arsine into an argon - hydrogen flame. This type of flame has remained one of the most popular detection systems, but there is a trend away from this flame and towards heated silica tubes. Silica tubes Electrically heated tubes. Chu et aL9 first reported the use of an electrically heated tube as an alternative to a flame. The liberated hydride was swept into the tube by a stream of argon carrier gas and, as no flame was employed, the background absorption was reduced.The longer residence time of the arsenic atoms in the tube, compared with the flame, produced an improvement in sensitivity. Vijan and Wood28 have described the use of an electrically heated glass cell, which was used in conjunction with an automated system. Several workers3*J7Jg have published papers based on the use of this type of detector, and Thompson and Thoresby40 reported its use in the determination of arsenic in soil and plant materials. A novel adaptation of this form of atomiser was described by Robinson and ~ o - w o r k e r s . ~ ~ ~ ~ They used spectrographically pure rods 10-20 mm in length heated with a 5000-W radio- frequency generator and a 10-turn copper coil. The quartz tube was heated by 700-W Nichrome resistance heaters.One of the advantages of this form of atomiser is the ability to control its temperature, and achieve an optimum temperature for each element. Flame-heated Tubes A novel hydride generation atomic-absorption technique for the deter- mination of selenium has been described by Siemer and Hagemann.16 Sodium tetrahydro- borate(II1) was used as a reductant and hydrogen was used to carry the released hydride to the silica tube. A small amount of oxygen was added to the hydrogen to support combustion and atomisation of the hydride. The first reported use of a flame-heated silica tube was by Thompson and Thomersonl4 in 1974. The liberated hydride was carried directly to a 17-cm silica tube mounted above an air - acetylene flame. The advantages claimed for the technique were that no collection vessel was required, flame background was virtually eliminated and better sensitivity was achieved than with the argon - hydrogen flame.One sample can be analysed every 40s when this apparatus is used. The paper describes an apparatus that is still commercially available, Intental -flame. Exterrtaljiame.1144 GODDEN AND THOMERSOX : GEEERATION OF COVALEST A?zalyst, VOl. 105 Similar systems have been described by various w 0 r k e 1 - s . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ Fig. 4 depicts the silica atomising tube used by Thompson and Tliomerson.14 The generated hydride, contained in a stream of nitrogen, is introduced through the side-arm (A) in the middle of the silica tube. Provision is made for cooling the outside of this side-arm via annulus B.An auxiliary nitrogen stream is injected into the transverse jets C and C’, which are designed to prevent the liberated hydrogen from igniting on the ends of the tube, but should instead burn on the ends of the jets L) and D’. A schematic diagram of the apparatus used in this method14 is shown in Fig. 5 . Graphite tube furnaces. JicDaniel et a1.22 introduced the liberated hydride into a proprietary graphite tube furnace. This very comprehensive paper compares various sample pre- concentration techniques, including liydride generation, for concentrating selenium for its determination by heated-graphite atomiser. I t is interesting that the authors conclude that published procedures (at least up to that date) may liberate as little as 1004 of the total inorganic selenium from the solution. Other workers37~*~ have reported the results of comparison studies between hydride genera- tion techniques and heated graphite furnace determinations.’‘I’rr-Fig. 4. Typical design A of flame-heated silica tube. Various detectors are worth mentioning. Molecztlar emission cavity analysis (MECA ) Belcher et aL45 have described this technique, and subsequently how to eliminate inter- ferences,46 for the determination of arsenic and antimony. In MECA the sample is placed within a small cavity a t the end of a rod, which is then positioned in a hydrogen - nitrogen flame so that the flame gases flow almost vertically past the entrance of the cavity. The sample in the cavity is vaporised and emits band emissions characteristic of the elements in the sample.Arsenic and antimony are reduced by sodium tetrahydroborate(II1) to arsine and stibine, respectively, and a stream of nitrogen is used to sweep the gases into the JIECA cavity, which is placed in a hydrogen - nitrogen diffusion flame. Atomic fluorescence Tsuju and Kuga4’ described a non-dispersive system, employing a solar-blind photo- multiplier, for the atomic-fluorescence determination of arsenic. Liberated arsine was carried to an argon - hydrogen flame by a stream of argon. The same workers subsequently reported a modification of this apparatus4* for the determination of arsenic and antimony. They employed a modification of the reaction cell described by Thompson and Thomerson,14 and used sodium tetrahydroborate(II1) to liberate the hydrides.An argon - hydrogen flame was employed, but instead of storing the liberated hydrides they were carried directly to the flame by a stream of argon. Although not strictly within the scope of this review, several non-atomic-absorptionDecember, 1980 HYDRIDES I N ATOhIIC-ABSORPTIOS SPECTROSCOPY 1145 - N2 in r- N2 in & NaBH4 solution Fig. 5 . Schematic diagram of apparatus used in flame-heated silica tube method. Kobayashi et al.4g reported the use of a non-dispersive system for the atomic-fluorescence determination of bismuth. Sodium tetrahydroborate(II1) was used to generate the gaseous hydride of bismuth, which was then carried to an argon - hydrogen flame by a stream of argon. The authors reported a detection limit of 5 pg ml-l or 0.1 ng of bismuth, but the reagent blank in a 20-ml sample was approximately 2 ng of bismuth.Analytical working graphs were drawn using both peak-height and peak-area measurements. The calibrations were found to be linear over a range of about four orders of magnitude from the detection limit. The presence of silver, gold, nickel, palladium, platinum, selenium and tellurium in a 1000-fold ratio to the bismuth caused pronounced depression of the signal. PIIercury and tin slightly enhanced the atomic-fluorescence signal. Nakahara et al.jo have also described a non-dispersive system for the determination of antimony, with particular reference to waste waters and lead. The system is identical with that described by Kobayashi et An improvement in sensitivity of at least one order of magnitude is claimed for non-dispersive atomic fluorescence compared with that of a dispersive system at the 217.6-, 231.1- and 259.8-nm antimony lines.The report includes a comprehensive comparative study of the zinc and sodium tetrahydroborate(II1) reduction systems, and of the interference effects of acids and diverse elements on the fluorescence signal of antimony. As a follow-up to this report, the same workersjl described the applica- tion of the same apparatus to the determination of antimony in phosphoric acid. has advocated a dispersive system whereby the hydrides of antimony, arsenic, selenium and tellurium were swept by a stream of argon into an argon - hydrogen flame supported on a 10-mm i.d. Pyrex tube. The top of this tube was mounted 2.5 cm below the bottom of a microwave cavity window.The microwave cavity was positioned so that the axis of the microwave lamp was 5 cni from the axis of the flame, which was positioned on the optical axis. The hydride generation system described by Thompson and Thomerson14 was used to produce the hydrides. Plasmas The idea of coupling a hydride generator to a plasma arc was first proposed by Lichte and S k ~ g e r b o e . ~ ~ This technique makes it possible to determine simultaneously several elements in a single solution.1146 GODDEN APiD THOMERSON GEKERATION OF COVSLENT Afzalyst, viol. 105 Thompson et a1.54J5 subsequently published a comprehensive study which showed that many of the operating parameters were critical to set up, and many interference effects were encountered.They generated the hydrides of arsenic, antimony, bismuth, selenium and tellurium by means of sodium tetrahydroborate(II1) and detected them simultaneously in an inductively coupled plasma by measuring the atomic line emission intensities. Many of the more severe interferences were overcome by separating the matrix from the elements to be determined, the matrix being precipitated with lanthanum hydroxide. A comprehensive study of the interference effects of a large number of elements was tabulated, as were the percentage recoveries after separation with lanthanum hydroxide. ’This technique was then applied to the analysis of geological samples. Other papers have described methods using plasma sources for the analysis of waste and sea waters,23 whole blood, flour and orchard leaves.56 D.c.a x discharge Braman et al.1° reported a system whereby arsenic and antimony compounds were con- verted into arsine and stibine, respectively, in 1% sodium tetrahydroborate(II1) solution. Arsine and stibine are swept out of the solution using helium carrier gas, passed through a calcium sulphate drying tube, then through a d.c. discharge detector. hlodifications of this system have been r e p ~ r t e d . ~ ~ ~ ~ ~ Light Sources Perhaps the most popular light source is the hollow-cathode lamp, and certainly it is commonly used where the light beam is required to pass through a (silica tube rather than a flame. Because of the relatively low output of hollow-cathode lamps in the ultraviolet region, they tend to give better signal to noise ratios with single-beam instruments with “straight through” optics.Fernandez13 reported the determination of arsenic and selenium using electrodeless dis- charge lamps (EDLs) as alternatives to hollow-cathode sources, and obtained better detection limits and sensitivities. The system used was sodium tetrahydroborate(II1) - acid with balloon collection and an argon - hydrogen flame. A deuterium EDL was used as a back- ground corrector. McDaniel ct aZ.22 reported the use of a special design of demountable hollow-cathode lamp, and claimed an energy output ten times that of conventional hollow-cathode lamps. Robinson et nl.42 used a commercially available demountable hollow-cathode lamp. Obviously the plasma and MECA systems require no light source; the fluorescence tech- niques use EDLs but not as atomic-absorption sources.Interferences The interferences associated with this technique are many and various and may be con- veniently broken down into five sections. Background Absorption This is a problem that occurs more often when a collection device is used in conjunction with the argon - hydrogen flame. When the collected hydride is released to the flame, the flame transparency decreases, and therefore a non-specific absorption signal is obtained. I t is necessary to subtract this signal from the total signal, by either continuous or sequential background correction. Fernandez and Manning6 first reported the use of a deuterium background correction system, described by K a h r ~ , ~ ~ applied to this technique.The reduc- tion in light throughput caused by automatic background correction systems has led to the common use of electrodeless discharge lamps with this type of system. Thompson and Thomerson14 reported that their system was free from this type of inter- ference unless liberated hydrogen was allowed to ignite in the opti’cal path of the spectro- photometer. The increasing use of tube devices and the specificity of the method have reduced the need for background correction systems.December, 1980 HYDRIDES IK ATOMIC-ABSORPTION SPECTROSCOPY 1147 Valency Several worker^^^^^^^ have reported that arsenic must be present in the trivalent form for the evolution of arsenic, whereas others10~1j~27 reported that arsenic and antimony may be present in the trivalent or pentavalent form.Thompson and Thomerson14 reported that by increasing the borohydride concentration from 1 to 4% rn/V the arsenic(V) signal was 90% of that obtained for arsenic(II1). However, they found that solutions of selenium(V1) and tellurium(V1) gave a negligible response compared with equal amounts of selenium(1V) and tellurium(1V). The selenium(V1) and tellurium(V1) solutions were simmered with an equal volume of aqua regia (3 + 1 V/V hydrochloric acid - nitric acid) for 15 min, allowed to cool and then diluted to twice the volume of sample originally taken. I t was then possible to add 1 ml of the resulting solution directly to the generator cell containing 2 ml of 2y0 m/V sodium tetrahydroborate(II1) solution. Under these conditions tellurium(V1) was reduced quantitatively to the tellurium(1V) state, but less than 50% of the selenium(T'1) was reduced.For solid samples they recommend dissolution in aqua regia followed by suitable dilution with distilled water. This technique should ensure that the resulting solution contains arsenic(V), selenium(1V) and tellurium(1V). It was assumed that all dilute tin solutions (concentrations of less than 0.1 pg ml-1) would be in the tin(1V) oxidation state. Fernandez and >fanning6 state that during wet digestion of samples the arsenic must be maintained in the quinquevalent form at all times to avoid losses. Aggett and Aspe1127 have produced a method that allows the selective determination of arsenic(II1) and arsenic(V) by carefully adjusting the pH of the solution before addition of sodium tetrahydroborate(II1). Acid Concentration The relationship between hydrochloric acid concentration [using sodium tetrahydro- borate(II1) as reducing agent] and sensitivity was studied by Fernandez.13 He reported that signal size, with increasing acid concentration, remained almost constant for As, Bi, Sb and Te.The optimum acid strength for Ge is 1-3 M, for Sn 0.2 JI and for Se 6-7 M. Thompson and Thomerson14 reported similar results except that a slight increase in signal size for tellurium was noted over the range 1-4 nr. For lead the sodium tetrahydroborate(II1) and acid concentrations were critical, and the optimum concentration was found to be 0.2 M. Miyazaki et al.23 confirmed the data for arsenic and selenium. By maintaining the solution a t pH 4-5, Aggett and Aspel12' reported that it was possible to determine arsenic(II1) and arsenic(V) selectively.Total arsenic was determined by increasing the acidity of the solution to 5 nr hydrochloric acid. This tends to confirm the earlier work of Knudson and Christian.15160 A method is given. A similar method has been described by Nakashima.jg Organic Binding Fernandez and Mannings reported that if the sample to be analysed contained organically bound arsenic, prior oxidation is necessary and a method reported by Howitzsl is cited. Braman and Foreback*j reported that methylarsenic acid, CH,AsO(OH),, and dimethyl- arsenic acid, (CH,),+sO(OH), are reduced to methylarsine, CH,AsH,, and dimethylarsine, (CH,),AsH, respectively, by sodium tetrahydroborate(II1) a t pH 1-2.The liberated hydrides were collected in a U-tube, which was cooled by liquid nitrogen. After the liquid nitrogen had been removed, the cold trap was warmed gently and the arsines were volatilised from the trap in order of their boiling-points: arsine (-55 "C), methylarsine (2 "C) and dimethylarsine (55 "C). The possibility of organic binding within the sample should always be borne in mind when environmental and biological samples are to be analysed, and the reader is referred to the section Applica- tions. Improvements to this technique have been published s u b ~ e q u e n t l y . ~ ~ ~ ~ ~ Inter-element Effects Perhaps the most important and widespread difficulty occurs with inter-element effects, which suppress, or enhance, the true signal.Braman et a1.10 mention that copper and silver interfere in the generation of arsine and stibine when these hydrides are generated from alkaline solution using sodium tetrahydroborate(II1). They also state that iron interferes1148 GODDEN AND THOMERSON : GESERATION OF COVALENT -4??,dySt, VOd. 105 with antimony but not arsenic, and that aluminium, cadmium, chromium, mercury, manganese, nickel, lead and zinc do not interfere with either antimony or arsenic. Dalton and hlalanoskij determined arsenic using a zinc - hydrochloric acid reduction after prior treatment with tin(I1) chloride - potassium iodide. They reported interference by nitric acid and antimony. Landsford et found that mercury, arsenic and nitric acid inter- fered in the determination of selenium. They used a zinc - hydrochloric acid reduction.Roulet et a1.G3 reported that arsine and stibine were decomposed by gold complexes. A major contribution to the documentation of interferences has been made by Smitli,GA who undertook a general study of the effects of 48 elements on the determination of arsenic, bismuth, germanium, antimony, selenium, tin and tellurium. This work was carried out using sodium tetrahydroborate(II1) and an argon - hydrogen flame. Certain trends in the interference effects were observed. For example, there was no interference on the six inetals studied from the alkali metals, alkaline earths, boron, aluminium, gallium, titanium, zirconium, hafnium, mercury, lanthanum, manganese, vanadium and yttrium. However, there was always some interference on the six metals from copper, silver, gold, platinum, rhodium, ruthenium, nickel and cobalt.I t was also found that nearly all of the volatile hydrides formed from the elements interfered with all the other volatile hydrides. hlaruta and Sudohs5 used various reducing agents for the gencration of arsine, and studied the interferences of various anions and cations. B6dard and KerbysonG6 reported the determination of bismuth in copper by precipitation of the bismuth with lanthanum hydroxide and subsequent conversion of the bismuth into its hydride with sodium tetrahydroborate(II1). The same authorsG7 have reported a separa- tion technique for As, Se, Sn and 're in the presence of copper. Kirkbright and TaddiaG8 have described a novel method of masking the interference of large amounts of Cu, Ni, Pt and Pd on the determination of arsenic in the argon - hydrogen flame.They used thiosemicarbazide and l,l0-phenanthroline as masking agents, and effectively eliminated the strong suppression of the arsenic signal caused by a 1000-fold amount of nickel and 100-fold amounts of platinum and palla.diuim. The effect of hydro- chloric acid concentrations between 1 and 4.5 11 was also studied. Pierce and Brown37~6g undertook a study to compare the interfering effects of several anions, cations and acids on arsenic and selenium. Three sample preparation - sample atomisation techniques were employed : manual liydride generation with argon - hydrogen flame ; automated hydride generation with heated quartz tube ; and ,graphite furnace atomisa- tion.The study showed significant differences in interfering effects between the three techniques. are the mo'st u.sefu1 reference sources of interferences for workers wishing to obtain background information before embarking on a new analysis. The papers by SmithG4 and Pierce and Detector Temperature Chu et aZ.