ISSN:0003-2654    

DOI:10.1039/AN97500FX025

出版商:RSC    

年代:1975

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97500BX027

出版商:RSC    

年代:1975

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97500FP077

出版商:RSC    

年代:1975

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97500BP081

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

JULY, 1975 The Analyst Vol. 100, No. 11 92 Chemiluminescence in Gas Analysis and Flame-emission Spectrometry A Review* J. H. Glover Consultant, 76 Craven Gardens, Wimbledon, London, S W19 8L U Summary of Contents Introduction Theory Gas-phase reactions Efficiency of chemiluminescence Ozone Oxides of nitrogen Measurement of nitrogen dioxide Ethylenic hydrocarbons Sulphur compounds Peroxyacetyl nitrate Carbonyls ame chemiluminescence Sulphur and phosphorus Halogens Nitrogen compounds Organic compounds Molecular-emission cavity analysis (M’ECA) Future developments F Introduction Although the phenomenon of chemiluminescence has been known since the middle of the 19th century, it has only recently been introduced as a technique for gas analysis. In this application it has solved a number of problems and has provided a simple and very elegant new method of analysis.Chemiluminescence can be defined as emission of light as a result of a chemical reaction. There are striking examples to be observed in nature, such as the “afterglows” observed in the upper atmosphere, which are the result of reactions between atomic gases. Another well known manifestation is the bioluminescence of the firefly and other insects. Studies of chemiluminescence have, until recently, been directed tbwards an understanding of the mechanism by which energy is transferred within the reacting molecules and elucidation of the energy states concerned. Gas-phase reactions have received a great deal of attention, particularly in relation to upper atmosphere studies, and there is a large volume of literature devoted to this topic.Young and Sharplesl give a good indication of the extent of previous work and add their own contribution on specific excitation processes that could be involved in the various chemiluminescence reactions of atomic oxygen and nitrogen. Analytical applications of chemiluminescence began to appear in the late 1950s and consist of solution- and gas-phase techniques. Also, certain flame-emission phenomena are caused by chemiluminescence reactions and many of these have found useful application in analysis. This review will be concerned with gas-phase and flame-emission techniques. The major factor promoting the use of chemiluminescence methods has been the greatly increased emphasis on air pollution and environmental analysis during the last few years.449 * Reprints of this paper will be available shortly. For details see summaries in advertisement pages.450 GLOVER: CHEMILUMINESCENCE IN GAS ANALYSIS Analyst, VoZ. 100 For example, the study of vehicle exhaust emissions exposed the lack of a method of analysis for oxides of nitrogen that would be suitable for use with continuous processes. The wet- chemical methods that were available when this environmental work was begun are more suitable for laboratory use, and there is a certain lack of confidence in the exact interpretation of results obtained by using them. Regulations limiting the composition of vehicle emissions were laid down by the Federal Authorities in America and motor car manufacturers were faced with the necessity of installing analytical equipment for the purpose of conducting exhaust-gas analysis on a production-control basis.The chemiluminescence technique was applied successfully to this problem in 196g2p5 and soon gained wide acceptance because of its considerable advantages of convenience, speed, reproducibility and selectivity; so much so, that it is now the standard technique for the measurement of oxides of nitrogen in vehicle emissions. The main advantages of the tech- nique have been indicated by Stevens and Hodgeson4; because it is an emission process it is more suited to trace analysis than is absorption spectroscopy, in that a small positive value is being measured against a low background in contrast to the much more difficult measurement of a small difference between two very large absorptions.Further, the require- ments for chemiluminescence to occur are such as to confer a high degree of specificity on the technique; and finally, the negative interference of quenching is predictable and can usually be accommodated. Instruments based on chemiluminescence are fairly easy to produce and extremely simple to operate, so that there is a large incentive to seek other applications for the technique. This review will attempt to present the progress that has been achieved so far. Theory The phenomenon of chemiluminescence is the result of the process of chemi-excitation, which involves increasing the total energy of a molecule by means of a chemical reaction. The concept of the excited state is essential to the theory of spectroscopy and is discussed thoroughly in standard textbooks, so that only a very limited treatment of aspects that are directly relevant to chemiluminescence is necessary here.When an excited molecule emits radiation it undergoes a permitted transition to a lower energy state. The relationship between the change in energy ( E ) and the frequency of the emitted radiation is given by the following equation : El - E , = AE = hv where 12 is Planck's constant and v is the frequency. Chemi-excitation is just another way of producing an excited molecule, as distinct from methods such as thermal excitation, irradiation or electronic excitation. The process, followed by emission of the corresponding luminescence, can be written5 : M .. * . (1) .... * (2) A+B---+C*+D .. .. c*-c + hv .. Indirect chemiluminescence can occur by a reaction such as: M A+B+C---+AB+C .. .. .. * - (3) Photolysis of certain substances may lead to the formation of excited species: AB + hv-A* + B .. .. .. (4) .. * ' (5) A + C D + h v - - + A C * + D .. .. The term photochemiluminescence has been used to define such reactions as (4) and (5).6 The process of excitation has some bearing on the type of radiation obtained and a simple view of the mechanism can be taken by considering the changes in potential energy that occur when an ideal diatomic molecule undergoes regular linear vibrations. Plotting potential energy against the distance between the nuclei gives rise to graph A , shown in Fig. 1, maximumJuly, 1975 AND FLAME-EMISSION SPECTROMETRY : A REVIEW 451 potential energy occurring when the two nuclei are nearest to each other and therefore subject to maximum repulsive forces.At the equilibrium position, E, the repulsive forces are balanced by the attractive forces and the latter increase, giving rise to more potential energy as the molecule is further stretched. Finally, dissociation occurs at point d. The horizontal lines on the graphs represent vibrational energy levels. The absorption of energy gives rise to an excited state that can be represented by graph B in Fig. 1. Here it is assumed that the time required for quanta to be absorbed is so small as to be negligible compared with the time required for the nuclei to move a significant distance. In this instance, therefore, the energy curve of the excited molecule can be drawn directly above that of the stable molecule and transitions to the ground state can be represented by vertical lines.AB is the most probable 0 -+ 0 transition, CD the most probable 3 -+ 3 and so on. A situation of this type would give rise to intense emissions of short wavelength (AE is large). Fig. 2 illustrates the instance when excitation is effected by a process that results in changes in the internuclear distances; for example, an excited molecule that has been pro- duced as a result of chemical change may be in a different state of vibration from that of a normal ground-state molecule. Transitions in this instance may occur as shown, AB being the most probable 0 -+ 4 and may give rise to changes when AE is small and the corre- sponding emission is weak and of long wavelength.In general, excitation by high-energy sources, such as discharge tubes, will produce a situation corresponding to that shown in Fig. 1, whereas collision excitation and chemi-excitation would be expected to produce the energy changes shown in Fig. 2. For example, Cormier et aL5 distinguish vibrationally excited molecules arising from exothermic gas-phase reactions from electronically excited molecules. The former emit in the infrared region in contrast with ultraviolet and visible-light emissions that result from electronic excitation. Fig. 1. Graphs of potential energy versus Fig. 2. Graphs of potential energy versus distance distance between nuclei ( A ) and absorption of between nuclei and absorption of energy when energy ( B ) for an ideal diatomic molecule. excitation is effected by a process resulting in changes in internuclear distances.Collision excitation can be expressed by constructing potential energy graphs, the simplest form being the association of two atoms, which can be illustrated by graphs of the type shown in Figs. 1 and 2. Reaction between molecules to give excited species is a much more complicated process as it frequently involves a three-body collision. Methods of studying such reactions involve the construction of three-coordinate models in which the collision complex is moving over a potential energy surface with respect to time. The various surfaces correspond to permitted energy states, and can intersect when vibrational energy is converted into electronic excitation.'452 GLOVER: CHEMILUMINESCENCE IN GAS ANALYSIS Analyst, VoZ.100 Efficiency of Chemiluminescence The amount of radiation emitted as a result of chemi-excitation is obviously a very important factor in considering the analytical usefulness of a particular reaction. The intensity and frequency of the emitted radiation is influenced by the factors referred to in the previous section, but the amount of radiation obtained for a given amount of reactant is dependent on the fate of the excited molecules. An excited molecule can also lose its energy by collision so that the greater the chances of collision in a system, the smaller will be the number of excited molecules available for radiation. There are therefore two opposing factors, the radiative life and the collisional life, which need to be considered in more detail.The probability (A) that a molecule in energy state 2 will undergo transition to energy state 1 with emission of radiation of frequency v is given by: = 8s x 3/Ldv cN1 p2 0 where N , is the number of molecules per unit volume in state 1 ; v is the wavenumber; P, and P, are the probabilities of the two states; a is the absorption coefficient ; and c is the velocity of light. If there is only one possible transition from state 2, then the radiative life, T , is given by 1 T = - 4 1 but if other transitions are possible, then the reciprocal of the mean life will be equal to the sum of the transition possibilities : 1 -= 442, + A2b + A,, - - - etc. Calculations of mean life can be made by using values for N obtained by use of the Boltzmann factor: where k is the Boltzmann constant and NE is the number of molecules in any state with energy E.By using calculations of this type, radiative lifetimes of excited species can be estimated, and a typical value is found to be 10-8 s. Collision life can be deduced from kinetic considerations and the following equation gives the number of collisions in unit time (C) that a molecule might be expected to undergo- where n is the number of molecules per millilitre; (T, the diameter of the molecule; and m the mass of the molecule (m = Mr x 1-66 x 10-24g; M , is the relative molecular mass). In the instance of nitrogen at n.t.p., this relationship indicates that a molecule would have 8 x log collisions per second, or a collision life of 1.3 x 10-lo s.The situation is therefore that an excited molecule can undergo at least 100 collisions during the time that it can lose its energy by radiation, and for less favoured transitions the number of collisions may be several orders of magnitude higher. Assuming that collisions will deactivate the molecule, the amount of radiation will be much reduced. Thus, two important conclusions emerge : one is that chemiluminescent measurements are best made under reduced pressure, so that the possibility of collisions is reduced; the second is that “quenching,” or collisional deactivation by a third body, will depend on the nature of the other molecules present in the system. Different diluents will have different effects on the amount of radiation emitted.An important factor that is omitted from the above simple treatment is the ability of an excited species to survive a number of collisions without deactivation. For example, the NO,* species is deactivated by one collision, whereas activated mercury atoms appear to survive many collisions. It is therefore more useful to consider the quenching power of a diluent gas, i.e.,July, 1975 AND FLAME-EMISSION SPECTROMETRY: A REVIEW 453 where Ip and I , are the radiation intensities at pressures of the quenching gas of p and 0, respectively; tr and tc are radiative and collisional lifetimes ; and a is the quenching coefficient. The over-all efficiency of a chemiluminescent reaction is the sum of the factors discussed, and according to Seitz and Nearys rarely exceeds 0.01, as expressed by the equation Number of photons emitted Number of molecules reacting Efficiency = - Bioluminescence, in contrast, is characterised by values for efficiency that approach unity.The reader is referred to Cormier et aZ.,5 Laidler' and Gaydong for a fuller treatment of the theory, and to Clough and Thrushlo for a more detailed discussion of quenching. Gas-phase Reactions Chemiluminescence in the gaseous phase has very important analytical applications and is widely used in the measurement of air p~llution.