ISSN:0003-2654    

DOI:10.1039/AN97500FX017

出版商:RSC    

年代:1975

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97500BX019

出版商:RSC    

年代:1975

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97500FP053

出版商:RSC    

年代:1975

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97500BP059

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

MAY, 1975 The Analyst Vol. 100, No. 11 90 Mass-spectrometric Analysis of Solutions Using an Atmospheric Pressure Ion Source A. L. Gray Applied Research Laboratories Limited, Wingate Road, Luton, Bedfordskive The use of an atmospheric pressure d.c. plasma as an ion source has been explored for the direct analysis of solutions introduced into it from a nebuliser. Ions are extracted from the plasma into a vacuum system and are focused into a quadrupole mass analyser. A high yield of singly charged ions with a small energy spread is obtained and clear spectra of the constituents of the solution are observed. The method is described and the results observed on simple solutions are given. The sensitivity of the method for a number of elements is indicated and appears to be comparable with other trace-analysis methods.The determination of trace elements in solutions by instrumental analysis has become an important requirement and a number of different methods have come into use, prominent among which is atomic-absorption spectrometry. More recently much attention has been paid to optical emission methods involving plasma excitation. Both of these methods enable low limits of detection to be achieved in routine applications. Atomic absorption, in the form in which instruments are at present marketed, is primarily a single-element technique, thus for multi-element routine analysis the use of an emission source for excitation combined with a conventional scanning or direct-reading spectrometer is attractive. A variety of plasma sources have been rep~rtedl-~ and one of these sources is now commercially marketed.* Although the high temperatures achieved in the most suitable of these plasmas lead to low limits of detection and relative freedom from inter-element effects and interferences, there are still requirements for higher sensitivity and flexibility that are not ideally met.Many of the most difficult problems that arise in trace analysis have been solved by recourse to mass spectrometry. The only suitable ion sources at present available for the determin- ation of most elements in the Periodic Table are the ion bombardment and spark sources, and in order to use either of these sources sample preparation into the preferred solid form is necessary. Although adequate methods are available for this purpose, neither source is well suited to routine analysis at large sample throughputs, and mass analysers compatible with these sources are necessarily costly.Consideration of possible ways in which sample introduction and analyser design could both be simplified while retaining the wide capability and sensitivity of the mass spectro- meter for element determinations led to the examination of the ionisation process that occurs in the atmospheric pressure electrical plasmas that were concurrently being studied as optical emission sources. The most convenient of these sources for this initial investigation was a small wall-stabilised d.c. plasma source that had been found to be very stable and reproducible as an optical source for the analysis of steel^.^ It was concluded that this plasma, when fed with the sample in a suitable form, produced substantial ion populations of the sample elements at atmospheric pressure and that if these ions could be representatively transferred to a mass analyser at its much lower operating pressure a useful technique might be developed.In particdar, the possibility of direct introduction of the sample at atmospheric pressure into the plasma, for example by means of a solution nebuliser, was thought to be especially attractive. Tech- niques of mass-spectrometric sampling of flames used in studies of combustion processes have been well established for some yearss*' and although rather higher temperatures occur in plasmas, it seemed reasonable to attempt sampling of this small d.c. plasma in the same way.This paper describes the production of ions in a plasma and the equipment used for the investigation, and presents the results obtained so far. 289290 GRAY : MASS-SPECTROMETRIC ANALYSIS OF SOLUTIONS Analyst, VoZ. 100 The Production of Sample Ions A wide variety of physical processes can be used to produce ions from sample atoms. Among these processes some of the best known are electron bombardment, photo-ionisation, field ionisation and thermal ionisation, and ion sources that involve the use of all these mechanisms are well known. In all of the sources commonly used, however, it is necessary for the sample, usually in the form of a gas, to be introduced into the vacuum system. When a solid sample is used two established methods are available for producing ions directly from it, either by bombardment by a primary ion beam, or by making the sample one electrode of a spark gap and initiating a discharge in the gap.The characteristics of the different ion source methods are extensively discussed in the literature, but it is sufficient to mention here that apart from requiring the sample to be inside the vacuum, those most useful for the analysis of solids and liquids produce ions with a wide distribution of energies and containing many ions with multiple charges. Both of these properties of the resulting ions require the use of a rather complex mass analyser system for their successful analysis, as high resolution is necessary to resolve the multiply charged ion peaks from the wanted spectrum and an energy analysing stage is required in order to restrict the energy range of the incoming ions so as to enable high resolution to be achieved.Ionisation at atmospheric pressure has recently been reported by Carol1 et aLs for organic vapours. Molecular ionisation is achieved in this source mainly by the addition or removal of protons as a result of ion - molecule reactions, and its attraction lies particularly in the high yield of M+, MH+ or (M-H)+ ions, which is obtained because of the small excess of energy available for fragmentation. Such a source, although of interest for organic applications, cannot be used for elemental analysis without some additional mechanism to enable the sample to be transformed into the dissociated vapour state before ionisation.One of the most convenient methods for vaporising and dissociating a sample, introduced either as liquid droplets or fine solid particles, is to feed it in a gas stream to a high-temperature plasma, and this is done in the plasmas used for optical excitation. At atmospheric pressure these plasmas may, under favourable electrical conditions, attain core conditions that approach thermal equilibrium, thus favouring rapid vaporisation and dissociation. Typically, in a small discharge in argon the core temperature may reach 5000 K or more. Under these conditions a small amount of fine solid particles or liquid droplets introduced into the carrier gas will, on entering the core, be vaporised and most molecules dissociated. The extent to which the resulting atoms become ionised is described by the Saha equation, which defines the ionisation constant at the given temperature for each component of the system.The degree of ionisation of each element present is then dependent on the relation- ship of the ionisation constant and the partial pressure of the atom in the plasma. A fuller discussion of this subject is not possible here but it is well treated by Bouman~.~ However, in practical terms, for a component in solution at 100 pg ml-l concentration, giving a partial pressure in the core of about atm at a core temperature of 5000 K, the degree of ionisation ranges from 100 per cent. for an element of ionisation potential of 5 V or less down to 15 per cent. for a potential of 10 V. At this partial pressure, even 15 per cent. ionisation represents an enormous number of ions, and as the partial pressure is reduced the degree of ionisation rapidly approaches 100 per cent. All but 13 elements of the Periodic Table have first ionisa- tion potentials below 11 V and only one, barium, has a second ionisation potential below this level.Thus, such a plasma represents a plentiful source of ions with single charges and contains very few ions with more than one charge. An additional advantage of operating at atmospheric pressure is that the ions produced very rapidly reach equilibrium with the surrounding gas molecules, mostly argon, and thus have kinetic energies of between 0.5 and 1 eV, corresponding to those of the gas molecules in thermal equilibrium in the plasma. The low electric field of about 50 V cm-l in the plasma is insufficient to affect their energies significantly, thus resulting in a low ion energy spread.The ions produced a t atmospheric pressure have to be transferred from the hot core to the mass analyser in a vacuum without significantly affecting their relative concentrations. The flow of carrier gas through the plasma core leaves the outlet of the arc as a small flame, carrying the ions with it. The transit time from core to flame is short enough for little ion recombination to occur, even though the gas cools considerably during this time. Techniques for sampling flameMay, 1975 USING AN ATMOSPHERIC PRESSURE ION SOURCE 291 gases at temperatures of up to 3000 K have been developed by several workers6*' who have studied combustion processes in flames, and such techniques can also be used here.During the course of the work described, the suggestion was also made by AlkemadelO that a flame of the type used for atomic absorption should be a useful analytical source of ions. In such flames, however, the rather lower temperatures obtained and the highly reactive species present may give rise to more complex spectra that are beyond the capability of a simple mass analyser to resolve. The process of sampling a flame or plasma through a small orifice in a boundary wall into a region of low pressure is complex and has been extensively studied.ll Provided that sampling conditions are correctly chosen, it is possible to avoid both mass selective effects in the flow through the orifice and distortion of the spectrum due to ion - molecule reactions in the boundary layer in front of the orifice, so that the sample expanded into the low-pressure region can be representative of the plasma composition.Once they are inside the vacuum system the mean free path of the ions is sufficiently large to freeze the composition effectively, and ions can be separated from the accompanying molecules and directed into the mass anal yser. Experimental Apparatus It consists of three main functional groups : The experimental system used for this investigation is shown in schematic form in Fig. 1. the capillary arc plasma, its power and gas supplies, and nebuliser for introduction of the the sampling orifice, ion-beam forming system and mass analyser; the ion detector, pulse counting system and signal read-out.sample and desolvator ; Nebuliser fi,Desoh iator Sampling Quadruple t Capi I law arc d.c. power Electrode bias supplies analyser Channel multiplier t I Quadrupole supplies and controls sensitive amplifier amplifier Ratemeter Fig. 1. Plasma sampling mass analysis system. Plasma and Sample Introduction The capillary arc plasma unit used for this work has been described,lJ2 and is shown in Fig. 2. The discharge occurs in the bore of the main insulator and is approximately 1 cm long and less than 3 mm in diameter. The tantalum cathode and copper anode are recessed so as to minimise contamination. Three separate argon flows metered by capillary tubes are fed to the arc: a small flow to cool the cathode; a main flow along the discharge channel; and the sample flow, which is introduced tangentially into the centre of the discharge.The main292 GRAY : MASS-SPECTROMETRIC ANALYSIS OF SOLUTIONS Analyst, VoZ. 100 body of the arc, which also forms the anode, and the block supporting the cathode, through which the tail flame emerges, are water cooled. The arc is fed from a well smoothed d.c. supply of 300 V through a ballast resistor of about 20 a. It is started by a high-voltage igniter. Spectroscopic measurements of the core temperature show that it rises from 2500 K Sample flow Fig. 2. Capillary arc plasma unit. at 6 A to 5700 K at 12 A. Operation is very stable and quiet and it will run for long periods without the need for attention. Aqueous solutions are introduced into the sample stream by direct nebulisation.Both pneumatic and ultrasonic nebulisers are used, as convenient. The ultrasonic nebuliser is similar to that described by Hoare and Mostyn,13 except that a concave crystal is used to agitate the surface of the sample, which is contained in a sample cell with a thin Mylar window that admits the energy from the transducer. The sample carrier gas is passed across the liquid surface and the mist is carried to the plasma. The ultrasonic nebuliser, although less con- venient to use than the pneumatic type, allows the sample gas flow to be varied without affecting the nebulisation efficiency. Whichever nebuliser is used the gas stream containing the sample droplets is passed through a glass chamber heated to about 470K and then through a water-cooled condenser.This condenses and removes much of the water from the sample. Typical operating conditions for the whole system are shown in Table I. TABLE I TYPICAL OPERATING CONDITIONS The arc is typically operated at between 10 and 12 A. Orifice diameter . . Operating pressures 1st chamber . . 2nd chamber . . Gas flows to plasma Main flow . . .. Sample flow . . Cathode purge . . Plasma current . . Pneumatic nebuliser Sample uptake . . Efficiency . . .. Frequency.. . . Ultrasonic nebuliser Power to transducer .. . . . . . . .. .. .. .. ,. .. 75 pm 2 x torr 5 x 10-4 t o n 1 1 min-l 1.5 1 min-l 0.1 1 min-l 12 A 3 ml min-' 5% approx. 1 MHz 15 W Sample size . . . . Sample consumption . . 1st chamber Collector electrode . . Electrode potentials Cylinder 1 .. .. Cylinder 2 .. . . Cylinder 3 .. . . Quadrupole body . . Quadrupole rods . . . . Type, Mullard . . . . EHT a t mouth . . .. 2nd chamber Channel electron multiplier 6 ml 0.25 ml min-1 -200 v -60 V ov -20 v - 7 v -10 V mean d.c. Ievei B 318 AL -2800 V Ion Sampling, Focusing and Analysis Optimum sampling conditions from the pIasma are obtained with an orifice diameter of between 75 and 125 pm and an orifice at the lower end of this range is usually used. This orifice admits a gas flow into the first vacuum stage (Fig. 3), which is pumped by a 9-in oil diffusion pump that maintains a pressure of less than torr.May, 1975 USING AN ATMOSPHERIC PRESSURE ION SOURCE 293 The orifice, which is drilled in a platinum insert, opens on the low-pressure side into a cone, which enables the effective wall thickness at the orifice to be comparable with the diameter of orifice.The insert is itself mounted in the tip of a metal cone, which projects into the tail flame and is mounted at its base on a small gate valve, which enables the cone to be isolated from the vacuum chamber for cleaning. Immediately inside the gate valve, ions are collected by a cylindrical electrode maintained at a negative potential of a few hundred volts and then focused into a beam so as to pass through a 2-mm aperture into the second chamber, the pressure in which is pumped to below torr by a 4-in diffusion pump. A further cylindrical electrode ensures that the ion beam is co-axial with the quadrupole mass analyser system, which is mounted in this chamber. Vacuum chambers 1s‘t/n\ 2nd r 2nd aperture Quad ru pol e 0-5 torr multiplier 9-in pump 4-in pump Fig.3. Ion sampling, focusing and analysis system. The quadrupole mass analyser or filter consists of four cylindrical electrodes mounted parallel to and equidistant from the beam axis. By means of appropriately controlled a.c. and d.c. potentials applied to the electrodes, the field along the axis is arranged so that for any particular field only ions of one mass to charge ratio (m/e) have a stable trajectory and emerge from the end of the system. The ratio of a.c. to d.c. field determines the “window” of ion mass that is transmitted and the a.c. level the mass centre of the window. This analyser is very compact and simple, and its operating para- meters are set by purely electrical levels.The electrode system used has rods that are 6 mm in diameter and 12.5 cm long. A mass range of 0-300 a.m.u. is covered with a resolu- tion up to 300 (- , 10 per cent. valley). A useful review of these analysers is given by Dawson and Whetten.14 A variety of such instruments are commercially available, although, as in the prototype used, long-term stability and reproducibility do not always meet the full requirement for quantitative measurements. However, short-term stability has been found adequate to explore the potential of the method, and the scope for improvement in quad- rupole performance is being studied. Those ions which are transmitted by the analyser emerge on the axis and can be passed directly into the detector. However, because in the equipment constructed the quadrupole axis is directly on a line of sight from the aperture, and the plasma forms a very intense ultra- violet light source, it is found necessary to mount the detector off the axis and to deflect the ions into it.Even this arrangement is not sufficient t o reduce the photon count to zero if the plasma core is located on the system axis because of the light scattered in the electrode system. However, the plasma can also be displaced slightly from the axis, provided that the tail flame plays on the sampling orifice, and this adjustment reduces the photon count to zero. All other ions are deflected away from the axis and are lost. M A M294 GRAY : MASS-SPECTROMETRIC ANALYSIS OF SOLUTIONS Analyst, VoZ. 100 Ion Detection and Signal Handling This detector, unlike other types of electron multiplier, tolerates the relatively poor vacuum of the second stage and can repeatedly be exposed to the atmosphere.Ions striking the mouth of the detector release electrons that are attracted along the conducting inside surface of the tube by the high electric field. Each time they strike the wall they release further electrons until an average gain of about 108 is achieved. The multiplier is operated in the saturated mode so that the output pulse height obtained for each incident ion is approximately the same. The pulses are fed to a charge- sensitive amplifier and the main amplifier and then through a conventional pulse height discriminator, thus rejecting electrical noise and presenting standardised output pulses to a linear rate meter and scaler.These instruments enable the arrival rate of pulses to be dis- played on a meter over a range of 10-105 pulses s-l and also the total number of pulses in a defined time interval to be integrated. The pulse rate at any given mass setting is a measure of the rate at which ions of that mass are entering the system. If this pulse rate is displayed as the Y deflection of an X - Y recorder and the X deflection made proportional to the mass setting, then a mass spectrum is obtained when the mass analyser is set to scan through the mass range, thus providing a very convenient display for a qualitative