$ have shown that the temperature of the detector bas a significant effect on the size of the signal produced by arsenic. The optimum temperature of their electrically heated tube appears to be in excess of 700 "C. I t is reasonable to assume that the detector temperature will have an effect on at least some of the other hydride-forming elements. However, it is obviously not possible to control the temperature of flames and flame-heated tubes over a wide range, and the innate temperature of these devices must be accepted.Nevertheless, it is important that they are allowed to reach a constant temperature before introducing the sample, and an optimum fuel to oxidant gas ratio may be found by experi- ment. The conclusion to be drawn from the above is that the generation of covalent hydrides, with subsequent determination by atomic-absorption spectroscopy, is subject to many interferences. These interferences may depend upon the method of reduction and the type of detector employed (e.g., flame or tube). No determination :;hou.ld be carried out without checking that the sample under analysis does not contain a potential interferent and, if it does, that interference must be determined and allowed for (e.g., by the method of standard additions).Dccenaber, 1980 HYDRIDES I 9 ATOMIC-ABSORPTIOK SPECTROSCOPY Detection Limits and Sensitivities 1149 Typical detection limits and sensitivities for the different methods are given in Tables 1-11.TABLE I TYPICAL DETECTIOS LIMITS FOR DIFFEREKT METHODS Values are given in micrograms Method* i Element A As . . . . 0.0012 Bi . . . . 0,0032 - Ge . . . . Pb . . . . Sb . . . . 0.0041 Se . . . . 0,0017 __ - Sn . . . . Te . . . . 0.0044 * See Table 11. B 0.01 0.005 0.2 0.005 0.005 0.004 0.005 - C D 0.003 0.0008 0.008 0.0002 0.5 0.1 - - - 0.0005 0.003 0.001 8 -. 0.0005 0.016 0.001 5 E 0.0001 - - 0.000 1 0.00006 0.00008 - -7 F 0.002 0.002 - - 0.005 0.002 - TABLE I1 KEY TO TABLE I Reducing Collection Light Column agent vessel Detector source* Reference h .. . . TiC1, - LIg Balloon Ar - H, flame HCL 12 B . . . . XaBH, Balloon Xr - I€, flame HCL 13 C . . . . . . NaBH, Balloon Xr - H, flame EDL 13 n . . . . KaBH, Kone Flame-heated HCL 14 silica tube Et . . . . KaBH, Sene Ar - H, flame EDL 52 F . . . . KaBH, Kone Electrically heated EDL Manufacturer's silica tube literature * HCL = hollow-cathode lamp; EDL = electrodeless discharge lamp t Atomic-fluorescence method. Applications Waters and Effluents As usual in atomic-ahsorption spectroscopy the analysis of waters and effluents is one of the most straightforward applications because the sample is already in a liquid form and matrix interferences are not usually severe. Schmidt and Royerll used the sodium tetrahydroborate(II1) reduction method on all types of water, from drinking water to heavily polluted water, including raw sewage, to determine arsenic and selenium.ilnalyses were compared of spiked samples (i) after digestion with sulphuric acid and nitric acid and (ii) untreated. The untreated samples gave a slightly larger response for both arsenic and selenium, and a blank for the arsenic of one third of the value for the digested sample. The blanks in all determinations appear to be very large (when compared with more recent rnethoils) and no attempt was made to explain or to rectify this. The usual cause of high blanks is impurities in the sodium tetrahydroborate(II1) and this has been found to vary between chemical manufacturers. Shaikh and Tallman7" found that a large portion of the arsenic in potable waters was organically bound because results for samples digested in sulphuric and nitric acids were higher than for untreated samples.They used a 50-ml sample volume and sodium tetra- liydroborate(II1) to reduce the arsenic to arsine, which was then trapped in a small volume of a chloroform - ephedrine solution of silver diethyldithiocarbamate. The arsenic content was then determined by graphite furnace atomic-absorption spectroscopy.1150 GODDEN AND THOMERSON : GEKERATIOS OF COVALEST Analyst, Vol. 105 TABLE I11 TYPICAL SENSITIVITIES FOR DIFFERENT XETHODS Values are given in micrograms Method* Element As . . . . Bi . . . . Ge . . . . Pb . . . . Sb . . . . Se . . . . Sn . . . . Te . . . . * See Table IV. A 0.006 0.015 - - 0.02 0.008 0.021 - B C 0.01 0.005 0.008 - 0.27 - 0.01 - 0.011 0.009 0.007 - 0.014 - - - D E- 0.0005 0.005 0.00004 - 1.0 - 0.08 - 0.000 6 - 0.002 0.001 5 0.0004 - 0.002 - TABLE IV KEY TO TABLE I11 Reducing* Collection Light Column agent vessel Detector !iourcet Reference A .. . . TiCI, - Mg Balloon Ar - H, flame HCL 12 B . . . . NaBH, Balloon Ar - H, flame HCL 13 C . . . . NaBH, Balloon Ar - H, flame EDL 13 D . . . . NaBH, None Flame-heated HCL 14 E: . . . . NaBH, None Electrically heated HCL 38 silica tube silica tube * Acid must be added in all instances. t HCL = hollow-cathode lamp: EDL = electrodeless discharge lamp. Automated system. An automated system has been describedz0 for the simultaneous determination of arsenic and selenium, based on reduction of the sample with tin(I1) chloride, potassium iodide and aluminium.Spiked samples gave recoveries of 98-101% for levels of arsenic between 2 and 10 pg 1-1 and levels of selenium between 0.5 and 5.0 pg 1-l. This method was found to be unsuccessful by Pierce et L z Z . , ~ ~ who replaced the aluminium reduction method with sodium tetrahydroborate(II1). The determination of selenium in natural waters using a tin(I1) chloride reduction system was subject to severe interference from arsenic and mercury, which were overcome only by using a tedious digestion process.62 Rlost methods now favour the sodium tetrahydro- borate(II1) reduction process, which suffers from fewer interferences. Recent rneth~ds’l-~~ have produced good results, mainly due to careful study of the possible interference effects and calibration by the method of standard additions.The determination of selenium(1V) in sea water was found to be straightforward by Thompson52 when using a fluorescence detection method. Samples were acidified with hydrochloric acid before reduction with sodium tetrahydroborate(II1). X negligible “back- ground” signal was reported when checked a t the 216.5 -nm selenium non-resonance line and the concentration was reported as less than 0.000 1 pg ml-l. The hydrides were decomposed in an electrically heated silica tube. Metallurgical Probably the most severe interferences encountered for any application in hydride genera- tion occur with metallurgical samples. For example, the interference of copper on selenium is very severe, almost completely suppressing the selenium absorption signal unless the copper is first removed.December, 1980 HYDRIDES Ih' ATOMIC-ABSORPTION SPECTROSCOPY 1151 The determination of &As, Bi, Se, Sn and Te in copper alloys has been achieved after precipitation of the bismuth with lanthanum h y d r o ~ i d e .~ ~ ? ~ ' A 10% wzjV solution of the sample, dissolved in nitric acid, is treated with lanthanum nitrate and ammonia solution, filtered and re-dissolved in nitric acid. The hydrides are then generated in the usual way. In this way the serious interference from copper is removed, and results for bismuth obtained by this method agree well with those from spectrographic methods. A less severe interference is caused by nickel on the absorption of bismuth and arsenic, where a 50% reduction in sensitivity occurs with a 1% mjV sample matrix.Nevertheless, the sensitivity for bismuth and arsenic is still sufficient to enable them to be determined by the method of standard additions in some alloys.74 However, the addition of EDTAi5 to the sample solution completely overcomes the matrix interference for the determination of bismuth in nickel alloys at low concentrations. The alloy was dissolved in hydrochloric and nitric acids, and an addition of 0.27 M EDTA solution was made. The EDTA solution had previously been adjusted to pH 9 with 30% m/V sodium hydroxide solution to prevent loss of hydrogen when the EDTA solution was mixed with the sodium tetrahydroborate(II1) reagent. There is negligible interference by iron or lead (using a 1% m/V matrix) on the absorption of arsenic, selenium and tin.Therefore, the determination of these metals in iron and mild steels can be carried out after dissolving the sample in a mixture of hydrochloric and nitric acids. Calibration is carried out using acidified aqueous standards of the metals.76 Table V shows some results for arsenic in steels using this method. The determination of arsenic, selenium and tin in a complex lead alloy is more difficult because of the problems associated with maintaining a stable solution of lead or tin. In this instance a method based on the use of fluoboric acid and hydrogen peroxide is the most suitable.77 A thorough evaluation of the interference effects of most of the elements present in steels on the absorption of antimony, arsenic, bismuth, selenium, tellurium, tin and lead showed TABLE V DETERMINATION OF ARSENIC I N STEELS BY HYDRIDE GENERATION -4ND ATOMIC-ABSORPTIOK SPECTROSCOPY Certified BCS No.Result, yo value, yo Range, yo 260/4 . . . . 0.010, 0.009 0.009 0.008-0.011 149/2 . . . . 0.0005, 0.0006 <0.001 - 324 . . . . . . 0.080, 0.085 0.084 0.079-0.085 321 . . . . . . 0.003, 0.003 0.003 0.002-0.003 218/3 . . . . 0.038, 0.035 0.0035 0.032-0.038 that many interferences occurred. However, by using a standard additions method for calibration good results could be obtained for most of the elements in a range of steels.7s Of particular interest was the ability to make use of the plumbane generated after reduction with sodium tetrahydroborate(II1). Plumbane had previously been detected using this technique14 but the sensitivity was insufficient to be of any practical value.The addition of potassium dichromate in a tartaric acid medium gives a useful enhancement of the lead absorption signal. This allowed lead in seven steels to be determined, and the reported values showed very good agreement with certified values. Arsenic and antimony can be determined in zinc powder by utilising the reducing power of the sample it~elf.~S Hydrochloric acid a t 80 "C is added to 1 g of the powdered sample in the reaction flask and the evolved hydrides are measured in the usual way. Foods The hydride generation technique has found wide application in the determination of toxic metals in food and drink products. The analysis of wine, sherry and beer samples presents no serious analytical problems.However, acidification and degassing of the samples must be carried out before the analysis. Water-soluble foods such as sugar, coffee and some organic food colourings can also be analysed directly once a suitable solution has been prepared.1152 GODDEN AND THOMERSON : GENERATION OF COVALENT Analyst, VoZ, 103 For other food samples a digestion process is usually necessary before reduction to the hydride, and most of the published work deals with these methods. Typically 2-5 g of food sample is digested in a conical flask with 5 in1 of concentrated sulpliuric acid and 500//, m/V hydrogen peroxide on a hot-plate until a clear digest is obtained.80 Alternatively, the sample can be treated in a Kjeldahl flask with a 4 + 1 + 1 mixture of concentrated nitric, sulphuric and perchloric acids.38 Reduction of the metal with sodium tetrahydroborate(II1) then gives excellent recoveries for both methods after calibration with standards that have been treated in the same way as the samples.The choice of acid digestion appears to be due to the particular authors’ preference rather than the type of foodstuff to be analysed. As many laboratories analysing foods often determine mercury in addition to some of the hydride-forming metals, one of the digestion methods developed was suitable for retaining both mercury and selenium in solution.s1 The digestion mixture used was 1 + 1 con- centrated nitric and sulphuric acids plus 0.1% nz/V of vanaclium(V) oxide added as a catalyst. The digested sample was divided into two parts ; the portion for the selenium determination was acidified with hydrochloric acid and then reduced with sodium tetrahydroborate(II1) in the usual way.Recoveries for selenium determined by this method were 99.5-101 .0::;. The aliquot retained for the mercury determination was treated with potassium permanganate before reduction to elemental mercury by means of hydroxylammonium chloride and tin( 11) chloride. Evans et aL.43 produced an extremely thorough paper on the determination of antimony, arsenic and tin in foodstuffs using the silica tube detection technique. Detailed information is given on both dry- and wet-ashing procedures, calibration ranges, interferences, recoveries and reproducibilities. As no review could do justice to the amount of information contained in this paper, we recommend that anyone faced with this type of analysis should read it for themselves.Detection was by the cold-vapour method. Clinical hrsenic is not a normal constituent in body tissue and fluids although no real information exists as to its toxic levels. A level of 0.05 pg in-l in human urine is considered “signifi- cant All existing work on the determination of any of the hydride-forming metals in body tissue and fluids agrees that the metal hydrides cannot be generated from the untreated sample because of severe signal depression and poor recoveries. The samples must therefore be wet aslied before reduction. Orheim and Bovees3 favoured a reduction using concentrated nitric and sulphuric acids, followed by addition of hydrogen peroxide, for the determination of selenium in blood, milk, hair, tissue, fat and urine.The method of standard additions was used for calibration, and the analysis of spiked samples gave recoveries of 95-105yo. An alternative digestion method for blood, before the determination of selenium, involves the use of nitric and perchloric acids.g4 Levels of selenium in the range 0.01--0.50 pg g-I were reported with a relative standard deviation of 4.3% and a mean recovery of 100.50/,. Bismuth compounds are currently used in the treatment of indigestion and stomach ulcers. As a result, the determination of bismuth in blood and urine is required. Xatural levels of bismuth in blood are quoted as being between 0.009 and 0.012 pg m-l, but the anticipated level after bismuth therapy would be much higher.Iioo11ey~~ compared the techniques of spectrophotometry and polarograpliy with flame, electrothermal and hydride atomic-absorption methods for blood and urine and found that the hydride generation method, with detection using a heated silica tube, was the most satisfactory. 