~ Well developed commercial analysers are now available, which are based on the application of chemiluminescent reactions. The method used is to mix an excess of reagent gas with the sample gas and then to measure the light produced; the two gas streams are fed into the reaction vessel at constant rates, thus facilitating continuous measurement.The light intensity ( I ) is proportional to the concentrations of the reactants : I = k [R] [XI where R is the reagent gas and X is the sample gas. [XI it can be assumed to be constant, then If [R] is made large compared with I = [XI The intensity of the emitted light is therefore directly proportional to the concentration of the constituent that is being determined provided that there is a large excess of reagent gas. The equipment used for the determination is simple and is shown in Fig. 3. It consists of an arrangement for ensuring constant flow-rates of the sample and reagent gases; this is usually a pressure regulator followed by a constriction.The two gases meet in the reactor, which commonly has an end window through which the light passes on, via the appropriate optical filter, into the photomultiplier tube. The reactor exhausts into a vacuum pump, which maintains the desired operating pressure. The electronic circuitry is not complicated Sample f- To Reagent U REG c = R = PM = PS = Fig. 3. Constriction A = Amplifier Reactor PR = Potentiometric recorder Photomultiplier REG = Pressure regulator Power supply Layout of chemiluminescence gas-phase analyser.454 GLOVER: CHEMILUMINESCENCE I N GAS ANALYSIS Analyst, Vd. 100 and consists of a high-voltage power supply to power the photomultiplier and an amplifier to facilitate the reading of light intensity. Ozone A gas-phase chemiluminescence analyser for ozone was first described in 1965,11 and was based on the reaction between ozone and ethylene, which gives rise to a continuum of radiation with a peak intensity at 440 nm.A portable analyser based on the same reaction was described in 1970.12 This instrument consists of a simple mixing chamber, constructed from a 100-ml beaker, that is attached to a photomultiplier assembly (EM1 95365) together with the associated circuitry. Two concentric tubes enter the mixing chamber, the sample air is drawn through the inner tube at the rate of 1 1 min-l and ethylene through the outer tube at the rate of 13 ml min-l. The two gases meet at the photomultiplier tube window, where the emitted light is measured. The instrument was found to give an output that was linear with ozone concentration, based on calibration tests performed against determinations by the iodide method. Under the conditions used the instrument gave a dark-current reading of 2 x 10-10A and was capable of detecting 1 part per lo8 of ozone.The chemiluminescence method has been accepted in the USA as a reference method for the determination of ozone concentrations following a number of field trialslS; modern instruments will detect as little as 0.05 part per 108 of ozone with less than 10 s response time and they display linearity over a concen- tration range of 4 or 5 orders of magnitude. Care should be taken with the disposal of the ethylene effluent from chemiluminescent ozone monitors because of its flammability. Lonnemanl4 describes an efficient catalytic disposal unit.Oxides of Nitrogen Before the introduction of the chemiluminescence technique, the determination of oxides of nitrogen was a difficult problem in gas analysis. The available methods have been reviewed by Allen15; he includes the various chemiluminescent reactions of analytical use. Of these, the reaction between nitric oxide and ozone has received most attention and an analytical application was described by Fontijn and co-worker~.~~~ This work has formed the basis for the design of the commercially available analysers which now find wide use in industry. The kinetics of the nitric oxide and ozone reaction had been studied1°J6J7 before Fontijn’s work was published; the reactions are .. - . (6) * . (7) NO + 0, -+ NO,* + 0, .. .. NO,* --+ NO, + hv . . . . .. The light intensity is given by: for the 600-875-nm region when M is air. M is the third entity, which may quench the activated NO,* by means of the reaction .. - * (8) NO,* + M ---+ NO, + M . . .. The rate constant for the over-all reaction is - --dr0,1 = -- d“ol = K”O][O,] dt dt = 1 x 10-7 1 mol-1 s-1 which is a low enough value to render the consumption of nitric oxide sufficiently small to be negligible and to obtain a uniform emission from the reactor. Clough and Thrush10 measured the rate constants of the reactions and found that only about 8 per cent. of the nitrogen dioxide is formed in the excited state (2B1). The effect of quenching varies with pressure and reaction (8) is the principal reaction at pressures greater than 0.1 mm.The sum effect is that light emission, reaction (7), occurs only with a very small fraction of theJzcly, 1975 AND FLAME-EMISSION SPECTROMETRY : A REVIEW 455 excited molecules. The fact that chemiluminescence is nevertheless a very sensitive method for the determination of nitric oxide illustrates the inherent sensitivity of the technique. Clyne et aZ.16 analysed the light emitted during the reaction and found it to be a continuum, starting at 600 nm and extending into the infrared region with a maximum at 1200nm. Although the emission extends over this wide range, only the region between 600 and 875 nm is analytically useful, the upper limit being set by the response of red-sensitive photomultipliers. In practice, a red cut-off filter is used to exclude any possible interfering emissions in the visible region.The expression for light intensity reduces to If [O,] is large and in excess, then the intensity is proportional to [NO] provided that [MI is also constant. M contributes to the total pressure in the system and it is evident that light emission will decrease as the pressure rises. This decrease is caused by quenching, or de- activation by inelastic collisions. The nature of the diluent molecule, M, affects its quenching efficiency and this can be a source of error in chemiluminescent measurements when large concentrations of other “neutral” molecules are present, particularly when using instruments that operate at a pressure near to atmospheric pressure. The layout of a typical low-pressure chemiluminescence analyser of nitrogen oxides is shown in Fig.4. Ozone is supplied at a constant flow-rate from a silent discharge cell and is fed via a constriction into the reactor, sample gas is introduced into the reactor at a controlled rate and the reactor is maintained at a constant low pressure by means of the vacuum pump. The sample-flow arrangement provides a rapid purge and a quick response time. I 10-1 R = Reactor SV1 and SV2 = Solenoid valves PM = Photomultiplier VR = Vacuum regulator REG = Pressure regulator Fig. 4. Low-pressure NO, analyser. A systematic study of the factors affecting the sensitivity of chemiluminescence * analysers for the determination of nitric oxide is the subject of a recent paper.ls The relationship between reactor pressure, pumping speed and residence time in the reactor is discussed, and the point is made that it is necessary to keep the residence time large compared with the radiative life.The importance of efficient mixing is stressed, also the need to choose a photomultiplier with extended response in the infrared region and with low noise and dark current. The authors conclude that the quality and characteristics of the photomultiplier may well be the most important factors in extending the sensitivity of a nitric oxide detector. The origin of dark current in photomultipliers is discussed by Sharpe,19 who shows that in tubes with extended red sensitivity cooling to -10 “C is usually all that is needed to456 GLOVER: CHEMILUMINESCENCE IN GAS ANALYSIS Analyst, VoZ. 100 obtain a worthwhile reduction in dark current.Thermoelectric cooling is a convenient method, although a circulatory system is equally satisfactory. Despite the fact that theoretical considerations favour the use of low pressure in chemi- luminescence determinations, there has been a trend in commercial instruments towards operation at near atmospheric pressure. The main reason for this is an attempt to eliminate the vacuum pump, and also to facilitate the production of a smaller, more easily portable unit. Fig. 5 shows the layout used in an ambient pressure monitor. Simple analysers such as these can measure nitric oxide concentrations between 0-1 p.p.m. and 0.1 per cent. with good linearity and a rapid response, but in order to achieve the full sensitivity of which the method is capable it is necessary to reduce the noise level by cooling the photomultiplier tube.Under these conditions a detection limit of 1 part in lo9 can be achieved. Ozoniser Oxygen Sample inlet Exhaust Digital voltmeter ~ I - T O electronic circuitry 4 Fig. 5 . Ambient-pressure NO, analyser. Measurement of Nitrogen Dioxide While chemiluminescence provides an elegant answer to the problem of nitric oxide deter- mination it is not applicable to nitrogen dioxide, and unfortunately it is the higher oxide that is frequently of more interest. Consequently, there has been a demand for a quick and easy method of converting nitrogen dioxide back into nitric oxide, thus enabling total oxides of nitrogen to be measured. A number of fairly successful methods have been offered, the original one being the thermal converter.This is a stainless-steel tube, heated to 650- 700 "C and its operation is based on the thermal decomposition of nitrogen dioxide: 2N0, + 2N0 + 0, Breitenbach and Shelef20 showed this reaction to be complete to the extent of 98 per cent. at 630 "C, provided that the partial pressure of oxygen is kept below 5 mm. At atmospheric pressure in air (oxygen partial pressure 150 mm), only 90 per cent. conversion is obtained. Consequently, the stainless-steel converter is useful only for instruments that operate at pressures below atmospheric pressure. Another factor influencing the operation of thermal converters is the concentration of nitrogen dioxide. The dissociation of nitrogen dioxide is a second-order reaction and its rate will therefore vary with the nitrogen dioxide concentration; Sigsby et aL21 have pointed out that because of this variation a much longer residence time, or a higher temperature, is needed for low than for higher concentrations of nitrogen dioxide.The stainless-steel converter is subject to serious errors in the presence of ammonia, as this compound is "con- verted" into nitric oxide; another error arises if carbon monoxide is present in the sampleJuly, 1975 AND FLAME-EMISSION SPECTROMETRY : A REVIEW and the oxygen content is low, when the following reaction is favoured- 457 2CO + 2N0 -+ 2C0, + N, These limitations led to the examination of other types of converter and Breitenbach and Shelef 2o examined many materials, concluding that a composite of carbon and molybdenum would convert nitrogen dioxide into nitric oxide with the minimum of interference from other components of the sample (e.g., ammonia and other nitrogen compounds).Modern instru- ments usually include converters packed with carbon, and in general these are satisfactory. Potential sources of error include adsorptive effects resulting from the use of an unsuitable grade of carbon and catalysis of the reduction of nitric oxide to nitrogen by use of the wrong material to enclose the carbon. Ethylenic Hydrocarbons The reaction between ethylene and ozone has already been mentioned, and it can also be applied to the measurement of ethylene. Precisely the same equipment is used as for the measurement of ozone, but with the addition of an ozoniser to provide the reagent.Kummer et aZ.22 report details of the spectral characteristics of chemiluminescence obtained from the reaction between various ethylenic compounds and ozone. Their results, which are presented in Table I, show that for a specific determination of ethylene a narrow-band filter would be needed in order to prevent interference from other ethylenic compounds. They also show that ethylene is not the best reagent for measuring ozone; a 50-fold increase in sensitivity could be gained by using one of the methyl derivatives. Finlayson et aZ.,3 extended these studies and distinguished three classes of olefins: 1, terminal olefins, giving a broad emission that peaks at 440 nm; 2, olefins such as cis-but-2-ene, giving a broad emission with a peak at about 465 nm plus narrow peaks at 520 and 565 nm; and 3, olefins such as tetramethylethylene, which are dialkyl substituted at a carbon atom of the double bond, giving narrow peaks at 520 nm and broad shoulders at 465 and 565 nm.TABLE I CHEMILUMINESCENCE OF SOME ETHYLENIC COMPOUNDS Compound Relative emission intensity Peak (ethylene = 1) wavelength/nm Ethylene . . .. .. .. 1 Trimethylethylene . . .. .. 50 Tetramcthylethylene . . .. .. 50 2,3-Dimethylbutadiene . . .. 8 2,5-Dimethylhexa-2,4-diene . . .. 30 440 520 520 - - Although no analytical methods based on the ozone - olefin reaction appear to have been developed, it can be seen that the chemiluminescence approach has possible applications in this field. Sulphur Compounds the spectra of emissions obtained by gas-phase reactions with ozone.reactions are given in Table I1 and show that sensitive measurements would be possible. A brief study of certain sulphur compounds is presented by Kummer et aZ.22; they examined Results for these TABLE I1 CHEMILUMINESCENCE OF SOME SULPHUR COMPOUNDS Relative emission intensity Peak Compound (ethylene = 1) wavelength/nm Hydrogen sulphide . . .. . . 25 370 Dimethyl sulphide . . .. .. 200 370 Methanethiol . . .. .. .. 