examination of the ions present in the plasma. Alterna- tively, quantitative measurements can be made by integrating counts on the scaler for known periods at each mass of interest.Because of the simplicity of electrical control, the quadrupole mass analyser and counting system permit electrical programming and data handling to be used in order to enhance greatly the operational convenience. The ion detector used is a channel electron multiplier. These two modes of operation represent the simplest available. Operation The vacuum pumps are usually left running so that start up is determined by the warm-up time of the counting electronics and of the heater of the desolvator. The arc can be started after briefly purging to clear the air from it and samples introduced as soon as the desolvator is hot. Samples can be exchanged in less than 1 min by using a pneumatic nebuliser but a slightly longer time is required in order to clean and change the window of the ultrasonic nebuliser. During operation the tip of the arc cathode becomes white hot and forms a molten hemi- spherical tip to the tantalum pin.The latter is usually replaced with a fresh, sharply pointed pin after operating for about 4 h. Longer periods of operation can usually be achieved on one electrode but the arc tends to burn unstably at the end of the life of the electrode. The change is made earlier in order to avoid the occurrence of this effect at the most interesting part of the day’s run. The orifice needs to be cleaned at intervals, depending on the material being analysed. With trace solutions cleaning is required after operation for about 10-20 h and can easily be performed in an ultrasonic cleaning bath.Sampling closer to the arc core, however, is thought to increase the interval between cleaning as this is also related to the incidence of atmospheric dust. It has been found to be convenient to position the orifice as close to the tail flame outlet of the arc cathode block as possible and, as described above, to incline the arc axis slightly to the system axis. Because of the use of a high-voltage igniter, it is necessary to withdraw the arc unit when striking it; it can, however, be quickly re-set. Positioning the orifice by eye in the centre of the tail flame is found to give a yield of ions close to the optimum; further transverse adjustment across the flame usually effects little improvement. The effective electrical potential in the tail flame is intermediate between the plasma electrode potentials.The ion yield is optimised by adjusting a bias potential between the earthed orifice and the arc supply. Results and Discussion The performance of the system on aqueous solutions was studied by running test solutions of a range of convenient elements in distilled and de-ionised water at levels of 100, 10 and 1 pg ml-l. Simple mixtures were also prepared from these solutions. A typical spectrum plotted on the X - Y recorder from a solution prepared by mixing equal volumes of aluminium The operation of the system described has been explored on aqueous solutions.May, 1975 USING AN ATMOSPHERIC PRESSURE ION SOURCE 295 and lead solutions, each containing 1 pg ml-1 of solute, is shown in Fig.4. In such a solution, containing 0.5 pg ml-1 of lead in total, the concentration of the isotope lead-204, which has an abundance of 1.48 per cent., is therefore 0.0075 pg ml-l. The peak due to this isotope can be clearly seen as the first of the lead peaks. On the scale shown, the height of this peak represents 400 counts s-1 and it can be seen that the background is extremely small; the ultimate sensitivity for lead is clearly very high. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 C l l l l 10 50 loo 150 200 m/e Fig. 4. Spectrum of solution containing 0.5 pg ml-1 of aluminium and 0-5 pg ml-I of lead. Metal Isotope Abundance, per cent. Aluminium 27 100 Lead 204 1.48 206 23.6 207 22-6 208 62.3 At the low mass end of the spectrum a complex collection of peaks occurs, but among them a substantial isolated peak of about 3000 countss-l is seen for 27Al+.Contaminant peaks from sodium (23Na+) and potassium (saK+ and 41K+; these are close to the correct isotopic ratio) are also clearly seen. The other peaks arise from a variety of causes and considerable assistance towards their identification can be obtained by comparing them with similar spectra obtained by other workers when sampling electrical discharges15 in which similar reactions are to be expected. They can most conveniently be examined on an expanded mass scale and Fig. 5 (6) shows such a spectrum obtained from AnalaR water. On the larger peaks the rate meter has become saturated so that they appear with square tops. The largest peak is due to NO+ at mass 30, which has so far always been found and is attributed to the presence of nitrogen in the argon used and possibly to slight air leakage.The large peak at mass 19 is identified as OH,+, an ion very familiar to mass spectroscopists. It is produced from trace amounts of water in the argon even when water is not being intro- duced. Trace amounts of sodium and potassium can again be seen and 40Ar+ is present, at a relatively low level owing to its high ionisation potential. Peaks of O,+ and NH,+ are also evident at masses 32 and 18 and there is a very small peak at mass 36, which is probably due to NH4+ with a water molecule attached.296 GRAY: MASS SPECTROMETRIC ANALYSIS OF SOLUTIONS ArtalySt, 'Vd. 100 Major peaks are found at masses 37 and 45, that at mass 37 being attributed to hydrated OH3+ and that at 45 to N20H+, an ion that is commonly found in discharges.Other small peaks (with mass numbers in parentheses) arise from HNO+ (31), N20+ (44), NO+.H,O (48) and OH3+.2H20 (55). The conflict between these peaks and those due to atomic ions of interest is less serious than might at first be thought. The peak due to OH,+ at m/e 19 would obscure any due to fluorine although, because of its high ionisation potential, the sensitivity for fluorine would be expected to be low, especially in the inevitable presence of elements of lower ionisation potential. The peak at m/e 30 of NO+ does not cause difficulty by direct coincidence with an ion of interest. Both of these peaks are, however, very large and could potentially cause interference on adjacent mass numbers owing to overlapping of peak fringes, which can occur because of faulty alignment or incorrect operation of the quadrupole. Some indication of this effect can be seen at the leading edge of some of the peaks in Fig.5. With good design and correct operation, however, the overlap contribution between adjacent mass numbers should be reduced to below 10". Interferences with P+ and S+ ions are caused by the peaks at masses 31 and 32. Phosphorus has no other isotope but sulphur has an isotope with 4 per cent. abundance at mass 34, which can be used for its detection at lower sensitivity and which is not subject to interference. Similarly calcium-40 coincides with the small argon peak but has an isotope at mass 42 with an abundance of 0.6 per cent., which is free from interference.Further peaks at masses 45, 48 and 55 interfere to some extent with scandium, titanium and manganese. 0 17 48 15 20 25 30 35 40 45 50 55 60 m/e Fig. 5. Comparative spectra of solution and water: (a), solution containing Al O-46, Mg 0.46, K 0.27 and Mn 0.38 pg ml-I; and (b), AnalaR water. Apart from these seven elements, the presence of the undissociated and molecular ions causes little significant interference, as can be seen in Fig. 5 (a), where the spectrum of a solution containing aluminium (0.46 pg ml-l), magnesium (0.46 pg ml-I), potassium (0.27 pg ml-l) and manganese (0.38 pg ml-l) is shown. This spectrum can be compared with the spectrum of AnalaR water [Fig. 5 (b)] where the peaks of 24Mg+, 25RIg+, 2sMg+, 27Al+, 39K+ and 41K+ are clearly distinguishable.The peak for Mn+ coincides with that for OH3+.2H20 although it is much larger. A small peak is evident at mass 57, probably due to 39K+.H20.May, 1975 USING AN ATMOSPHERIC PRESSURE ION SOURCE 297 Above mass 55 the background is very small (Fig. 4) and no evidence is seen of doubly ionised lead at mass 104, nor do there appear to be any ions corresponding to hydrocarbons from the vacuum pumps. A spectrum was obtained from a solution of cadmium at 100 pg ml-l concentration (shown in Fig. 6) with the mass analyser set at a rather lower resolution than that used for the spectrum in Fig. 4 and the cadmium peaks are not fully resolved. In addition, at this lower resolution the analyser transmission is higher and the molecular peaks at the lower masses are consequently larger.A small group of peaks is also visible at about mass 206 due, pre- sumably, to a trace amount of lead. Although such spectra are useful for qualitative purposes, they are not suitable for quantitative measurements under the conditions used to plot these examples. 112 ?4 , 1 , 1 1 1 1 I l l 1 I 1 1 1 1 # I I I 10 50 100 150 200 m/e Fig. 6. Spectrum of 100 pg ml-l cadmium solution. Cadmium isotope Abundance, per cent. 106 1.22 10s 0-88 110 12.39 111 12-76 112 24.07 113 12.26 114 28.86 116 7.58 At the scan rate employed (approximately 1 a.m.u. s-l) the dwell time on each peak is correspondingly short. It has been found that there is a significant fluctuation in the plasma flame, thought to be caused by fluctuations in the sample gas flow, with a period of rather less than 1 s and with such a short dwell time on the centre of each peak that it results in a significant fluctuation in peak height.In addition, the response required from the X - Y plotter for a large peak approaches the limit of its performance at this scan rate and therefore the peak may not be fully developed. For both of these reasons large peaks do not show the correct ratios for elements with several isotopes. Although both of these effects can be greatly reduced by limiting the scan to a smaller mass range and by increasing the dwell time, it has been found to be more convenient to set the analyser manually to the peak of interest and integrate the signal obtained on the scaler. A counting period of 30 s has been found to provide satisfactory reproducibility.Successive measurements can be made on the unknown and then on a blank water sample at the same mass setting so as to provide a measure of the background level. The background signal, in the absence of the element concerned, in the blank water is small and mostly non-ionic,298 GRAY: MASS-SPECTROMETRIC ANALYSIS OF SOLUTIONS Analyst, VoZ. 100 arising from such sources as stray photons, electrical noise of all sorts and random multiplier noise pulses. Ionic contributions to the background may arise from sources such as memory effects or contamination, which can provide ions of the mass of interest, or from ions of a mass other than that selected, which are still transmitted by the analyser owing to inadequate resolution or significant contributions from the fringes of the analyser response.Interference signals may also arise from molecular ions, as shown in Fig. 5, although so far the effect of these interferences has been found to be limited to only a few elements. Measurements of the integrated count obtained for a series of standard solutions can be used to give a measure of the sensitivity of the method, which can conveniently be expressed for the element concerned as the count-rate obtained for a solution of 1 pg ml-l concentration. With the equipment at present used these sensitivities depend on a number of variables in the system, among which are the nebuliser efficiency, the position of the sampling orifice in the flame, the analyser transmission related to resolution settings and the ion mass and ion optics settings chosen.The over-all ion transmission from the vacuum side of the sampling orifice to the detector is determined by the electrical parameters, which can reproducibly be set so as to optimise the over-all performance. The performance of the nebuliser and behaviour during the plasma sampling are less predictable at the present stage of the investigation and require the most attention in order to make the over-all system satisfactorily reproducible over long periods. However, even in its present experimental form, reasonably stable quantitative performance is obtained during a working day and can be repeated on successive occasions. The sensitivities obtained when using an ultrasonic nebuliser for a range of elements under constant operating conditions is shown in Table 11.The isotopes observed are listed in order of decreasing ionisation potential and the sensitivity in thousands of counts per second for a solution of 1 pg ml-l concentration is shown, first as ST, the mean count-rate observed on the chosen isotope over a 30-s integration period for 1 pg ml-l of the naturally occurring element. This level is the practical sensitivity that can be used, the most abundant isotope normally being selected. The second value shown, SI, is the sensitivity normalised to 100 per cent. abundance of the selected isotope, a more convenient parameter for comparing the perform- ance on different elements. For elements of ionisation potential below 8.0 V a high count- rate is achieved.TABLE I1 SENSITIVITIES FOR A RANGE OF ELEMENTS Isotope measured 75As *OSe 1Wd 6sFe ssC0 2*Mg 107Ag 2oaPb Ionisat ion potential/V 9.81 9-75 8.99 7.87 7-86 7.64 7-57 7.42 Abundance, per cent. 100 49.8 24.1 91-7 78.7 51.8 52.3 100 Coun t-rate sensitivity* ST 1-30 0.10 1-24 68.14 16.92 34-27 182.2 232.3 SI 1-30 0.21 5.14 198.7 68.14 295.2 32.6 65.5 Effective -, detection limittlpg ml-1 0.002 0.03 0.003 0-000 02 0.000 06 0.000 02 0.0002 0*0001 * ST,count-rate, in counts s-1 x 1000 for 1 pg ml-' concentration of element; &,count-rate, in t Effective detection limit (2a value), assuming uniform background standard deviation of 50 counts over counts s-1 x 1000 for 100% abundance of isotope a t 1 pg ml-l concentration. 30 s, expressed in micrograms per millilitre of the element.The sensitivity is, however, influenced by reference to a fixed solution concentration because for elements of higher relative atomic mass proportionately fewer atoms are present to be ionised and the lower count-rates for lead and silver partly reflect this fact. For the three elements of high ionisation potential the lower sensitivity suggests that incomplete ionisation occurs. Although the signals obtained are high, their usefulness for detection of trace levels is related to the background achieved and therefore the usual definition of limit of detection is difficult to apply because of the low background levels obtained. In all these measurements the integrated background was below 1000 counts and in some instances below 200 counts. At these very low levels the background counts depend more on random electrical noise pulses and stray photons than on true background ions and it is thought to be unrealistic toMay, 1975 USING AN ATMOSPHERIC PRESSURE ION SOURCE 299 use the standard deviation of background as a basis for the detection limit.This is especially sn as the standard deviation values obtained over ten successive background integrations show o values that vary from 9 to 40 counts for the various elements, as illustrated by the 95 per cent. confidence limits of detection shown in the final column of Table 11, which were calculated from the measured levels for the isotopes shown (at the natural abundance) and a uniform value for background (T of 50 counts in each instance. When the element concerned is normally absent in the blank, so that the background is very low, the performance can be more usefully judged from the count-rate sensitivity.The limit of detection can then be reserved strictly for particular analytical problems in which the background level due to ionic background, arising either from isotopic interference or the level of the element in the blank, is at least one order of magnitude greater than the noise. Lower sensitivities are shown in Table I1 for the elements with ionisation potentials above 8 V, suggesting that incomplete ionisation occurs, possibly associated with less effective penetration of the plasma by the sample, but more probably as a result of the effect on the ionisation equilibrium of the presence of a significant concentration of nitric oxide, which has an ionisation potential of 9.4V. At present, no steps are taken to purify the argon, but clearly this should be investigated in order to increase the ionisation of these elements.Conclusion The investigation of the technique of plasma sampling mass analysiP has reached a stage at which it appears to have considerable interest for trace analysis. Direct introduction of solution samples into the plasma is practicable by use of conventional nebulisers and the transfer of ions from the plasma to a mass analyser in order to produce qualitative spectra has been demonstrated. Satisfactory quantitative performance requires further development of ion production and sampling techniques and also of mass analyser stability but the sensitivity appears potentially to be very high and the background low.The instrumental configuration necessary to realise the potential performance, while retaining the simple sample handling and rapid throughput of the plasma source, is being studied. From the initial concept to the practical realisation of the system described invaluable help and advice has been generously given by many people working in the field of flame and plasma mass spectrometry. In particular, the author gratefully acknowledges the advice and assistance given by Dr. A. N. Hayhurst, University of Sheffield, Dr. P. F. Knewstubb, University of Cambridge, Dr. J. L. Moruzzi, University of Liverpool, and Professor F. M. Page, University of Aston, and also the practical assistance of colleagues at Applied Research Laboratories Limited, especially that of Mr. D. Hagger, in achieving the results reported above. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. References Greenfield, S., Jones, I. L., and Berry, C. T., Analyst, 1964, 89, 713. de Boer, F. J., and Boumans, P. W. J. M., Acta Colloq. Spectrosc. Id. X V I I , 1973, 1, 107. Fassel, V. A., and Knisely, R. N., Analyt. Chem., 1974, 46, l l l 0 A . Jones, J. L., Dahlquist, R. L., Knoll, J. W., and Hoyt, R. E., Paper presented at the 1974 Pittsburg Jones, J . L., Dahlquist, R. L., and Hoyt, R. E., Appl. Spectrosc., 1971, 25, 628. Knewstubb, P. F., “Mass Spectrometry of Organic Ions,” Academic Press, New York, 1963, Chapter 6, Hayhurst, A. N., and Sugden, T. M., PVOG. R. SOG., 1966, A293, 36. Caroll, D. I., Dzidic, I., Stillwell, R. N., Homing, M. G., and Homing, E. C . , Analyt. Chem., 1974, Boumans, P. W. J. M., “Theory of Spectrochemical Excitation,” Adam Hilger Ltd., London, 1966, Alkemade, C. Th. J., PVOC. SOC. Analyt. Chem., 1973, 10, 130. Hayhurst, A. N., and Telford, N. R., Proc. R. SOG., 1971, A332, 483. Applied Research Laboratories Ltd., British Patent 1,261,596, 1969. Hoare, H. C., and Mostyn, R. A., Analyt. Chem., 1967, 39, 1153. Dawson, P. H., and Whetten, N. R., Adv. Electrovzics Electron Phys., 1969, 27, 68. Knewstubb, P. F., “Mass Spectrometry and Ion Molecule Reactions,” Cambridge University Press, Applied Research Laboratories Ltd., British Patent 1,371,104, 1971. Received June 17th, 1974 Amended December 9th, 1974 Accepted December 16th, 1974 Conference, Cleveland, Ohio. pp. 255-307. 46, 706. Chapter 7, pp. 156-232. Cambridge, 1969.