14 digestion using nitric and perchloric acids was also used in this method and recoveries of bismuth from both blood and urine were good. Agricultural Because of the widespread occurrence of arsenic in minerals and pesticides, its dispersal in the environment requires monitoring in soils and plants. hlost current methods deal only with arsenic and selenium, although there seems no reason why the methods could not be extended to the other hydride-forming metals if required.Dcccmbev, 1980 HPDRIDES IN ATOMIC-ABSORPTION SPECTROSCOPY 1153 S'ijan and c o - ~ o r k e r s ~ ~ ~ ~ ~ used a mixture of perchloric and nitric acids for soil and vegetation.,4fter reduction with sodium tetrahydroborate(II1) and detection using an electrically heated quartz cell, results and interferences were reported for the determination of arsenic and selenium. Nany of the interferences (by now commonly reported) were confirmed. Results were obtained using the standard additions technique. Recoveries for arsenic using this method were 86.7-126.7%. \Vau~hope~~ compared the results obtained on solutions of samples prepared by two different methods. The first method was simply to soak the samples in hydrochloric acid for 24 11.The second method involved the digestion of the samples in a mixture of nitric and perchloric acids. Thompson and Thoresby40 treated soils by leaching with a mixture of nitric and sulphuric acids, or by fusion with potassium pyrosulphate followed by oxidation to arsenic(T:) with potassium permanganate. Plants were wet oxidised with nitric and sulpliuric acids. Samples of the digests thus produced were reduced with sodium tetrahydroborate(II1). An electrically heated tube was used as a detection device. The results obtained were consistently higher than those obtained with a molybdenum blue spectrophotometric method. However, the authors believed the atomic-absorption method to be the most accurate, as collaborative reports had indicated erratic recoveries with the spectrophotometric method.The determination of both arsenic(II1) and total arsenic in waters and orchard leaves has been r e p ~ r t e d . ~ ' The leaves were digested in nitric acid at 70 "C for 5 min and the mixture was cooled, treated with dilute sulphuric acid and heated at 70 "C for 15 niin. The acidity of the solution was then adjusted with 5 11 hydrochloric acid, from which the total arsenic was determined, or with a citrate buffer at pH 4.5-5.0, for the arsenic(II1) determination. The arsine was then evolved from the particular solution by means of sodium tetrahydro- borate(II1) and atomised in an argon - hydrogen flame. The determination of germanium, arsenic, selenium, tin and antimony in orchard leaves and flour,j6 carried out on acid-digested samples, showed good recoveries and close agreement with certified values.The determination of selenium in feed and forage samples was carried out by drying the samples at 200-300 "C for 30 min followed by dissolution in concentrated hydrochloric acid.16 Selenium recoveries averaged 101 yo with a relative standard deviation of 5.0% using this method. I t was reported that attempts at wet ashing with mixtures of nitric and sulphuric acids and sulphuric acid and hydrogen peroxide resulted in a loss of selenium. The sensitivities for antimony, arsenic, selenium and tellurium by hydride generation, with detection by atomic absorption, are still inadequate for some samples. ThompsonjZ linked the hydride generation technique with atomic fluorescence, and showed that not only were calibra- tion graphs more linear, but that sensitivities were improved by factors of between 5 and 30.This method was also applied to the determination of selenium in animal feedstuff and sea water. This technique was then applied to the determination of selenium in soil digests obtained by an overnight treatment with concentrated nitric acid and followed by the addition of percliloric acid and heati11g.8~ Several sample digestion methods have been proposed. Both methods gave similar results. The detection limit of this method is 10 ng of selenium. Geological Rocks and minerals are complex samples chemically, and dissolution methods are often tedious and time consuming. The improved sensitivity of the hydride generation method for arsenic, selenium and tellurium has found applications with these samples.Arsenic is used to indicate the presence of some precious metals in rocks and minerals, and several methods for the determination of arsenic have been proposed. As flame and electro- thermal atomisation methods for these metals suffer from several interferences, the hydride generation method is favoured by most worker^.^^,^^,^^ Sample digestion procedures vary, but one of the most popular is treatment with concentrated hydrochloric and nitric acids, followed by reduction with sodium tetrahydroborate(II1). Glass samples were dissolved in a mixture of hydrochloric, hydrofluoric and nitric acids in a sealed polyethylene bottle at 70 "C. After cooling, boric acid was added and hydrogen selenide generated with sodium tetrahydroborate(II1) .87 The results showed good agreement with standard values, and the detection limit was 0.030 pg of selenium.1154 GODDEN AND THOMERSON : GEIiERATIOIi OF COVALENT Analyst, VOl.105 Tellurium, at concentrations down to 5 ng g-l, has been determined in silicate rocks by treating the sample with hydrofluoric acid and evaporating the resulting solution to dryness. The residue was dissolved in a mixture of hydrochloric, nitric and perchloric a ~ i d s . 1 ~ The samples were reduced with sodium tetrahydroborate(II1) and atomised in an electrically heated tube. Calibration was effected by the method of standard additions. Tobacco A 4-g amount of tobacco was treated with nitric acid before refluxing with a mixture of nitric and perchloric acids.88 Detection was by reduction with potassium iodide and tin(I1) chloride, followed by collection of the arsine (as a white solid) in a U-tube immersed in liquid nitrogen.\\'hen the reduction was complete the liquid nitrogen was removed and the evolved arsine measured by atomic-absorption spectroscopy. A method for the determination of antimony, arsenic, bismuth, selenium and tellurium in tobacco and cigarette smoke was proposed by S p i n ~ e r . ~ ~ ~ ~ ~ Samples are digested in nitric acid and diluted with EDTA, followed by reduction with sodium tetrahydroborate(II1) and detection using the silica tube method.14 The maximum levels observed by this method were arsenic 500 ng g-l, antimony, bismuth and tellurium 50ngg-l and selenium 90 ngg-I. The metal content of smoke from 40 cigarettes was trapped a t -80 "C and extracted with nitric acid.Environmental -4ir particulates, after collection on a fibre-glass filter and digestion with a 1 + 2 mixture of nitric and sulphuric acids, were treated with sodium tetrahydroborate(II1) to reduce the arsenic.28 Studies showed that the presence of either selenium, cobalt, nickel or copper at 200 pg ml-l causes severe interferences but as they were seldom present at levels greater than 5 pg ml-1 in the prepared solutions no inter- ferences occurred in practice. Petroleum Products No published work exists in which atomic-absorption detection has been used for hydride- forming metals in petroleum products. However, it is worth recording that arsenic has been determined in stocks and catalysts by generation of the arsenic as a r ~ i n e .~ ~ The pre-treatment method for the stocks involved refluxing samples with sulphuric acid and hydrogen peroxide. The catalysts were fused with sodium peroxide. A spectrophotometric determination was used to measure the arsenic, but there seems to be no reason why atomic-absorption spectro- scopy could not be applied to this method. Future Trends Arsenic has been determined in tobacco after a wet-ashing procedure. Recoveries for actual samples were 99.5-103y0 . The technique appears to have reached a plateau, and no major improvements are likely in the near future. Most commercial manufacturers offer hydride kits as accessories for their atomic-absorption spectrophotometers. As might be expected, there is a wide variety in the designs offered, some employing the argon - hydrogen flame and others flame-heated or electrically heated silica tubes.The simplest design is a reaction cell, which is coupled to the spray chamber of the instrument, to supply liberated hydride to an argon - hydrogen flame. The most sophisticated has an electronic control box which, at the touch of a button, purges the with inert gas, starts a stirrer, initiates the recorder chart drive and adds the reducing agent to the reaction cell. Two possible areas of future development are improved automation and methods of lowering detection limits. These may improve as commercial manufacturers introduce successive designs, but the really major improvements to sensitivity and detection limits may lie in the combination of hydride generation and techniques such as atomic fluorescence, inductively coupled plasmas and d.c.arc discharges. Although these techniques have already been mentioned they are not yet in widespread use. A further area in which much useful work may yet be done is in the reduction or elimina- tion of interference effects. An electrically heated silica tube detector is employed.December, 1980 HYDRIDES IN ATOMIC-ABSORPTION SPECTROSCOPY 1155 The development of an inexpensive specific detector system could be a useful idea for the future. This would release the more expensive atomic-absorption spectrophotometers for other purposes, and eliminate the time wasted in fitting and dismantling the hydride accessory. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45, 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. References Kahn, H. L., and Schallis, J . E., At. Absorpt. Newsl., 1968, 7, 5. Holak, W., Anal. Chem., 1969, 41, 1712. Madsen, R. E., At. Absorpt. Newsl., 1971, 10, 57. Hundlcy, €-I. K., and Underwood, J . C., J . Assoc. Off. A m l . Chem., 1970, 53, 1176. Dalton, E. F., and Malanoski, A. J., .4t. Absovpt. Xewsl., 1971, 10, 92. Fernandez, F. J . , and Manning, D. C., At. Absovpt. ,Vewsl., 1971, 10, 86. Manning, D. C., A t . Absorpt. Newsl., 1971, 10, 123. Pollock, E. X , , and West, S. J . , A t . Absorpt. Sewsl., 1972, 11, 104. Chu, R. C., Barron, G. I’., and Baumgarner, P.A. \V., Anal. Chem., 1972, 44, 1476. Braman, K. S., Justen, L. L., and Foreback, C. C., Anal. Chem., 1972, 44, 2195. Schmidt, F. J , , and Royer, J . L., Anal. Lett., 1973, 6, 17. Pollock, E. K., and \Vest, S. J . , A t . Absorpt. Newsl., 1973, 12, 6. Fernandez, F. J . , A t . Absovpt. Sewsl., 1973, 12, 93. Thompson, I(. C., and Thomerson, D. R., Analyst, 1974, 99, 595. Iinudson, E. J . , and Christian, G. D., A t . Absorpt. S e z d . , 1974, 13, 74. Siemer, D. D., and Hagemann, L., Anal. Lett., 1975, 8, 323. Greenland, L. P., and Campbell, E . Y., Anal. Chinz. Acta, 1976, 87, 323. Rooney, R. C., Aqzalyst, 1976, 101, 749. Knechtel, J . I < . , and Fraser, J . L., .4+zalyst, 1978, 103, 104. Goulden, P. D., and Brooksbank, P., Anal. Chem., 1974, 46, 1431.Freeman, H. C., and Uthc, J , F., A t . Absorpt. Newsl., 1974, 13, 75. McDaniel, M., Shendrikar, A . D., Iiciszner. K. D., and West, P. \V., Anal. Chem., 1976, 14, 2240. Miyazaki, A., Kimura, A,, and Umezaki, Y . , Aizal. Chim. Acta, 1977, 90, 119. Crecelius, E. A . , Anal. Chem., 1978, 50, 826. Braman, R. S., and Foreback, C. C., Science, 1973, 182, 1247. Kokot, 11. L., A t . Absorpt. Kewsl., 1976, 15, 105. Xggett, J., and Aspell, A. C., Analyst, 1976, 101, 341. \?]an, P. K., and lYood, G. R., A t . Absorpt. h’ewsl., 1974, 13, 33. Collett, D. L., Fleming, D. E., and Taylor, G. A., Analyst, 1978, 103, 1074. Fleming, D. E., and Taylor, G. A,, Analyst, 1978, 103, 101. Hwang, J . Y., Cllucci, P. A,, and hlokeler, C. J . , and Smith, S. B., Am. Lab., 1973, 5, 43.Kan, K. T., Anal. Lett., 1973, 6, 603. Schmidt, F. J., Royer, J . L., and Muir, S. X, Anal. Lett., 1975, 8, 123. Pierce, F. D., Lamorcaux, T. C., BroTvn, H . R., and Frazer, R . S., Appl. Spectrosc.. 1976, 30, 38. Vijan, P. K., Rayner, A. C., Sturgis, D., and Wood, G. R., A i d . Chinz. Acta, 1976, 82, 329. Vijan, P. N., and Wood, G. R., Talanta, 1976, 23, 89. Pierce, F. D., and Brown, H. Ii., Anal. Chem., 1977, 49, 1417. Fiorino, J . A . , Jones, J . W., and Capar, S. G., Anal. Chem., 1976, 48, 120. Wauchope, R. D., A t . Absorpt. h’ewsl., 1976, 15, 64. Thompson, A. J., and Thoresby, P. A,, Analyst, 1977, 102, 9. Robinson, J . U,, Slevin, I-’. J , , Hindman, G. D., and Wolcott, D. K., Anal. Chim. Acta, 1972, 61, 431. Robinson, J . \V., Garcia, I<., Hindman, G., and Slevin, P., Anal. C h i m .Acta, 1974, 69, 203. Evans, W. H., Jackson, F. J . , and Dellar, D., Analyst, 1979, 104, 16. Freeman, H. C., Cthe, J . F., and Fiemming, B., A t . Absorpt. iVewsl., 1976, 15, 49. Belcher, R., Bogdanski, S., Ghonaim, S., and Townshcnd, A , , Anal. C h i m . Acta, 1974, 72, 183. Belcher, R., Bogdanski, S., Hcndcn, E., and Townshend, A,, Analyst, 1975, 100, 522. Tsujii, K., and Kuga, K., Anal. Chim. Acta, 1974, 72, 85. Tsuju, K., and Kuga, I<,, Anal. Chim. Acta, 1978, 97, 51. Icobayashi, S., Sakahara, T., and Jlusha, S., Talanta, 1979, 26, 951. Kakahara, T., Kobayashi, S., and Musha, S., Anal. C h i m . Acta, 1978, 101, 375. Nakahara, T., Kobayashi, S., and Musha, S., Anal. Chem., 1979, 51, 1589. Thompson, K. C., Anal3ist, 1975, 100, 307. Lichte, F. E., and Skogerboe, R. I<,, A m l . Chew., 1972, 44, 1480. Thompson, M., Pahlavanpour, B., lValton, S. J . , and Kirkbright, G. F., Amlyst, 1978, 103, 568. Thompson, M., Pahlavanpour, B., \Valton, S. J , , and Kirkbright, G. F., Analyst, 1978, 103, 705. Robbins, W. 13., Csruso, J . A,, and Fricke, F. L., Analyst, 1979, 104, 35. Braman, R. S., Johnson, D. L., Foreback, C. C., Ammons, J . hI., and Bricker, J . L., Anal. Chem., Kahn, H. L., A t . Absorpt. A’ewsl.. 1968, 7, 40. Nakashima. S., Analyst, 1979, 104, 172. Knudson, E. J . , and Christian, G. D., Anal. Lett., 1973, 6, 1073. Horwitz, W., “Official Methods of Analysis of the Association of Official Analytical Chemists,” Landsford, M., RlcPherson, E. M., and Fishman, M. J , , A t . Absorpt. Newsl., 1974, 13, 103. 1977, 49, 621. Eleventh Edition, Association of Official Analytical Chemists, Washington, D.C., 1965.1156 GODDEX AND THOMERSON 63. 64. 65. 66. 6 i . 68. 69. i 0 . 71. 72. 73. 74. 76. 76. 77. i 8 . i 9 . 80. 81. 82, 83. 84. 85. 86. 87, 88. 89. 90 91 Roulet, li., Ngyuyen Quang Lan, Nason, \Ir. R., and Fenske, G. P., j u n . , Helv. Chim. Acta, 1973, Smith, A. E., Analyst, 1975, 100, 300. Maruta, T., and Sudoh, G., Anal. Chinz. Acta, 1976, 77, 3 i . BBdard, hl., and Kerbyson, J . D., A~zal. Chem., 1975, 47, 1441. EBdard, XI., and Kerbyson, J . D., Can. J . Spectrosc., 1976, 21, 64. Kirkbright, G. F., and Taddia, AT., Anal. Chim. Acta, 1958, 100, 145. Pierce, F. D., and Brown, H. It,, Anal. Chew., l 9 i 6 , 48, 693. Shaikh, A. U., and Tallman, D. E., Anal. Chew., 1977, 49, 1093. Corbin, D. R., and Barnard, \i'. M., A t . Absovpt. ATewsl., 1976, 15, 116. Guimont, J., Pichette, M., and Rheaunie, X, At. .4bsorpt. ,Vcrrs/., 1977, 16, 5 3 . Mesman, 13. B., and Thomas, T. C., A w l . J A t . , 1975, 8, 449. Baird Applications Report, X o , 461, 1977, Uaird-Atomic, Braintree, Essex. Drinkwater, J . E., Analyst, 1976, 101, 672. Baird Applications Report, KO, 460, 19i7, Baird-.\tomic, Hraintrce, Essex. Baird Applications Report, KO. 459, 1977, Uairtl-Atomic, Braintree, Essex. Fleming, H. D., and Ide, K. G., A n d . Ciiznz. .4cta, 1976, 83, 67. Milner, U. A , , Whiteside, P. J . , and Price, LV. J . , .4ria/yst, 1979, 104, 474. Rees, D. I . , J . Assoc. Public Anal., 1978, 16, 71. Egaas, E. J . , and Julsharnn, K., At. Absoupt. A'eu'sl., 1978, 17, 135. Berman, E., Appl. Spectrosc., 1975, 29, 1. Orheim. R. M.. and Bovee. H . H.. Anal. Chew.. 1974. 46. 921 56, 2405. -.... ~~ Clinton, 0. E.,'Analyst, 1977, 102; 187. Azad, J . , Kirkbright, G. F., and Srlook, 11. D., Analyst, 19i9, 104, 232. Rubeska. I., and Hlavinkova, V., .4t. .4bsovpt. Sewsl., 1979, 18, 5 . Hermami, R., A t . Absovpt. A'ewsl., 197i, 16; 44. Griffin, H. R., Hocking, kI. B., and Lowery, D. G.. Anal. C h n ? , , 1!)75, 47, 229, Spincer, D., Paper presented a t 30th Tobacco Chemists Research Conference, Xashville, Tenn., Octobcr 1976. Spincer, D., Proc. Anal. Div. C h e m SOC., 1979, 16, 131. Liederman, D., Bowen, J , E., and Milner, 0. I., Anal. Chew., 1959, 31, 2052. Received March 6th, 1980 Accepted J w z e 30th, 1980