2000 -458 GLOVER: CHEMILUMINESCENCE IN GAS ANALYSIS Analyst, VoZ. 100 Peroxyacetyl Nitrate Pitts et aZ.24 have studied the reactions of peroxyacetyl nitrate and of ozone with triethylamine and suggest a design for an atmospheric monitor for these two oxidants. Peroxyacetyl nitrate is one of the important oxidants formed in photochemical smog, it is a known eye irritant and has been reported in concentrations of about 30 parts per 109 in Los Angeles.The chemiluminescence spectrum obtained from the reaction between peroxyacetyl nitrate and triethylamine is a broad emission with a maximum at about 650nm, as opposed to the ozone - triethylamine reaction, which gives a peak emission at 520 nm. The authors reported results of trials with various filter combinations and finally selected two cut-off filters for their prototype instrument, one transparent only to wavelengths less than 550nm and the other transparent only to those greater than 665nm. By using this combination the ratio of two intensities can be used to give the concentration ratio of the two oxidants, and then calibration with respect to one of them is all that is needed for an absolute measurement of both.It was found during this work that the chemiluminescent efficiency of the peroxyacetyl nitrate reaction was about ten times that of the corresponding ozone reaction ; another feature of the peroxyacetyl nitrate reaction is the long afterglow, which persists for several minutes, even at the lowest concentration. Carbonyls Two interesting examples of gas-phase chemiluminescence involve the carbonyls of iron and nickel. These have been shown to take part in chemiluminescent reactions with ozone,25 the emissions being due to the excited FeO and NiO species and exhibiting maxima in the regions 565, 590 and 620nm. While no analytical technique based on these emissions has so far been proposed, they are included because metal carbonyls have been found to give rise to interferences in other chemiluminescence methods.FIame Chemiluminescence The use of flames for analytical purposes is well established in techniques such as flame- emission spectrometry and atomic-absorption spectroscopy ; perhaps less well appreciated are the applications of chemiluminescence occurring in flames. The distinction between chemiluminescence in a flame and radiation due to thermal excitation is not always clear-cut and it is generally accepted that both phenomena can occur simultaneously. However, for the present purpose, consideration will be given only to cool flames in which light emission is well in excess of that which would be expected from thermal excitation, and in which chemiluminescence is the accepted mechanism.Emissions of this type were described by Gilbert,26 who used the air - hydrogen flame in the presence of organic compounds, such as alcohols, added to the sample or a hydrocarbon mixed with the hydrogen. Gilbert found that line intensities were enhanced approximately 1000-fold by the addition of the organic species to the flame. That the enhancement was not caused by a higher flame temperature was shown by the addition of oxygen to the flame; this gave only slight intensification at the expense of a higher background emission. Gilbert recorded enhancement with many elements including tin, lead, arsenic, antimony, bismuth, nitrogen and phosphorus. A good introduction to this subject is given by Gibson et aZ.,27 who adopted asystematic approach to the elucidation of factors that affect the intensity of flame emissions.After mentioning the contradictory findings that have appeared in earlier literature, these authors pointed out the complexity of the processes that occur when aqueous solutions are atomised into flames in order to produce radiation. They used the Boltzmann distribution equation as a criterion for thermal excitation in a flame: Emission intensity cc N* cc NOe-AEIkT where No corresponds to the concentration of species in the ground state and T is the tem- perature of the flame. Ground-state concentrations (No) were measured by atomic-absorption spectroscopy and T by the sodium line reversal technique as described in an earlier paper.28 The enhancementJ d y , 1975 AND FLAME-EMISSION SPECTROMETRY : A REVIEW 459 of emissions by organic solvents was investigated with these methods.The results showed that enhancement with sodium and calcium was due to higher concentrations of ground- state atoms as a result of the more efficient evaporation brought about by the presence of an organic solvent; in the instance of tin, however, the enhancement by propan-2-01 could only be explained by an increase in excited-st ate concentration. This conclusion confirmed the work of Gilbert,,s who found that tin gave only a very weak emission in the hydrogen - oxygen flame both with and without the addition of propan-2-01, although with the cooler hydrogen - air flame a considerable enhancement was obtained with propan-2-01.The enhancement was difficult to determine because the normal emissions were too weak t o measure, but a minimum value of 40 was reported. The mechanism was ascribed to the chemiluminescent reaction : This compares with Gaydon and Wolfhard’s proposed mechanism for lead29 : Sn + CH + OH+%* + CO + H, Pb + H + OH-+Pb* + H,O A comprehensive study of enhancement phenomena has been presented by Buell30; he used limited area techniques to investigate the effect of organic solvents on emission intensities. Buell’s paper gives excitation potentials and dissociation energies corresponding to over 600 atomic lines, and correlates these with the height in the flame corresponding to maximum emission intensity and the solvent enhancement factor. His results show that as excitation potentials increase, solvent enhancement increases and the height of maximum emission decreases.Solvent enhancing factors are considerable, commonly between one and two orders of ten, and in many instances are recorded as infinite when no emission is obtained without the use of the solvent. Buell puts forward the argument that if the only function of the solvent is to increase the flame temperature or the evaporation rate, then all enhancements should be of the same order of magnitude. This is obviously not so. In addition, Buell found many spectral lines with excitation potentials in the range 7-9 eV, far in excess of that expected from thermal energy. He concluded that chemiluminescence was the only feasible mechanism that would explain his results.The papers cited so far have shown the general possibilities of solvent enhancement and a more detailed contribution provides the basis of a sensitive emission method for determining tin.31 These authors used a direct thermocouple method to determine the temperature of the flame and demonstrated an intense emission at 284 nm with flame temperatures of about 300°C. Chemiluminescence occurred in the presence of alcohols, 40 per cent. propan-2-01 giving the best conditions. Sulphur and Phosphorus Sulphur and phosphorus have been determined by chemiluminescent flame emission ; these elements, together with the halogens, are beyond the range of atomic-absorption methods because their principal resonance lines are in the far ultraviolet region. This prompted Dagnall et aL3, to investigate the molecular emission of sulphur in cool flames.These workers found that a hydrogen - nitrogen mixture was the most useful fuel gas and they measured the S, emission at 384 nm; emission intensity was shown to be dependent on flame temperature, being reduced as the temperature was increased. It was found that all of the sulphur compounds investigated gave an S, emission, including sulphuric acid, sulphates, sulphites, thiosulphates, thiocyanates, sulphur dioxide, hydrogen sulphide and mercaptoacetic acid. The maximum response per molecule was obtained from sulphur dioxide and hydrogen sulphide, while sulphates and sulphuric acid gave only very weak emissions (about 1/600th of that given by sulphur dioxide). The best flame for the analysis of sulphates was found to be a shielded, pre-mixed, cooled hydrogen - air flame.The S, emission displayed maximum intensities a t 384 and 394 nm. As the S, species is formed via sulphur atom recombination s + s -+ s, + hv two atoms are needed, and the intensity of emission is proportional to the square of the concentration of the sulphur compound in the flame.460 GLOVER: CHEMILUMINESCENCE IN GAS ANALYSIS Analyst, VoZ. 100 The hydrogen-rich flame has been used extensively as a detector of sulphur-containing species, particularly in gas chromatography, and was first described in a West German patent.33 An improved version is described by Brody and C h a n e ~ , ~ ~ and more recent work is summarised in reference 5. There are now a number of commercial forms of this sulphur detector, known as the flame-photometric detector, and it has been used in gas-chromatographic studies of atmos- pheric pollution.35 The detector consists of a small flame, fed with a hydrogen and inert gas mixture, the primary combustion zone of which is shielded and the cooler secondary zone displays the emission, which is viewed via an interference filter by the photomultiplier.This detector has high sensitivity, but also has a limited range owing to deactivation of the excited species by self-collisional quenching. The exponential response of the sulphur flame-photometric detector prompted Crider and Slate? to suggest the flame luminescence intensification and quenching detector, which operates on the most sensitive part of the response curve. This is achieved by adding sulphur dioxide to the carrier gas, thus obtaining a background luminescence.Eluates from the chromatographic column are detected by their quenching or intensifying effect on the emission. These workers used a column consisting of a 5-ft length of 0-1-in i d . PTFE tubing, packed with 9.1 per cent. squalane on 60-80 mesh Gas-Chrom 2, and they measured the detector res- ponse to several compounds at various base loadings of sulphur dioxide. Compounds with and without sulphur in their molecules were used and a useful degree of specificity was claimed from the response characteristics. The behaviour of phosphorus-containing compounds was studied in another paper by Dagnall ef aL3' When using the nitrogen - hydrogen diffusion flame the best conditions were found to be in the coolest zone at temperatures between 280 and 300 "C.The emission at 528nm was used and ascribed to the H-P-0 species. Emission intensity was found to vary in a linear manner with concentration when using phosphoric acid over the range 0.2-200p.p.m. As with sulphur, the response varied with different phosphates and a pre- liminary cation-exchange separation is advised. The flame-photometric detector can, of course, be used for detecting phosphorus-containing compounds in gas-chromatographic eluates just as for sulphur compounds, except that a different filter must be used. Aldous et aL3* described a method for determining sulphur and phosphorus in organic and aqueous media by using a simple filter photometer. They pointed out that because the emission is banded, sensitivity is lost with a highly resolving monochromator, and also demonstrated the advantages to be gained by using a simple filter photometer.Detection limits of 0.08 p.p.m. for sulphur and 0.007 p.p.m. for phosphorus were obtained. Important features of the method are the need to prevent oxygen from reaching the centre of the flame and quenching the emission, and also the finding that hydrocarbon fragments can quench the emission. Sample treatment methods were recommended in order to avoid these interferences and calibration graphs constructed for sulphur in the range 1-100 p.p.m. and phosphorus between 0.09 and 30 p.p.m. Halogens The well known Beilstein test for halogens, based on the green flame obtained by heating on a copper wire, was the basis of a special burner known as the Van der Smissen burner.39 This device is sensitive, stable, gives a linear response and uses the emission obtained with copper in the presence of a halogen.However, the spectrum consists of emissions due to the species Cu, CuH, CuO and CuOH and is therefore not specific for halogens. Gilbert40 sought to increase the sensitivity of this technique by using indium instead of copper and expected to measure the resonance line of indium at 451.1 nm. Instead, he observed a very intense emission due to InC1, which shows an intensity maximum at 359.9 nm. The appear- ance of this line not only conferred even higher sensitivity on the method, but also made it specific for halogens. Gilbert concluded that the InCl was abnormally excited by chemi- luminescence because the total power output of the InCl spectrum was found to exceed that of the indium spectrum in the lower part of the flame by one order of magnitude.Gilbert's burner consisted of two jets, the upper completing the combustion of the hydrogen burned at the lower jet. Indium foil was suspended between the jets and the InCl emission was observed in the primary reaction zone of the upper flame.July, 1975 AND FLAME-EMISSION SPECTROMETRY : A REVIEW 461 The indium method was shown to have a very high sensitivity and possess good linearity, and Gilbert suggested that a detection limit of 0.001 pg of chlorine per litre of air should be possible. Gilbert's method has been applied to the measurement of chlorinated pesticides by Herrmann and G ~ t s c h e .