ISSN:0003-2654    

DOI:10.1039/AN9750000289

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

300 Analyst, May, 1975, Vol. 100, pp. 300-306 Interferences in the Determination of Elements that Form Volatile Hydrides with Sodium Borohydride Using Atomic-absorption Spectrophotometry and the Argon - Hydrogen Flame A. E. Smith ICI Mond Division, Research and Development Department, Winnington Laboratory, Northwich, Cheshire A study has been made of the determination of arsenic, bismuth, germanium, antimony, selenium, tin and tellurium by conversion into the hydrides by reaction with sodium borohydride in dilute hydrochloric acid followed by measurement with atomic-absorption spectrophotometry in an argon - hydro- gen flame. A general study, involving 48 elements, of interferences has been carried out and i t has been shown that significant interference occurs in many instances but that the procedure is simple to carry out and gives a considerable increase in sensitivity and detection limits for the elements listed, with the exception of tin, for which high blank values were obtained owing to the presence of tin in the sodium borohydride reagent.Reducing agents have long been used to form volatile hydrides for subsequent use in analysis, but interferences in the determination of the elements that form volatile hydrides in reducing media have not been studied thoroughly. Vogell mentions that in the detection of arsine (ASH,) in the Gutzeit test, copper, cobalt and nickel reduce the rate of formation of arsine as do other metals that can be precipitated by zinc ( i e . , noble metals). In this method tin and hydrochloric acid were used as reducing medium as in the Marsh reaction.Braman et a1.2 mention that copper and silver interfere in the generation of arsine and stibine (SbH,) when these hydrides are generated from alkaline solution using sodium borohydride, that iron interferes with antimony but not arsenic and that aluminium, cadmium, chromium, mercury, manganese, nickel, lead and zinc do not interfere with either antimony or arsenic. They used d.c. discharge emission in helium as the analytical technique and the hydrides were evolved Over a period of about 1 min. In the U.S.A. this type of analysis is recommended3 for antimony and arsenic, using tin(I1) chloride - potassium iodide - hydrochloric acid with a colorimetric finish, but no mention is made of possible interferences.Several papers have recently appeared that give methods for the generation of these hydrides and for their analysis by atomic-absorption spectrophotometry in the argon - hydrogen flame but interferences are not mentioned. Pollock and West4 used titanium(II1) chloride - hydrochloric acid and magnesium - zinc to produce stibine (SbH,). Fernandez5 used sodium borohydride to study all of the seven elements with which this paper is concerned, namely, arsenic, bismuth, germanium, antimony, selenium, tin and tellurium. He also studied the effects of pH on sensitivity and used electrodeless discharge tubes as well as hollow-cathode lamps to examine detection limits. Fernandez and Manning6 published the first paper in which the use of the Perkin-Elmer balloon reservoir accessory was described. They studied arsenic and selenium with tin(I1) chloride - hydrochloric acid as the reducing medium and Manning7 has described a modification of this method. Pollock and Wests used titanium(II1) chloride - hydrochloric acid - magnesium to produce the hydrides of arsenic, bismuth, selenium, antimony and tellurium, and sodium borohydride to produce germane (GeH,), for the determination of these elements.Schmidt and Royers used sodium borohydride in the determination of arsenic, bismuth, antimony and selenium with atomic- absorption spectrophotometry. KanlO has automated the determination of antimony and arsenic using sodium borohydride and atomic-absorption spectrophotometry with an argon - hydrogen flame, Thompson and Thomersonll investigated the generation of eight volatile hydrides, the seven with which this paper is concerned and also lead, using sodium boro- hydride solution.The evolved gases were passed into a silica tube heated in an air - acetylene flame. None of these papers3-11 mentioned possible interferences.SMITH 30 1 Dalton and Malanoski12 determined arsenic by using zinc - hydrochloric acid after treatment with tin(I1) chloride - potassium iodide. They used no carrier gas other than the hydrogen evolved and found that nitric acid and antimony interfered. Lansford et aZ.,13 using tin(I1) chloride - hydrochloric acid, found that mercury, arsenic and nitric acid interfered in the determination of selenium. Roulet et aZ.14 found that arsenic and antimony hydrides were decomposed by gold complexes.Macklenl5 stated that some germanium metal can be produced in the preparation of germane from germanium(1V) chloride in solution using sodium borohydride, that sodium borohydride can also produce the free metal with silver, bismuth, arsenic and antimony from their ions in solution and reduces Ce4+, Crs+, Hg2+, Fe3+ and Tl3+ to lower valency states. Jollyls mentions that arsenic, germanium, an- timony and tin form hydrides with sodium borohydride under alkaline conditions but bisniuth does not. Yamamoto et aE.I7 studied interferences in the determination of arsenic using zinc - potassium iodide - tin(I1) chloride in hydrochloric acid as reducing agent with an atomic-absorption finish. They found that Mn2+, Cr3+, Fe3+, AP+, Co2+, Ni2+, Cu2+, Zn2+, Ag+, Cd2+, Hg2+, Na+, K+, Mg2+ and Ca2+ at 250p.p.m.(5mg), lead and antimony at 10 p.p.m. and selenium at 0.4 p.p.m. did not interfere but that the first fifteen metals did inter- fere above 250 p.p.m, (5 mg) ; lead above 150 pg, antimony above 220 pg and selenium above 7 pg were found to interfere in the determination of 1 pg of arsenic. Vijan and Woodla investigated a few interferences in the determination of arsenic using sodium borohydride and passing the evolved hydrides into a heated wire-wound quartz cell instead of a flame. They found that copper, cobalt, nickel and selenium severely interfered in the determination of arsenic but that aluminium, cadmium, chromium, iron, potassium, manganese, magnesium, sodium, lead, vanadium and zinc did not (greater than 96 per cent.recovery of arsenic). Daher and Saleh19 studied the interaction of arsine with metal films. They studied iron, nickel, palladium, tungsten, silver and lead and found that with the first three of these metals rapid dissociative chemisorption at -80 "C and evolution of hydrogen occurred. On tungsten, arsine exhibited weak dissociative chemisorption and weak molecular adsorption. Physical adsorption occurred mainly with silver and with lead. Saleh20 also studied arsine and hydrogen selenide on copper film and found both to exhibit dissociative chemisorption. The work described in this paper was carried out, therefore, to evaluate the usefulness of this technique, which seemed to have great potential for the determination of arsenic, bismuth, germanium, antimony, selenium, tin and tellurium as it is simple to operate and has high sensitivity.However, information is lacking both on the interference of these elements with one another and on the interference from elements that do not form hydrides with sodium borohydride. It was decided to investigate these interferences. Apparatus Experimental The work described in this paper was carried out with a Perkin-Elmer atomic-absorption spectrophotometer, Model 303, with the recorder read-out accessory, deuterium background corrector and a Bryans recorder, Series 27000. The reduction vessel used is a modification of that distributed by Instrumentation Laboratories Inc. with a three-way tap on the argon inlet to allow argon to by-pass the reduction vessel and a two-way tap to allow argon into the vessel and to allow flushing of the liquid contents from the vessel.The Instrumentation Laboratories reduction vessel was slightly modified in that the rubber or glass stopper was replaced with a gas tap with wide-bore glass tubing (12 mm i.d.) and the hole in the PTFE tap was bored out to just less than the diameter of the glass tube. (This large hole extends only halfway through the PTFE tap so as to form a cup. The other half of the original hole was blocked with a PTFE plug.) The gas-tight seal was achieved with O-rings placed around the outside of the bottom wide-bore glass tube. Cylinder pressures Argon: 40 lb in-2 at cylinder head; 30 lb in-2 at gas regulator box. Hydrogen : 8 lb in-2 at cylinder head. Argon : 9 divisions on flow meter (15 1 min-l) .Hydrogen: 4.5 divisions on flow meter (3.5 1 min-l). Gas JIows302 SMITH INTERFERENCES IN THE AAS DETERMINATION OF ELEMENTS Analyst, VOl. 100 Burner height The burner height was adjusted so as to give the greatest signal while nebulising an aqueous standard of sufficient concentration to give a reasonable signal into the argon - hydrogen flame by using the Boling burner (Part No. 0303-0401, Perkin-Elmer). This height was found to be at setting 1 on the burner mount. The centre of the image of the hollow-cathode lamp at the middle of the burner was usually about 12 mm above the burner top. Instrument parameters The wavelength and slit width used for each element are given in Table I. The hollow- cathode lamp current used was that stated on the lamp.Scale expansion x l was used for the work involving the study of interferences, but scale expansions x 1 and x 3 were used for the determination of sensitivity and detection limits. Noise suppression 1 was used throughout for the sodium borohydride interference experiments. The work described and tabulated in Table I1 wac carried out at noise suppression 2. Reagents Distilled water was used. The solutions of standards and the interfering elements studied were prepared by dilution of standard solutions purchased from BDH Chemicals for gold, arsenic, boron, cobalt, germanium, indium, lithium, platinum, silicon, titanium, vanadium, tungsten and zirconium, and from Hopkin and Williams for aluminium, silver, barium, bismuth, calcium, cadmium, chromium, copper, iron, mercury, potassium, magnesium, manganese, molybdenum, sodium, nickel, lead, antimony, tin, strontium and zinc.The remaining standard solutions were prepared from Johnson Mat they Specpure materials as described by Smith.21p22 The sodium borohydride was obtained from Hopkin and Williams as a powder and pelleted in the laboratory. (Pelleted sodium borohydride is available from Alfa Inorganics.) For the interference work using sodium borohydride, the concentrations of the volatile elements in the solutions used as standards were as follows: bismuth, 0.5; arsenic and anti- mony, 1.0; germanium and selenium, 2.0; and tellurium, 10.0 pg ml-l. All reagents used were of AnalaR or Aristar grades. Procedure The auxiliary oxidant gas line is disconnected from the burner chamber and connected to the inlet of the reduction vessel.The outlet of this vessel is connected to the auxiliary oxidant inlet point of the burner chamber. The air line to the gas control box is replaced with an argon supply and the acetylene supply is replaced with a hydrogen supply. The burner height is optimised by nebulising an aqueous standard solution of the element of interest into the flame and adjusting it so as to give the maximum signal. (The set-up of the spectrophotometer is described in the manufacturer's handbook and is not described in detail here.) With the two-way tap closed and the three-way tap set so as to divert the argon flow around the vessel direct to the flame, the modified tap at the top of the vessel was removed and 10 ml of distilled water were placed in the vessel by pipette, followed by 1.0 ml of concentrated hydrochloric acid and 1.0ml of the standard solution of the element being studied. If an interferent was being studied, 1.0 ml of the 1 mg ml-l standard of that element was also added.The modified tap was replaced and the three-way tap turned so that the air in the vessel was purged. A pellet of sodium borohydride (about 0.30 zt 0.02 g) was inserted into the modified tap and the tap was then turned through 180" so that the pellet fell into the reduction vessel and rapidly dissolved with the evolution of hydrogen. A peak of about 2-2, duration was seen on the recorder (set at noise suppression 1 on the recorder read-out accessory, i.e., no noise suppression). When the signal returned to zero, the modified tap was removed and the vessel washed several times with water, replacing the tap each time and allowing the water to drain off by opening the two-way tap.A value for the reagent blank was obtained by following the same procedure, but excluding the standard solution of the element being studied. No problems were encountered withMay, 1975 THAT FORM VOLATILE HYDRIDES WITH SODIUM BOROHYDRIDE 303 blanks except for the tin which was present in the sodium borohydride. The work was carried out in the following order. A standard (with no interferent present) was analysed and the signal recorded. This was followed by the analysis of two standards with interferents present, and then another standard with no interferent present. Reagent blanks were determined during the run and treated as if they were standards with interferents present.In this way, a standard was always analysed immediately before or after those with interferents present and also a large number of standards with no interferents present were analysed to give a good estimate of precision. Results Interferences in the Formation and Evolution of the Hydrides The results, given in Tables I and I1 for the work carried out with sodium borohydride reduction, were obtained with a single mass of sodium borohydride (0.30 g), at a single pH of 0, using a single burner height and a single set of gas flows. pH 0 was chosen for this work as a compromise based on the work of Fernande~.