ISSN:0003-2654    

DOI:10.1039/AN9800501137

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

A fzalyst, Decembev, 1980, Vol. 105, p p . 1157--1162 1157 Simultaneous Determination of the Noble Metals in Geological Material by Radiochemical N eu t ro n -act i va t i o n An a I ysi s Susan J. Parry Uuioevsity of Loizdon Reactov Centve, Siltoood Pai'k, Sunningizill, Ascot, Bwkshire, SL5 I P Y A simple procedure is described for determining the noble metals in geological saniples rich in chromium and copper. Powdered rock samples (0.1-0.3 g ) arc irradiated in an epithernial neutron flux, under a cadmium filter, to reduce interference from "Cr and 64Cu. After digestion witli a peroxide fusion, the noble nietals are separated on Srafion NRIliIZ ion-exchange resin. -4ctivity from chromium(VI), retained on the resin, is eliminated by reduction to chroniiuiii(II1) nitti iron(I1) sulphate before separation.The gamma-ray activity is measured ivitli a lithium-drifted germanium detector for palladium (losPd), platinum ( 199Xu), iridiuni (lS4Ir) and gold (1'38Xu). The separation yield (iridium 40y/,, palladium SOYA, gold 90%) is determined for eacli sample by irradiating the resin to activate a stable noble metal carrier. Multi-element noble metal standards (0.1-8 pg) are prepared from dilute solutions (100 pl) evaporated on resin (1 nil). The practical detection liniits are 2 . 5 ng of palladium and platinum, 0.1 ng of iridium and 0.01 ng of gold. The reliability of the procedure was confirmed by analysing sulphide standards (PThl, PTC) and standard rocks (PCC-1, DTS-1). The method has been applied t o the analysis of chrome-spinels.Keywovds : Soble metal determination ; epitizermal neutvon-activation analysis ; ion-exchange sepnvation ; refevence vocks ; chwome-spinel Methods for the determination of the noble metals have been reviewed by Beamish and Van L0on.l The most common method for geological samples is atomic-absorption spectro- metry with a fire-assay pre-concentration step. Emission spectrography and spectrophoto- metry are also useful techniques for ore analysis, where representative samples of up to 50 g are measured, and therefore sensitivity is not a problem. One of the most sensitive methods for the analysis of rock samples for which the supply of material is limited is neutron-activation analysis. Instrumental activation analysis has been applied successfully to the determination of iridium and gold in ores,* but generally it is necessary to separate the noble metals from interfering activity of the rock matrix.There are many examples of radiochemical separation schemes to isolate one or more of the noble metals in the neutron- activation analysis of geological samples. These schemes include solvent extraction, distilla- tion, ion exchange and pre~ipitation.~-~ A relatively rapid method has been described by Kadkarni and Morrison,6 based on the ion-exchange separation of the noble metals as a group. The resin Srafion SMRR is #an anion exchanger that exhibits quantitative retention of the noble metals.' The resin has been used for the separation of gold,*--ll iridi~m,1~,13 silver,14 copper,15 molybdenum16 and mercury.17 It has been shown, in addition, to adsorb vanadate and chromate ions.l8 Nadkarni and Morrison6 found that the noble metals were collected quantitatively on Srafion NhIRR but that the solution had to be passed through a second column for complete collection of iridium. I t has been demonstrated that iridium is not easily removed from the resin once adsorbed13 and preliminary studies in this work confirmed this to be the case for the other noble metals.The high background activity due to the retention of copper on the resin was a problem in ore material and solvent extraction was introduced for its removal prior to the separation step.lB Chromium was not considered to be an interferent in the analysis of silicates and oresIg but this study showed that, following an oxidising fusion, the chromate ions on the resin caused interferences in samples containing high concentrations of chromium, such as chrome-spinels.Previous worked0 have used epithermal irradiations to enhance the activation of the noble metals and reduce background interference from the matrix. In this work a cadmium filter reduced the activity from copper and chromium by1158 PARRY: DETERMINATION OF NOBLE METALS IN GEOLOGICAL Analyst, Vol. 105 a factor of 30. This paper describes a modified procedure of Nadkarni and Xorrison6 to overcome these problems, applied to the determination of platinum, palladium, iridium and gold. After an alkaline peroxide fusion and dissolution in acid, the sample solution is reduced with iron(I1) sulphate to convert chromium(V1) into chromium(III), which is not collected by the resin.As it is difficult to remove the noble metals from the resin, they are counted on it and the yield of the separation is measured for each sample by re-irradiating the resin to activate the carrier. The yields were 90-100°/o of gold and platinum, 80-90% of palladium and 40% of iridium. Experimental Apparatus The samples were irradiated in the University of London reactor at Silwood Park, Ascot. The 30-h irradiations were made in the core-tube (10l2 n cm-2 s-l), in an irradiation can lined with cadmium 1 mm thick (RCd Au = 2.8). The 5-min irradiation to determine the yield was made in a pneumatic irradiation system in the reactor core (2 x lo1' n cm-' s-l). The gamma-ray analyses were performed with a high-resolution lithium-drifted germanium detector (nominal active volume, 42 cm3; resolution, 1.81 key IWHN at 1.33 MeV; peak to Compton, 36.3 a t 1.33 MeV; Princeton-Gamma Tech), and a 4096-channel analyser (Link, Model 290).The lead shielding surrounding the detector was covered with cadmium sheet to avoid interference from lead X-rays. Reagents Analytical-reagent grade materials were used unless specified otherwise. Hydrochloric acid, 12, 1 and 0.05 M. Nitric acid, 16 M. Ammonia solution, 15 M. Sodium peroxide, powder. Sodium hydroxide, pellets. Scandium oxide. Specpure. Iron(I1) sulphate. Ammonium tetrachloropalladate(III). Tetraminineplatinzum(II) diclzloride. Ammonium hexachloroiridatc(II.'). Potassium dicyanoaurate(III) solution. Srafion NMRR ion-exchange resin. Specpure, Johnson Matthey Chemicals Ltd.Specpure, Johnson Matthey Chemicals Ltd. Specpure, Johnson Matthey Chemicals Ltd. Specpure, Johnson Matthey Chemicals Ltd. hyalon LYater Conditioning Co., Haifa, Israel. Noble Metal Solutions Dissolve accurately weighed amounts of the noble metal coinplexes in 1 31 hydrochloric acid to give stock solutions of 1-5 mg m1-l. Dilute the stock solutions with 1 11 hydro- chloric acid to give a carrier solution containing 10 pg ml-L of iridiutn and gold and 500 pg nil-1 of palladium and platinum. Further dilute these solutions 100-fold to give standards containing 0.1 pg of iridium and gold and 5 [~g of palladium and platinum in 100 pl of solution. Ion-exchange Resin Equilibrate the Srafion NIIRR resin in 0.05 11 hydrochloric acid for several hours before packing in columns (120 mm x 7 mm i.d.) (5-ml disposable plastic pipettes) with a glass- Irradiation Prepare the standards for quantitative analysis by evaporating exactly 100-pl aliquots of the standard solution on to 1 ml of Srafion NMRR resin, in a sizc 2 polythene irradiation capsule (15 mm diameter x 10 mm high).b'eigli the powdered rock samples (0.1-0.3 g) into size 1 polythene irradiation capsules (15 mm diameter x 6 mm high). Place four rock samples, two multi-element standards and one reference rock sample in the cadmium-lined wool plug.Deccmbev, 1980 3IATERIAL BY RADIOCHEXIICAL SEUTROK-ACTIVATION ANALYSIS 1159 irradiation can and irradiate it for 4 d in the reactor core-tube. When the samples have been unloaded count the multi-element standards immediately and digest the rocks for separation of the noble metals.Radiochemical Separation Transfer 0.1 ml of carrier solution into a new nickel crucible (capacity 25 ml) and leave the solution to evaporate. Add a few milligrams of solid scandium oxide hold-back carrier and 2 g of sodium hydroxide pellets. Transfer the active rock sample into the crucible and cover the contents with 5 g of sodium peroxide powder. Heat the mixture with an electro- thermal Bunsen burner and keep it a t red heat for 15 min, swirling the mixture occasionally to ensure isotopic equilibration between the carrier and the sample, and allow the melt to cool to below 100 "C. The melt dissolves violently and, in order to avoid losses from effervescence, the crucible should be placed i n a glass filter-funnel ol-er a round-bottomed flask and covered with a watch-glass.Transfer the solution into the flask and repeat the dissolution step several times. Rinse the nickel crucible with 12 zI hydrochloric acid and the watch-glass and funnel with de-ionised water, and transfer all washings into the round-bottomed flask. Add 12 M hydrochloric acid to the dissolved sample until the solution is clear green. Add glass beads to the flask and boil the solution for 5 min. If the sample is rich in copper, add 2 ml of 16 31 nitric acid before boiling. Cool the flask i n cold water and, if the sample contains more than a few micrograms per gram of chromium, add 1 g of solid iron(I1) sulphate. Adjust the solution to pH 1-2 with ammonia solution, taking care not to precipitate the iron(I1) ions.Make the solution up to 100 ml with de-ionised water and transfer the sample to the ion-exchange system. In this work an automated system was used but the separation can be carried out manually. Pass the sample solution 011 to the column at 2 1111 min-l and subsequently wash the resin with 0.05 M Iiydrocliloric acid. Then transfer the resin sample into a plastic counting vial (5 cm diameter x 4 cm high). Counting Transfer the standard-loaded resin into a disposable plastic counting vial immediately after it has been unloaded from the reactor. Add 4ml of fresh Srafion NNRR resin and mix well to give a uniform distribution of the activity. Count the standards for 1 h on the lithium-drifted germanium detector with the sample as close to the detector as possible in the I'erspex sample holder of the sample changer.Determine the counts from the gamma- ray energies of lO9Pd, l99Au, lRJIr and 19SA4u, given in Table I, using the total peak-area method. \I'hen the first rock sample has been processed, transfer the resin to the sample changer and count for 3 11. Add 10 ml of de-ionised water to the melt while it is still hot. The solution will effervesce as any excess of peroxide boils off. Repeat for the remaining samples. Repeat for the remainder of the samples, including the reference rock. TABLE I NUCLEAR DATA FOR THE NOBLE METAL NUCLIDES MEASURED I N THIS WORK^^ Suclide Half-life Gamma-ray energy1keV Interferences lo8Pd . . . . 13.4 h 88.1 Pb X-rays at 87 keV from photon interaction l 9 9 P t , .. . 30.8 min - - 198Au . . . . 3 . 1 4 d 158.3 47Sc a t 159.5 keV from -"Ti(n,p); lg8Au from lg*Au(n,y) where gold content >> platinum content 1941r . . . . 19.2 h 328.0 - 19*ASu . . . . 2.70 d 411.8 - with detector shielding (decays to lssAu) Yield Determination Leave the resin samples to decay for 1 week before re-irradiation to determine the yield. Transfer the resin to polythene capsules (15 mm diameter x 60 mm high) and dry a t 80 "C. Place the capsules in outer irradiation capsules (18 mm diameter x 70 mm high). Prepare standards for the yield determination by evaporating 1.0 ml of carrier solution on 5 ml of1160 Afzalyst, Vol. 105 Srafion NMRR resin in the polythene capsules used for the resin samples, packed in outer irradiation capsules. Irradiate the samples and standards for 5 min in the pneumatic irradiation system at 10-min intervals sequentially.Leave them to decay overnight in order to remove the 38Cl activity from the resin. Transfer the samples and standards to plastic counting vials and count them on the lithium-drifted germanium detector for Y ? d , 194Ir and lYSAu for 5 min each. Platinum is separated on the resin as 1y9Au and therefore the yields for both gold and platinum are derived from the measurement of 198Au. PARRY: DETERMINATION OF KOBLE METALS I N GEOLOGICAL Results and Discussion Analysis of Sulphide Standards The accuracy of the method was checked with reference material that had recommended values for palladium, platinum and gold and experimental values for iridium.PTC22 and PTW3 are produced by the Canadian Certified Reference Material Project. PTC is a noble metal-bearing sulphide concentrate containing 5.160/o of copper, 9.42% of nickel, 23.5% of sulphur and 26.9% of iron. PTM is a noble metal-bearing copper - nickel matte containing 30.2% of copper, 44.896 of nickel, 1.58yl of iron and 21.6Y0 of sulphur. The activity that resulted from the copper content of the sulphides was reduced by epithernial neutron activa- tion. The active sample solution was left in the oxidised state, as preliminary experiments showed that the retention of copper on the Srafion NMRR resin was lo!, of copper(I1) com- pared with 25% of copper(1). The noble metals are present in the sulphides at the micro- grams per gram level and therefore the sensitivity was not a problem in the determination.Table I1 shows that the analyses of the reference material gave good agreement with the recommended values for the noble metals except for gold and platinum in PTC, which were low. This may be attributed to inhomogeneities, which are enhanced by the small sample size used, reflected by the large standard deviation on the results. Low results can be caused by losses due to incomplete decomposition of the sample, but it would be expected that if such a problem occurred it would also be seen in the analysis of PTM. TABLE I1 DETERMINATION OF THE NOBLE METALS I N STANDARD SCLPHIDES r- Sample Parameter Pd PTM . , , . Mean value (this work) 8.8 Standard deviation 1.6 Recommended valuez3 8.1 95% confidence interval 7.4-8.8 Standard deviation 1.8 Recommended valuezz 12.7 95% confidence interval 12.0-13.0 PTC ., . . Mean value (this work) 11.8 * For information (not recommended) Concentration;pg g-1 A > Pt l r All 5.97 0.38 1.78 0.53 0.07 0.32 5.8 0.25, 0.36, 0.81* 1.8 5.5-6.2 1.6-1.9 1.8 0.20 0.39 1.1 0.06 0.23 3.0 0.12, 0.24, 0.01, 0.19* 0.65 0.55-0.72 2.8-3.2 Analysis of US Geological Survey Standard Rocks The US Geological Survey reference rocks PCC-1 and DTS-1 were also analysed to deter- mine both the sensitivity and the accuracy of the technique. ‘There are no certified values for the noble metals in these rocks but they have been measured by other workers and the values are comprehensively listed in the l i t e r a t ~ r e . ~ Palladium and platinum could not be measured below the detection limit of 12 pg kg-l, determined from 2.3 standard deviations on the background.Table I11 shows that the values obtained for iridium and gold were within the range of results obtained by other workers. An unexpected source of error was found to be due to 47Sc interference a t 159.5 keV, which contributed to the 158.3-keV peak of lg9Au, used in the determination of platinum. The 47Sc is the result of an n,p reaction on irradiation of 47Ti, which is enhanced by the epithermal neutron activation. The inter- ference was removed by the addition of scandium oxide hold-back carrier. The detection limits for iridium and gold were 1 and 0.1 pg kg-l, respectively.Deccnaber, 1980 MATERIAL BY RADIOCHEhIICAL NEUTRON-A4CTIVATION ANALYSIS I161 TABLE I11 DETERXIKATIOS OF THE NOBLE METALS IN STANDARD ROCKS Sample I’CC-1 .. DTS-I Concentrationlpg kg-’ r- A Parameter Pd Pt I r Mean value (this work) < 1 1 < 20 4.2 Standard deviation 2.2 Results of other workers 3.0-7.2 3.5-15 2-6.7 Mean value (this work) < 12 < 12 2.4 Standard deviation 1.0 Results of other workers 1.8-6.1 4.7-11 0.3-4.5 -7 Au 1 .o 0.3 0.63-3.4 1.1 0.5 0.45-1.4 Analysis of Chrome-spinels The method was then applied to the analysis of chrome-spinels, containing about 30% of chromium. The reduction step witti iron(I1) sulphate converted the chromate ions, which gave ’750;, retention on the resin, into chroniium(III), which was collected a t only the lo/o level. However, tlie further reduction of the chromium activity by a factor of 30 with epithermal irradiation was essential before the noble metals could be determined in the chrome-spinels, The spinels were separated and purified prior to analysis and in most instances the sample size was limited to less than 0.1 g.Table IV shows that the detection limits for palladium and platinum in chrome-spinels are about 40 pg kg-’. The noble metal concentrations in the chrome-spinels are presented as the means of three determinations in most instances. The large variation of the values reflects the inhomogeneity of their distri- bution within the sample. TABLE IV DETERMINATION OF THE KOBLE NETALS I N CHROME-SPIXELS Concentration/pg kg-1 <- l y p e of sample Location Code Pd Stratiform complexes . . Bushveld 1930, 66jl’i) 64 i- 40 Stillwater 1970, P20(1) <60 Rhum RWR9 13i + 86 IlWR13 118 * 49 Rhodesia (Zimbabwe) RWRh2 150 56 Oregon RWThl 85 * 50 Alpine-type deposits .. Greece Pakistan Turkey Unst RWGR2 RWGR3 KWPl RWP3 RWT14 RWT16 SUll < 20 430 <40 <60,350 170 3.7 <30,14n Pt 215 i- 7 147 92 330 I 180 <110 55,435 <so, 83 < 30 170 < 40 < 40 190 < 30 < 14 Ir 74 =t 18 18 i i 30 1- 10 3 4 * 3 140 * 26 47 * 11 27 I 8 26 & 17 3 9 t 6 49 * 23 59 2 3 + 4 540 ~ ~~ AU 2.4 0.5 2.2 I 1.4 2.1 1.7 2.2 2.7 2.2 = 1.6 0.29 + 0.20 5.7 0.9i 5 0.49 2.3 = 3.8 1.1 & 0.8 4.1 1.6 * 0.6 23 * i Conclusions A rapid and simple procedure has been established for the routine determination of the noble metals in geological samples. Palladium, platinum, iridium and gold are measured simultaneously in five samples within 18 11 of being unloaded from the reactor.Except for the initial digestion of the samples, the procedure can be automated for unattended operation. The precision of the method has been confirmed by analysis of standard rocks and the sensitivity is adequate for samples containing iridium and gold a t the micrograms per kilo- gram level and platinum and palladium at levels above 50 pg kg-l. The scheme is particu- larly suited to the analysis of samples that contain very high concentrations of copper and chromium. The author gratefully acknowledges P. Henderson, British Nuseum (Natural History), and R. J. Wood for supplying the samples of chrome-spinel.1162 PARRY References 1. 2. 3. 4. 5. 6 . 7 . 8. 9. 10. 11. 12. 13. 14. 15.16. 17. 18. 19. 20. 21. 22. 23. Beamish, F. E., and Van Loon, J. C., “Recent Advances in the Analytical Chemistry of the Noble Le Roux, H. J., Beukes, P. J . L., Turkstra, J . , and De Wet, 1%’. J., J . S. Afr. Chem. Inst., 1974, 27, Crocket, J. H., Keays, R. R., and Hsieh, S., J . Radioanal. Chenz., 1968, 1, 487. Millard, H. T., and Bartel, A. J., i n Brunfelt, A. O., and Steinnes, E., Editors, “Activation Analysis Ahmad, I., Ahmad, S., and Morris, D. F. C., Analyst, 1977, 102, 17. Nadkarni, R. A,, and Morrison, G. H., Anal. Chenz., 1974, 46, 232. Koster, G., and Schmuckler, G., Anal. Chinz. ilcta, 1967, 38, 179. Green, T. E., and Law, S. L., U.S. Bur. Mines, Rep. Invest., RI 7358, 1970. Green, T. E., Law, S. L., and Campbell, 14’. J.2 AIzal. Chem., 1970, 42, 11749. Sundberg, L. L., Anal. Chem., 1975, 47, 2037. Blount, C. W., Leyden, D. E., Thomas, T. L., and Guill, S. M., Anizl. Chenz., 1973, 45, 1045. Das, H. A,, Janssen, R., and Zonderhuis, J., Radiochem. Radioanal. L e t t , 1971, 8, 257. Baghai, A,, Unpublished PhD Thesis, Reading University, 1977. Nadkarni, R. A., and Morrison, G. H., Anal. Chem., 1975, 47, 2285. Gladney, E. S., Anal. Chim. Acta, 1977, 91, 353. Kadkarni, R. A,, and Morrison, G. H., Anal. Chem., 1978, 50, 294. Warshawsky, A., and Ehrlich-Roginski, S., Microchem. J., 1977, 22, 362. Muzzarrelli, R. A. A,, and Rocchetti, R., Anal. Chim. Acta, 1974, 70, 465. Xadkarni, R. A,, and Morrison, G. H., J . Radioanal. Chem., 1977, 38, 435. Tredoux, M., NPRU-76/1, University of Witwatersrand, Johannesburg, 1976. Lederer, C. M., and Shirley, V. S., “Table of Isotopes,” Seventh Edition, John Wiley, Kew York, McAdam, R. C., Sutarno, R., and Moloughney, P. E., Technical Bulletin No. TB 176, Information RPcAdam, R. C., Sutarno, R., and Moloughney, P. E., Technical Bulletin No. TB 182, Information Metals,” Pergamon Press, Oxford, 1972. 28. in Geochemistry and Cosmochemistry,” Universitetsforlaget, Oslo, 19’71, p. 353. 1978. Canada, Department of Energy, Mines and Resources, Ottawa, 1973. Canada, Department of Energy, Mines and Resources, Ottawa, 1973. Received June 23rd, 1980 Accepted July 17th, 1980