~ ~ These workers increased the surface area of the indium by using indium-coated copper - beryllium coils, which were raised above the primary combustion zone and held at a temperature of 200 "C. A ten-fold increase in sensitivity was obtained. In a second paper42 a further modification was made, copper - beryllium sheet being used instead of coils of wire. This method was applied to the determination of bromide in urine samples. Sodium gave a strong luminescence background and a prior ion-exchange separation was required. A detection limit of 0-0062 pg of bromine was claimed when using the 375-8-nm InBr emission. Dagnall et aZ.43 had noted a similar chemiluminescence when solutions of tin(I1) halides were aspirated into cool flames and demonstrated the blue SnCl and red SnH emissions.These emissions were shown to occur in the coolest part of the flame, where the droplets were still evaporating. However, the emissions from tin(I1) halides were found to be much less intense than those from the corresponding indium species; the latter were systematically studied in a later paper44 as a basis for an analytical method for determining halides. These workers record details of the spectra of the three indium halides that exhibit intensity maxima as follows: InC1, 360 nm; InBr, 376 nm; and InI, 410 nm. They used a similar burner to that used in their studies on sulphur and p h o s p h ~ r u s ~ ~ ~ ~ ~ and established conditions and interference data for chlorides, bromides and iodides.Detection limits were found to be in the range 1-2 p.p.m., and linear calibration graphs were obtained. Phosphate and sulphate interfered seriously and had to be removed, and interhalogen effects were noted; for example, a large excess of bromide completely destroyed the In1 emission. Fluoride cannot be measured by this means as no InF emissions occur; this failure was explained on the basis of the greater stability of the InIIIF species. It was also found that chloride enhanced the In1 emission. Nitrogen Compounds A paper describing the detection of nitrogen compounds by means of flame chemilumines- cence was presented by Krost et aZ.45 They used a scanning monochromator to examine emissions from the oxygen - hydrogen flame due to the reaction H + NO -+ HNO* HNO* --+ HNO + hv The emission obtained is in the range 660-770 nm.The effect was observed from both nitric oxide and nitrogen dioxide, and the reaction is first order for both reactants in that the emission intensity is proportional to the hydrogen atom and nitric oxide concentrations. I = I . [HI [NO] where I is the emission intensity. The reaction mechanism postulated above is supported by the fact that the spectrum obtained with the hydrogen-rich flame is identical with that obtained by Clyne and by the reaction of nitric oxide with atomic hydrogen produced by a radiofrequency discharge. The fact that the same spectrum is obtained with both nitric oxide and nitrogen dioxide was explained by the very rapid reaction NO, + H -+ NO + OH which takes place in the hydrogen-rich flame.The paper describes detector design and pro- poses a shielded, hydrogen-rich flame viewed by two photomultipliers, one with a narrow- band interference filter at 690nm for measuring the nitrogen luminescence, and the other for measuring the sulphur response at 394 nm. Nitrogen added to the system as a third entity gave the lowest background signal and the most favourable flame conditions. Studies of the responses showed that ammonia and monomethylamine gave equivalent nitrogen responses but that sulphur dioxide was a source of serious interference in the measure- ment of nitrogen. The nitrogen compounds that were examined did not interfere in the detection of sulphur.462 GLOVER: CHEMILUMINESCENCE IN GAS ANALYSIS Analyst, VoZ. 100 This detector was shown to have a linear response for nitrogen oxides up to 60 p.p.m.V/V with a detection limit of 0.16 p.p.m. V/V. The sulphur detection system gave a lowest detection limit of 0.004 p.p.m. V/V for sulphur dioxide. Organic Compounds In the course of his work on the flame-photometric detector, Crider reported that the addition of chlorinated hydrocarbons intensified the S, emission at 405 nm. This effect was investigated further4' when it was shown that these compounds can emit separately and do not necessarily act as catalysts for the sulphur emission reaction. The spectra obtained when chloroform, ethylene dibromide and methyl iodide were aspirated into cool flames were recorded at various hydrogen to air ratios and optimum conditions were established for maximum emission intensity.The results suggested that low detection limits (0.01 p.p.m. of methyl iodide) were capable of being achieved. Preliminary results were recorded for other simple halogenated aliphatic hydrocarbons. The technique of aspirating solutions of various organic compounds into cool flames is described in a general study by Dagnall et ~ 1 . ~ ~ They used low-temperature, laminar-flow nitrogen - hydrogen diffusion flames and took measurements in the coolest region of the flame (about 280 "C). Most of the emissions recorded in this paper are chemiluminescent in origin and a characteristic feature is the dependence of emission intensity on experimental variables. Many bands that might have an analytical application were recorded and the authors suggest that cool hydrogen flames could find wider application as gas-chromatographic detectors.Molecular-emission Cavity Analysis (MECA) This is a relatively new technique that was pioneered by Belcher et aZ.49 in which chemi- luminescent flame reactions are involved. It presents an alternative to nebulisation into a flame and uses a cavity at the end of a rod into which the sample is deposited. The rod is inserted into the flame so that the cavity is in line with the detector and the spectrum is recorded with respect to time lapse after insertion. Conditions are adjusted so that emission is confined to the cavity and the result is a concentrated and sustained effect giving very high sensitivity. Emissions are sometimes obtained that do not occur with conventional flame techniques, for example from selenium and tellurium,50 much higher sensitivities are possible, certain sulphur compounds can be detected at the picogram level using the S, emissions as with the flame-photometric detector, and phosphorus compounds can be measured by use of the HPO, emission.Other examples of the use of MECA are the determination of halo- gens,51 arsenic and antimony,52 and boron and silicon.49 Injection of oxygen into the cavity can extend the usefulness of the technique when oxides and hydroxides are the emitting species. 63 Future Developments A very interesting possibility for the development of new techniques in gas analysis based on chemiluminescence is the use of atomic gases as reagents. One of the best feasibility studies in this field is that by Snyder5* who examined the possibility of using atomic oxygen to measure air pollutants by means of the general reaction XO+O-+XO,+hv where X is nitrogen, carbon or sulphur.In such reactions the light intensity of the emitted radiation ( I ) is given by I = k [XO] [O]. An excess of atomic oxygen was produced by means of a microwave discharge in molecular oxygen, and recordings were taken of the relevant spectra. Both nitric oxide and sulphur dioxide were shown to be capable of being determined with a detection limit of 1 part per 109; carbon monoxide gave less favourable detection conditions and a lowest detection limit of about 150 p.p.m. V/V was reported. However, there is little doubt that this limit could be improved by at least a factor of ten.The spectral distributions obtained were studied in some detail and the peak transmission wavelengths were as follows: nitric oxide, 650 nm; carbon monoxide, 400 nm; and sulphur dioxide, 270nm. The reaction between nitric oxide and atomic oxygen has been studiedJm?y, 19 75 AND FLAME-EMISSION SPECTROMETRY : A REVIEW 463 extensively and has been used in “titration reactions” of many atomic and free radical species. Cormier et aZ.5 give a good illustration of the techniques involved and show how the reaction 0 + NO,--+NO + 0 2 can be used to determine the amount of oxygen atoms in a stream of gas. In this instance, when oxygen atoms are in excess, the indicator reaction gives rise to a whitish green glow. When the flow of nitrogen dioxide is equal to or exceeds that of the oxygen atoms this glow contracts to a very small area round the mixing nozzles.The reaction between nitric oxide and atomic oxygen has been adapted as a laboratory standard for chemilumine~cence,~~ and detailed studies of rate constants and quantum yield have been made.66 Reeves et aLS7 also examined the mechanism of this important reaction. Another method involving oxygen atoms was described by Krieger et aL5*; they showed the possibility of determining unsaturated hydrocarbons by radiative reaction with atomic oxygen. These workers reported a better signal to noise ratio for emissions from the ethylene - atomic oxygen reaction than for the corresponding reaction with ozone. Olefins were found to emit between 700 and 900nm and acetylene at 600nm.The kinetics and mechanism of the acetylene reaction have been well studied59 and in this work the oxygen atoms were produced by the reaction of nitrogen atoms with nitric oxide M + NO + O-+NO, + M + IZV N+NO-+N,+O The nitrogen atoms were produced in a microwave discharge. The necessity to use such a discharge in order to produce atomic gases has so far inhibited the application of this type of technique; if an alternative source could be found, then the wider use of atomic gases could give rise to some very attractive methods in gas analysis. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. References Young, R. A., and Sharples, R, L., J . Chem. Phys., 1963, 39, 1071. Fontijn, A., Sabadell, A.J., and Ronco, J., Analyt. Chem., 1970, 42, 575. Fontijn, A., Clearinghouse for Federal Scientific and Technical Information, Report No. PB188104, Stevens, R. K., and Hodgeson, J. A., Analyt. Chem., 1973, 45, 44314. Cormier, M. J ., Hercules, D. M., and Lee, J ., Editors, “Chemiluminescence and Bioluminescence,” Wayne, R. P., Photochem. Photobiol., 1966, 5, 889. Laidler, K. J., “The Chemical Kinetics of Excited States,” Clarendon Press, Oxford, 1955. Seitz, W. R., and Neary, M. P., Analyt. Chem., 1974, 46, 188A. Gaydon, A. G., “Spectroscopy and Combustion Theory,” Second Edition, Chapman and Hall, Clough, P. N., and Thrush, B. A., Trans. Famday Soc., 1967, 63, 915. Nederbragt, G. W., van der Horst, A., and van Duijn, J., Nature, Lond., 1965, 206, 87. Warren, G.J., and Babcock, G., Rev. Scient. Instrum., 1970, 41, 280. Mamantov, G., and Shults, W. D., Editors, “Determination of Air Quality,” Plenum Press, London Lonneman, W., Envir. Lett., 1973, 4, 117. Allen, J. D., J . Inst. Fuel, 1973, 123. Clyne, M. A. A., Thrush, B. A., and Wayne, R. P., Trans. Faraday Soc., 1964, 60, 359. Clough, P. N., and Thrush, B. A., Trans. Faraday SOL, 1969, 65, 23. Steffenson, D. M., and Stedman, D. H., Analyt. Chem., 1974, 46, 1704. Sharpe, J ., “Dark Current in Photomultiplier Tubes,” Publication Ref. R/P021, EM1 Electronics Breitenbach, L. P., and Shelef, M., J . Air Pollut. Control Ass., 1973, 23, 128. Sigsby, J. E., Black, F. M., Bellar, T. A., and Klosterman, D. L., Envir. Sci. Technol., 1973, 7 , 51. Kummer, W. A., Pitts, J.N., and Steer, R. P., Envir. Sci. Technol., 1971, 5, 104s. Finlayson, B. J., Pitts, J. N., and Akimoto, H., Chem. Phys. Lett., 1972, 12, 495. Pitts, J. N., Fuhr, H., Gaffney, J. S., and Peters, J. W., Envir. Sci. Technol., 1973, 7 , 550. Morris, E. D., J. Am. Chem. Soc., 1970, 5742. Gilbert, P. T., Paper presented at Pittsburgh Conference on Analytical Chemistry and Applied Gibson, J. H., Grossman, W. E. L., and Cooke, W. D., Analyt. Chem., 1963, 35, 266. Gibson, J. H., and Cooke, W. D., Paper presented a t Pittsburgh Conference on Analytical Chemistry Gaydon, A. G., and Wolfhard, H. G., “Flames, Their Structure, Radiation and Temperature,” U.S. Department of Commerce, Springfield, Va., 1969. Plenum Press, London and New York, 1973. London, 1948. and New York, 1970. Ltd., Hayes, 1970. Spectroscopy, March, 1961. and Applied Spectroscopy, March, 1961. Chapman and Hall, London, 1960.464 30. 31. 32. 33. 34. 36. 36. 37. 38. 39. 40. 41. 42. 43. 44. 46. 46. 47. 40. 49. 60. 61. 62. 63. 64. 66. 56. 67. 58. 59. GLOVER Buell, B. E., Analyt. Chenz., 1963, 35, 372. Dagnall, R. M., Thompson, K. C., and West, T. S., Analyst, 1968, 93, 618. Dagnall, R. M., Thompson, K. C., and West, T. S., Analyst, 1967, 92, 606. Draegerwerk, W. Ger. Pat. 1,133,918, 1962. Brody, S. S., and Chaney, J. E., J. Gas Chromat., 1966, 4, 42. Stevens, R. K., Mulik, J. D., O’Keefe, A. E., and Krost, K. J., Analyt. Chem., 1971, 43, 827. Crider, W. L., and Slater, R. W., Analyt. Chem., 1969, 41, 531. Dagnall, R. M., Thompson, K. C., and West, T. S., Analyst, 1968, 93, 72. Aldous, K. M., Dagnall, R. M., and West, T. S., Analyst, 1970, 95, 417. Draegerwerk, W. Ger. Pat. 1,095,552, 1960. Gilbert, P. T., Analyt. Chem., 1966, 38, 1920. Herrmann, R., and Gutsche, B., Analyst, 1969, 94, 1033. Gutsche, B., and Herrmann, R., Analyst, 1970, 95, 805. Dagnall, R. M., Thompson, K. C., and West, T. S., Analyst, 1968, 93, 518. Dagnall, R. M., Thompson, K. C., and West, T. S., Analyst, 1969, 94, 643. Krost, K. J., Hodgeson, J. A., and Stevens, R. K., Analyt. Chem., 1973, 45, 1800. Clyne, M. A. A., and Thrush, B. A., Discuss. Faraday SOC., 1962, 33, 139. Crider, W. L., Analyt. Chem., 1969, 41, 634. Dagnall, R. M., Smith, D. J., Thompson, K. C., and West, T. S., Analyst, 1969, 94, 871. Belcher, R., Bogdanski, S. L., Townshend, A., Analytica CRim. Acta, 1973, 67, 1. Belcher, R., Kouimtzis, T., and Townshend, A.. Analytica Chim. Acta, 1974, 68, 297. Stiles, D, A., PYOC. SOC. Analyt. Chem., 1974, 11, 141. Belcher, R., Bogdanski, S. L., Ghonaim, S. A., and Townshend, A., Analytica Chim. Ada, 1974, Ghonaim, S. A., Proc. SOC. Analyt. Chem., 1974, 11, 167. Snyder, A. D., Clearinghouse for Federal Scientific and Technical Information, Report No. Fontijn, A., Meyer, C. B., and Schiff, H. I., J. Chem. Phys., 1964, 40, 64. Vanpee, M., Hill, K. D., and Kineyko, W. R., A.I.A.A. J Z , 1971, 9, 136. Reeves, R. R., Harteck, P., and Chace, W. H., J. Chem. Phys., 1964, 41, 764. Krieger, B., Maiki, M., and Kummler, R., Envir. Sci. Technol., 1972, 6, 742. Arrington, C. A., Brennen, W., Glass, G. P., Michael, J. V., and Niki, H., J. Chem. Phys,, 1966, Received February 6th, 1975 Accepted February 20th, 1975 72, 183. PB188103, U.S. Department of Commerce, Springfield, Va., 1969. 43, 1489.