~ pH does not seem to have much effect on sensitivity except for tin, the sensitivity for which is at a maximum in the pH range 0-1, hence the very low pH value chosen.The interferences described in Table I1 were obtained for a single level of interferent (1 mg) added to a single level of the element to be determined. This work was carried out with no scale expansion and no noise suppression. For certain elements, e.g., germanium, scale expansion x 3 was used for the determination of detection limits. Germane evolved at a much slower rate than the hydrides of the other six elements studied. It was therefore possible to use noise suppression 2 for this element in order to improve the detection limit. TABLE I INSTRUMENT PARAMETERS AND COMPARISON OF SENSITIVITIES AND DETECTION LIMITS FOR THE DETERMINATION OF THE SEVEN ELEMENTS GENERATED AS HYDRIDES There is an apparent reagent blank for arsenic and selenium of about 0.1 pg.This may be due to absorption caused by air entrained as the pellet is introduced. Element As Bi Ge Sb Se Sn Te Wavelength/ nm 193.7 223-1 265.1 217.6 196.0 224.6 214-3 Sensitivity/pg per Slit 1 per cent. (spectral width/nm) absorption 3 0.010 3 0.006 3 0.05 3 0.01 4 0.02 3 -0.015 3 0.1 (0.4 nm) (0.7 nm) Detection limit/ tLg 0.03 0-01 0-05 0.02 0-05 0*1* 0.2 * The detection limit for tin is poor owing to the large blank value in the sodium borohydride of approxi- mately 1.0 pg of tin for 0.30 g of sodium borohydride. Detection limit is defined as that concentration of the element equivalent to twice the standard deviation of the reagent blank, or base-line noise if there is no reagent blank (twice the peak to trough signal).Reproducibility is about 5 5 per cent. at a concentration well above the detection limit. It should be noted that there is a significant tin blank in the sodium borohydride (corresponding to 1.0 pg of tin for 0.30 g of sodium borohydride) , which hindered later work on interferences in the determination of tin. It is thought that the tin is picked up from the tin plate on the can in which the material is supplied. It was found at an early stage of the investigation that the signal due to the element being determined increased as the mass of sodium borohydride increased, owing to more rapid evolution of the hydride. Experimentally, it was found that 0.30 g of sodium borohydride gave a signal very close to the maximum signal observed and it must be emphasised that a constant mass of sodium borohydride should be used in work involving this reagent.No attempt was made either to examine the effect of the valency states of the metals being deter- mined or to study other acids or cations. As can be seen from Table I, sensitivities are very good.304 SMITH: INTERFERENCES I N THE AAS DETERMINATION OF ELEMENTS AnaZyst, Vd. 100 Certain trends are evident in the interferences observed (Table 11) : (i) Elements of Groups I, 11, 111, IIIA and IVA ( i e . , the alkali metals, alkaline earths and boron, aluminium, gallium, titanium, zirconium and hafnium) and also mercury, lanthanum, manganese, vanadium and yttrium do not interfere in the determination of the six elements at the levels studied.(ii) Copper, silver, gold, platinum, palladium, rhodium, ruthenium, nickel and cobalt always interfere (except for silver with germanium) in these determinations at the levels studied. (iii) All the volatile hydrides formed from the elements studied interfered with all the other volatile hydrides (except for the hydrides of bismuth and tellurium with those of antimony and selenium). A small amount of work was carried out on interferences in the determin- ation of tin, but owing to the large blank and probable heterogeneity of the tin in the sodium borohydride, little useful information resulted. However, germanium and nickel were found to interfere severely. TABLE I1 INTERFERENCES OBSERVED FOR THE DETERMINATION OF THE SIX* ELEMENTS AS HYDRIDES GENERATED WITH SODIUM BOROHYDRIDE IN THE ARGON - HYDROGEN FLAME Interference (suppression of signal) A .Severe to moderate (greater than 60 per cent. suppression) Au, Gel Nil Pt, Pd. Rh, Ru Ag, Au, Co, Cu, Ni Pd, Pt, Rh, Ru, Se, Te As, Au, Cd, Co, Fe, Ni, Pd, Pt, Rh, Ru, Sn, Sb, Se Au, Co, Ge, Ni, Pt, Pd, Rh, Ru Ag, Cu, Ni, Pd, Pt, Rh, Ru, Sn Ag, Au, Cd, Co, Cu, Fe, Ge, In, Ni, Pb, Pd, Pt, Re, Rh, Ru, Se, Sn, Te Moderate to slight (10-60 per cent. suppression) Ag, Bi, Co, Cu, Sb, Se, Sn, Te As, Cd, Cr, Fe, Ge, Ir, Mo, Sb, Sn Bi, Cu, Ir, Te Ag, As, Cr, Cu, Re, Se, Sn Au, As, Cd, Co, Fe, Ge, Pb, Sb, Zn As, Bi, Ir, Mo, Sb, Si, W Not significant (less than 10 per cent. suppression) Al, B, Ba, Be, Ca, Cd, Cr, Cs, Fe, Ga, Hf, Hg, In, Ir, K, La, Li, Mg. Mn, Mo, Na, Pb, Rb, Re, Si, Sr, Ti, TI, V, W, Y, Zr, Zn Al, B, Ba, Be, Ca, Cs, Ga, Hf, Hg, In, K, La, Li, Mg Mn, Na, Pb, Rb, Re, Si, Sr, Ti, T1, V, W, Y, Zn, Zr All Ag, B, Ba, Be, Ca, Cr, Cs, Ga, Hf, Hg, In, K, La, Li, Mg, Mn, Mo, Na, Pb, Rb, Re, Si, Sr, Ti, T1, V, W, Y, Zn, Zr All B, Ba, Be, Bi, Ca, Cd, Cs, Fe, Ga, Hf, Hg, In, Ir, K, La, Li, Mg, Mn, Mo, Na, Pb, Rb, Si, Sr.Te, Ti, TI, V, W, Y, Zn, Zr Al, B, Ba, Be, Bi, Ca, Cr, Cs, Ga, Hf, Hg, In, Ir, K, La, Li, Mg, Mn, Mo, Na, Rb, Re, Si, Sr, Ti, T1, V, W, Y, Zr (Te) Al, B, Ba, Be, Ca, Cr, Cs, Ga, Hf, Hg, K, La, Li, Mg, Mn, Na, Rb, Sr, Ti, V, Y, Zn, Zr * Results for tin are not quoted because of the high blank. Discussion The results obtained show that this technique of analysing the volatile metal hydrides by atomic-absorption spectrophotometry is not as simple as many workers and instrument manufacturers may have implied.Many papers have recently appeared on this type of analysis but only a few mention the possibility of interferences and this is the first major attempt to examine metal interferences comprehensively. The method as described in this.May, 1975 THAT FORM VOLATILE HYDRIDES WITH SODIUM BOROHYDRIDE 305 paper may well exaggerate or even maximise interferences but it is similar to other systems widely used. No attempt has been made to overcome the interferences. The causes of the interferences are probably manifold but the two most likely to be serious are as follows. (i) Preferential redaction of the interferent Preferential reduction of the metal ion interferent in solution to a different valency state or to the free metal can cause precipitation of that valency species that can either co-precipitate the metal of interest, adsorb the volatile hydride formed, catalytically decompose it or slow down or completely stop its evolution from solution.Certainly many of the elements that interfered formed precipitates after the addition of sodium borohydride or formed coloured solutions, if only transiently. As some of the sodium borohydride is used up in this way less is available for reduction of the element of interest to the hydride and a lower signal is ob- tained. In this context, easily reducible organic compounds might be expected to interfere. (ii) Compound formation in the $awe Compound formation can occur in the cool argon - hydrogen flame.This interference explains the mutual interference of virtually all the volatile hydrides on each other, the only exceptions being for bismuth and tellurium with antimony and selenium. Cause (i) also applies to the seven elements studied as they are reduced to the hydrides or even to the metal or other valency states when present in large excess. Thompson23 has found that the presence of copper in solution causes a considerable decrease in the signal due to arsenic and antimony and completely suppresses the signal due to selenium when the Shandon-Southern accessory is used.ll This accessory uses the hotter air - acetylene flame. In this instance the inter- ference cannot be said to be due to the low-temperature argon - hydrogen flame.I have defined an interIerence as occurring if the signal obtained for any element of interest in the presence of a possible interfering element is different by two standard deviations of the signal obtained for that element with no possible interferent present. As the reproducibility is about 5 5 per cent., an interference is said to have occurred if the signal differs by more than &lo per cent. of the signal obtained for the standard with no interferent present. For ease of tabulation, I have divided the interferences into three arbitrary classes: (a), no inter- ference (less than 10 per cent. difference); ( b ) , slight to moderate interference (10-50 per cent. difference) ; (c), severe interference (greater than 50 per cent. difference). It must be remembered that only one level of interferent (1 mg in a total of 13 ml) was used for each element, itself present at only one concentration.I t may be that there is a threshold value for each element present at only one concentration and that there is a threshold value of concentration, different for each element added as interferent, above which it will interfere with the element forming the volatile hydride being studied. This threshold value may vary with concentration of the element forming the volatile hydride (among other variables). In other words, elements that have been found to interfere under the conditions described in this paper might not interfere at different concentrations of both the interferent and the element forming the volatile hydride. I t is also possible that elements found not to interfere might interfere (at higher concentrations of potential interferent and probably lower concen- trations of the element forming the volatile hydride). Also, only one pH value was examined (pH 0).It may be that different amounts of acid could alter the level of interference as the acidity will alter upon the addition of sodium borohydride. Tellurium at first appeared to enhance the signal due to selenium, but this enhancement was found to be due to a low level of selenium in the tellurium standard, hence its appearance in the table in parentheses. No accurate value could be assigned to it as a potential interferent. No work was done on other reagents that can be used for hydride evolution e.g., zinc - hydrochloric acid, tin(I1) chloride - hydrochloric acid, other metals and mineral acids, and titanium( 111) chloride and hydrochloric acid with magnesium or zinc.As an apparent interference could be caused by a reduced rate of evolution, but complete evolution of the metal hydride could still take place, i.e., a broader peak would be obtained, the use of an integrated signal or some form of collection as in the Perkin-Elmer balloon system might overcome this apparent interference. Collection would, however, give a lower sensitivity. The author is certain, however, that complete evolution does not occur with many of the interferences studied as in some instances no signal was observed. The use of306 SMITH standard addition techniques would also probably overcome some of those interferences.Thompson23 has found this to be the case for arsenic and antimony in the presence of copper. Conclusions Determination of elements by the method of generation of volatile hydrides with subsequent use of atomic-absorption spectrophotometry or any other analytical finish is subject to many interferences. Before this type of accessory for atomic absorption for the determination of arsenic, bis- muth, germanium, antimony, selenium, tin or tellurium is used, it is necessary to ensure that the matrix being analysed does not interfere. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. References Vogel, A. I., “A Text Book of Quantitative Inorganic Analysis,” Third Edition, Longmans, Green & Braman, R. S., Justen, L. L., and Foreback, C. C., Analyt. Chem., 1972, 44, 2195. “Official Methods of Analysis of the Association of Official Analytical Chemists,” Eleventh Edition, Association of Official Analytical Chemists, Washington, D.C., 1970. Pollock, E. N., and West, S. J., Atom. Absorption Newd., 1972, 11, 104. Fernandez, F. J., Atom. Absorption Newsl., 1973, 12, 93. Fernandez, F. J., and Manning, D. C., Atom. Absorptioa Newsl., 1971, 10, 86. Manning, D. C., Atom. Absorption Newsl., 1971, 10, 123. Pollock, E. N., and West, S. J., Atom. Absorption Newsl., 1973, 12, 6. Schmidt, F. J., and Royer, J . L., Analyt. Lett., 1973, 6, 17. Kan, K.-T., Analyt. Lett., 1973, 6, 603. Thompson, K. C., and Thomerson, D. R., Analyst, 1974, 99, 595. Dalton, E. F., and Malanoski, A. J., Atom. Absor$tion Newsl., 1971, 10, 92. Lansford, M., McPherson, E. M., and Fishman, M. J.. Atom. Absorption Newsl., 1974, 13, 103, Roulet, R., Ngyuyen Quang Lan, Mason, W. R., and Fenske, G. P., jun., Helv. Chim. Acta, 1973, Macklen, E. D., J . Chem. Soc., 1959, 1989. Jolly, W. L., J . Awzer. Chem. SOC., 1961, 83, 335. Yamamoto, Y., Kumamaru. T., Hayashi, Y. and Kamada, T., Bull. Chem. Soc. Japan, 1973, 46, Co., London, 1961, pp. 796-797. 56, 2405. 2604. Vijan, P. N., and Wood, G. R., Atom. Absorption Newsl., 1974, 13, 33. Daher, I. M., and Saleh, J. M., J . Phys. Chcm., 1972, 76, 2851. Saleh, J. M., J . Chem. SOC., 1972, 1620. Smith, A. E., Analyst, 1973, 98, 66. Smith, A. E., Analyst, 1973, 98, 209. Thompson, K. C., personal communication. Received September Accepted December 16th, 1974 17th, 1974