ISSN:0003-2654    

DOI:10.1039/AN9800501157

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

Analyst, Decenzbev, 1980, Vol. 105, Pp. 1163-1170 1163 Determination of Aluminium in Wood Pulp Liquors Using Graphite Furnace Atomic-absorption Spectrometry J.-A. Persson, W. Frech, G. Pohl and K. Lundgren Departme?zt of Analytical Chevnistvy, University of Umed, S-901 87 Umed, Swedeit X graphite furnace atomic-absorption spectrometric method for the deter- mination of 0.5-100 pg ml-l of aluminium in matrices containing chlorine and high concentrations of sulphur is described. The precision of the over-all procedure was 40/. Graphite tube conditions and the composition of the matrix were found to change the sensitivity for aluminium. Keywovds : Aluiiziniuin detevntination ; graphite furnace atomic-absorption spectrometry ; szrlpliuv intevfevence ; chlovine intevfevence ; graphite tube conditions Cellulose pulp is produced by cooking wood in a solution of sodium sulphite and sodium hydrogen sulphite.In order to minimise emissions and production costs, the chemicals used in the process have to be recycled as much as possible. The recycling of the cooking liquor includes neutralisation and evaporation. Part of the condensate is removed from the process in order to prevent accumulation of compounds originating from the raw materials. The remaining part is combusted and re-used in the process. The wood used as raw material contains calcium and oxalate compounds at such levels that severe problems caused by calcium oxalate incrustations arise during condensation in the evaporation plant. These problems can be avoided by adding small amounts of aluminium to the liquors before neutralisation.The added aluminium forms strong complexes with oxalate ions, thus preventing calcium oxalate depositions, The concentration of aluminium in the liquors has to be controlled carefully, as high concentrations will lead to precipitation of aluminium silicate and aluminium hydroxide. At the Donisjo plant, aluminium was determined in wood pulp liquors by precipitation followed by colorimetry1 or flame atomic-absorption spectromety2 and by a procedure involving extraction of aluminium as a chelate followed by flame atomic-absorption spectro- n ~ e t r y . ~ The high sensitivity of graphite furnace atomic-absorption spectrometry has been exploited to develop simple and rapid methods for the direct determination of aluminium in a variety of matrices, e.g., steel^,^.^ aerosols,6 petroleum products' and biological materials.8,g The purpose of this work was to develop a graphite furnace atomic-absorption spectrometric method for the direct determina- tion of aluminium in wood pulp liquors.Experimental These methods are laborious and time consuming. Apparatus A Perkin-Elmer 372 atomic-absorption spectrometer provided with background correction and a research spectrophotometer were used. Both instruments were equipped with HGA-74 graphite furnaces. The HGa4-74 furnaces were connected to home-made power supplies with facilities for temperature-controlled heating of the graphite tubes.1° Tempera- ture settings referring to the inner surface of the graphite were calibrated with an NiCr - Ni thermocouple below 1000 "C and with an optical pyrometer (Keller Spezialtechnik Pyro IVerk GmbH, Model PBO6AF3) at higher temperatures.Samples mere injected with a Perkin-Elmer automatic sampler, Xodel AS-1, or manually with Oxford micropipettes. The recorder output of the spectrometer was connected to a peak-reader modulell which provided simultaneous recordings of the peak height and the peak area. The results given in this paper are based on the peak-height absorbance values because better precision was obtained in this way. The peak reader was connected to a printer (Newport Laboratories Inc., Model 810) and a strip-chart recorder (Perkin-Elmer, Model 56). The instrumental parameters are given in Table I.1164 PERSON et al. : DETERMINATION OF ALCMINIGM IN TABLE I Analyst, Vol.105 ISSTRUMENTAL PARAMETERS Time/s Temperaturel'C Drying . . . . 35 100 Wavelength/nni . . . . 309.3 Ashing . . . . 60 1200 Siitlnm . . . . . . . . 0.7 Delay . . . . 4 - Metal lamp current/m.l . , 7 Atomisation . . . . 6 2 400 Deuterium lamp , . . . On Clean-out . . . . 6 Maximum Hydrogen lamp/mX . . . . 20* Sample volume/pl . . . . 1-1oot Argon flow-rate/l min-l- Internal . . . . . . 0.35: External . . . . . . 1 3 * Only used with research spectrophotometer for the determinatioii of the background. t Depending on the concentration of aluminium. standards and unknown samples. $ Gas-stop during atomisation. The same volumcs have to bc used for Dry Ashing of Evaporated Liquor and Dried Pulp Samples dried a t 100 "C for 4 h and then placed in a furnace at 1000 "C for 4 11.residue was dissolved in 2 ml of concentrated nitric acid and diluted to 100 in1 with water. One gram of evaporated liquor was transferred into a platinum crucible. The sample was After cooling the Reagents and Materials Kitric acid (65% V / V ) and sodium chloride were hIerck Suprapur grade products. Aluminium metal, sodium sulphide, sodium sulphate and sodium carbonate were llerck analytical-reagent grade materials. Sodium thiosulphate and sodium sulphite were obtained from Riedel de Haen (analytical-reagent grade) and sodium 111-drogensulpliite from hlallin- kcrodt (analytical-reagent grade). All absorbance values obtained in the investigation of matrix components were corrected for blank values. Solutions were stored in polyethylene bottles.All apparatus was washed with 4 M nitric acid. Stock nltmi?zium solution, 100 pg ml-l of alzwziizizuiz. Ilissolve (1.100 g of aluminium in 5 ml of concentrated nitric acid, transfer the solution into a 1-1 calibrated flask and dilute to volume with water. Transfer 40 g of sodium sulphate, LO g of sodium carbonate and 2 g of sodium chloride into a 1-1 calibrated flask, adjust to pH 1 with nitric acid and dilute to volume with water. A rtijcial soldions of sulphide liquor mad grw6 liquor. Prepare solutions containing the salt concentrations given in Table 11. Transfer 1 ml of each solution into two separate 100-ml calibrated flasks and dilute to volume with water. Stock nrtijcinl matrix solution. TABLE I1 .k'ROXIMATE COMPOSITIOK OF THE LIQUORS AYALYSED Green liquor Sulphite liquor from scrubber l:\aporateil liquor ......7r--p----h-__ -7 -.---L r ~ - ~ - A p - - Concentra- Conccntra- Concentra- tion/g I-' Component tion/g 1-I ConlpoIlent tion/g 1-' Component Na . . . . , . 80 S a . . . . . . 50 S a . , . . . . B 0 Total sulphur . , , , 40 Total sulphur . . 35 Total sulphur . . 60 Na,S . . . . . . 85 NaHSO, . . . . 19 s;ic1 . . . . 2 Na,CO, . . . . . . 45 Na,CO, . . . . 3 Organic material . . 680 Na,SO, . . . . . . 7 Nn,SO, . . . . 10 Na,S,O,. . . . . . 5 Sa,S,O, . . . . 8 NaCl . . . . . . 2 Na,SO, . . . . 102 XaOH . . . . 1 NaCi . . . . 1December, 1980 WOOD PULP LIQUORS BY GRAPHITE FURKACE AAS 1165 Procedure Preparation of Sample Solutions Liqaors containing 0.5-10 pg ml-l of aluminium Ecaporated liq.tior.nitric acid (650/,) and dilute to volume with water. Sztlphite liquor f~oin scrubber. add 0.5 ml of nitric acid (6546). Green liquor. with water. Liquors containing 10-100 pg ml-1 of alzwninium Weigh 1 g of sample into a 100-ml calibrated flask, add 0.5 ml of Transfer 1 ml of sample into a 100-ml calibrated flask and Dilute to volume with water. Transfer 1 ml of sample into a 100-ml calibrated flask and dilute to volume Proceed as described above but dilute the samples to 1 1 instead of 100 ml. Preparation of Standard Solutions Aluini?ziu~i~ staizdnrd solattiom flasks. Standaids prepaid 1'72 an artijcinl 9mtri.x foi. deteriiainatioiz of nlimi&uin in the range 0.5- 10 pg ml-I Transfer I..O-ml aliquots of each of the aluminium standard solutions into 100-ml cali- brated flasks each containing 1 in1 of the stock artificial matrix solution.Dilute to volume with distilled water. These solutions contain 0, 0.02, 0.05, 0.07 and 0.1 pgml-' of aluminium. Stmidaids jwc@zred in distilled water f o r determinntioiz of alzuniizizi~z iit the range 10-100 pg nil-l Transfer 1 .O-nil aliquots of the aluminium standard solutions into 100-ml calibrated flasks and dilute to volume with distilled water. These solutions contain 0, 0.02, 0.05, 0.07 and 0.1 pg nil-l of aluminium. Measurement I:se tlie instrumental parameters given in Table I and determine the aluminium contents of green liquor, sulphite liquor from the scrubber, standards and blank solutions, a t least in duplicate. For every fourth sample a standard should be run.The aluminium concentration in green liquor and sulphite liquor from scrubbers is determined by evaluating the mean of the peak-height absorbances of standards, blanks and samples. For evaporated liquor samples, only the mean of the second and tliird determinations is evaluated. Transfer 0, 2.0, 5.0, 7.0 and 10.0ml of stock aluminium solution into 100-ml calibrated Add 0.5 ml of nitric acid (6504) and dilute to volume with water. For evaporated liquor, samples have to be run in triplicate. Results and Discussion As the aluminium concentration has to be monitored during tlie whole process, determinations have to he made in solutions representing different matrices. Mean values of tlie main constituents of the different types of liquors involved in the processing of wood pulp are given in Table 11.Depending on the composition of the raw materials used and the processes involved, the contents of sulphur and chloride in the different liquors may also vary from day to day. The effect of this variation on the graphite furnace atomic-absorption spectrometric deter- minations had to be investigated in detail in order to obtain an accurate procedure. In previous publications12J3 it was shown that the determination of aluminium by the graphite furnace technique suffers from interferences in the presence of both sulphur and chloride. Optimisation of Ashing Temperature Fig. 1 sl~ou~s the effect of the ashing temperature on the sensitivity for aluminium in the three types of liquors investigated (upper curves) and the magnitude of the non-specific The composition of the liquor changes during the processing of wood pulp.1166 PERSSON et al.: DETERMISATION OF A L u m s I u k i IX 100 ae 0- E e % p 50 m + m m .- - a I I I I Analyst, Vol. 105 ~~~ 0 50 1000 1500 Final ashing ternperature/'C Fig. 1. Relative peak absorbance for aluminium as a function of the final ashing temperature: A, evaporated liquor; B, sulphite liquor from scrubber; C, green liquor; and D, the background in green liquor. All samples were diluted 1 1 99. The aluminium contents for evaporated liquor, sulphite liquor from scrubber and green liquor were 0.2, 0.3 and 0.14 pg ml-l, rcspectively. absorbance for green liquor (lower curve). The high salt concentration of the liquors gave rise to a disturbing non-specific absorbance a t temperatures below 800 "C.For example, the non-specific absorbance for a green liquor sample diluted 1 + 99 and charred at 500 "C was 0.4. The background can be removed almost completely by asliing a t 1100-1250 "C and in addition optimum sensitivity for aluminium is obtained. As the sensitivity with the evaporated liquor is only slightly decreased at 1200 "C, this temperature was used for all matrices. At the ashing temperature chosen the reproducibility was found to be typically about 30/, compared with 5% a t 500 "C. The impaired reproducibility and tlie lower sensitivity at 500 "C are probably caused by a time overlap in the furnace between the volatilisations of analyte atoms and the interfering species thereby increasing the number of gaseous aluminium molecules formed.Sensitivity and Graphite Tube Conditions During this work it was found that the sensitivity for aluminium varied from tube to tube. The magnitude of this variation was normally 2504 but could be as much as 100%. However, the relative change in sensitivity between tubes was found to be equal in the liquors and aqueous solutions. In order to be able to compare results obtained with different tubes, all solutions had to be run relative to 0.01 b~ nitric acid. However, for a particular tube the sensitivity changed only slightly during its lifetimc. This is illustrated in 1;ig. 2 , in which the results of determinations of aluminium in 0.01 nitric acid as well as in a sulphide liquor sample are shown. Tlie useful lifetimc for a graphite tube was found to be dependent on the type of solution determined.For liquor samples diluted 1 -'- BY, a tube could be used for 50-60 determinations. U'ith more dilute solutions, up to 100 detcrminations could be per- formed. Before use each graphite tube was conditioned by three blank firings at maximuni temperature. Interferences from Sodium Sulphate and Sodium Sulphide The sensitivity of the graphite furnace atomic-absorption spectrometer made it possible t o dilute samples 100- to 1000-fold. After these dilutions samples contained less than 0.35 mg ml-I of sodium sulphate and less than 0.85 mg ml-* of sodium sulphide. Table I11 gives the relative sensitivity for aluminium in tlie presence of various concentrations of sodium sulphate. It can be seen that the signal for aluminium was enhanced in 0.01 $1 nitric acid solution for increasing salt concentrations. An ashing time of at least 60 s was needed to obtain these optimum conditions.December, 1980 WOOD PULP LIQUORS BY GRAPHITE FURNACE AAS 1167 0.35 0" . .. . . . . . . . v . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vv W V -v _ _ _ _ _ _ _ - - -vVv- _ _ _ _ _ _ _ - -- - - - - x - so? 4.6 n 27 +' F 0.400 V vv. -s vv v vv vvv n a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - V I I Fig. 2. A, 0.1 pg ml-l of aluminium in 0.01 M nitric acid; and B. 0.07 pg ml-' of aluminium in a sulphite liquor from scubber sample diluted 1 + 99. Sensitivity for aluminium during the lifetime of a graphite tube.However, in the liquors and liquor-like matrices the aluminium signals decreased a t the corresponding sulphate concentrations. With sodium sulphide similar interferences were obtained. These results show the importance of investi- gating interferences in solutions with compositions matching the samples as closely as possible. Tables 111 and IV also show that the magnitude of the interference effects obtained is dependent on the type of sulphur compound, as the effect of sodium sulphate is larger than that of sodium sulphide. However, the sensitivity for aluminium is decreased by a maximum of 15% in the presence of the highest tested concentrations of these sulphur compounds. Minor variations in the concentration of sodium sulphate and sodium sulphide in the liquors will therefore have a negligible effect on the analytical results provided that the procedure given above is followed.TABLE 111 RELATIVE INTERFERENCE EFFECTS OF SODIUM SULPHATE ON THE SENSITIVITY The results are given in Table IV. FOR ALUMINIUM I N DIFFERENT MATRICES Relative absorbances corrected for reagent blank values are given. Sulphite liquor from Na,SO, added to Evaporated rrA-, sample/mg ml-1 0.01 M HNO, liquor* Artificial Sample k r t i f i c i a l e Green liquor* scrubber* 0 100 100 100 100 100 100 1 105 94 91 94 99 96 2 109 93 87 86 95 94 4 112 90 85 85 95 92 * Evaporated liquor was diluted 1 + 649, green liquor 1 + 49 and sulphite liquor from scrubber 1 + 199. Interferences from Sodium Chloride The effect of 0-1.5 mgml-l of sodium chloride on the determination of 60 ngml-1 of aluminium was first investigated in artificial solutions of green liquor and sulphite liquor from the scrubber.The experiment was repeated with real samples diluted 1 + 99 and the results are shown in Fig. 3. It can be seen that the sensitivity for aluminium is unaffected by chloride in green and evaporated liquor samples, whereas in the sulphite liquor sample from the scrubber the sensitivity for No interferences were observed in these solutions.1168 PERSSON et al.: DETERMINATION OF ALUMINIU~I IX *4.1zalyst, Vol. 105 TABLE IV RELATIVE INTERFERENCE EFFECTS OF SODIUM SULPHIDE ON THE SENSITIVITY FOR ALUMINUM* Relative absorbances corrected for reagent blank values are given. Green liquor 0 . 0 1 ~ HNO, -- ;\;a$ added to sample/ mg ml-l 0 100 100 100 1 95 97 101 2 105 96 95 4 106 96 96 * For the dilution factors used, see Table 111.aluminium is decreased by 50y0. The interference with the sulphite liquor is probably caused by a combined action of the chlorides added and some minor component not listed in Table 11. The minor component is assumed to form stable chlorides which cannot be removed by ashing a t 1200 "C. During atomisation these chlorides must be volatilised simultaneously with aluminium in order to favour the formation of gaseous aluminium monochloride molecules. In this way the free atom concentration will be decreased, resulting in lower absorbance signals. From the results given in Fig. 3 it can be seen that even a 10-fold variation in the sodium chloride concentration will change the sensitivity for aluminium by only 5%.This means that the analytical results are relatively unaffected by large variations in the sodium chloride content. Liquor samples diluted 1 + 99 will contain 0.01 mg nil-' of sodium chloride. 0.5 1 .o 1.5 Sodium chloride addedirng ml-' Fig. 3. Interference effects of sodium chloride on the peak absorbance of aluminium. A, Sulphite liquor from scrubber; B, green liquor; and C, evaporated liquor. Combined Effects For the determination of aluminium the green and sulphite liquors using the instrument parameters given in Table I, typical relative standard deviations of 30/, were obtained whereas for the evaporated liquor a typical value was 10%. The reason for this poor precision can be understood by studying the peak height signals given in Fig.4. After conditioning the tube, triplicate determinations of aluminium were made alternately in 0.01 M nitric acid (thin peaks) and evaporated liquor (thick peaks). Initially aluminium was determined in the nitric acid solution. The evaporated liquor was then tested. It can be seen that the sensitivity is markedly decreased after the first deter- mination of the evaporated liquor after which the sensitivity is stabilised. In 0.01 M nitric acid the signal is depressed for the first determination and the original sensitivity is then regained. It is also shown in Fig. 4 that these variations in sensitivity were reproducible. For both solutions the over-all precision is poor; however, if the first determination in each series is excluded, relative standard deviations of about 3% are obtained. The evaporated The precision was better than 2%.Decenaber, 1980 WOOD PULP LIQUORS RY GRAPHITE FURNACE A.4S n c __ 0 10 1169 Determination number Fig.4. Illustration of memory effects on the determination of aluminium caused by evaporated liquor. Thin peaks, 0.1 p g ml-1 of aluminium in 0.01 M nitric acid; thick peaks, 0.07 p g nil-1 of aluminium in evaporated liquor. liquor contains a much greater amount of organic matter than the other liquors investigated and this was initially assumed to cause the irregularities described above ; however this was shown not to be the case because results similar to those shown in Fig. 4 were obtained for evaporated liquor samples in which the organic matter had been removed by ashing.The relative sensitivities obtained in the three types of liquors, an artificial standard solution and 0.01 M nitric acid were also investigated. This was accomplished by using the standard-addition technique. The slopes obtained in the linear range were taken as a measure of the sensitivity. Table T.' shows the slopes for samples diluted 1 + 99 and the artificial standard solution relative to 0.01 M nitric acid. As can be seen in the table the ratios are 10% less than unity and hence calibration against standards in 0.01 11 nitric acid will give results that are too low. I t was also shown that correct analytical results for the three types of liquors would be obtained by standardisation against the artificial standard solution. No interferences were observed in samples diluted 1 + 999 and therefore aqueous standards could be used for calibration.TABLE V SEXSITIVITY FOR ALUMINIUM IN THE DIFFERENT SAMPLE TYPES RELATIVE TO 0.01 IlI NITRIC ACID Slope of sample graph Slope of graph for 0.01 M HNO, Sample Green liquor . . . . . . . . 0.89 Evaporated liquor. . . . . . . . 0.88 Sulphite liquor from scrubber . . . . 0.92 Artificial standard solution . . .. 0.88 Analytical Results In order to test the validity of the proposed method, determinations of aluminium in sulphite liquor from scrubber, green and evaporated liquor samples were made by two other methods. In addition a number of liquor samples from other stages in the processing of wood pulp were included. The results obtained for the different methods can be seen in Table VI and the agreement is No liquor samples with certified values for aluminium were available.1170 PERSON, FRECH, POHL AND LUNDGREN acceptable.The typical precision of the over-all graphite furnace atomic-absorption spectro- metric procedure is 4%. TABLE V I COMPARISON OF RESULTS OBTAINED FOR THE DETERMINATION OF ALUMINUM BY GRAPHITE FURNACE ATOMIC-ABSORPTION SPECTROMETRY WITH RESULTS OBTAINED BY OTHER METHODS Sample Green liquor I . . . . . . . . Green liquor I1 . . . . . . White liquor . . . . . . . . Evaporated liquor . . , . . . Spent sulphite liquor I . . . . Spent sulphite liquor I1 . . . . Dried pulp: . . . . . . . . Sulphite liquor from scrubber I Sulphite liquor from scrubber I1 Hot water from scrubber .. . . Water from chloride scrubber . . . . . . Graphite furnace Flame atomic- method/pg ml-1 methodlpg ml-1 atomic-absorption spectrometric absorption spectrometric 72 22 106 335 4.5 3.9 7.9 3.5 0 . 5 3.6 9.65 loot 32*§ 4.5* 3.9* 7.8* 2.8* 0.6* 3.1* 9.1*§ * Determination of aluminium by atomic-absorption spectrometry with dinitrogen oxide - acetylene t Extraction of aluminium as a chelate followed by flame atomic-absorption spectrometry.3 $ Dry ashing of the sample. § Concentration in micrograms per gram. flame after precipitation.2 The authors thank Dr. B. Eriksson, Swedish Forest Products Research Laboratory, for providing samples, Mr. K. Nilsson and Mr. K. Bergner, MoDoCell AB, for details of the processes used in the factory and Dr. M. Sharp for linguistic revision of the manuscript. 1. 2. 3. 4. 5 . 6 . 7. 8. 9. 10. 11. 12. 13. References Willard, H. H., and Tang, N. K., Ind. Eng. Chem., 1937, 9, 357. Chan, K. M., and Riley, J. P., Anal. Chim. Acta, 1966, 34, 337. Svenska Traforskningsinstitutet, Stockholm, personal communication. Shaw, F., and Ottaway, J . M., Analyst, 1975, 100, 217. Persson, J.-A., Frech, W., and Cedergren, A., Anal. C h i m . Acta, 1977, 89, 119. Pilate, A,, Geladi, P., and Adams, F., Talanta, 1977, 24, 512. Brodie, K. G., and Matousek, J . P., Anal. Chem., 1971, 43, 1557. Garmestani, K., Blotcky, A. J., and Rack, E. P., Anal. Chem., 1978, 50, 144. Langmyhr, F. J., and Tsalev, D. L., Anal. Chim. A d a , 1977, 92, 79. Lundgren, G., Lundmark, L., and Johansson, G., Anal. Chem., 1974, 46, 1028. Lundberg, E., A$$. Spectrosc., 1978, 32, 276. Persson, J.-A., Frech, W., and Cedergren, A,, Anal. Chim. Acta, 1977, 92, 85. Persson, J.-A., Frech, W., and Cedergren, A,, Anal. C h i m . Acta, 1977, 92, 95. Received October loth, 1979 Accepted June 25th, 1980