ISSN:0003-2654    

DOI:10.1039/AN9750000449

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

Analyst, July, 1975, Vol. 100, pp. 465-470 466 The Determination of Barium in Unused Lubricating Oils by Means of Atomic-absorption Spectrop hotometry S. T. Holding and J. J. Rowson Shell Research Limited, Thomton Research Centve, P.O. Box 1, Chester, CH1 3SH The use of a mixed-solvent system for the determination of zinc and calcium in unused oils by means of atomic-absorption spectrophotometry has been extended to the determination of barium. The excessive interference that occurs in the determination of barium because of high concentrations of calcium in certain oils has been overcome by modifying the mixed solvent in order to increase sensitivity. The work could then be carried out a t much lower concentrations of barium, at which the interference has been shown to be eliminated.The procedure has been applied t o a wide range of samples, and results are in good agreement with those obtained by means of X-ray fluorescence. The development and use of an acidic mixed-solvent system for the determination of zinc and calcium additives in unused lubricating oils by means of atomic-absorption spectro- photometry has been described previous1y.l This solvent system allows inorganic salts to be used as standards and also eliminates the systematic errors that sometimes occur when an air - acetylene flame is used. These errors are largely due to the different chemical con- stituents of the metal-containing adhtives in the sample and of the organometallic com- pounds used as standards. The use of inorganic salts as standards in the determination of zinc and calcium additives had the advantage of permitting simple and inexpensive checks on the validity of the aqueous salt standard whenever it was thought necessary and had the additional advantage of the salts being generally available in pure forms.The same advan- tages would be obtained if this approach could be extended to the determination of barium additives in lubricating oils. For the determination of barium it is necessary to use the hotter nitrous oxide - acetylene flame, and many worker^^-^ have reported on the determination, using this flame, of barium in lubricating oils. However, all the methods have depended on the use of oil-soluble organo- barium compounds as standards and have employed solvents such as white spirit or wholly aromatic solvents such as xylene.Our experience is that solvents such as white spirit, which give moderate sensitivity for barium, generally give accurate results but that the precision may not be adequate for blending quality control. In addition, with such solvents spectral interference may occur when analysing formulations in which there is a large concentration of calcium in relation to barium. The magnitude of the interference is dependent on flame stoicheiometry. Although Peterson and Kahn2 analysed oils, including some with a high calcium to barium ratio, they used xylene as solvent, which gives very good sensitivity for barium and so spectral interference would be minimal. They reported single results only for the direct determination of barium using atomic-absorption spectrophotometry but for the small number of samples analysed they obtained good agreement with results obtained by use of X-ray fluorescence.Our early experience of the use of wholly aromatic solvents in atomic-absorption spectro- photometry for blending quality control was that repeatability was not usually adequate. The burners formerly available were prone to carbon build-up along the slit, which necessitated frequent shutdown for cleaning. Furthermore, the nitrous oxide - acetylene flame is some- what hazardous and our early assessment was that it should not be used in routine work. How- ever, many instrument manufacturers now incorporate safety features such as switching control units in the gas supply line and bursting discs in the burner assemblies.In addition, on commercial instruments greater emphasis has now been placed on providing fine and steady control of the fuel and support gases, which is essential for good repeatability.466 Analyst, VoZ. 100 The success of the methods for zinc and calcium, together with the almost universal adoption of the instrumental improvements described above, prompted attempts to extend the use of an inorganic salt standard to the determination of barium additives in lubricating oils by means of atomic-absorption spectrophotometry. This has led to the development of an atomic-absorption procedure for the determination of barium in unused lubricating oils by a direct solvent dilution procedure utilising barium chloride as standard. HOLDING AND ROWSON : DETERMINATION OF BARIUM IN UNUSED Experimental Development of Solvent System The mixed solvent used for earlier work had the following composition : cyclohexanone - butan-1-01 - ethanol - concentrated hydrochloric acid - water (10 + 6 + 4 + 1 + 1, V/V).This solvent gives a clear-burning flame and seemed likely to be suitable for the determination of barium. However, the sensitivity for barium was so poor that up to 3 g of oil per 100 ml of solvent had to be used for oils of low barium content (e.g., 0.05 per cent. m/m). The addition of this amount of oil together with aqueous standard solutions or water balance as well as the ionisation suppressant (1000 p.p.m. of potassium as chloride) produced heterogeneous solu- tions. An alternative solvent that would overcome this difficulty was therefore sought.From a limited survey of possible alternatives the most promising appeared to be mixtures of either butan-2-01 or 2-methylpropan-2-01 with white spirit and toluene. Compatibility trials indicated that a mixture, designated solvent N.l and having the following composition, could accommodate in homogeneous solution 4 g of oil and up to 10 ml of water per 100 ml of standard solution: 2-methylpropan-2-01 - toluene - white spirit (3 + 1 + 1). Use of solvent N.l had an additional advantage in that it burned with a steady flame with minimum carbon formation along the burner slit. It was also found that the incorporation of mineral acid in the solvent was not necessary, because the high flame temperature caused complete dissociation of the metal additive.Potassium naphthenate was added as ionisation suppressant instead of potassium chloride because of its better solubility in solvent N.l. It is not necessary to know the suppressant metal concentration accurately, provided that the same amount is added to each solution. To prevent separation of the components the aqueous phase had to be added in the final stage of preparing the calibration and sample solutions; water is added to the sample solutions in order to balance the solvent matrix in relation to the aqueous inorganic salts used in the calibration solutions. In burner trials of various solvents it had been observed that an increase in the toluene concentration also increased sensitivity. Experiments were also carried out, therefore, with a second solvent, designated solvent N.2 and having the composition : 2-methylpropan-2-01 - toluene (3 + 2).Results of experiments using solvent N.l and of experiments using solvent N.2 to obtain greater sensitivity, are reported in Table I. Analysis of Samples and Interference of Calcium A number of samples were analysed using solvent N.l. The results obtained are given in Table I, column 5. The standards were prepared from aqueous barium chloride solution, 5 ml, plus the solvent mixture, 95 ml. It can be seen that for samples 1, 2 and 3, which had high concentrations of calcium, there is poor agreement with the results obtained by means of X-ray fluorescence (column 3) and it is inferred that the poor agreement for these samples was associated with the higher calcium concentrations.It is known that at the wavelength (554nm) of the barium resonance line used for these measurements there is a narrow band due to calcium oxide or calcium hydroxide and therefore it seemed likely that the calcium was responsible for the interference. Although the use of very rich flames reduced the interference, it was not practicable because these flames result in fluctuating signals and hence in poor precision. Furthermore, excessive carbon formations occur along the burner slit. In order to study the effect of calcium in more detail, lean-flame conditions were used to accentuate the interference at the concentrations of interest ; concentration versws absorb- ance graphs were plotted (Fig. 1) for both calcium and barium in mixed-solvent solutions containing 1000 p.p.m.of potassium added as naphthenate. The shape of the graph for calcium produced under lean-flame conditions was surprising but shows why lower concen- trations of calcium had no visible effect on the result for barium when the normal, richerJuly, 1975 LUBRICATING OILS BY COMPARATIVE ATOMIC-ABSORPTION SPECTROPHOTOMETRY TABLE I RESULTS OF ANALYSIS FOR BARIUM Barium, per cent. m/m, obtained by- Calcium, obtained by X-rav fluorescence, I 1 atomic-absorption spectrophotometry with- X-ray a wet chemical (-*-------, Sample" - per cent. m/m fluorescence method solvent N.l solvent N.2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0.533 0.533 1-17 1.16 2.27 2.26 0-203 0-204 0-056 0.056 0.127 0-126 0.101 0.100 0.023 0.023 0-303 0.257 0.052 0.350 0.392 0-391 0.046 - - - - - 0.349 0.347 0.173 0.171 0.594 0.586 0.186 0.186 0.037 0.037 1.01 0-997 0.092 0.092 0.123 0.122 0.021 0.356 1.09 0,014 0-422 0.436 0.021 - 0.133 0.056 0.057 0.104 0.106 1.07 1.08 - - - - - - 0.189 0.039 0*040$ 1.02 0.100 0-126 0.126; 0-191$ 1-02: 0.102$ - - - I - - - 0.209t 0-216 0.213 - 0.111 0.108 1.06 1-04 0.404 0.415 0.283 0.290 0.763 0.783 0.205 0.203 0-043 0.043 1.03 1.05 0-098 0.100 0.132 0.130 - - - - - - - - - - - - 0-355 0.355 0.179 0.174 0.635 0-624 0.185 0.188 0.038 0.036 0-964 0.981 0.086 0.087 0.116 0.117 0.019 0.019 0.353 0.359 1-09 1.10 0.014 0.014 0.440 0.435 0.454 0.456 0-020 0.020 0.210 0-214 0.127 0.131 0-055 0.055 0.104 0.104 1-00 1.01 467 * Samples 1-8 were samples that had been circulated for chemical analysis in an IP correlatioii 7 IP 110 (Method B).1 Modified IP 110 (Method B). programme. flame was used. The results are given in Table I, column 5 . Although the graphs in Fig. 1 were produced when a non-standard lean flame was used, they serve to show that if such a flame were used for the barium determination, then at the normal concentration of the analyte in the sample solution (15 mg 1-1 of barium) any calcium present in a concentration in excess of 15 mg 1-1 would significantly increase the absorbance reading. However, it is also evident that, even with the lean flame, interference from calcium can be markedly reduced by diluting and worlring at lower concentrations of calcium and barium. Thus, if the normal flame could be used together with high dilutions of the metal under test then interference from calcium might be eliminated, certainly for concentrations up to the level present in samples 1 , 2 and 3.It was found that although solvent N.2 gave a significant increase in sensitivity, the sensi- tivity was nonetheless lowered at both the low and high acetylene flow-rates used to obtain the lean and rich flames, respectively. The optimum setting is therefore readily found by468 HOLDING AND ROWSON: DETERMINATION OF BARIUM I N UNUSED AHa&Sf, VOl. 100 Concentration of bariumhg I” 0 20 40 60 80 100 120 Concentration of caIciurn/mg I-’ Fig. 1. Absorbance versus concentration graphs for A, calcium and B, barium in mixed-solvent solution (Solvent N.l) at the barium resonance line (553.5 nm), using a non-standard lean nitrous oxide - acetylene flame.Acetylene flow-rate, 2-8 1 min-l. adjusting the acetylene flow to find the signal of greatest magnitude using a standard con- taining a reasonably high barium concentration. The calibration graph now obtainable for barium is shown in Fig. 2, together with a concentration veysus absorbance graph for calcium plotted for identical conditions. At the 5 mgl-1 of barium level interference in samples containing exceptionally high calcium to barium concentration ratios is negligible. The improvement in sensitivity is such that for oils of low barium content only 0.5 g of oil per 100 ml of solution (solvent N.2) is required in contrast to the 3 g of oil per 100 ml of solution required with the original acidic mixed-solvent system.The improved results for samples 1, 2 and 3 and the results obtained for a number of other oils with solvent N.2 are given in Table I, column 6, and can be compared with those obtained by X-ray fluorescence, column 3, and by wet chemical methods, column 4. Details of the method are given below. 0 2 - 6 8 10 12 14 4 Concentration of bariurnhg I-’ 0 10 20 30 40 50 60 Concentration of calciumhg I-’ Fig. 2. Absorbance veYsus concentration graphs for A, calcium and B, barium in mixed-solvent solution (Solvent N.2) a t the barium resonance line (553.5 nm), using a standard nitrous oxide - acetylene flame. Acetylene flow-rate, 3-6 1 min-1. Determination of Barium in Unused Lubricating Oils by Atomic-absorption Spectrophotometry Using a Mixed-solvent System The sample, which is contained in a mixed-solvent - aqueous solution, is analysed by means of atomic-absorption spectrophotometry using a nitrous oxide - acetylene flame, and the results are compared with those obtained with barium standards.July, 1975 LUBRICATING OILS BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 469 Apparatus An atomic-absorption spectrophotometer fitted with nitrous oxide - acetylene facilities was used; Techtron AA3 and Unicam SP90 instruments are satisfactory.The other equipment required consists of 100-ml calibrated flasks, 5- and 10-ml pipettes, and an Agla 0.5-ml syringe-burettk. All glassware should be cleaned with concentrated hydrochloric acid before use. Reagents All reagents should be of analytical-reagent grade and de-mineralised water or water of equal purity should be used throughout.Standard barium solution, 2000 mg I-1. Dissolve 3.557 g of barium chloride (BaC1,.2H20) in water and make up to 11. Mixed solvent. Potassium naphthenate solution, 10 000 mg 2-1 of potassium. Weigh 167 g of potassium naphthenate (obtainable from Durham Raw Materials Ltd., London), which contains 6 per cent. m/m of potassium, and make up to 1 1 with the mixed solvent. Preparation of Standards Add, from an Agla syringe-burette, 0, 0.1, 0.2, 0.3, 0.4 and 0.5 ml of the 2000 mg 1-1 barium standard solution to successive 100-ml calibrated flasks. Pipette 5 ml of water into each flask, and dilute to approximately 80 ml with the mixed solvent. Then, pipette 10 ml of the 1000Omgl-l potassium solution into each flask and make up to volume with the mixed solvent.Preparation of Sample Weigh into a 100-ml calibrated flask an amount of sample containing between 0.3 and 0.7 mg of barium and add approximately 80 ml of the mixed solvent. Pipette 10 ml of the 1OOOOmg1-1 potassium solution and 5ml of water into the flask and dilute to the mark with the mixed solvent. Procedure Determine the barium content of the sample by means of atomic-absorption spectro- photometry (using a nitrous oxide - acetylene flame) by comparison with the standards in the usual manner. The sensitivity for barium is dependent on flame stoicheiometry. The optimum setting is found by adjusting the acetylene flow, while spraying a standard, until a maximum absorbance reading is obtained. Typical settings are in the region of 3-6 1 min-l of acetylene and 5.5 1 min-l of nitrous oxide.Calculation Mix 2-methylpropan-2-01 with toluene in the proportions 3 + 2. Calculate the barium content of the sample by means of the following equation Barium, a w x 100 per cent. mlm = where a mg 1-1 of barium is the reading obtained for the sample solution from the calibration graph and w g is the mass of sample. Discussion It is evident from the analytical results that when oils contain a high ratio of calcium to barium the determination of barium by use of atomic-absorption spectrophotometry can be subject to severe interference. The magnitude of the interference is dependent both on the solvent used (i.e., the effect is large for a solvent giving poor sensitivity for barium) and on the stoicheiometry of the flame.By using the solvent N.2, which provides very good sensitivity for the barium determination, the interference is virtually eliminated. Table I shows that there are still some small discrepancies between the results obtained by use of atomic-absorption spectrophotometry and by X-ray fluorescence, although agreement is generally good. The method developed for barium using the nitrous oxide - acetylene flame complements the acidic mixed-solvent system developed for the determination of zinc and calcium, which470 HOLDING AND ROWSON is carried out using the air - acetylene flame. In practice, this means that two different solvent systems and two different flames are used in order to analyse an oil for barium, calcium and zinc. Apart from other considerations, solvent N.2 has poor burning characteris- tics in the air - acetylene flame and so could not be directly substituted for the acidicmixed- solvent system used in the determination of zinc and calcium. Solvent N.2, however, because of its qualities in the hotter flame, can be used with this flame to determine all three elements, zinc, calcium and barium, with adequate precision.The majority of the oils in which we are interested contain only zinc or calcium additives, either alone or in combination, and we have shown that these elements can be determined precisely using the relatively safe air - acetylene flame. It is felt at present, therefore, that there is not sufficient justification to carry out a programme of work on the application of the 2-methylpropan-2-01- toluene solvent in the determination of zinc, calcium and barium in the nitrous oxide - acetylene flame.The development of a satisfactory procedure for barium means that atomic-absorption spectrophotometric procedures for the three additive elements, zinc, calcium and barium, can be recommended. The results obtained indicate that the methods are suitable for the analysis of samples of unknown composition and also for blending quality control. However, their general applicability can only be adequately tested on an inter-laboratory basis. Conclusions Barium alone or in the presence of calcium can be readily determined by atomic-absorption spectrophotometry using a nitrous oxide - acetylene flame. The oil must be diluted in a solvent consisting of 2-methylpropan-2-01- toluene (3 + 2) plus small amounts of water. Although calcium and zinc could probably be determined by use of the new procedure for barium a satisfactory atomic-absorption spectrophotometric procedure already exists for their determination and further work is not at present justifiable. References 1. 2. 3. Kaegler, S., Perkin-Elmer News, 1966, 17 (l), 1. 4. Rimmer, A., 3rd International Atomic and Fluorescence Spectrometry Congress, Paris, 27th September-1st October 1971, Paper 809. 5. Holding, S. T., and Matthews, P. H. D., Analyst, 1972, 97, 189. Peterson, G. E., and Kahn, H. L., Atom. Absorption Newsl., 1970, 9, 71. Mostyn, R. A., and Cunningham, A. F., J . Inst. Petrol., 1967, 53 (519), 101. Received May lst, 1974 Amended January 28th, 1975 Accepted February 7th, 1975