ISSN:0003-2654    

DOI:10.1039/AN9750000300

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

Analyst, May, 1975, Vol. 100, pp. 307-310 307 The Atomic-f luorescence Determination of Antimony, Arsenic, Selenium and Tellurium by Using the Hydride Gene ration Tech n iq u e K. C. Thompson Shandon Southern Instruments Limited, Frimley Road, Camberley, Surrey, G U16 5ET The method of determination of antimony, arsenic, selenium and tellurium, by using a sodium borohydride reduction with subsequent atomic fluorescence, was found to be very sensitive. The calibration graphs were linear over a wide range of concentrations and the method was 6-30 times more sensitive than the corresponding atomic- absorption technique. It was found that good detection limits could be obtained by using the method of hydride generation with subsequent atomisation in a flame-heated silica tube for the determination of antimony, arsenic, selenium and tellurium.1 However, the detection limit for selenium of 0.0018 pg was not considered adequate for certain determinations (e.g., the determination of selenium in blood, effluents and some plant and soil samples). As is the case with most absorption techniques, the calibration graph was linear over a limited range only. This limitation necessitated careful dilution of many samples when the standard additions method of calibration that is normally necessary for this technique was used. Tsujii and Kuga2 have reported the atomic-fluorescence determination of arsenic by using a hydride generation technique. The arsenic was reduced by zinc to arsine, which was then passed into an argon - hydrogen - entrained air ff ame.A non-dispersive solar-blind detection system was used in order to detect the fluorescence. The detection limit was 0.002 pg and the reagent blank was 0.025 pg. The method for the determination of mercury that involves atomic fluorescence in con- junction with the cold-vapour technique has been found to be more sensitive and to give a larger linear calibration range than the corresponding absorption method,3 and it appeared that a fluorescence detection method for the determination of elements generated as hydrides should also have these advantages. The hydrides were generated by using sodium borohydride and passed directly into an argon - hydrogen - entrained air flame. The atomic fluorescence was excited by using modulated microwave sources and detected by using a dispersive measuring system.The detection limits (measured as twice the noise level on the base-line) for arsenic, antimony, selenium and tellurium were all 0.0001 pg or less. Experimental Results were obtained by using a Shandon Southern Instruments A3400 atomic-absorption - fluorescence spectrophotometer (fitted with an EM1 9783R photomultiplier), a slightly modi- fied (see below) Shandon Southern Instruments A3490 hydride generator, a 2450-MHz micro- wave generator (Electro-Medical Supplies Ltd.) with an electronic modulation unit4 and a three-quarter wave Broida-type cavity (Southern Spectral Sources Ltd.). The lamps were electronically modulated a t a frequency of 325 Hz and initiated by using the voltage generated by the burner ignition system.The spectrophotometer was used in the fluorescence mode with phase-sensitive detection of the fluorescence. Stock solutions containing 1000 pg ml-l of arsenic(II1) , antimony(III), selenium(1V) and tellurium( IV) were prepared. A 1 per cent. m/V aqueous solution of Alfa Inorganics sodium borohydride, 10/32-inch pellets," was freshly prepared for each set of measurements. The experimental arrangement is depicted in Fig. 1. The hydrides generated in cell A were passed through tube C into an argon - hydrogen diffusion flame maintained on a 10 mm i.d. Pyrex tube D. The top of this tube was mounted 2.5 cm below the bottom of the microwave- All &luted solutions were freshly prepared. * Obtainable from Ralph N. Emanuel Ltd., 264 Water Road, Wembley, Middlesex.308 Analyst, VoZ.100 cavity window. The microwave cavity, F, was positioned so that the axis of the microwave lamp, G, was 5 cm from the axis of the flame E, which was positioned on the optical axis, H, 8-5 cm from the monochromator entry slit. THOMPSON : ATOMIC-FLUORESCENCE DETERMINATION OF sb, AS, se H - d ii Fin. 1. Arrangement of i. IB atomic-fluorescence spec<rophotometervand hydride generator (for key, see text). The hydride generation system (A3490) was similar to that previously describedl except that the inlet of the auxiliary transverse gas flow meter was connected to a hydrogen supply, the outlet being connected to the gas line that links the exit of cell A to the Pyrex tube, D. This arrangement ensured that the flame remained alight when the cell was removed.Optimisation of the Operating Conditions Hydride generator cell. This was as previously described.l Source operating conditions, wavelength and gas $ow-rates. Table I lists the optimal oper- ating conditions for the lamp, the wavelengths used and the optimal gas flow-rates. TABLE I LAMP OPERATING CONDITIONS, WAVELENGTHS AND GAS FLOW-RATES The hydrogen flow-rate used for each element was 1.3 1 min-l, which corresponded to a reading of 0.6 1 min-' on the A3490 flow meter. Flow-rate of cooling Modulation air applied Argon Incident Reflected switch to base of Wavelength/ flow-rate/ Element power/W power/W position' lamp/l min-1 nm 1 min-l Arsenic . . .. . . 40 15 11 3 193.7 2 Antimony .. . . 48 12 9 2.4 231.1 2 Selenium .. ,. 38 13 8 0 196.1 3 Tellurium .... 46 15 11 4 214.3 3 Spectral band pass. A spectral band pass of 3 nm was used for all studies. Damping. In order to attain the optimal signal to noise ratio the A3400 spectrophoto- meter was operated at damping position 1 (time constant = 0.5 s). Measurement Procedure All samples and blanks were acidified with hydrochloric acid (Le., 15 ml of concentrated hydrochloric acid (36 per cent. m/m) were diluted to 100 ml with sample or blank). A 2-ml amount of 1 per cent. m/V sodium borohydride solution was placed in cell A and the acidified sample that was contained in a l-ml MLA pipette (Shandon Southern Instruments) was inserted into the side-arm, B, of cell A (Fig. 1). After a delay of 5 s, to allow any entrained air to be drawn through the system, the sample was injected and the resulting peak recorded.May, 1975 AND Te BY USING THE HYDRIDE GENERATION TECHNIQUE Results Detection Limits and Calibration Graphs 309 It was considered impractical to determine the true detection limit of the method because of its extreme sensitivity, the small reagent blanks and the unknown stability of those solutions that contained less than 0.001 pg ml-1 of antimony, arsenic, selenium or tellurium. Table I1 lists, for each of these elements, twice the noise level on the base-line expressed as a concentration.5 (The concentration used in order to obtain these values was 0.001 pg ml-l or 0.002 pg ml-1.) The blank values, due to the hydrochloric acid, sodium borohydride and distilled water, are-also given.Fig. 2 shows some typical results for selenium.TABLE I1 VALUES FOR BASE-LINE NOISE, REAGENT BLANK LEVELS AND UPPER LIMITS TO THE LINEAR REGIONS OF THE CALIBRATION GRAPHS l-ml sample volume. 2 x noise level Element Pg Id-' blank/pg ml-l on base-line/ Reagent Arsenic . . .. . . 0~0001 0*0005 Selenium .. . . 0.00006 O*OOO 08 Tellurium .. . . 0-00008 0-000 16 Antimony .. . . 0.0001 0~0001 Upper limit to linear region of calibration 0.1 0-16 0.2 0.1 graPh/Pg A negligible blank signal was observed when using the 225.9-nm tellurium line. This line was approximately three times more intense than the 214.3-nm line and gave a fluorescence signal intensity of 5 per cent. of that observed at 214.3 nm. Thus the blank signal observed at 214.3 nm was almost wholly due to tellurium in the reagents and not caused by scattering of the source radiation.For arsenic, the blank was lower than in a previous study,l a different batch of sodium borohydrides being used for the present work. +- Time Fig. 2. Typical traces for selenium; l-ml sample volume. Table I1 also shows the upper concentration limits to the linear parts of the calibration graphs (i.e., the concentration at which 5 per cent. deviation from the best straight line drawn through the calibration points occurs). This large range of concentration over which the calibration graphs are linear illustrates another advantage of atomic fluorescence compared with atomic absorption. Applications Five grams of the feedstuff, which had been dried at 50 "C, were digested with 40 ml of nitric acid - perchloric acid until just fuming.A 5-ml volume of 6 M hydrochloric acid was then added, the beaker covered with a watch-glass and the mixture boiled for 5 min. This digestion reduced The determination of selenium in some animal feedstuffs was attempted.310 THOMPSON selenium(V1) to selenium(1V). The solution was allowed to cool and then made up to 50 ml with distilled water. Standardisation was achieved by the method of standard additions. Table I11 shows the results, together with some results obtained by a spectrofluorimetric TABLE I11 ANALYSIS OF ANIMAL FEEDSTUFFS Selenium, found by- A I > atomic fluorescence Sample p.p.m. generation, p.p.m. spectrofluorimetry*, and hydride Grassmeal . . .. . . 0.067 (0*012)t 0.075 Cattle nuts . . . . . . 0.24 (0.04)t 0.25 Weatings .. . . . . 0-65 (0.07)t 0.61 * Average of 23 results. t The standard deviation is given in parentheses. method, which involved the use of 2,4-diaminonaphthaleneJ a similar digestion procedure and solvent extraction of the selenium( IV) - 2,4-diaminonaphthalene complex. The determination of selenium(1V) in sea water was straightforward and there was a negligible source scatter signal (checked at the 216.5-nm selenium line). The sea water was acidified with concentrated hydrochloric acid (ie., 15 ml of acid were diluted to 100 ml with sea water). The sensitivity found, by using the method of standard additions, was similar (+5 per cent.) to that for standard solutions in 15 per cent. hydrochloric acid. The concen- tration of selenium(1V) was found to be less than 0.0001 pg ml-1. Conclusions The hydride generation technique, using a sodium borohydride reduction with subsequent detection by atomic fluorescence, is a very sensitive method for the determination of antimony, arsenic, selenium and tellurium. It should also be applicable to the determination of bismuth, germanium, tin and possibly even 1ead.l The author thanks Mr. C. K. Bishop of the Ministry of Agriculture, Fisheries and Food, Derby, for the animal feedstuffs and for the results of the spectrofluorimetric determination of selenium, and also the Directors of Shandon Southern Instruments Limited for permission to publish this paper. References 1. 2. 3. 4. 5. 6. Thompson, K. C., and Thomerson, D. R., Analyst, 1974, 99, 595. Tsujii, K., and Kuga, K., Analytica Chim. A d a , 1974, 72, 85. Thompson, K. C., and Reynolds, G. D., Analyst, 1971, 96, 771. Thompson, K. C., and Wildy, P. C., Analyst, 1970, 95, 562 and 776. Price, W. J., “Analytical Atomic Absorption Spectrometry,” Heyden and Sons, London, 1972, p. 5. Knudson, E. J., and Christian, G. D., Atom. Absorption Newsl., 1974, 13, 74. Received December 4th, 1974 Accepted January 2nd, 1975