ISSN:0003-2654    

DOI:10.1039/AN9800501163

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

Analyst, December, 1980, Vol. 105, pp. 1171-1175 1171 Determination of 2-Aminobutane in Potatoes Using High-performance Liquid Chromatography K. A. Scudamore ATIzizzstr~ o j A4graculture, Fzsherzes and Food, Slough Laboratory, London Road, Slough, Bevkshare, SL3 IH J Development xvork on the use of 2-aminobutane to control certain potato- tuber diseases necessitated the development of a convenient and sensitive analytical method for residue determination. The amine was distilled from potatoes, dansplated and determined using reversed-phase high-performance liquid chromatography with fluorescence detection. Recovery of 2-amino- butane by distillation was about 957; from standard solutions and 92% from treated potatoes. Keywords : d-ilininobutnne deterwzinatiox ; potatoes ; residues ; high-performance The lower limit of detection is below 0.2 pg kg-'. liquid chromatography ; fluorescence detection The control of decay in citrus fruits by treatment with an aqueous solution of 2-aminobutane was described by Eckert.l Later, application of the compound as a vapour was found to control the potato-tuber diseases gangrene, skin spot and silver scurf2y3 and a method for the treatment of small bulks of potatoes in fumigation chambers was developed.* The Ministry of Agriculture, Fisheries and Food and the Department of Agriculture and Fisheries for Scotland have prepared recommendations for safe and efficient use of 2-aminobutane for fumigation of p ~ t a t o e s .~ Recently, more extensive trials have shown that large bulks of up to 1000 tons or more of potatoes can be treated successfully ir, situ in store.Heavy sorption of 2-aminobutane vapour occurs particularly in the eel.^,^ Residues, mainly in the form of amine salts, can persist for lengthy periods. However, achievement of a certain residue level is necessary for protection of the potatoes from disease. Sampling and analysis of 2-amino- butane in potatoes in bulk fumigations can give an indication of the likely success of the treatment as well as the safety of the process. Investigation of these factors in relation to the residue contents during storage and processing has necessitated the development of a simple and sensitive analytical method for their determination. Day et aL7 described a method for steam distillation of 2-aminobutane from fruit and this was later found to be equally applicable to potatoes3 In the method described here distilled amine extracts are dansylated and determined using high-performance liquid chromato- graphy (HPLC) with a fluorescence detector.Experimental Distillation Finely chop one quarter of each potato and mix the four bulked portions thoroughly. Transfer 50 g of the bulked potato sample into a 500-ml round-bottomed flask and add 50 ml of distilled water, 40 ml of 1 >I calcium chloride solution, 5 g of magnesium oxide slurried with 25 ml of water and a drop of anti-foaming agent. Use a further 50ml of water to wash down the sides of the flask. Attach a distillation head and water condenser to the flask and heat the mixture to boiling. Collect the first 70 ml of the distillate in 10 ml of 0.3 &.I sulphuric acid.Pour the whole into a 250-ml beaker and adjust the pH to 9.0 using 0.5 iv sodium hydroxide solution. Transfer into a 100-ml calibrated flask and make up to volume with distilled water. Derivatisation Reagents Borate buffeer, pH 9.0. Add 12.4 g of boric acid to 100 ml of 0.1 M sodium hydroxide solution and dilute to 1 1 with distilled water. Mix 85 ml of this solution with 15 ml of 0.1 M hydrochloric acid. A 3 mg ml-l solution in dry analytical-reagent grade acetone. Take four potato tubers from the bulk sample and quarter them. Dansyl chloride (l-dime~hylamir,ona~hthalene-5-sul$honyl chloride). Store in the dark at 0 "C.1172 SCUDABIORE : DETERMINATION OF Analyst, Vol. 105 Standard solutions of 2-aminobutane in water, 20, 40 and 100 pg ml-l.Solution of analytical-reagent grade aniline (0.2 g) in aizalytical-rsLigcnt gifade acetone (50 nil). Hydrochloric acid, 1 M. Preparation of dansylated aniline stock solution chloride in a 100-ml calibrated flask. ture. to volume with acetone and store in the dark at 0 "C. aniline stock solution has been found to be stable for a t least 3 months. Add 1 ml of the aniline in acetone solution to 5 ml of borate buffer and 10 ml of dansyl Shake and leave for 4 h in darkness at room tempera- Make up Under these conditions the dansylated Add 1 ml of 1 hf hydrochloric acid to hydrolyse excess of dansyl chloride. Preparation of dansyl 2-aminobutane standards Mix 2-ml aliquots of the different standard solutions of 2-aminobutane with 2 ml of borate buffer and 2 ml of dansylated aniline stock solution, as an internal standard, in a 10-ml calibrated flask.Add 2 ml of dansyl chloride solution and leave for 1 h at room temperature in darkness. Destroy excess of chloride by the addition of about 0.1 ml of 1 ;2.1 hydrochloric acid and make up to volume with acetone. Dansylation of samples of dansyl aniline to each as an internal standard. Chromatography Set any suitable isocratic constant-flow HPLC system to deliver a de-gassed solvent mixture of methanol (doubly distilled or ultra-pure grade) and water, 65 + 35 O,(-,V/V, at a flow-rate of 0.8 ml min-l to a 10 cm x 4.6 mm i.d. column packed with Hypersil 5 SAS. Inject 1-20 p1 aliquots of standards and samples. Measure the fluorescence of eluting peaks with a suitable fluorescence detector. A Gilson Spectra GLO fitted with a broad-band 390 mm cut-off input filter and a 460 nm cut-on output filter was used in this work.Feed the output signal to a recording integrator or recorder. Under these conditions dansylated aniline elutes at 3.2 min and dansyl2-aminobutane at 3.9 min. Area counts or peak heights obtained for the samples are compared with those for 2-aminobutane standards using internal standards correction. Derivatise 2-ml aliquots of the sample distillates as in the previous section, adding 2 ml Calculation A p x Ns x V, x 100 A s x N , x I x W x M- 2-aminobutane/mg k g ' = where A , = area count (peak height) for potato sample; A , = area count for 1 ng of 2-aminobutane standard; N , = area count for aniline in the standard; NP = area count for aniline in the sample; VR = nominal volume of derivatised solution in millilitres; I = injection volume in microlitres; W = mass of potato sample in grams; and .V = volume of sample distillate taken for derivatisation in millilitres. Results and Discussion Two solutions containing 103 and 51.5 pg ml-l of 2-aminobutane were derivatised as described below.Aliquots of these solutions were taken during the reaction period and injected directly into the chromatograph. After 2 h a further 1 ml of dansyl chloride solution was added and further aliquots were taken for analysis. The amount of derivatised product reached a maximum value within 1 h and then remained constant. The addition of further dansyl chloride did not increase the relative size of the derivative peak, taking into account changes in total solution volume, compensated for by use of the internal standard.A logarithmic plot of detector response against amount of dansyl 2-aminobutane (Fig. 2) shows that a linear calibration was obtained over the range 0.5-200 ng of dansyl 2-aniinobutane The rate of formation of the dansyl derivative of 2-aminobutane is shown in Fig. 1.December, 1980 2-AMIBOBUTAXE IN POTATOES BY HPLC 1173 m 80 E * 60 3 Q m +d .E 40 - 2 20 1 2 3 Time/h Fig. 1 . Rate of dansylation of 2-aminobutane solution a t room temperature: A, 103 pg ml-*; and B, 51.6 pg ml-l. using either peak-area or peak-height measurements. Although no absolute standards were available, the evidence presented in Figs.1 and 2 indicates that all the amine has reacted, or at least a very consistent percentage of it. Both the dansyl derivatives of 2-aminobutane and that of aniline were stable for several months when stored at 0 "C in the absence of light. When two derivatised solutions from distilled potato extracts were stored for 3 months at 0-2 "C in darkness, the peak-area ratio of dansyl 2-aminobutane to internal standard changed by only 1 and 40/, respectively, over the period. u 5 105 ou m 2 1 10 100 Amount of 2-amino- butane dansylatedhg Fig. 2. Relationship between detector response and amount of 2-aminobutane dansylated: A, peak area; and B, peak height. A print-out obtained from a computing integrator of a typical chromatogram for a derivatised sample of distilled potato extract is shown in Fig.3. The printed report gives peak retention times (RT), the method of integration used (Type), peak area, a peak- reference number (ID) and the residue of 2-aminobutane in potatoes in milligrams per kilo- gram (AMT). No additional peaks, other than that due to the 2-aminobutane derivative, are seen when compared with the pattern obtained with a sample of untreated potatoes. The peak at 2.19 min produced by both treated and untreated samples was shown to correspond, in retention time, to the dansyl derivative of ammonia (l-dimethylaminonaphthalene-5- sulphonamide). The peak heights shown represent injection of approximately 100 ng of aniline internal standard (3.21 min) and 20 ng of 2-aminobutane (3.94 min).1174 SCUDAMORE : DETERMIXATION OF Analyst, Vol.105 7 Inj. I I - /- * 2.19 - - 3.21 3.94 7 ISTD RT Type Area ID AMT 2.19 IM 815356 3.21 T 102618 1 3.94 TM 139162 2 30.96 Total 30.96 Fig. 3. Reversed-phase liquid chromatography of dansylated 2-amino- Internal standard (ISTI)), butane (3.94) extracted from fumigated potatoes. dansylated aniline (3.21). Although the method as described will determine residues below 1 mg kg-l, sensitivity may be further increased, if necessary, by using a larger sample of potato, e.g., lOOg, by increasing the volume taken for derivatisation up to 5 ml or by evaporating the derivatised solution to dryness and re-dissolving in as little as 200 pl of methanol. Recoveries from standard solutions of 2-aminobutane through the distillation and derivatisation stages averaged 95% (Table I).This rate of recovery remained unaffected by the presence of 50 g of chopped potatoes in the distillation flask. On several occasions a further 30 ml of distillate were collected but no 2-aminobutane could be detected in this, confirming that all the compound is distilled over in the initial 70 ml, as outlined in the procedure. In further recovery experiments, 2-5 ml of 2-aminobutane in water were dripped on to 50 g of chopped potatoes in 500-ml distillation flasks. The flasks were stoppered and stored for 24 h a t room temperature, after which time no excess of liquid was visible. The standard reagents were then added and the 2-aminobutane was distilled out and determined. No drop in recovery was recorded at the 5 mg kg-1 level and above (Table I).However, recovery wa5 somewhat lower a t the 1 mg kg-l level but this might be due in part to loss of 2-aminobutane from the TABLE I RECOVERY OF 2-AMINOBUTANE BY DISTILLATION 2-Aminobutane Residue added/mg equivalent/mg kg-l Recovery, 76 Mean, % Standard solutions- 5.15 103 4.65 3 2.06 43 1.86 37 0.93 18.6 Standard solutions -+ potatoes- 4.65 93 0.93 18.6 0.51 10.2 0.256 5.1 0.068 1.4 Treated fiotatoes- 4.65 93 0.256 5.1 0.068 1.4 95.7, 92.0 96.2. 99.0 101.0, 94.8, 91.8 96.5, 99.9, 94.8 85.3, 89.5, 94.5 Mean of 13 samples 98.5, 93.4, 92.4 99.9, 99.0, 90 2 92.2, 95.1, 96.5 99.9, 98.9, 93.2 94.7, 87.6, 88.2 Mean of 15 samples 99.2, 93.5, 93.5 96.7, 95.1, 93.6 88.1, 84.0, 82.6 Mean of 9 samples 93.8 95.9 97.6 97.3 89.6 94.9 94.8 96.4 94.6 97.3 90.2 94.6 95.9 95.1 84.9 92.0December, 1980 2-AMINOBUTAKE I N POTATOES BY HPLC 1175 apparatus during the treatment period.Typically, a potato sample could be prepared, distilled, derivatised and chromatographed in about 3 h and in practice 4-5 samples could be handled together. Table I1 gives results showing the disappearance of 2-aminobutane from two 30-ton bulks of treated potatoes stored below 10 “C for 8 months in two metal bins. Samples were taken from three positions in each bin. In bin A the residue dropped from 24 to 2 mg kg-1 but in B only from 28 to 9 mg kg-l. However, in A intermittent through-ventilation was carried out while in B a system using continuous re-circulation of air was in operation. TABLE I1 LOSS OF 2-AMINOBUTAKE RESIDUES FROM POTATOES DURING STORAGE Sampling Bin position A 1 A 2 A 3 B 4 B 5 B 6 Residue/mg kg-I - October June 27 2 18 1 28 3 27 11 29 7 28 9 During development of the method both normal-phase and reversed-phase systems were used.The short alkyl chain reversed-phase packing material was found to give better separation of the dansyl 2-aminobutane peak from dansyl aniline than on the octadecasilyl packing materials used. \Vith normal-phase operation the elution order of these peaks is reversed. Reversed-phase separations were found to be more reproducible and did not suffer from the interference experienced with normal-phase packings owing to the late emergence of the peak for the sulphonamide derived from ammonia present in all distilled potato samples. Columns were found to have a long life, but were occasionally flushed out with 1000,o methanol to remove any impurities that had accumulated on the packing.Although dansyl derivatives also absorb in the ultraviolet region, fluorescence detection is more selective. It also gives a better signal to noise ratio and hence better sensitivity. Conclusions The analytical technique described provides a quick and sensitive method for determina- tion of 2-aminobutane in distillates from potatoes. The dansylation of 2-aminobutane in acetone and water solution a t pH 9.0 has been shown to be completed in 1 h a t room temperature, providing the basis for a convenient analytical method for studying residues in potatoes after treatment. References 1. 2. 3. 4. 5 . Eckert, J . W., World Rev. Pest Cowtrol, 1969, 8, 116. Graham, D. C . , and Hamilton, G. d., Nutuve (London), 1970, 227, 297. Graham, D. C., Hamilton, G. A , , Quinn, C. E., and Ruthven, A. D., Potato Res., 1973, 16, 109. Graham, D. C., Hamilton, G. A, Xash, M. J., and Lennard, J . H., Potato Res., 1973, 16, 234. Ministry of Agriculture, Fisheries and Food and Department of Agriculture and Fisheries for Scotland, “The Safe Use of 2-Xminobutane for Fumigation of Potatoes,” Ministry of Agriculture, Fisheries and Food (Pesticides Branch), Great Westminster House, Horseferry Road, London, SWIP 2AE. 1977. 6. 7. Graham, D. C., Hamilton, G. A , , Quinn, C. E., and Ruthven, A. D., Potato Res., 1975, 18, 410. Day, E. \V., Holzer, F. J . , Tepe, J , B., Eckert, J. W., and Kolbezen, A L J , , J . Assoc. Off. A n a l . Chem., 1968, 61, 39. Received February 2nd, 1980 Accepted June 9th, 1980