ISSN:0003-2654    

DOI:10.1039/AN9750000465

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

Analyst, July, 1975, Vol. 100, $p. 471-475 47 L The Determination of Silver in Animal Tissues by a Wet-oxidation Process Followed by Atomic- absorption Spectrophotometry R. C. Rooney Rooney and Ward Ltd., Blackwatey Station Estate, Camberley, Surrey Silver in animal tissues can be determined at fairly low levels by conventional atomic-absorption spectrophotometry, following wet oxidation. The wet- oxidation stage is difficult when large samples are to be analysed and the final solution must be adjusted in order to minimise losses of silver by adsorp- tion or precipitation. The method has been applied to samples of urine, faeces, individual organs, skin and whole-animal homogenates. There is little information in the literature on the normal levels of silver in animal tissues or on its toxicity.Thus, the work described below was undertaken as part of a toxicological study involving silver taken orally. The experimental animals included mice, rats and monkeys, and the distribution of the silver in the body was studied. Atomic-absorption spectrophotometry was the technique chosen for this work; the detection limit is good, less than 0.01 pg ml-1, and because interferences are few, sample preparation is fairly simple. Experimental The samples to be analysed varied widely in nature and in mass. They included faeces, urine soaked into filter-paper, individual organs from mice, such as liver, kidney or gut, whole-mouse homogenates, whole-rat homogenates and pieces of monkey skin with varying amounts of subcutaneous fat. Samples were wet oxidised by a method similar to that described by hTangniot,l in which the sample is heated gently with a 1 + 1 mixture of nitric and perchloric acids, using about 5 ml of the acid mixture per gram of organic material present.The smaller samples, containing up to 2-3 g of organic material, responded well to the method, oxidising smoothly and rapidly to give clear solutions and small white final residues. Recoveries of silver added to portions of rat homogenate tended to be low (see Table I) and this was ascribed to the formation of chloride during the final stages of oxidation. The silver was added to the sample as silver nitrate solution before the addition of the acid mixture. The addition of ammonia solution to the final solutions gave quantitative recoveries within the limits of experimental error, i e ., &3 per cent. over the range 1-100 pg of silver (Table I), TABLE I RECOVERY OF ADDED SILVER Additions were made to a rat homogenate sample that was prepared from an animal that had not been dosed with silver. SampIe size/g Procedure Silver added/pg Silver found/pg 0.2 HNO, - HC10, 1.0 0-54, 0.63 0-2 HNO, - HC10, 10 8.6, 7-5 0.2 HNO, - HCIO, 100 95, 92 0-2 HNO, - HCIO, - NH,OH 1.0 0.98, 1.03 0.2 HNO, - HC10, - NH,OH 10 9.9, 10.4 10 HNO, - HC10, - NH,OH 1.0 0.96, 0-99 10 HNO, - HCIO, - NH,OH 100 96, 99 10 HNO, - H,SO, 100 65, 56 10 HNO, - HESO, - NH,OH 100 72, 68 10 HNO, - HClO, - NH,OH Nil 1.0, 0.95 0.2 HNO, - HCIO, - NH,OH Nil (0.1, <0.1 0-2 HNO, - HC10, - NH,OH 100 102, 98 50 HNO, - HC10, - KCN Nil 4, 7, 5472 Analyst, VoZ.IOU Oxidation of whole-mouse homogenates and monkey skin samples caused more difficulty because of the larger amount of organic material present, i.e., up to 10 g, and the relatively high fat content. This procedure relies for safety on the formation of a homogeneous solution, which is not obtained with high-fat materials. Fat usually separates and forms a separate upper phase, so that it reacts only to a limited extent with the nitric acid and remains relatively inert until the situation arises in which a fat - hot perchloric acid interface exists; this situation is hazardous. When such an interface occurred the solutions blackened and began a vigorous ebullition as the last traces of nitric acid boiled off. This necessitated continuous supervision of the oxidation and the addition of further nitric acid as this stage began, which was con- sidered potentially hazardous; alternative methods of oxidation were therefore investigated.Oxidation mixtures that involved the use of sulphuric acid were found to be unacceptable, as copious amounts of calcium sulphate were deposited at the final evaporation stage and arge losses of silver by adsorption were incurred (Table I). Dry ashing was rejected because of the large amounts of objectionable fumes evolved and because the residues obtained were difficult to dissolve. The nitric acid - perchloric acid digestion was therefore retained, despite being recognised as hazardous, as with reasonable care it gave acceptable results. Whole-rat homogenates caused considerable difficulties in that the animals as supplied were not particularly well homogenised and it appears to be difficult to homogenise effectively an animal of this size.It was agreed with a toxicologist that at least 10 per cent. of the whole body mass would have to be taken in order to ensure a representative sample, and as a male rat weighs up to 600g, sample masses of 50-1OOg were necessary. This range usually corresponded to 150-300 g of homogenate. Initial attempts to use the nitric acid - perchloric acid oxidation were obviously extremely dangerous, although we have had considerable experience in the wet oxidation of many types of organic material. Five grams of organic matter have usually been regarded as a desirable upper limit and we have never experienced an actual explosion from mixtures of organic matter and perchloric acid.The worst uncontrolled reaction to date has been spontaneous ignition in a beaker, when a reaction mixture had been allowed to blacken and boil to dryness. When using the large, high-fat rat samples it soon became apparent that there was a real danger of an explosion involving tens of grams of material, and a new approach was obviously essential. Many animal tissues can be dissolved in hot, concentrated nitric acid to give true solutions, and even fat can be distributed by swirling or shaking the vessel to give a relatively homo- geneous suspension. If such a solution is added to a large excess of hot perchloric acid in small portions, the organic material will be rapidly oxidised without bringing large amounts of it into contact with perchloric acid.Such a procedure is hazardous in that it requires the handling and transfer of fairly large volumes of hot, concentrated acids, but is much safer from the point of view of risk of explosion. The procedure given below describes this approach; it should be carried out by an ex- perienced analyst, aware of the dangers, and wearing suitable protective clothing. Gloves and safety glasses are a minimum requirement and all beakers should be handled with tongs. It is very important that the perchloric acid should be boiling, not merely fuming, and that it should be brought back to the boil after each addition of the nitric acid solution and before the next. It is easy to mistake the ebullition of reaction for boiling and if too much organic material is introduced at once, the perchloric acid solution will boil over as a vigorous oxidation takes place.The solutions obtained from the large rat samples did not respond well to treatment with ammonia solution in order to dissolve the silver chloride. Too much calcium phosphate was present to permit working with final volumes smaller than 500 ml and there was a considerable loss of silver by entrainment (Table I). The use of acidic solutions in order to maintain calcium salts in solution also caused a loss of silver, however, so that it was finally decided to use a cyanide complex, which is soluble and resistant to hydrolysis by dilute acids. There is usually sufficient iron present to complex any excess of cyanide, so that there is little hazard in acidifying the cyanide solution.Recoveries of added silver were found to be good by this method, and the results are given in Table I. Three procedures are described, for various sample sizes; the smallest sample size compati- ble with the provision of a representative sample should be used. The results obtained with all three methods are given in Table 11. ROONEY: DETERMINATION OF SILVER IN ANIMAL TISSUESJdy, 1975 BY WET OXIDATION FOLLOWED BY AAS TABLE I1 TYPICAL RESULTS ON TISSUE SAMPLES Whole-mouse homogenates- Total silver/pg Control . . .. .. .. .. <1 Control . . .. .. .. .. <1 Whole-mouse homogenates and corresponding excreta (all body masses were 20 f 2 g)- Silver content of Silver content of faeces in 24 h/pg urine in 24 h/pg Silver in whole body, 24 h after dosinglbg 11 12 29 22 36 6.8 Individual ovgans- Time after dosing 8h 48 h 72 h 7 d Monkey skin samples- Monkey Sample site M1 Chest Hand Foot Muzzle M2 Chest Hand Foot Muzzle M3 Chest Hand Foot Muzzle 10 2-4 6-8 6-6 0.4 4.2 Silver in Silver in gut/ Pg liverlpg 7.4 1.2 2.0 3.8 3.4 8.2 1.4 1.8 Rat homogenates-- Animal 1 2 3 4 6 6 7 8 Body mass/g (10-20% taken) 426 616 269 666 266 287 246 246 Silver/ pg 3.6 0.6 4.0 18 1-6 0.6 1.0 0.6 2.7 4.4 1.2 0.8 2.8 1.0 3.3 0.9 1.3 2.4 Silver in 17 24 0.8 0.8 kidneY/Pg Silverlpg 8-1 0.5 0.3 1.0 4.7 0.2 0.2 0.2 0.1 0.4 1.6 0.5 0.2 Silver in whole bodylpg 3280 4710 21 126 6080 60 880 4000 473 Method Reagents Nitric acid, sp.gr. 1-42. Perchloric acid, sp. gr. 1.54. Nitric acid (1 + 1 ) .Mix equal volumes of nitric acid (sp. gr. 1.42) and water. Ammonia solution (1 + 1 ) . Mix equal volumes of ammonia solution (sp. gr. 0.89) and water. Potassium cyanide solution, 1 per cent. Dissolve 10 g in water and dilute to 1 1. Tartaric acid solution, 2 per cent. Dissolve 20 g in water and dilute to 1 1. Procedure 1. Tissue Samples up to 0-5 g (e.g., Individual Organs) Weigh the sample into a 150-ml squat beaker and add 5 ml each of concentrated nitric and perchloric acids. Cover the beaker and evaporate the contents on a hot-plate at a rate such that the nitric acid is boiled off in 10-15 min and the perchloric acid begins to fume. If the solution darkens at this stage add a few drops of nitric acid in order to clear it. When oxidation has ceased, and the solution is colourless, remove the cover and evaporate the solution to dryness.474 Analyst, VoZ.100 Re-dissolve the residue in 2 ml of nitric acid (1 + l), cool and transfer it to a 25-ml cali- brated flask. Add 2 ml of 2 per cent. tartaric acid solution, rinse the beaker with 5 ml of ammonia solution (1 + l ) , add the rinsings to the flask and dilute to volume with water. Determine the amount of silver in this solution by use of atomic-absorption spectrophoto- metry, using the 328.0-nm resonance line for highest sensitivity; a stoicheiometric air - acetylene flame has been found to give the best results in our laboratory. Standardise the instrument against aqueous silver solutions containing the same concentrations of ammonia and nitric acid.ROONEY: DETERMINATION OF SILVER IN ANIMAL TISSUES Procedure 2. Tissue Samples up to 10 g (e.g., Whole Mice) Weigh the sample into a 600-ml squat beaker and add 30 ml of concentrated nitric acid. Cover the beaker. Boil the mixture gently until the initial reaction and effervescence has subsided, then evaporate it to as small a volume as possible without allowing the solution to boil dry. Add 30 ml of concentrated nitric acid and 30 ml of perchloric acid and evaporate again until incipient fumes of perchloric acid appear. At this stage the solution will probably darken; if darkening occurs, add 1-2 ml of nitric acid to the fuming solution. A vigorous ebullition will occur and as the nitric acid boils off the solution will darken again; the addition of nitric acid should then be repeated. Continue with this procedure until the solution does not darken beyond a clear brown colour; at this stage there must be at least 10 ml of per- chloric acid still present, and more should be added if necessary.Make the same additions of extra acids to the blank determination and continue to evaporate to fumes until the solution is colourless or very pale yellow. Remove the cover from the beaker and evaporate the solution to dryness. Re-dissolve the residue in 6 ml of nitric acid (1 + l ) , cool the solution and transfer it to a 50-ml calibrated flask. Add 6 ml of 2 per cent. tartaric acid solution, rinse the beaker with 10 ml of ammonia solution (1 + 1) and add the rinsings to the flask. Make the solution up to volume and determine the silver as in Procedure 1. Procedure 3.Tissue Samples in Excess of 10 g Weigh the frozen homogenate and cut portions corresponding to 10-20 per cent. of the homogenate; an amount of sample corresponding to 50-60 g of the animal is the most that can be handled. As many animals cannot easily be homogenised, as large a sample as possible should be taken in order to ensure that it is representative. Transfer the sample to a 600- or 800-ml squat beaker, add 100 ml of concentrated nitric acid and heat. Reduce the level of heating as necessary if frothing becomes excessive; it may be necessary in extreme cases to use surface heating with an infrared lamp. When the volume has been reduced to about 20m1, monitor the evaporation continuously and remove the beaker from the heat as soon as the residue begins to darken.Then add 100 ml of concentrated nitric acid and heat the mixture to boiling. Heat 120 ml of perchloric acid (sp. gr. 1.54) to boiling in a 600- or 800-ml conical flask. Swirl the solution in nitric acid to disperse the fat throughout, and cautiously pour a few millilitres into the boiling perchloric acid. A violent ebullition will take place, and the solution may darken. If the solution darkens appreciably, add a further 50 ml of nitric acid to the squat beaker in order to dilute the organic matter further. Allow the oxidation to reach completion and the solution to clear, and add a further few millilitres of nitric acid solution. Care should be taken that only a few millilitres are added at a time, or the oxidation may become uncontrolled.When all of the solution has been transferred and the final oxidation completed, allow the perchloric acid to cool. Then, add 20 ml of nitric acid to the beaker, transfer the solution in perchloric acid to it and evaporate the whole solution first to fumes of perchloric acid and finally to dryness. Re-dissolve the residue in 5 ml of nitric acid (1 + l), and transfer the solution to a 100-ml calibrated flask. Rinse the beaker with water, then with 10 ml of ammonia solution (1 + l), and add the rinsings to the flask. At this stage calcium phosphate and metal hydroxides will precipitate. Next rinse the beaker with 10 ml of 1 per cent. potassium cyanide solution and again add the rinsings to the flask, and then add 1 + 1 nitric acid dropwise until the precipitate just re-dissolves.Finally, cool the solution, dilute it to 100 ml with wateranddeter- mine the silver as in Procedure 1.Jdy, 1975 BY WET OXIDATION FOLLOWED BY AAS 476 Discussion The recoveries of added silver suggest that the method is accurate to about the same degree as the reproducibility, ie., 10-15 per cent. relative at the blank level, which is usually less than 10 pg of silver with the volumes of reagents required for 50g of sample. The detection limit is set by the variance of the blank, and is about 0-1 pug g-l of sample. It can be seen that the level of silver in unexposed animals is less than 0.1 pg g-1, and that large amounts of silver are absorbed into the tissues following ingestion. The number of distribution and elimination experiments was small, but it can be seen that in mice a large proportion of the silver is eliminated in the faeces. It also passes fairly rapidly from the gut to the kidneys, where a further substantial fraction is lost via the urine; there is some evidence of retention in the liver for a period of a few days. The distribution experiments on monkeys were inconclusive, which may be a result of the difficulty of removing skin samples with a constant amount of subcutaneous fat, or it may be caused by the small number of animals involved. The interpretation of the results has more toxicological than analytical significance. Reference 1. Nangniot, P., J . Electroanalyt. Chem., 1964, 7 , 50. Received July 19th, 1974 Amended January Sth, 1975 Accepted February 7th. 1975