ISSN:0003-2654    

DOI:10.1039/AN9750000307

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

Analyst, May, 1975, Vol. 100, pp. 311-315 31 1 Spectrophotometric Determination of Molybdenum in Steel with Thiocyanate and Tetraphenyl- arsonium Chloride A. G. Fogg, J. L. Kumar and D. Thorburn Burns Department of Chemistry, Loughborough University of Technology, Loughborough, Leicestershire, L E l l 3TU An alternative procedure is described for the colorimetric determination of molybdenum in steel. Molybdenum(V1) is reduced to molybdenum(V) with ascorbic acid and titanium(II1) before being caused to react with thiocyanate and extracting it with tetraphenylarsonium chloride into chloro- form that contains quinol. The procedure is sensitive (emax. = 17 400 1 mol-1 cm-1 at 470 nm) and precise. A forty-fold excess of tungsten over molybdenum can be tolerated. Results for a series of low-alloy and tungsten tool steels are reported.Thiocyanate extraction has been used for analytical purposes for over 100 years. In 1863 Braunl showed that the thiocyanate complex formed when molybdic acid was reduced with zinc in the presence of thiocyanate could be extracted into diethyl ether. Four years later he used the technique, with tin(I1) chloride as reductant, to detect molybdenum in minerals2 Despite this early work thiocyanate extraction methods for the determination of molybdenum have not found wide application. The extraction of metal - thiocyanate complexes has recently been reviewed by Sultanova et al.3 In previous work on the determination of tungsten with thiocyanate4p5 it has been shown that the precision of the method is greatly improved when the tungsten(V) - thiocyanate complex is extracted with an onium cation prior to spectrophotometric measurement.In the recommended procedure tungsten(V1) was reduced to tungsten(V) with tin(I1) and titanium(II1) before reaction with thiocyanate and extraction of the yellow complex into chloroform with tetraphenylarsonium chloride. A reducing agent, quinol, was added to the chloroform to prevent the formation of unwanted coloured oxidation products on standing. Attempts to improve the precision of spectrophotometric methods involving the formation of thiocyanate complexes for the determination of other metals, for example, molybdenum, rhenium and niobium , were not immediately successful. In developing any procedure involving the use of thiocyanate for the determination of molybdenum, a major concern is to prevent interference from tungsten by using mild reducing conditions in which molybdenum(V1) is reduced but tungsten(V1) is not.The nature of the reductant also affects the precision of the determination of molybdenum. A probable cause of low sensitivity and precision is the continuation of the reduction of molybdenum(V1) to the molybdenum(II1) state.6 Indeed, in the determination of tungsten, tin(I1) chloride can be used to reduce tungsten(V1) and molybdenum(V1) completely to tungsten(V) and molybdenum(III), respectively; molybdenum( 111) does not interfere in the determination of tungsten(V) by extraction with thiocyanate. Tin(I1) chloride is commonly used under milder conditions to form molybdenum(V) in molybdenum thiocyanate methodsJ7,* but it is not an entirely satisfactory reductant .6 y 9 - 1 1 Ginzburg and LurJef2 proposed the use of iodide in the presence of sulphite (to oxidise the iodine formed) for the reduction of molybdenum(V1) to molybdenum(V). Nabivanetss compared the performance of several reducing agents, including tin(II), iodide with sulphite, thiourea and ascorbic acid , for this reduction. Maximum colour development and stability were obtained with iodide and sulphite and with ascorbic acid. Solutions containing sulphuric acid were found to be more stable than those containing hydrochloric acid; maximum molar absorptivities at 465 nm were found to be 12 8001mol-1 cm-1. The effects of iron(II1) and copper( 11) were discussed. Several workers have applied iodide9J2,13 and ascorbic acid procedure^.^^^^^ Copper(1) was used by Crouthamel and Johnson,6 titanium(II1) by Fer’yanchich16 and tin(I1) plus titanium(II1) in the British Standard method.17 The interference from large amounts of tungsten can be masked by the addition of citric,17 tartaricls or oxalic acid.15312 FOGG et al.: SPECTROPHOTOMETRIC DETERMINATION OF MOLYBDENUM Analyst, Val. 100 The method described herein is a modification of Lazarev and Lazarev’s16 ascorbic acid reduction method, with the addition of titanium(II1) as co-reductant or catalyst and an onium salt to facilitate extraction into quinol-treated chloroform prior to spectrophotometric measurement. Experimental In comparing a number of procedures for the spectrophotometric determination of molyb- denum by use of thiocyanate, the precision, sensitivity, colour stability, interferences (parti- cularly from tungsten) and the results of extraction in the presence of tetraphenylarsonium chloride were considered to be important criteria in the final choice of a recommended pro- cedure.Preliminary studies were made of several procedures described in the literature. The current British Standard method1’ was examined by using British Chemical Standard Steel No. 323. The results obtained (0.099 per cent. of molybdenum) were within 1 per cent. of the certificated value (0.100 per cent.) and the coefficient of variation (10 results) was 1.7 per cent. It was found that Emax, was 8550 1 mol-l cm-l at 470 nm and that up to a 100-times excess of tungsten could be tolerated.The results obtained by using onium extraction were less reproducible than those obtained with the basic method. Crouthamel and Johnson’s6 procedure using copper(1) as reductant was found to be satisfactory for pure molybdenum(V1) solutions; copper(1) and iron(II1) were found to interfere with the onium extraction. Copper(1) alone could be removed by precipitation with benzoin a-oxime and coefficients of variation of 1.2 per cent. were obtained. In the presence of iron, its removal, together with the precipi- tation of copper under alkaline conditions, resulted in a loss of molybdenum by co-precipita- tion. The reduction using iodide in the presence of sulphitelS was found to be sensitive (gmaX. = 17 000 1 mol-l cm-l at 470 nm) and could be combined with extraction with an onium salt for pure solutions of molybdenum.The serious interference caused by iron and tungsten could not be overcome, despite extensive masking studies. The ascorbic acid reduction in the Lazarev and Lazarev procedure has been modified and successfully com- bined with an onium extraction. Reagents Ammonium thiocyanate solution, 25 per cent. m/V. Dissolve 125 g of ammonium thiocyanate in water and dilute to 500 ml with water. This solution should be prepared fresh, daily. Sulphuric acid, 50 per cent. V/V. To 400 ml of water carefully add, with stirring, 500 ml of concentrated sulphuric acid (sp. gr. 1.84). Cool the mixture, dilute it to 1 1, and mix. Hydrochloric acid, 20 per cent. V/V. To 800 ml of water add 200 ml of concentrated hydro- chloric acid and mix thoroughly.Ascorbic acid reagent solution, 10 per cent. m/V. Dissolve 10 g of ascorbic acid in 60 ml of water, add 2 ml of 0.1 M disodium EDTA solution and 10 drops of 98 per cent. formic acid, mix and dilute the mixture to 100 ml with water, then re-mix. Dissolve 1-05 g of tetraphenylarsonium chloride in water and dilute to 100 ml. Prepare a 1 per cent. m/V solution of quinol in ethanol and dilute 20 ml of this solution to 250 ml with chloroform. The stock ethanolic quinol solution is stable for about 1 week when stored in the dark, but should be discarded if a pink colour develops. This solution is available com- mercially. The solution used in the present work was found to contain 12.0 per cent. m/V of titanium(II1) chloride, by titration with cerium(1V) sulphate.Dissolve 0.75 g of molybdenum- (VI) oxide (analytical-reagent grade) in 5 ml of 0.1 M sodium hydroxide solution. Carefully add to this solution 10 ml of 50 per cent. V/V sulphuric acid and 10 ml of 50 per cent. V/V nitric acid and dilute to 1 1 in a calibrated flask. Dilute 50 ml of concentrated standard molybdenum solution to 500 ml with water in a calibrated flask. Tetra~henylarsonium chloride solution, 0.025 M. Chloroform containing 0.08 per cent. m/V of quinol. This solution is stable for about 2-3 weeks. The solution in chloroform should be prepared daily. Titanium(III) chloride solution, about 15 per cent. m/V. Concentrated standard molybdenum solution, 500 pg ml-1. Dilute standard molybdenum solzttion, 10 pg ml-1.Nitric acid, concentrated. Analytical-reagent grade. Oxalic acid dihydrate. Analytical-reagent grade. Iron(l11) chloride solution. Dissolve 72-50 g of iron(II1) chloride hexahydrate in 6 MMay, 1975 IN STEEL WITH THIOCYANATE AND TETRAPHENYLARSONIUM CHLORIDE 313 hydrochloric acid and make the solution up to 500 ml with 6 M hydrochloric acid in a calibrated flask. 1 ml of solution = 0.05 g of iron. Dissolution of Steel Samples Dissolve an appropriate amount (0.1 g) of steel by gently warming it with 10 ml of 20 per cent. V/V sulphuric acid and a minimal amount of concentrated nitric acid (10-15 drops). Heat until fumes appear, cool the solution and dissolve the residue in 50 ml of water by warming. Dilute the solution to 100 ml with water in a calibrated flask.NOTES- 1. In the analysis of tungsten steels, the dissolution is carried out by adding to an appropriate amount of steel 10 ml of concentrated hydrochloric acid and the minimum amount of concentrated nitric acid (1 ml) The mixture is heated to dryness and cooled. Water (50 ml) is added, followed by 4 g of oxalic acid. The mixture is then heated until all of the solids have dissolved. This solution is finally transferred to a 100-ml calibrated flask and made up to the mark with water. 2. The dissolution of niobium steels is carried out in the same way as for tungsten steels, except that the mixture is heated for half an hour after the addition of oxalic acid, in order to convert niobium into its hydrated oxide. The oxide residue is filtered off and washed six times with hot 5 per cent.sulphuric acid, adding the washings to the 100-ml calibrated flask before dilution. The presence of 500 pg of niobium did not affect the recovery of molybdenum at niobium to molybdenum ratios of up to 5 : 9. Procedure Pipette a portion of sample solution containing about 70 pg of molybdenum, corresponding to an absorbance of 0.4-0.5, into a separating funnel. Add sufficient 50 per cent. V/V sulphuric acid to give a final sulphuric acid concentration of 2-0 M, followed by 2 ml of water, then mix and cool. Add 5 ml of ascorbic acid solution per 100 mg of iron present, 0.2 ml of titanium(II1) chloride solution and mix thoroughly. To the mixture add 3 ml of ammonium thiocyanate solution, again mix and allow it to stand for 20 min for maximum colour to develop.Next add 1 ml of tetraphenylarsonium chloride solution and 10 ml of chloroform (quinol treated), shake the funnel for 30 s and allow the phases to separate. Pass the chloro- form layer through a Whatman No. 1 filter-paper into a 25-ml calibrated flask. Extract the aqueous phase with a further 5-ml portion of chloroform and collect as before. Then dilute the contents of the flask to volume with quinol-treated chloroform. Finally, mix the solution thoroughly and measure the absorbance at 470 nm against a chloroform blank, prepared by extracting all of the reagents in the absence of sample, in 1-cm cells. The calibration graph is prepared by using portions of the dilute (10 pg ml-l) standard molybdenum solution containing 20, 40, 60, 80, 100, 120 and 150pg of molybdenum plus 100 mg of iron(III), each treated by the procedure described above. Results The recommended procedure has been used in order to analyse standard molybdenum solutions in the presence of potential interfering ions and to analyse samples of British Chemical Standard steels.The addition of titanium(II1) chloride solution at the reduction stage was found to increase the sensitivity of the method by 10 per cent. compared with the use of ascorbic acid alone and to speed up colour development. The amount added, 0.1-1.0 ml, was not critical; any excess forms an insoluble precipitate with the tetraphenylarsonium cation and does not interfere so that the volume of 0.2 ml was chosen for ease of manipulation. The effect of variations in the amount of ascorbic acid was examined by using the recom- mended procedure with different concentrations of ascorbic acid in the range 2.0-20 per cent.with 100mg of iron and 1OOpg of molybdenum. Above a concentration of 8 per cent. recovery was constant. It was therefore decided that 5 ml of a 10 per cent. ascorbic acid solution would be used, and this is satisfactory except for those steels which contain small amounts of molybdenum; in such instances it is necessary to increase the volume of ascorbic acid by 5 ml per 100 mg of iron present. The effect of variation in sulphuric acid concentration was examined in a similar manner over the range 0.5-5 M. Above 1-5 M constant recoveries were obtained; 2.0 M was selected as being a convenient concentration. Recoveries were found to depend on the amount of iron(II1) initially present, as is shown in Table I.These results indicate that a ratio of iron to molybdenum of 100: 1 is necessary for maximum recovery, hence iron must be included in the solutions used to prepare the calibration graphs.314 FOGG et a!?. SPECTROPHOTOMETRIC DETERMINATION OF MOLYBDENUM Afidyst, 'Vd. 100 TABLE I EFFECT O F AMOUNT OF IRON PRESENT ON THE RECOVERY OF MOLYBDENUM 100 pg of molybdenum in 26 d of sample solution. Volume of 10 per cent. Absorbance Molybdenum * Iron added/mg ascorbic acid/ml (at 470 nm) recovery, per cent. 0 6 0.588 82 5 5 0.672 87 10 5 0.710 100 20 5 0.708 100 50 10 0.715 100 100 10 0.705 100 Under the recommended conditions the absorbances of extracts are very stable and remain constant for at least 24 h; Emax.= 17 400 1 mol-1 cm-1 at 450 nm. Possible interferences were examined by use of the recommended procedure, extracting 25-ml volumes containing 100 mg of iron and 100 pg of molybdenum, together with 500 pg of another metal ion. Results are given in Table I1 and show no major interference effects. TABLE I1 RECOVERY OF MOLYBDENUM I N THE PRESENCE O F DIVERSE METALS 100.0 pg of molybdenum added. Molybdenum Element Source found/pg Zn ZnSO, 99.8 100.1 98.9 Sn SnC1, 100.5 cu CUSO, 100.0 Cd Cd 2 Al A12(s04) 3 Pb Pb(C,H,O,)Z 98.8 V NH,V03 100.0 Cr cr2(s04)3 99.9 Mn MnSO, 100.1 Ni NiC1, 100.0 In the absence of oxalic acid, tungsten and niobium were found to interfere seriously at about the 1 : 1 tungsten or niobium to molybdenum ratio. The addition of oxalic acid was found to mask up to a 40: 1 ratio of tungsten to molybdenum and a 10: 1 ratio of niobium to molybdenum. Niobium, if present, precipitates as the hydrated oxide after the addition of oxalic acid and subsequent heating stage.co C O P , ) 2 99.8 The results obtained for a series of steels are given in Table 111. TABLE I11 ANALYSIS OF BRITISH CHEMICAL STANDARD STEELS Coefficient of Molybdenum Molybdenum variation Tungsten present, found, ( 10 results), present, BCS No. per cent. per cent. per cent. per cent. 25 1 0.18 0.18, 0.8 None 320 0-22 0.22, 0.4 0.17 22012 4.92 4.96 0.8 6.97 434 0.01 1 0*010, 1.0 None 325 0.16 0.16, 0.6 0.12 24 1 /2 0.63 0.53, I *2 19.9 Niobium present., per cent. None None None None None 0.10 Discussion The calibration graph for the recommended procedure is rectilinear over the range 0-150pg The pro- of molybdenum and corresponds to Emax.= 174001mol-1cm-1 at 470nm.May, 1975 IN STEEL WITH THIOCYANATE AND TETRAPHENYLARSONIUM CHLORIDE 315 cedure is more sensitive than the British Standard method and has been found to be more precise, particularly for low levels of molybdenum. The copper( I) reduction was sensitive and precise for pure solutions of molybdenum but interference could not be overcome in order to allow its use for the analysis of steels. A similar conclusion was reached for the iodide reduction procedure. The results obtained for a series of steels when using the modi- fied Lazarev and Lazarev procedure15 show excellent precision and accuracy. The mechanism for the determination of molybdenum(VI), after the addition of a reducing agent, with thiocyanate is still the subject of speculation. The increased sensitivity observed when using single-electron reductants, or the presence of copper(I1) and iron(II1) to form such reductants, has been explained on the basis of avoidance of the formation of molybdenum(III).s The increase in intensity has also been explained on the basis of the formation of hetero- polynuclear complexe~,~~ indeed the degree of interference from molybdenum in the deter- mination of tungsten depends on the amount of iron pre~ent.~ A recent paper20 puts forward evidence that the coloured species may in fact be molybdenum(V1) and not molybdenum(V), as is usually assumed.Reduction of molybdenum(V1) thiocyanate, MoO,(SCN),(H,O),, yields a five co-ordinate molybdenum(V), which by two half-Berry pseudo-rotations21 forms a molybdenum(V) species with the two thiocyanate ligands in non-equivalent positions.This complex is readily oxidised to a six co-ordinate molybdenum(V1) species with the thio- cyanate ligands remaining in non-equivalent positions. Despite the uncertainty of the nature of the coloured species, molybdenum - thiocyanate reactions can form the basis of excellent analytical methods as is shown above. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. References Braun, C . D., 2. Analyt. Chem., 1863, 2, 36. Braun, C. D., 2. Analyt. Chem., 1867, 6, 86. Sultanova, 2. Kh., Chuchalin, L. K., Iofa, B. A., and Zolotov, Yu. A., Zh. Analit. Khim., 1973, 28, Fogg, A. G., Marriott, D. R., and Burns, D. T., Analyst, 1970, 95, 854. Fogg, A. G., Jarvis, T. J., Marriott, D. R., and Burns, D. T., Analyst, 1971, 96, 475. Crouthamel, C. E., and Johnson, C . E., Analyt. Chem., 1954, 26, 1284. Ward, F. N., Analyt. Chem., 1951, 23, 788. Wilson, A. M., and McFarland, 0. K., Analyt. Chem., 1964, 36, 2488. Nabivanets, B. I., Ukr. Khim. Zh., 1958, 24, 775. Dick, A, T., and Bingley, J. B., Nature, Lond., 1946, 158, 516. Lillie, E. G., and Greenland, L. P., Analytica Chim. Ada, 1974, 69, 313. Ginzburg, L. B., and Lur’e, Yu., Zav. Lab., 1948, 14, 538. Hope, R. P., Analyt. Chem., 1957, 29, 1053. Williams, C. H., J . Sci. Fd Agric., 1955, 6, 104. Lazarev, A. I., and Lazarev, V. I . , Zav. Lab., 1958, 24, 896. Fer’yanchich, F. A., Zav. Lab., 1937, 6, 289. “Determination of Molybdenum, ” British Standards Handbook No. 19, British Standards Institution, Buss, H., Kohlschutter, H. W., and Miedtank, S., 2. Analyt. Chem., 1960, 178, 1. Nabivanets, B. I., Zh. Neorg. Khim., 1959, 4, 1797, Greenland, L. P., and Lillie, E. G., Analytica Chim. Acta, 1974, 69, 335. Springer, C. S., J . Amer. Chem. Soc., 1973, 95, 1459. 369. London, 1970. Received July 16th, 1974 Accepted October 15th, 1974