ISSN:0003-2654    

DOI:10.1039/AN9800501171

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

1176 A?Lalyst, December, 1980, Vol. 105, $9. 1176-1181 Pyrolysis - Gas Chromatography of Proline, Hydroxyproline and Related Peptides* Roger M. Smith, Sanaa A. K. Shawkat and William P. Hayes Department of Chemzstry, Unzverszty of Technology, Loughbouough, Lezcester ,hive, LEI 1 3TL On pyrolysis using a Curie point analyser at 770 "C proline and hydroxy- proline (cis- and trans-) yielded as their major products a previously un- reported fragment, tentatively identified as 1-pyrroline, and pyrrole, respec- tively. h number of dipeptides and a tripeptide containing these hetero- cyclic amino acids were also examined. All the degradations were found t o be dependent on the pH of the solution that was examined. Keymords : Prolane ; hydroxyproline; pyvolysis - gas chroinatogra,phy; di- peptides Pyrolysis - gas chromatography is increasingly being used to compare peptides and proteins as it is a convenient, direct method, avoiding the hydrolysis and derivatisation required by conventional gas - liquid chromatography (GLC) .Is2 The pyrograms formed are often complex and are used as fingerprints, few of the components being identified.As part of an investigation to examine the feasibility of identifying the amino acids in a peptide from characteristic pyrolysis fragments, we have previously examined histidine and 3-methyl- histidine.3 In this paper we report our studies on proline, cis- and trans-hydroxyproline and on a number of peptides containing these amino acids, and show that the free amino acids have distinctive pyrograms, contrary to earlier reports.These amino-acids are of particular interest because of the presence of a high percentage of hydroxyproline in collagen and as a marker of the metabolic disorders hydroxy- prolinaemia and hyperprolinaemia. As a result a range of alternative analytical methods have been reported including spectr~photometry~~~ and gas - liquid chromatography following derivatisation.6 A number of previous workers have examined the pyrolysis of proline and hydroxyproline using different conditions. The first studies were by Winter and Albro, who used filament pyrolysis and reported peaks from proline corresponding to ammonia, methyl or dimethyl- amine and tripropylamine or benzene.' In 1964 Simon and Giacobbo found pyrrolidine to be the major product from proline,8 although in a later paper both proline and hydroxy- proline were reported to give similar pyrograms at 700 "C with pyrrole as the major p r o d ~ c t .~ This conclusion was confirmed by Vollmin et al. in a pyrolysis - GLCjMS study.10 If large sample sizes were used N-alkylpyrroles were also formed. However, in 1967 Merritt and Robertson, again using GLC - hIS (gas - liquid chromatography - mass spectrometry) and a Barker - Coleman pyrolyser, reported the unique products to be pyrrole and N-methyl- pyrrole from proline and hydroxyproline, respective1y.l' Some previous work has also been carried out on proline containing dipeptides, including glycylprolinelo+ and prolylphenylalanine,lO and S l a ~ k ~ ~ p l ~ reported that low temperature (280-320 "C) pyrolysis of collagen gave a large pyrrole peak due to the presence of proline and hydroxyproline.Experimental Apparatus Pyrolyses were carried out by using a Pye Unicam 104 gas chromatograph fitted with a Philips PV 4080 Curie point pyrolysis unit; except where stated iron pyrolysis wire (main- tained a t 770 "C for 5 s) was used. The fragments were separated on glass columns (1.5 m x 4 mm i.d.) packed with either Chromosorb 103 (100-120 mesh) a t 155 "C or 10% Carbo- wax 2031 with potassium hydroxide on Chromosorb W (100-120 mesh) a t 75 "C. The carrier gas was nitrogen, flowing a t a rate of 60 ml min-l, and the peaks were detected by using a flame-ionisation detector a t an attenuation of x 500. * Presented at the 5th SAC International Conference on Analytical Chemistry, Lancaster, July 20-25th, 1980.SMITH, SHAWKAT ASD HAYES 1177 Pyrolysis - mass spectra were measured on a direct-insertion probe Curie-point pyrolyser with VG mass spectrometer system.Reagents Amino-acids a d di$e$tides. Company Ltd., Poole, Dorset. The materials were supplied by Sigma (London) Chemical Procedure The pyrolysis wires were cleaned before use by heating to a red heat. The standard poly- (methyl methacrylate) in acetone (1 or 2 p1, O.lyo m/V) and samples in aqueous solutions (1% m/V) were applied by use of a microlitre syringe and dried in air. The pH values of the aqueous solutions were adjusted to 2 or 9 (pH paper) before application by using 1 hi hydrochloric acid or aqueous sodium carbonate solution (10%). Results and Discussion Trial pyrograms suggested that although hydroxyproline yielded pyrrole, proline gave a very different pattern with a major peak similar to pyrrolidine.As these results differed from previous reports, it was decided to undertake a detailed study of these two amino acids and related compounds. In order to be able to make accurate retention time comparisons with possible pyrolysis products it was necessary to use a pyrolysable internal standard, which would yield a single major product that would be used in direct chromatograms. After a number of trials poly(methy1 methacrylate) was selected, methyl methacrylate, its single major pyrolysis product,I4 having a suitable retention time on Chromosorb 103. However, it could not be dissolved in the aqueous solutions used to apply the amino acids to the Curie point wire, and was therefore applied separately in acetone.The relative retention times of methyl methyacrylate and pyrrole, and a number of related compounds, are given in Table I. In all the following examples runs were carried out both with and without the standard in order to check that peaks were not being masked. In the limited studies using Carlowax 20M with potassium hydroxide the pyrrole peak had such a long retention time that it could not be readily observed. Although initial studies made use of aqueous solutions of the amino acids (denoted by pH 7 in the tables) we found that adjustment of the pH to approximately 2 or 9 with hydro- chloric acid or sodium carbonate solution, respectively, usually resulted in more reproducible chromatograms, particularly with the dipeptides.TABLE I RETENTION TIMES OF POTENTIAL PYROLYSIS FRAGMENTS FROhl PROLINE AND HYDROXYPROLINE Relative retention times (relative t o methyl methacrylate). Compound Acetaldehyde . . Acetone . . . . Acetonitrile . . Pyrrolidine . . 3-Pyrroline . . (1-Pyrroline) * . . 3-Hydroxypyrrolidine Methyl methacrylatet Pyrrole . . . . Column A I Chromosorb 103, Carbowax 20M - KOH, 165 "C 75 "C . . . . 0.13 0.23 . . . . 0.28 0.37 . . . . 0.37 1.02 . . . . 1.00 0.83 . . . . 1.02 1.26 . . . . 1.12 - . . . . 1.91 6.27 . . . . 3.65 - . . . . 1.00 1.00 * Major fragmentation peak from proline; tentative identification. f Typical retention times: Chromosorb 103 6.9 min and Carbomax 20M - potassium hydroxide 5.4 min.1178 SMITH et al.: PYROLYSIS - GAS CHROMATOGRAPHY OF TABLE I1 AFzalyst, Vol. 105 PYROGRAMS OF PROLINE AND Cis- AND tYanS-HYDROXYPROLINE ON CHROMOSORB 103 AT 155 "c Peak areas of major components t o pyrrole* (relative retention timet) - 7 Sample Proline . . . . trans-H ydroxyproline cis-Hydroxyproline . . Proline + trans- Polyhydroxyproline hydroxyproline (1 : 1) . . Approximate Number PH of runs 2 4 7 4 9 4 2 3 7 6 9 3 7 7 Acetonitrile Unknown Pyrroline Pyrrole (0.37) (0.54) (1.12) (1.91) . . 7 1 7 - 81 100 2 1 100 3 100 . . 2 3 11 (5) 9 3 > 20 7 3 4 (1) 1 10 (2) 100 * Pyrolysis at 770 "C for 5 s (standard deviation). t Relative to poly(methy1 methacrylate) pyrolysate. Typical retention time 6.9 min. The pyrograms of a number of runs of proline and cis- and trans-hydroxyproline are reported in Table I1 with typical examples given in Fig.1. Pyrrole presumably formed by dehydration and decarboxylation was the major product from the hydroxyprolines and the orientation cis- or trans- of the hydroxy and carbonyl groups did not appear to affect the results, suggesting a lack of interaction. In all the later studies only the more readily available trans- form was used. The pyrogram of proline was very different, pyrrole being a minor component (Fig. 1 (a)). The major peak (RRT 1.12) was unexpectedly not pyrrolidine Frelative retention time (RRT) 1.001, nor the commercially available 3-pyrroline (RRT 1.02), and was clearly resolved from the standard peak. We can postulate that the expected product should be 1- or 2-pyrroline, formed by decarboxylation.However, both of these compounds are unstable as 2-pyrroline spontane- ously rearranges to the 3-isomer and the 1-isomer condenses to a trimer in the liquid phase15; 0 15 0 15 Timeim i n Timeimin Pyrograms of: (a), L-proline at pH 2 ; ( b ) , L-hydroxy- Sample (20 pg) pyrolysed a t 770 "C for 5 s and Position of Fig. 1. proline at pH 2. products separated on Chromosorb 103 a t 155 "C. standard peak from poly(methy1 methacrylate) marked by S.December, 1980 PROLIXE, HYDROXTPROLINE AND RELllTED PEPTIDES 1179 thus, direct comparisons cannot be made. Electron impact mass spectrometry (EIMS) of proline yields a major fragment ion at wz/z 70 corresponding to the loss of CO,H but no peak at ~ n j z 69 for a pyrroline,16 although Posthumus has suggested that EIMS may be a poor guide to pyrolysis fragmentation pathways.17 However, pyrolysis - mass spectrometry using a Curie point pyrolyser in the direct inlet source yielded a spectrum with m/z 69 (30%), 68 (219/0), 67 (50%) and 41 (looyo) as the major peaks, suggesting the possible presence of a pyrroline.Because of the instability of the %isomer to isomerisation, this peak and hence the pyrolysis - gas chromatographic product was tentatively assigned to 1-pyrroline. The use of the solid chromatographic phase Chromosorb 103, rather than the more usual liquid phase, may be one reason why this peak from proline has not previously been observed. Confirmatory studies using a Carbowax 2011 - potassium hydroxide column (Table I) were inconclusive because although 3-pyrroline (RRT 1.26) and pyrrolidine (0.83) were now separated, the only major pyrolysis peak (1.00) was unresolved from that of acetonitrile (1.02), which was also present in the pyrolysis.The pH of the sample had little effect on the hydroxyprolines but low pH increased the proportion of 1-pyrroline from proline. The proportion and yield of the two major products from proline and hydroxyproline a t different pyrolysis temperatures were determined (Table 111). This determination showed a maximum yield a t 770 “C, the highest tempera- ture examined. The decrease in yield of the standard a t the higher temperatures is in agreement with previous reports.14 Although most amino acid pyrolysis studies have been qualitative, the potential of this method for quantitative analysis was examined using the internal standard as reference (Table IT,.). The major peaks, 1-pyrroline from proline and pyrrole from hydroxyproline, showed a reasonable correlation with relative peak areas and suggest that with care direct pyrolysis could give satisfactory results but that further work needs to be carried out to evaluate the effects of matrices and sample application techniques.In order to be able to use this technique to examine the composition of peptides it is necessary to determine whether the same pyrograms are obtained when these amino acids are pyrolysed as a mixture of individual compounds or when bound by peptide links. .ilthough a mixture of proline and hydroxyproline gave a pyrogram very similar to the sum of their individual patterns, the peak assigned to the 1-pyrroline was very small in the pyrogram of prolylhydroxyproline (Table 11).This is a surprising result as the prolyl ring is unsubstituted and might have been expected to behave as in the free acid. Polyhydroxy- proline behaved very much as the free amino acid (Table 11). In order to explore these relationships further a series of dipeptides and one tripeptide with glycine and alanine were studied (Table V). In separate studies neither glycine nor alanine yielded peaks on pyrolysis with retention times longer than that of acetonitrile. TABLE I11 VARIATION OF PEAK AREAS OF THE hlAJOR COMPONBNTS O F PROLIKE AND HYDROXYPROLIKE WITH PYROLYSIS TEMPERATURE Peak areas relative to the standard poly(methy1 methacrylate) peak a t 770 OC* Sample Pyrolysis temperature/”C Standard 1-Pyrroline Pyrrole A f > Proline .. .. . . . . tuans-Hydroxyproline 480 510 610 770 358 480 510 ~~ 610 770 52 17 2 105 27 3 109 146 14 100 202 35 4 1 24 79 4 53 90 5 61 .~ 101 IS 133 100 18 191 * 1 pg of poly(methy1 methacrylate) and 20 p g of amino acid, pH 7.1180 SMITH et al. : PYROLYSIS - GAS CHRORIATOGRAPHY OF Analyst, VOL. 105 TABLE IV QUANTITATIVE ANALYSIS OF PROLINE .4ND HYDROXYPROLINE* Proline: tvans-Hydroxyproline : Amino acidlpg peak area ratio 1-pyrroline to standard peak area ratio pyrrole to standard 0 0 0 5 - 1.21 5 - 1.35 10 2.65 2.09 10 2.52 1.33 15 3.85 4.24 15 4.49 4.54 20 5.57 5.31 20 4.87 5.31 Slope 0.265 0.275 Correlation coefficient 0.991 0.971 Standard deviation 0.281 0.504 Intercept 0.007 -0.208 * Pyrolysis at 770 "C for 5 s and separation on Chromosorb 103; standard polyjrnethyl methacrylate) (1 r g ) .Glycylhydroxyproline yielded only pyrrole but the proline-containing dipeptides showed variable results, depending both on the position of the linkage and the pH of the solution, Glycylproline, alanylproline and glycylprolylalanine gave low yields of I-pyrroline. In contrast, pyrolysis of prolylglycine and prolylalani~ie yielded a significant peak for 1 - pyrroline. The formation of 1-pyrroline from these dipeptides was enhanced by low pH in a similar way to the fragmentation of proline (Table 11). However, the results showed great variability, which was unrelated to sample storage or drying time on the wire before pyrolysis. A consistent feature of all the peptide pyrograms was an extra peak (RRT 0.54) which could not be identified.TABLE V HYDROXYPROLINE ON CHROMOSORB 103 AT 155 "C PYROGRAMS OF DIPEPTIDES AND TRIPEPTIDES CONTAISING PROLINE AND Peak areas of major components relative to pyrrole* (relative retention timet) I . -___ Approxi- Number ilcetonitrile Unknown Unknown 1-Pyrroline Pyrrole Sample mate pH of runs (0.37) (0.54) (1.00) (1.12) (1.91) Prolylglycine . . . . Glycylproline . . . . Glycylhydroxyproline . . Prolylalanine . . . . Alanylproline . . . . Glycylprolylalanine . . 2 9 2 9 9 2: 4 3 6 4 4 4 3 4 3 4 2 5 84 (27) 63 (15) 39 (10) 26 (2) > 64 26 (10) 102 (27) >37 > 100 >65 > 200 > 200 18 f7\ 4 jij 41 (34) 15 (2) 63 (10) 28 (32) 73 28 (14) 102 (64) 17 (13) 13 (10) 11 (7) 0 0 140 (98) 78 (74) 16 (9) 9 ( 7 ) [521 7 (3) 100 100 100 100 100 100 100 100 100 100 100 100 * Pyrolysis at 770 "C for 5 s (standard deviation).3 Results poorly reproducible, peak at a relative retention time of 1.10 has an anomalous shape and is Relative t o methyl methacrylate. possibly not due to 1-pyrroline. Thus, it appears that although the ratio of 1-pyrroline to pyrrole could be used to deter- mine the proportion of proline in a mixture of free amino acids, once the proline is bound in a peptide its individuality is lost unless it is a t an N-terminus and a low pH is used. We thank Mr. G. Solabi for initial trials and Mr. B. B. Wheals and Nr. I. Jane of the Metropolitan Police Forensic Science Laboratory for pyrolysis - mass spectra.December, 1980 PROLISE, HYDROXYPROLINE AND RELATED PEPTIDES References 1 .2. 3. 4. 5. 6 . 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 1181 Irwin, W. J . , J . Anal. A p p l . Pyrolysis, 1979, 1, 1 . May, R. LV., Pearson, E. F., and Scothern, D., “Pyrolysis - Gas Chromatography,” Analytical Smith, R. M., Solabi, G. A., Hayes, LV. P., and Stretton, R. J . , J . Anal. Appl. Pyrolysis, 1980, 1, Science Monograph No. 3, The Chemical Society, London, 1977. 197. Bergman, I., and Loxley, R., Anal. Chem., 1963, 35, 1961. Rergman, I., and Loxley, R., Analyst, 1969, 94, 575. IkIacKenzie, S. L., and Tenaschuk, D., 1. Chromatogr., 1975, 104, 176. Winter, L. N., and Mbro, P. W., J . Gas Chromatogr., 1964, 2, 1 . Giacobbo, H., and Simon, W., Pharm. Acta Helv., 1964, 39, 162. Simon, W., and Giacobbo, H., Angew. Chem. Internat. Edn., 1965, 4, 938. Vollmin, J , , Kriernler, P., Omura, I., Seibl, J., and Simon, W., .Wicrochem. J . , 1966, 11, 73. Merritt, C., and Robertson, D. H., J . Gas. Chromatogr., 1967, 5, 96. Slack, M. V., Biochem. J . , 1965, 96, 56P. Slack, M. V., J . Gas. Chromatogv., 1967, 5, 22. Lehrmann, L. A,, and Brauer, G. H., Anal. Chem., 1961, 33, 673. Coffev. S.. Editor. “Rodd’s Chemistrv of Carbon ComDounds.” Second Edition, Volume IVA. Elsevier; Amsterdam, 1973, p. 368. ‘ Junk, G., and Svec, H., J . A m . Chem. Soc., 1963, 85, 839. Posthumus, M. A , , and Nibbering, N. M. M., Ovg. Mass. Spectrom., 1977, 12, 334. Received June 18th, 1980 Accepted August 4th, 1980