ISSN:0003-2654    

DOI:10.1039/AN9750000471

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

476 Analyst, July, 1975, Vol. 100, pp. 476-481 The Determination of Unsulphonated Primary Aromatic Amines in Water-soluble Food Dyes and Other Food Additives E. J. Dixon and D. M. Groffman Department of Industry, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, SE19NQ EEC directives prescribe limits for the amount of unsulphonated primary aromatic amines present in food colouring matters in general and prohibit the presence of particular carcinogenic amines in food dyes and other food additives. Integrated methods based on solvent extraction and thin-layer chromatography have been devised for identification of these amines, followed by the spectrophotometric determination of their condensation products with 4-dimethylaminocinnamaldehyde. The method is also extended to the determination of primary aromatic amines in biphenyl and to the determina- tion of aminoazobenzenes in the dye Fast Yellow. An additional thin-layer chromatographic method is given for the differentiation between the isomers of naphthylamine as their tosyl derivatives.Directives in force within the European Economic Community (EEC)l require synthetic organic colouring matters for use in foods not to contain more than 0.01 per cent. of free aromatic amines nor more than 0-5 per cent. of intermediate synthetic products other than free aromatic amines. Further, they should not contain 2-naphthylamine, benzidine or 4-aminobiphenyl (xenylamine) or their derivatives. In Britain, the use of these three com- pounds is prohibited,2s3 except for special research purposes, and as a result authentic reference samples are difficult to obtain.In general, for purposes of identification the analyst has to rely on the physical data for these substances of unspecified purity quoted in the litera- ture.194.5 Examples of such data are given in Tables I-IV. Few up-to-date methods for the analysis of food dyes have been published in recent years and those previously available6-s are considered to lack sensitivity and specificity. Several gas-chromatographic methodsg910 are available but are unsuitable because of the difficulty of expressing mixtures of amines as “aniline,” as required in the EEC directives, and because gas - liquid chromatography by itself would not give an unambiguous identification. Methods that depend on the production of a diazonium compound, followed by addition of a coupling agent in order to produce a colour, require specific conditions for each amine under test and for each individual coupling agent.Several thin-layer chromatographic method~~9~~912 are suitable for the separation of the amines under consideration; some of these methods have been adapted to the determination of these amines in dyes, as also have some spectrophoto- me tric4J3 met hods. The spectrophotometric methods adopted depend on the formation of Schiff’s bases and afford quantitative determinations of individual amines that had previously been identified by thin-layer chromatography. Mixtures of amines are more difficult to analyse spectro- photometrically but a semi-quantitative estimate can be made following separation with thin-layer chromatography by comparing the intensities of individual spots with standards.However, the method is not sufficiently sensitive for a fully quantitative determination to be made by scraping off the spots and eluting the amines with solvent in order to give a solution of a known volume. Section 1 of this paper describes a basic procedure for the identification and determination of unsulphonated primary aromatic amines. However, it is not possible by this procedure to distinguish between 1- and 2-naphthylamines; a suitable method for this purpose is given in section 4. Section 2 describes the modifications necessary to make section 1 applicable to the examination of biphenyl, and section 3 extends the general procedure to the deter- mination of aminoazobenzenes in Fast Yellow.Crown Copyright.DIXON AND GROFFMAN 477 1. Determination of Unsulphonated Primary Aromatic Amines in Apparatus Water-soluble Food Dyes Separating funnels. Capillary tubes or micropipettes. PiPettes. 1- and 5-ml capacity. Pasteur ca9illary pipettes. Filter-paper. Strips of a Whatman low-ash grade filter-paper, approximately 20 x 50 mm. Cellulose thin-layer chromatographic Plates. Prepare the plates as follows : shake 20 g of microcrystalline cellulose powder with 60 ml of methanol for 2 min, then immediately spread the mixture on 200 x 200-mm glass plates to a thickness of 0.25 mm. After air-drying them for 30 min, dry the plates for 10 min at 105 "C. Up to five plates can be prepared in one application by using the above amount of mixture.Equally suitable are 200 x 200-mm ready-made plates of 0-1-mm thickness, which are obtainable from E. Merck. Test-tubes. 10-ml capacity, pointed and graduated, with ground-glass necks and stoppers. Recording spectrophotometer. Spectrophotometer cells. 10-mm path length. Sodium hydroxide solution, 1 M. Toluene. Analytical-reagent grade. Hydrochloric acid, 3 M. Methanol. Analytical-reagent grade. Propan-2-ol. Analytical-reagent grade. Mobile solvent for thin-layer chromatography. 100-ml capacity, with a PTFE stopcock and glass stopper. chromatographic plates, and 100-pl capacity, for dispensing 3 M hydrochloric acid. 5-p1 capacity, for spotting samples on to thin-layer Suitable for scanning in the range 350-600 nm.Reagents Dissolve 1.0 g of sodium acetate in about 10ml of distilled water and add 0.10ml of glacial acetic acid. Transfer the mixture, with aqueous washings, into a 100-ml calibrated flask and make the volume up to the mark with water. Dissolve 0.1 g of DAC in about 50 ml of methanol, carefully add 5 ml of concentrated sulphuric acid, transfer the mixture plus washings with methanol into a 100-ml calibrated flask and make the volume up to the mark with methanol. DAC solution B. Dissolve a mixture of 0.1 g of DAC and 0.1 g of P-toluenesulphonic acid in about 10ml of propan-2-01, transfer the mixture plus washings with propan-2-01 into a 100-ml calibrated flask and make the volume up to the mark with propan-2-01. 4-Dimethylaminobenzaldehyde solution (DAB).Dissolve 1 g of DAB in about 10 ml of methanol, add 4 ml of concentrated hydrochloric acid, transfer the mixture plus washings with methanol into a 100-ml calibrated flask and make the volume up to the mark with methanol. Procedure Weigh accurately about 3 g of dye into a separating funnel containing 50ml of water. Shake the funnel in order to dissolve the sample. Add 10 ml of 1 M sodium hydroxide solution and mix, then add 5 ml of toluene and shake the funnel for 1 min. Allow the layers to separate and discard the lower aqueous layer. Add another 5-ml portion of 1 M sodium hydroxide solution and shake the funnel for 1 min. If on separation the lower layer is coloured (caused by partition of the dye between the organic and inorganic phases) discard the lower layer and repeat the above extractions with alkali. Finally, wash the toluene extract with two 10-ml portions of water and discard the washings.Dry the inside of the stem of the separating funnel, then insert a 20 x 50-mm rolled-up strip of Whatman low-ash filter-paper. Transfer the toluene extract into a graduated, pointed ground-glass test-tube containing 100 & 2 pl of 3 M hydrochloric acid, stopper the tube and shake it vigorously for about 30 s. Allow the acid to sink to the bottom of the tube (it may be helpful to tap the side of the tube) and carefully remove as much as possible of the toluene with a Pasteur capillary pipette, the toluene being discarded. Spot two 5-pl portions of the remaining acidic solution about 100 mm apart, and each 20 mm from the bottom of a cellulose thin-layer chromatographic plate, and spot 5p1 of appropriate standards alongside each spot under test.Run the 4-Dimethylaminocinnamaldehyde (DAC) solution A .478 DIXON AND GROFFMAN : DETERMINATION OF UNSULPHONATED PRIMARY Analyst, VoE. 100 chromatogram in a tank containing the prescribed mobile solvent until the solvent has travelled about 150 mm above the base-line. Allow the plate to dry, cover one set of standards and sample with a glass plate and spray the remaining set with DAC solution A. Cover the other half of the plate, then spray with DAB solution. Compare the positions and colours of spots in the sample with the standards. The chromatographic behaviour of some amines that may be present is shown in Table I. TABLE I CHROMATOGRAPHIC PROPERTIES OF DAB AND DAC DERIVATIVES OF SOME AMINES RF value on cellulose TLC plate Aniline .. . . .. .. 0.88 p-Toluidine . . .. .. 0.86 l-Naphthylamine . . .. 0.47 2-Naphthylamine . . .. 0.46 4-Aminobiphenyl . . .. 0.34 Benzidine . . .. .. 0.32 Colour produced with 1 ~ A B spray DAC spray Yellow Pink Yellow Orange - pink Yellow Purple - pink Yellow Bluish pink Bluish pink Yellow Orange - pink Blue By using this technique it is not possible to differentiate between 1- and 2-naphthylamines, for which purpose a separate method is described in section 4. The absolute detectionlimitsfor the DAC spray are 10 ng of 4-aminobiphenyl and 2-naphthyl- amine and 2 ng of benzidine, which are equivalent to 0.07 and 0.02 pg g-l, respectively, when 5 pl are spotted from 3-g samples.Spectrophotometric Determination of Amines The identity of the amines may have been indicated by the thin-layer chromatographic method described in the second paragraph of Procedure and the operator can then refer to the appropriate standards, the spectrophotometric properties of which are shown in Table 11. Confirmation of identity can be made by spectrophotometric analysis as described below. To the remainder of the acidic solution (or to a 5-p1 aliquot of this solution if the presence of large amounts of aniline or toluidine has been indicated) and to another tube containing 100 p1 of 3 M hydrochloric acid (to be used as a blank), add 1 ml of DAC reagent B and make the volumes up to a total of 6 ml with methanol. Mix and allow the solution to stand for 20 min.Scan the absorption spectrum of the colour produced (if any) in the range 350-600 nm by means of a recording spectrophotometer, using 10-mm cells with the reagent blank as reference. Note the absorbance at the wavelength of maximum absorption and determine the amount of amine present by comparison with appropriate standards, as shown in Table 11. TABLE I1 SPECTROPHOTOMETRIC PROPERTIES OF THE AMINE STANDARDS Amax. for DAC derivativelnm Aniline . . .. .. .. 618 l-Naphthylamine . . .. 612 2-Naphthylamine . . .. 634 Benzidine . . .. .. 676 p-Toluidine . . .. .. 62 1 4-Aminobiphenyl . . .. 534 Absorbance of 10 pg in 6 ml of final solution (10-mm cell) 0-46 0.76 0.33 0.79 0.62 0-64 The results should be expressed as aniline unless another amine is known to be present.With a mixture of amines the results should be expressed in terms of the predominant amine. Discussion of Experimental Details In the final stage of the extraction procedure it was established that the amines should be extracted as their hydrochlorides so as to prevent losses caused by the relatively high volatilities of free aniline and toluidine. Extraction into a large volume of hydrochloric acid followed by evaporation down to a small volume proved laborious and gave rise to several unknown interferences in both the thin-layer chromatographic and spectrophotometric determinations. As far as possible it is necessary to exclude water from the amine - DAB and amine - DAC reactions, which are of the condensation type. The spectrophotometric methodJZCZY, I975 AROMATIC AMINES IN WATER-SOLUBLE FOOD DYES AND ADDITIVES 479 with DAC can tolerate the presence of up to 2 per cent.of water, above which content the sensitivity4 is markedly reduced. Several acids were tried in the spectrophotometric procedure before finally deciding on the use of 9-toluenesulphonic acid. Mineral acids such as hydrochloric or sulphuric acid caused great changes in sensitivity with different acid concentration^,^ as well as producing unstable complexes. They were, however, satisfactory for use in the DAB and DAC spray reagents. Trichloroacetic acid produced more stable colours but only at very high acid concentrations. 9-Toluenesulphonic acid, although producing less sensitive colours, formed very stable solutions at low acid concentrations with all the amine complexes tested, reaching a maximum colour intensity within 20 min.2. Determination of Unsulphonated Primary Aromatic Amines in Biphenyl Biphenyl, which is often used as a preservative for oranges and other citrus fruits, is liable t o contain trace amounts of amines. An EEC directive14 requires that biphenyl shall contain not more than 2 mg k g l of aromatic amines, expressed as aniline. The proposed method is essentially the same as that described in section 1 for dyes, apart from a variation in the extraction procedure. Procedure Dissolve 3 g of biphenyl in 5 ml of toluene and transfer the solution into a separating funnel with the aid of a further 5 ml of toluene. Add 2 ml of 3 M hydrochloric acid and 50 ml of water and shake the mixture for 1 min.Allow the layers to separate and run the lower aqueous layer into a second separating funnel containing 10 ml of 1 M sodium hydroxide solution and 5 ml of toluene. Proceed as in the first and second paragraphs of the procedure in section 1. 3. Determination of 2- and 4-Aminoazobenzenes in Fast Yellow Fast Yellow dye (C.I. Number 13015), also known as Acid Yellow 9 and Food Yellow 2, is manufactured by the disulphonation of 4-aminoazobenzene which, with the 2-isomer, occurs as an impurity in the dye. Apart from the general requirement, the EEC directive1 specifically requires that Fast Yellow (E105) shall contain not more than 10 mg k g l of unsulphonated aromatic amines. In the procedure for the determination of amines in dyes (section l ) , these substances exhibit some properties that may mistakenly be taken as evidence for the presence of benzidine. The following method permits individual spectrophotometric determinations to be made of the extracted isomers after separation by column chromato- Several improvements in the method described in the EEC directive1 have been made, namely an approximate ten-fold reduction in the amounts of sample and solvent, and the use of toluene instead of chlorobenzene.Toluene has two advantages over chlorobenzene for use in this method. Firstly, it does not form emulsions when shaken with aqueous solutions as does the latter. Secondly, it is more convenient to use a solvent that has a lower density than water when several washes with water are required. Apparatus at least 200 mm and internal diameter approximately 15 mm.of the chromatographic column. Reagents Brockmann activity 11, equivalent to a 3 per cent. water content. For prepar- ation of the alumina column for chromatography, weigh accurately 20 g of alumina into a beaker, add a few millilitres of toluene, mix and pour the slurry into the chromatographic column fitted with a small cotton-wool plug at the bottom. Run off the toluene until the meniscus coincides with the top of the column of alumina. Procedure Dissolve 2.0 g of Fast Yellow in 20 ml of water and transfer the solution into a separating funnel with the aid of a further 30 ml of water. Proceed thereafter as in the first paragraph graphy- Chromatographic column. Glass, with PTFE stopcock and Quickfit joint at the neck, length Separating funnel.Cylindrical, 100 ml in capacity, with ground-glass spout to fit the neck AZumina.480 DIXON AND GROFFMAN : DETERMINATION OF UNSULPHONATED PRIMARY Analyst, V d . 100 of the procedure in section 1. Transfer the toluene extract into a test-tube and reduce the volume of the solution to about 0.5 ml by evaporation. Transfer the solution and a small amount of toluene used for washings on to the prepared alumina column and elute it with toluene. Collect separately the first 30 ml of eluate, which contains the 2-aminoazobenzene, and the next 40 ml containing the 4-isomer. Make the volume of each fraction up to 50 ml with toluene and scan each solution on a spectrophotometer over the range 300-500nm. Note the wavelength of maximum absorbance of each solution and the absorbance at that wavelength and compare the values with the standards shown in Table 111.TABLE I11 SPECTROPHOTOMETRIC PROPERTIES OF THE AMINOAZOBENZENES Amax./nm 2-Isomer .. .. 414 40 4-Isomer .. .. 376 110 Absorptivity (1 mg ml-l, 10-mm path length) 4. Detection and Identification of 1- and 2-Naphthylamines in Water-soluble Food Dyes and Other Food Additives With neither the cellulose thin-layer chromatographic method nor the spectrophotometric method is it possible to distinguish satisfactorily between 1- and 2-naphthylamines in the presence of each other or in the presence of large amounts of other amines. However, their tosyl derivatives can be resolved by the following method, which affords further confirmation of the identity of amines determined by the thin-layer chromatographic and spectrophoto- metric methods.Of other amine derivatives that were investigated for suitability for thin- layer chromatography, only the dansyl derivative gave reasonable separation of 1- and 2-naphthylamines. The spots produced by the dansyl derivatives were more compact and more strongly fluorescent than those of the tosyl derivatives but the presence of hydrolysis and other products gave a chromatogram that was too complex for positive identification. Apparatus Alumina thin-layer chromatographic plate. Dimensions 200 x 200 mm and 0.1 mm thick, with a pH of approximately 9 (obtainable from Eastman Organic Chemicals). Reagents Mobile solvent. n-Hexane - chloroform (2 + 1 V/V). p-Toluenesul~honyl puoride.Procedure Prepare the amine mixture in toluene as described in the first paragraph of the procedure in section 1. Transfer this solution into a graduated, pointed test-tube containing approxi- mately 10 mg of 9-toluenesulphonyl fluoride. Hold the tube over a steam-bath and slowly pass air into the tube until the volume of the solution is reduced to about 0.1 ml. Spot 6 p1 of this solution about 20 mm from the bottom of an alumina thin-layer chromatographic plate and spot 5 p1 of available appropriate standards alongside. Run the chromatogram in the hexane - chloroform mobile solvent until the solvent front has travelled about 150 mm from the base-line. Allow the plate to dry and then observe the spots under ultraviolet radiation. Note the positions and colours of the spots and compare them with standards (Table IV).TABLE IV PHYSICAL PROPERTIES OF AMINE TOSYLATES RF value on alumina - Aniline . . .. .. .. l-Naphthylamine . . .. 0.62 2-Naphthylamine . . .. 0.57 4-Aminobiphenyl . . .. 0.59 Benzidine . . .. .. 0.11 - p-Toluidine . . .. .. Fluorescence under ultraviolet radiation Not observed Not observed Sky blue Dark blue Dark blue Dark blueJdY, 1975 Discussion Although the tosyl derivatives of 1- and 2-naphthylamine are not completely resolved by the alumina thin-layer chromatographic method, either isomer can be observed in the presence of at least a ten-fold excess of the other as they give different fluorescent spectra. Large excesses of aniline or j+toluidine, which are liable to interfere in 1- and 2-naphthylamine determinations by the cellulose thin-layer chromatographic method, do not interfere under the reaction conditions described in section 4; 4-aminobiphenyl or benzidine would not normally be expected to occur with the naphthylamines and in any event would have already been identified by the latter method.We thank the Government Chemist, Department of Industry, for permission to publish this paper. AROMATIC AMINES IN WATER-SOLUBLE FOOD DYES AND ADDITIVES 481 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References EEC Directive 2645/62, 23 October, 1962, Official Journal No. 115, 11.11.62. “Factories Act, The Carcinogenic Substances Regulations 1967,” S.I. 1967 No. 879, H. M. Stationery “Factories Act, The Carcinogenic Substances (Prohibition of Importation) Order 1967,” S.I. 1967 Sakai, S., Mori, H., Suzuki, K., and Fujino, M., Ja#an Analyst, 1960, 9, 862. Parihar, D. B., Sharma, S. P., and Tewari, K. C., J . Chromat., 1966, 24, 443. Butt, L. T., and Strafford, N., J. A$$l. Chem., 1956, 6, 525. Caemmerer, A. B., J . Ass. 08. Agric. Chem., 1948, 31, 592. Caemmerer, A. B., J . Ass. Off. Agric. Chem., 1949, 32, 613. Pisano, J . J ., “Theory of Application of Gas Chromatography to Industrial Medicine,” Hahnemann Clarke, D. D., Wilk, S., and Gitlow, S. E., J . Gus Chromat., 1966, 4, 310. Seiler, N., J . Chromat., 1971, 63, 97. Yasuda, K., J. Chromat., 1971, 60, 144. Feigl, I?., “Spot Tests in Inorganic Chemistry,” Seventh Edition, Elsevier Publishing Company, EEC Directive 67/428/EEC, 27 June, 1967, Official Journal No. l48/10, 11.7.67. Office, London. No. 1675, H.M. Stationery Office, London. Symposium, First Edition, Grune and Stratton, New York, 1966 (published 1968), p. 147. Amsterdam, London and New York, 1967, p. 243. Received January 13th, 1975 Accepted February 5t12, 1975