ISSN:0003-2654    

DOI:10.1039/AN9750000311

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

316 Analyst, May, 1975, Vol. 100, pp. 316-321 Determination of Nitrate in Soil and Water by an Adaptation of an Orange I Method D. L. Heanes Department of Agriculture, Box 1671 G.P.O., Adelaide, South Australia 5001 A procedure is described for the determination of nitrate in air-dry and field- moist soil and in water, by an adaptation of Middleton’s Orange I method. The molar absorptivity for nitrate-nitrogen in solution is greater than 2 x lo4 1 cm-1 mol-1 by the proposed method, and without further modification to the method 0.04-40 mg kg-l of nitrate-nitrogen in soil can be determined. The procedure is simpler and quicker than Middleton’s method and is free from interferences, provided that procedures carried out before colour development are carefully standardised. Comparative analyses by a distillation method are included.Middletonl described a method for the determination of nitrate-nitrogen in water and later2 adapted it for the analysis of soil. The method involves quantitative reduction of nitrate to nitrite by zinc powder followed by colorimetric determination of an azo dye (Orange I) produced by the diazotisation and coupling of nitrite, 1-naphthol and sulphanilic acid.2 Although the method2 is sensitive to low nitrite concentrations it has basic weaknesses: some nitrite is lost during the extraction and filtration steps owing to further reduction; the colour intensity may be affected by variation in the ionic concentration of the electrolytes present ; and the procedure is slow and may give unreliable results with both wet and dry soils.In this paper modifications to Middleton’s method are described that extend the range of concentrations to which the method is applicable, make it quicker and more accurate, and enable it to be used successfully with both wet and dry soil and with water. Recommended Method for Analysis of Dry Soil Reagents ( a ) Calcium acetate solution. A 0.5 per cent. m/V solution of anhydrous calcium acetate in ammonia solution (sp. gr. 0.91) - water (4 + 96), prepared as described by Middleton.2 (b) Manganese(II) sulphate solution. A 1 per cent. m/V solution of MnS0,.4H20 in 5 per cent. V/V aqueous acetic acid. (c) Zinc metal powder. Each batch of powder needs to be tested before it is used routinely as zinc powder varies in particle size and in the concentration of metallic impurities and therefore its reducing properties vary.Zinc powder is satisfactory if a net absorbance reading exceeding 0.7 is obtained for a concentration of 5mg1-1 of nitrate-nitrogen in the extract or diluted standard. AnalaR (90 per cent. m/m) zinc powder is sometimes satisfactory. Ninety-eight per cent. m/m zinc powder* containing less than 0.004 per cent. m/m of iron is preferred for improved reproducibility and precision. (d) Diazotisation reagent. A solution containing 0.1 per cent. m/V of sulphanilic acid and 0.08 per cent. m/V of 1-naphthol in 75 per cent. V/V aqueous acetic acid is prepared by dissolving sulphanilic acid and 1-naphthol in distilled water and glacial acetic acid, respec- tively, mixing and making up to volume.The reagent should be stored in an amber bottle away from direct light. Procedure Weigh 10 g of air-dry soil into a 100-ml Erlenmeyer flask. Add 50 ml of reagent (a), then 1 ml of reagent (b). Add a volume of reagent (c) corresponding to 0.05-0.2 g (standardise to &O.O2g) and close the flask with a rubber stopper. Shake the flask vigorously for the * Specially supplied by Hopkin and Williams.HEANES 317 chosen time* in a mechanical flask-shaker. Pour the suspension immediately into a Whatman No. 44 filter-paper and collect the filtrate in a 25-ml calibrated flask or test-tube containing 5 ml of reagent (d). When the level of the filtrate reaches the calibration mark, stopper and shake immediately. Allow the colour to develop for at least 30 min before measuring the absorbance with a spectrophotometer using a l-mm flow cell.Read the absorbance at a wavelength of 475 nm if the nitrate-nitrogen concentration in the soil is 0-25 mg k g l and at 425 nm for a concentration of 2540 mg k g l . TABLE I EFFECT OF VARIATION IN SHAKING TIME ON NET ABSORBANCE DUE TO 250pg OF NITRATE-NITROGEN Timelmin Absorbance at 475 nm 2 0,724 3 0.731 3.8 0.734 4 0.735 6 0.692 Standard graphs are prepared by substituting for the dry soil various amounts of stock solution containing 25 or 100 mg 1-1 of potassium nitrate nitrogen, the volume of reagent ( a ) used being reduced correspondingly. Stock solutions are prepared in ammonia solution (sp. gr. 0.91) - water (4 + 96). Testing the Procedure Nitrate redaction with zinc The colour intensity of the Orange I complex depends on the concentration of nitrite- nitrogen when the extract is mixed with reagent (d).This concentration is a function both of the formation and loss of nitrite, as analytical-reagent grade zinc powder, unlike pure zinc, reduces nitrite in addition to nitrate. The rates of reduction depend on the amount and quality of the zinc powder used. Some effects are illustrated in Fig. 1. Nitrite disappeared as more zinc was used, and so the absorbance due to nitrate was also reduced by large amounts of zinc. In practice, optimum amounts of zinc powder and shaking times can be determined empirically, as described above. 0.2 I I I I I I I 0 0.1 0-2 0.3 Mass of zindg Fig. 1. The effect on absorbance measured of the amount of zinc powder added to A, 250 pg of nitrite-nitrogen and B, 250 pg of nitrate-nitrogen.Shaking time, 3 min. Standard graphs By using the procedure described above, the absorbance measured at either wavelength increased linearly with the amount of nitrate-nitrogen added (Fig. 2). * The shaking time, usually 3 4 min, is determined empirically for each batch of zinc powder by selecting the time in which the maximum absorbance reading for a given concentration of nitrate-nitrogen is produced (Table I).318 HEANES: DETERMINATION OF NITRATE IN SOIL AND WATER Analyst, VOl, 100 i 0.7 - 0.6 - 0 40 80 120 160 200 240 280 320 360 400 N itrate-nitrogedpg Fig. 2. Standard graphs for nitrate-nitrogen measured at A, 475 nm and B, 426 nm. Values on abscissa indicate amounts of nitrate- nitrogen mixed with reagents (a)-(c) before filtration. Filtration and colour development Middleton2 stated that the shaken suspension should be filtered immediately and then reagent (a) added to produce the colour.Because of the dynamic nature of the nitrate - nitrite system, illustrated in Fig. 1, it is important that procedures carried out before colour de- velopment should be carefully standardised. Errors are most likely to result from variations in the time taken to filter the suspension and to add reagent (a). The effect of variations in the filtration time on the absorbance produced is shown in Table 11. As most soil suspensions filter in 3-5-6 min, it is clear that errors from this source are not serious provided that the standardised conditions described above are used.TABLE I1 THE EFFECT OF FILTRATION TIME ON NET ABSORBANCE DUE TO NITRATE SOLUTIONS Amount of Filtration Net absorbance a t nitrate-nitrogen in solution/pg time*/min 475 nmt 100 2 0-318 3.6 0.320 11 0-3 17 200 2 3.6 11 0.635 0.636 0.623 * Filtration through one circle of No. 1 filter-paper, one circle of No. 44 filter- paper and two circles of No. 42 filter-paper for 2, 3.5 and 11 min, respectively. t 0.013 absorbance unit = 0-4 mg kg-1 of nitrate-nitrogen in the soil. Further delay in the development of the colour is avoided by filtering the suspension into a flask that already contains reagent (d). This procedure is also quick and convenient. Stability of the Orange I dye The colour developed from a standard nitrate solution and from extracts of acidic and alkaline soils were stable for 2 d (Table 111).Hence extraction and colour development may be continued for 1 or 2 d and absorbance measured up 2 d later. TABLE I11 STABILITY OF THE ORANGE I DYE Absorbance a t 476 nm Sample Iniiial 24 h 48 h 7i h Acidic soil .. . . 0.322 0-323 0.326 0,340 Alkaline soil . . . . 0.176 0.175 0.178 0.183 Working standard . . 0.164 0.166 0.166 0.170May, 1975 319 Inter ferences Organic matter. Filtrates prepared so far, from a wide range of acidic and alkaline soils, have been free of coloration from organic matter. Clay inJiZtrate. With heavy soils, clay may cause some cloudiness. This can be overcome by using a finer grade of filter-paper (e.g., Whatman No. 42) or less soil. Nitrite content. The prescribed method measures the concentration of nitrate-nitrogen and nitrite-nitrogen.A separate measurement of the nitrite-nitrogen content, which is usually insignificant, can be made by omitting zinc from the procedure. Chloride. Sodium chloride concentrations as high as 0.5 per cent. mlm in soil caused no interference in the analysis for nitrate. Dilution of the extract Middleton2 found it necessary to dilute soil extracts up to 20-fold in order to provide a satisfactory range of concentrations to which the method was applicable. With the tech- nique recommended here, no dilution is needed when analysing dry soil, Analysis of wet soil or water involves up to 20 per cent. V/V dilution of reagent (a) but this was found to have no effect on either the pH of the buffered extract or on the measured concentration of a nitrate solution, neither did the replacement of up to 20 per cent.of the volume of reagent (a) with equivalent amounts of stock nitrate solution affect the linearity of the standard graph, Dilution errors, therefore, can be avoided with the technique recommended here. Recovery of nitrate-nitrogen A range of soils containing 0-34 per cent. m/m of calcium carbonate were analysed for nitrate-nitrogen before and after the addition of 20 mg k g l of nitrogen as potassium nitrate. Two batches of AnalaR zinc powder were used: zinc(A) was satisfactory according to the test described earlier (absorbance with 5 mg 1-1 of nitrate-nitrogen = 0.704) but zinc(B) was un- satisfactory (absorbance = 0.598). The results (Table IV) show that with zinc(A) recovery was consistently high (98-100 per cent.m/m). The full recovery of nitrate reported here for a variety of Australian soils contrasts with an incomplete recovery (as low as 92 per cent. m/m) reported by Middleton.2 Interfering ions were apparently extracted from the Nigerian soil by mixed reagents ( a ) and ( b ) ; the possibility that this may occur more generally needs to be examined. Sensitivity and range The sensitivity of the technique depends on the relative rates of reduction of nitrate and nitrite (see above). When a solution containing 250 pg of nitrate-nitrogen gives a net absor- bance of 0.700 the molar absorptivity is 2 x lo4 1 cm-1 mol-l; using a soil sample of 10 g this is equivalent to 0.02 mg k g l of soil. Sensitivity can be increased by using purer and more finely divided zinc powder or a longer optical path length. With a 10-g soil sample, the analytical range is 0.0440 mg k g l .Still smaller concen- trations can be measured accurately if longer optical path lengths are used. The upper limit can be extended by taking smaller amounts of soil. Flow cells with optical path lengths of 1 mm extend the range of analysis 20-fold without further dilution. Middleton2 was probably handicapped by the unavailability of such flow cells in 1959. Comparison with a Distillation Method of Nitrate-nitrogen Determination A range of soils containing 0-58 per cent. m/m of calcium carbonate and varying in texture from sand to clay were analysed for nitrate-nitrogen content by the method described and by an automated distillation method.s (The distillation analyses were conducted by Messrs.S. McLeod and G. Whitehead of the C.S.I.R.O. Division of Soils, Glen Osmond.) In general, the results for the two methods agreed satisfactorily (Table V) and analyses by the method described were consistent. Each procedure can be used to analyse soil samples at the rate of 400 per week. With Middleton’s procedure2 only about one third of this number of analyses are possible per week. The proposed method is of particular use in laboratories with limited resources. BY AN ADAPTATION OF AN ORANGE I METHOD Analysis of Field-moist Soils and Water Changes that occur in nitrate-nitrogen during drying4 indicate that some soils should be analysed in the field-moist state. Water samples have very low nitrate concentrations that320 Analyst, VoZ.100 require highly sensitive methods of analysis. The method for analysing air-dry soils was modified so as to measure nitrate-nitrogen in field-moist soil and in water. Changes in electrolyte concentration caused by adding up to 10 ml of water did not affect the results obtained by using this method; neither has the pH of any extracts analysed in this laboratory been outside the range (10.2-1 1.2) stipulated by Middlet0n.l However, as a precautionary measure in the analysis of water samples, additional ammonia can be con- veniently added to reagent (a) in order to compensate for (i) the amount lost by dilution with water and (ii) the presence of nitrate-nitrogen or nitrite-nitrogen in the ammonia. HEANES: DETERMINATION OF NITRATE IN SOIL AND WATER TABLE IV RECOVERY OF NITRATE-NITROGEN (20 mg k g l ) ADDED TO A RANGE OF SOILS, USING TWO BATCHES OF ZINC POWDER Ni trate-ni trogen concentration t Img kg-1 Soil* Acidic soils- Sandyloam ..0-10 cm Sand .. .. 10-30 cm Loamyclay . . 50-60 cm Alkaline soils- Clay loam . . 0-15 cm Loamy clay . . 10-30 cm Loamyclay .. 30-45 cm Loamy clay .. 45-60 cm Sandyloam . . 0-15 cm Calcium carbonate, per cent. mlm . . 0 .. 0 .. 0 7 .. 10 * . 20 . . 25 .. 34 , 1 Zn (4 - Stage'; Mean S.D. Zn (B), mean I I1 Recovered I I1 Recovered I I1 Recovered I I1 Recovered I I1 Recovered I I1 Recovered I I1 Recovered I I1 Recovered 10.7 30.3 19.6 9-0 28-8 19-8 1.0 20-9 19.9 3.2 23.2 20.0 1.6 21.4 19.9 5.7 25.7 20.0 11.0 30.8 19.8 8.6 28.6 20.0 0.1 0.1 0 0.1 0.2 0.1 0.1 0.1 9.8 26.6 16.8 8.9 26.7 17.8 1.0 19.8 18.8 3.2 23.2 20.0 1.5 21.0 19.5 5-2 24-6 19.4 10.9 30.1 19.2 8.6 28.2 19.6 * Ranges given are sampling depths.t Triplicate determinations. '; I, before addition of nitrate-nitrogen; 11, after addition of nitrate-nitrogen. Field-moist Soil Weigh 15 g of field-moist soil into a suitable dish and dry in an oven at 105 "C. Record the oven-dry weight (S). Meanwhile, weigh an identical amount of the soil into a 100-ml Erlenmeyer flask. Add 50ml of reagent (a) and proceed as recommended for dry soils. Calculate the results as follows: (66 - S) Nitrate-nitrogen concentration in soil (mg k g l ) = n x - 51 x S where n is the mass of nitrate-nitrogen (pg) read from the abscissa of the standard curve and 66 - S is the volume of the extraction solution (51 ml) plus the volume of water in 15 g of field-moist soil [the volume in millilitres equivalent to (15 - S) g ] .Water Modify reagent (a) by adding 8 ml 1-1 of aqueous ammonia (sp. gr. 0.91). To a 10-ml volume of the sample add 40 ml of modified reagent (a). Then add 1 ml ofMay, 1975 321 reagent (b). Proceed as recommended for dry soils. Determine the amount (in micrograms) of nitrate-nitrogen in the 10-ml volume from the standard curve for dry soil (Fig. 2). In order to convert the result into milligrams per litre of nitrate-nitrogen in the sample, multiply by 0.1. BY AN ADAPTATION OF AN ORANGE I METHOD TABLE V COMPARISON OF CONCENTRATIONS OF NITRATE-NITROGEN FOUND IN SOILS BY THE PROPOSED METHOD AND BY A DISTILLATION METHOD Calcium carbonate, / Soil* per cent.mlm Acidic soils- Sand .. .. .. 0 Sand .. .. .. 0 0-10 cm 0-10 cm 0-10 cm 0-10 cm 40-50 cm Loamysand .. .. 0 Sandyloam . . .. 0 Clay . . .. .. 0 Nitrate-nitrogen concentrationt /mg kg-l Proposed method Distillation method -7 & Mean S.D. Mean S.D. 6.2 0.1 6.7 0.8 8-3 0.1 8.3 0-6 38.0 1.1 37.7 0.9 1.2 0.2 1.7 0.5 8.2 0-3 7.8 0-7 A I I Alkaline soils- Sand .. .. .. 58 1-1 0 1.7 0.2 Loamysand .. .. 34 74.5 1.5 73.0 0.4 0-10 cm 0-5 cm 0-10 cm * Ranges given are sampling depths. t Triplicate determinations. Sandyloam .. .. 30 7.1 0.1 7.9 0.1 Conclusion In the proposed procedure, errors that previously were likely to arise from delays in colour development have been eliminated, and large dilutions are no longer needed. These changes have made the technique simpler and quicker. Furthermore, some processes that affect the accuracy and sensitivity of the method, such as the simultaneous reduction of nitrate and nitrite, are now better understood. The technique can be used to determine nitrate concentrations in soil accurately and sensitively and is suitable for routine analysis. The method is free from the interferences that occur in some other routine procedures, such as the phenoldisulphonic acid method. Its high sensitivity makes it ideal for measuring low concentrations, such as occur in water samples. The technique can now be used with greater confidence, convenience and accuracy than was possible previous1y.ls2 The author thanks Dr. A. Clarke, Mr. R. French, Mrs. E. Bennett and Miss J. Noble (Department of Agriculture) for their assistance in preparing the manuscript and Messrs. S. McLeod and G. Whitehead of the C.S.I.R.O. Division of Soils for undertaking comparative analyses. He also thanks Messrs. D. Harvey and D. Lewis of the Department of Chemistry, Adelaide, South Australia, for initial help in his investigations. References 1. 2. 3. 4. Middleton, K. R., Chemy Ind., 1957, 1147. Middleton, K. R., J . Sci. Fd Agric., 1959, 10, 218. Keay, J., and MenagB, P. M. A., Analyst, 1970, 95, 379. Storrier, R. R., J . Aust. Inst. Agric. Sci., 1966, 32, 106. Received August 20th, 1974 Accepted November 25th, 1974