ISSN:0003-2654    

DOI:10.1039/AN9800501176

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

1182 Analyst, December, 1980, Vcd. 105, $9. 1182-1187 Use of a Sulphur Dioxide Gas Sensing Membrane Probe in Wines and Juices at Their Natural pH D. J. Spedding and G. M. Stewart Chemistry Department, University of Auckland, Auckland, New Zeala+zd h sulphur dioxide gas sensing membrane probe has been used to determine the concentration of “free” sulphur dioxide in aqueous solutions of pH 3-4 in the concentration range 10-2-10-4 mollkl. The accuracy and repro- ducibility suggested that this was a new non-destructive method suitable for the analysis of sulphur dioxide in juices and beverages. Keywovds Sulphur dioxide deternzination ; gas sensing membrane pvolje ; juice and beverage analysis ; wine analysis The theory and use of gas sensing membrane probes for the determ:ination of sulphur dioxide in solution has been previously described.192 These probes use a membrane permeable to sulphur dioxide to separate a hydrogen ion sensing electrode immersed in an aqueous hydrogen sulphite solution from the aqueous sample under test.In order to elevate the concentration of sulphur dioxide in solution to its maximum, measurements are usually made a t or below pH 1. This is because of the equilibria associated with the hydrolysis of sulphur dioxide in aqueous solutions [equations (1)-(3)] Fig. 1 includes a diagrammatic representation of the proportion of the total sulphur species present that exist as molecular sulphur dioxide (H,O.SO,) a t vairious solution pH values. For clarity, the sum of the concentrations of all sulphur species represented by equations (1)-(3) will be referred to as the S(1V) concentration.As sulphur dioxide is a major food preservative most of the expected applications of the probe are for sulphur dioxide determinations in foodstuffs and be~e:rages.~9~ The adjustment of the pH to very low values makes this, of necessity, a destructive analytical method. A sulphur dioxide sensing membrane probe offers the possibility of continuous, automatic monitoring of free sulphur dioxide in a flow system through which a beverage is being passed during its preparation. This in turn allows a constant concentration of free sulphur dioxide in the beverage to be maintained by automatic sulphur dioxide additions to the beverage flow in response to the probe signal. Measurements in the absencse of added acid are there- fore essential if they are to be made within the beverage flow itself.It should be emphasised that this investigation refers only to the possibility of measuring free sulphur dioxide in beverages. The techniques for measuring bound sulphur dioxide and hence total sulphur dioxide are destructive to the sample and cannot be made otherwise by the proposed use of the probe. The use of the sulphur dioxide probe in basic research involving aqueous sulphur dioxide solutions would be enhanced if the solution pH did not require adjustment prior to measure- ment. There is no theoretical reason why the sulphur dioxide probe may not operate at pH values in excess pH 1, the only limitation being the concentration of free sulphur dioxide in solution.Guthrie5 has calculated that a solution 1 x mol 1-1 in hydrogen sulphite contains 8 x 10-4mo11-1 of free sulphur dioxide at pH 11, which decreases to 0 x 10V mol 1-1 at pH 4. A typical lower limit of detection for sulphur dioxide with the probe is about 10-6 moll-l.l An upper limit of solution pH for sulphur dioxide detection in many samples is probably 4 although’ this may be higher at high S(1V) concentrations. Many of the beverages and fruit juices for which the sulphur dioxide probe could be usedSPEDDING AND STEWART 1183 300 > E 1 285 6 0 n .- +- 270 Fig. 1 . Effect of pH changes on a sodium sulphate - sodium sulphite solution 0.095 moll-' in S(1V) at 33 "C (for details see text). A, Proportion of S(1V) solution present as H,O.SO, (scale on left side of figure) and B, sulphur dioxide electrode potential (scale on right side of figure).have a pH in the range 3-4; this upper pH limit therefore does not exclude these major areas of use. This paper describes experiments to determine the potential of a commercial sulphur dioxide sensing membrane probe for the determination of sulphur dioxide in aqueous solutions with pH values up to 4. Experimental Apparatus The sulphur dioxide sensing membrane probe used was an Orion sulphur dioxide electrode, Model 95-64. This was connected t o a Radiometer 26 pH meter having an expanded scale capable of being read to +0.5 mV when operated in the millivolt detection range. All determinations were performed on 50-ml aqueous samples in a 100-ml beaker, the contents being stirred by means of a magnetic stirrer with a PTFE-coated follower.A Radiometer 28 pH meter with a combination pH electrode (Jena, Model N61) was used for all pH measurements. Chemicals Solutions of sulphur dioxide were prepared fresh daily from analytical-reagent grade sodium sulphite, which was standardised against a standard iodine solution. All other chemicals were of analytical-reagent grade and aqueous solutions were prepared using distilled water. Results and Discussion Effect of pH in distilled water. A solution approximately 0.1 moll-1 in S(1V) was prepared by dissolving sodium sulphite To 50 ml of this solution was added 1.9 g of anhydrous sodium sulphate.1184 SPEDDING AND STEWART: SULPHUR DIOXIDE GAS SEKSING Analyst, Vol. 105 The pH and potential of the sulphur dioxide probe were measured continuously as sulphuric acid was added dropwise to give the graph illustrated in Fig.1. It can be seen that the response of the probe closely parallels the concentration of molecular sulphur dioxide in solution. A more detailed graph for the pH range 2.5-5 (Fig. 2) indicates an essentially 270 - A A A PH Fig. 2. Sulphur dioxide electrode potential versus solution pH for a 0.074 moll-' S(1V) solution a t 25 "C. linear relationship between potential and pH up to pH 4 in these solutions. Theory predicts1 a linear relationship between potential and the logarithm of the sulphur dioxide concentration. When the free sulphur dioxide concentration a t each pH in the S(1V) solution shown in Fig. 2 is calculated this is found to be so (Fig.3). The work described here was initiated to test the efficiency of the sulphur dioxide probe in the study of oenological problems. As wines are essentially aqueous solutions containing A 270 280 290 PotentiaVrnV Logarithm of the concentration of H,O.SO, in a 0.074 moll-' S(1V) solution a t different solution pH versus sulphur dioxide electrode potential a t 25 "C. Fig. 3.December, 1980 1185 approximately loo/, of ethanol, such a solution was used in place of distilled water. Satis- factory probe responses were obtained with solutions of pH 2 4 , containing or I t was also found that the addition of sodium sulphate was unnecessary, the osmotic properties of the ethanolic solutions being suitable for adequate operation of the probe. Variations in the standard graph were less than 5% for measurements in aqueous solutions containing between 8 and 12% of ethanol.Response at Constant pH Aqueous solutions containing 10% of ethanol were adjusted to pH 3.0 or 4.0 by the addition of sulphuric acid. Weighed portions of solid sodium sulphite were added, the pH re-adjusted to the same value and the potential of the sulphur dioxide probe recorded. Graphs of these potentials against the logarithm of the S(1V) solution concentration are shown in Fig. 4. I t can be seen that these graphs are linear and that determination of S(1V) in solutions of pH as high as 4 is possible. MEMBRANE PROBE IX WINES AND JUICES AT THEIR NATURAL PH moll-' of S(1V) and 10 5 2% of ethanol. I I 25 30 40 50 PotentiaVmV Logarithm of the concentration of S(1V) in a 10% ethanol - water solution versus sulphur dioxide electrode potential a t 25 "C when the solution pH was (a) 4.0 and ( b ) 3.0.Fig. 4. Effect of Temperature Changes Temperature effects on the potential measured by the sulphur dioxide probe have been discussed in detail by Bailey and Riley.2 In this work their recommendations were followed and both sample and standard measurements were made at the same temperature. Solution temperatures ranged from 20 to 35 "C with the majority of measurements being made a t about 25 "C. Response Time The response time of a gas sensing membrane probe is an important consideration for its routine laboratory use. Ross et a1.l have suggested that an adequate electrode would reach 99% of its final reading in 2-3 min, after a 10-fold increase in concentration of the species sensed. When operating in solutions of pH 3 4 our electrode required about 5 min for such a change, but never more than 10 min elapsed before a steady potential was attained.A response time of 2-3 min for a sulphur dioxide probe in solutions as dilute as lop4 mol 1-1 at pH 1 has been f0und.l I t is likely that the longer response time found at pH 3 4 is directly related to the lower concentration of molecular sulphur dioxide in solution. Whether or not the longer response time a t higher pH is of any consequence in routine laboratory use will depend upon the alternative methods available for the determination.1186 SPEDDING AND STEWART: SULPHUR DIOXIDE GAS SENSIXG Analyst, VOl. 105 Accuracy and Reproducibility The use of the sulphur dioxide probe for the determination of sulphur dioxide in S(1V) solutions is dependent upon the reproducibility of the electrode potential with changing pH while using the same standard S(IV) solution.In order to test this reproducibility, solutions containing mol 1-1 of S(1V) were prepared and a graph of electrode potential veysus pH was plotted for each solution. A straight-line graph was obtained for each solution by linear regression analysis and the 95% confidence shell was constructed about each line (Fig. 5). After 4 h, fresh S(1V) solutions of the same concentration were prepared and a further scan of electrode potential versus pH for each solution was obtained. Fig. 5 shows that each of the new measured points falls within the 95% confidence (twice the standard error) shell about the original data obtained 4 h previously.At the 95% confidence limit, therefore, the standard graphs are reproducible. and 420 - 400 - > E 1 .- w W 4- n 380 - 360 - L / __- PH 3.0 3.5 4.0 Fig. 5. Sulphur dioxide electrode potential at 25 "(3 zlei'sus solution pH for A, 10-4 and B, lO-~rnoll-' of S(1V). Initial readings are represented as a straight line, obtained by linear regression analysis, surrounded by the 95% confidensce limit shell. Readings obtained 4 h later are indicated by v. Determination of Sulphur Dioxide in 10% Ethanol Solutions of pH 3-4 Initially the pH and potential of the sulphur dioxide probe in the unknown sample were measured. An estimate of the sulphur dioxide concentration in the sample was made and two fresh standard S(1V) solutions were prepared to bracket the estimated concentration.Each of these standards was separately tested for pH and sulphur djoxitle electrode potential, as the pH was adjusted by stepwise addition of concentrated sulphiiric acid, so that the unknown solution pH was bracketed. A graph of potential v e ~ s u s pH for each of the standards was plotted [Fig. 6 ( a ) ] , enabling the potential of each standard a t the unknown sample solution pH to be read off. Each of these potentials was plcltted veysus the logarithm of the S(1V) concentration [Fig. 6 ( b ) ] and a straight line drawn between the two points. The concentration of the unknown sulphur dioxide solution could then be obtained from the originally measured sulphur dioxide probe potential.Fig. 6 illustrates the use of this method for a wine sample of pH 3.51 with a free sulphur dioxide concentration, determined by the standard aspiration method, of 3.8 x moll-1.Decetnber, 1980 MEMBRANE PROBE IN WINES AND JUICES AT THEIR NATURAL PH 1187 3.4 3.6 3.8 4.0 PH 160 180 PotentialimV Fig. 6 . (a) Sulphur dioxide electrode potential a t 25 "C vevsus solution pH for standard S(IV) solutions of and moll-'. (b) Logarithm of S(1V) solution concentration vevsus sulphur dioxide electrode potential a t pH 3.51. The wine sample had a sulphur dioxide electrode potential of 172 mV. The method was further tested by using it to determine the S(1V) concentration arising from additions of known masses of solid sodium sulphite to a white wine.Table I shows that the sulphur dioxide electrode was able to determine the S(1V) concentration to within 5% of the known addition. Similar data were obtained with a red wine, but only after the sulphite binding capacity of the constituents of the wine had been satisfied. The sulphur dioxide gas sensing membrane probe may thus be used for the determination of S(1V) present in aqueous solutions at concentrations above 10-4mo11-1 and with pH values up to 4.0. The accuracy of the method is suitable for most beverage analyses and probably also for fruit juice analyses, but the response time may be a limiting factor in its routine operation. TABLE I ELECTRODE RESPOXSE TO KNOWN S(1V) ADDITIOSS TO A WHITE WINE S(IV) added/ Total S(1V) present/ Measured S(IV)/ moll-' x moll-1 x moll-1 x 0 2.7 2.7 3.0 5.7 5.8 5.0 7.7 7.4 6.5 9.2 8.8 8.1 10.8 10.5 The authors thank the New Zealand Department of Labour for financial assistance to G.M.S. during the course of this work. References Ross, J. W., Riseman, J. H., and Krueger, J. A,, Pure Appl. Chem., 1973, 36, 473. Bailey, P. L., and Riley, M., Analyst, 1975, 100, 145. Faulkner, S . V., Process Biochem., 1976, 47. Binder, A,, Ebel, S., Kaal, hl., and Thron, T., Dtsch. Lebensm. Rundsch., 1975, 71, 246 Guthrie, J . P., Can. /. Chem., 1979, 57, 454. 1. 2. 3. 4. 5. Received May 19th. 1980 Accepted July 3rd. 1980

ISSN:0003-2654    

DOI:10.1039/AN9800501182

出版商:RSC    

年代:1980

数据来源: RSC