ISSN:0003-2654    

DOI:10.1039/AN9750000316

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

322 Analyst, May, 1975, Vol. 100, $@. 322-324 Interference of Mercury(l1) in the Colorimetric Determination of Inorganic Phosphate in Water R. W. Tillman and J. K. Syers Department of Soil Science, Massey University, Palmerston North, New Zealand Mercury(I1) caused a significant positive interference in the colorimetric determination of low levels of inorganic phosphate by the procedure of Murphy and Riley. The interference resulted from the formation of a precipitate that varied in particle size and was not always visible to the naked eye. With higher levels of inorganic phosphate a coarse precipitate formed t h a t partially removed the molybdophosphate complex ion from solution. Elimination of mercury(I1) interference was achieved by formation of a complex that resulted from the addition of chloride or a metabisulphite - thiosulphate reagent, as used in a conventional procedure for the removal of arsenate prior to the determination of inorganic phosphate.Mercury(I1) chloride is frequently recommended as a preservative for water samples1 that are required to be analysed for one or more forms of pho~phate.~,~ Although in the book “Methods of Chemical Analysis of Water and waste^"^ it is indicated that mercury(I1) can interfere in the colorimetric determination of inorganic phosphate in solution by the method of Murphy and Riley,* no reference to the literature on this subject is given and information on the nature and extent of the interference is not presented. Furthermore, the problem of interference by mercury is rarely mentioned in analytical manuals.It appears that steps to remove this interference are either not usually taken or are taken fortuitously, as in the reduction procedure used with water samples in which the presence of arsenate is suspected. Preliminary data indicated that mercury(I1) could interfere appreciably in the colori- metric determination of phosphate by the method of Murphy and Riley.4 This paper reports on the nature and extent of this interference and on its removal by the addition of an excess of chloride ions or a metabisulphite - thiosulphate reagent. Experimental The effect of mercury(I1) interference in the Murphy and Riley4 procedure was evaluated by adding various amounts of a mercury(I1) chloride standard to inorganic phosphate standards contained in the 50-ml calibrated flasks used for colour development.In order to investigate the extent of interference at realistically low phosphate concentrations, 0, 440, 880 and 1320pg of mercury(I1) were added to 0, 1, 2 and 4pg of inorganic phosphate, giving concentrations of 0, 8.8, 17.6 and 26.4 pg ml-l of mercury(1I)l and 0, 0.02, 0.04 and 0.08 pg ml-1 of inorganic phosphate, respectively, in the final 50 ml of solution used for the colorimetric determination. The range of mercury concentrations used covers that recom- mended by Howe and Ho1ley.l The effect of the addition of 26.4 pg ml-l of mercury(I1) on the recovery of two levels of inorganic phosphate (0.080 and 0.800 pg d-l) was evaluated as a function of time, using an appropriate control [no addition of mercury(II)] .The effect of including a conventional arsenate reduction step6 on the recovery of added inorganic phosphate in the presence of mercury(I1) was investigated by adding 5ml of a reductant solution (prepared by mixing 40 ml of a 20 per cent. sodium metabisulphite solution, 40 ml of a 1 per cent. sodium thiosulphate solution and 20 ml of 5 N sulphuric acid) to inorganic phosphate standards contained in 50-ml calibrated flasks, prior to the addition of the Murphy and Riley “mixed reagent.” The levels of inorganic phosphate used in the experiment reported in Table I1 are proportionate to those of mercury(I1) and can be regarded as representing a range of aliquots of a water sample to which mercury(I1) chloride has been added as a preservative. In all instances the source of inorganic phosphate was potassium &hydrogen ortho- phosphate.Inorganic phosphate was determined by the method of Murphy and Riley,&TILLMAN AND SYERS 323 All reported using a Unicam SP1800B spectrophotometer and a 10- or 1-cm path length. values represent the mean of at least duplicate determinations. Results and Discussion Addition of mercury(I1) caused a significant positive interference in the colorimetric determination of inorganic phosphate in the range 0-0.080pgml-1 by the procedure of Murphy and Riley4 (Table I). With one exception, the recovery of added inorganic phosphate increased with increasing mercury( 11) addition at each level of added inorganic phosphate, recovery values ranging from 100 to 265 per cent. TABLE I EFFECT OF MERCURY(II) CHLORIDE ON THE RECOVERY OF ADDED INORGANIC PHOSPHATE BY THE MURPHY AND RILEY PROCEDURE Inorganic phosphate HgC1, added/pg ml-l added/pg ml-1 0.080 0 0.080 8.8 0.080 17.6 0.080 26.4 0.040 0 0.040 8-8 0.040 17.6 0.040 26-4 0,020 0 0.020 8.8 0.020 17-6 0.020 26.4 0 0 0 8.8 0 17.6 0 26.4 Inorganic phosphate Determined/pg ml- 1- Recovery, per cent.0.080 100 0.1 18 147 0.129 161 0-163 204 0.041 102 0.054 135 0-080 200 0-068 170 0.020 100 0.036 175 0.040 200 0-053 265 o*ooo 0.007 0.024 0.053 It was observed that the positive interference resulted from the formation of a precipitate. This precipitate, however, varied in particle size and was not always visible to the naked eye. At low phosphate concentrations (0.080 pg ml-1 in the final 50 ml of solution), the precipitate was finely divided and gave a relatively stable colloidal suspension that enhanced the absorbance. At a higher phosphate concentration (0.800 pg ml-1 in the final 50 ml of solution), a coarser precipitate formed and this settled rapidly in the flask, causing a decrease in absorbance.This decrease in absorbance as the precipitate settled in the mercury(I1)- containing system is attributed to the removal of the molybdophosphate complex ion from solution, presumably by a sorption reaction. This conclusion was confirmed by determining the absorbance of 0.080 pg ml-1 inorganic phosphate standards in the presence and absence of 26.4 pg ml-l of mercury(I1) following centrifugation. A blue precipitate was separated by centrifugation from the mercury( 11)-containing system and the absorbance was reduced by an amount similar to that obtained for the mercury(I1)-containing system that was not centrifuged.The chemical composition of the precipitate obtained in the presence of mercury(I1) is not known, although molybdate and ascorbate ions are required for its formation. Because the precipitate forms in the absence of inorganic phosphate (Table I), it is clear that the molyb- dophosphate complex ion is not involved. Interference from mercury(II), at the highest concentration used in this study (26.4 pg ml-l in the final 50 ml of solution), was eliminated completely by the addition of at least 10 pg ml-l of chloride, as sodium chloride, to 0.080 pg ml-1 of inorganic phosphate in the final 50 ml of solution.This observation confirms a previous finding3 that excess of chloride ions can prevent interference from mercury(I1). The formation of a mercury(I1) - chloride complex ion probably prevents the development of a precipitate when the molybdate reagent is added to a sample containing mercury(I1) in the presence of excess of chloride ions. Such a complex could involve HgC13- (reference 6) or HgC142- (reference 7). The results given in Table I1 indicate that the recovery of added inorganic phosphate in the presence of mercury(I1)324 TILLMAN AND SYERS varied from 98 to 100 per cent. when the metabisulphite - thiosulphate reagent was added. In the absence of the latter reagent, the recovery varied from 130 to 178 per cent. The ability of thiosulphate to form very stable, soluble complexes with mercury(I1) is well established,’ and it thus prevents the formation of a precipitate when the mixed molybdate reagent is added to a sample containing mercury(I1) in the presence of the metabisulphite - thiosulphate reagent.It is fortuitous that mercury( 11) interference is eliminated during the conventional reduction step for the removal of arsenate. Because arsenate is a significant component only of waters in rather restricted areas of the world, this reduction step is not expected to be commonly used. TABLE I1 PHOSPHATE BY THE MURPHY AND RILEY PROCEDURE WITH AND WITHOUT A EFFECT OF MERCURY(I1) CHLORIDE ON THE RECOVERY OF ADDED INORGANIC METABISULPHITE - THIOSULPHATE REDUCTION STEP Inorganic phosphate r-- Determined (with Inorganic phosphate HgC12 reduction) / addedlpug ml-l added/pg ml-1 pg ml-l 0-050 22.0 0.049 0.040 17.6 0.040 0-030 13-2 0.030 0.020 8.8 0.020 0.010 4.4 0.010 7 Determined (without Recovery, reduction) / Recovery, per cent.pg ml-l per cent. 98 0.089 178 100 0.061 163 100 0.046 152 100 0.027 136 100 0.013 130 The results obtained in this study indicate that erratic and erroneously high results can be obtained in the determination of low levels of inorganic phosphate in waters when mercury(I1) is used as a preservative. This interference can be removed by complexing the mercury(I1) with chloride or thiosulphate. This research was supported by a grant from the Ministry of Agriculture and Fisheries and the Department of Scientific and Industrial Research, and by the Faculty of Agricultural and Horticultural Sciences, Massey University. 1. 2. 3. 4. 6. 6. 7. References Howe, L. H., and Holley, C . W., Envir. Sci. Technol., 1969, 3, 478. Jenkins, D., J . Wat. Pollut. Control Fed., 1967, 39, 159. “Methods of Chemical Analysis of Water and Wastes,” Environmental Protection Agency, Cincinnati, Murphy, J., and Riley, J. P., Analytica Chim. Acta, 1962, 27, 31. Jackson, M. L., “Soil Chemical Analysis,” Constable, London, 1958. Garrett, A. B., J . Amer. Chem. Soc., 1939, 61, 2744. Sidgwick, N. V., “Chemical Elements and their Compounds,’’ Clarendon Press, Oxford, 1962. Ohio, 1971. Received September 23rd, 1974 Accepted Novembev 20th, 1974

ISSN:0003-2654    

DOI:10.1039/AN9750000322

出版商:RSC    

年代:1975

数据来源: RSC