ISSN:0003-2654    

DOI:10.1039/AN9750000476

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

482 Analyst, July, 1975, Vol. 100, pp. 482-484 Determination of 3,5=Dinitro=o=toluamide in Feedstuffs and Pre-mixes M. Severijnen and F. G. Buizer Rijkslandbouwproefstation, Kruishereizgang 21, Maastricht, The Netherlands A spectrophotometric method is described for the determination of 3,5-dinitro- o-toluamide (dinitolmide) in feedstuffs and pre-mixes. Dinitolmide is ex- tracted from the sample with acetone - water (1 + 1). After purification procedures, which include liquid - liquid extraction and chromatography on an alumina column, a coloured complex is formed with sodium hydroxide, which is measured spectrophotometrically. 3,5-Dinitro-o-toluamide (dinitolmide) is a coccidiostat that is incorporated in feedstuffs, usually at a concentration of 125 mg k g l . The current spectrophotometric methods for determining dinitolmide in feedstuffs and pre-mixe~l-~ have the disadvantage that they require the use of several highly toxic agents, such as acetonitrile, dimethylformamide and met h ylamine.In a search for less toxic agents, we found that under certain conditions Janovsky’s colour reaction495 could be applied to the determination of dinitolmide. We further found that for a suitable range of concentrations this colour reaction obeyed Beer’s law. Based on Smith’s method,6 we extracted dinitolmide from the sample with a water - acetone mixture. Dinitolmide can be extracted from this medium with organic solvents, and of these solvents dichloromethane proved to be the most suitable for our purpose because of its relatively low toxicity.’ However, Janovsky’s reaction could not be directly applied to these extracts ; further purification was necessary.Previous experience with chromatography had shown that dinitolmide when dissolved in methanol will pass through a column of alumina without being absorbed and without loss, while substances that may interfere are adsorbed. Method Apparatus ChromatograPhic columns. These were made of glass, of internal diameter 10mm and length 200 mm, narrowed at one end, which was plugged with cotton-wool. For the prepara- tion of the columns, transfer into a column a slurry of 4-5 g of alumina with dichloromethane. Prepare a separate column for each sample less than 5 min before use. Rotary evaporator. Spectrophotometer. Reagents All reagents should be of analytical-reagent grade.Acetone. Acetone - water (1 + 1). Dichloromethane. Methanol. Alumina. E.g., Merck, Darrnstadt, 1097, activity grade 11-111. Sodium hydroxide solution. Hydrochloric acid, 5 mol I-I. Dinitolmide reference standard. Dissolve 15 g of sodium hydroxide in water and dilute to 100 ml with water. Procedure Weigh accurately 2-25g of sample containing 0.25-5Omg of dinitolmide into a 300-ml glass-stoppered conical flask. For feeds containing 125 mg kg-l of dinitolmide, a sample size of 5 g is usually taken. Add 100.0 ml of acetone - water (1 + 1) and agitate the mixture for 1 h. Filter it through filter-paper, rejecting the first 15 ml of filtrate. Prepare from the filtrate a solution in acetone - water (1 + 1) containing 2-5-20 mg 1-1 of dinitolmide [dilution factor (F) = final volume divided by initial volume].SEVER1 JNEN AND BUIZER 483 Transfer 20.0 ml of this solution into a 100-ml separating funnel.Add 5 ml of hydrochloric acid (5 moll-l), mix and add 25 ml of dichloromethane. Agitate the mixture gently by inverting the funnel ten times. After the phases have separated, draw off the dichloromethane layer through a cotton-wool plug into a 250-ml evaporating flask. Extract the aqueous layer three times more with 25-ml portions of dichloromethane, shaking the funnel well for 1 min each time. Reduce the volume of the combined extracts to about 5 ml at 35 "C in a rotary evaporator. Pour this solution on to a freshly prepared chromatographic column, washing the evaporating flask and column four times with 4-ml portions of dichloromethane. Elute the dinitolmide with 30 ml of methanol.Collect the eluate in a 100-ml evaporating flask and evaporate it to dryness at 50 "C in a rotary evaporator. Dissolve the residue in 5 ml of acetone by heating the mixture for a few seconds in a water-bath at 50 "C. Cool and transfer the solution into a 25-ml calibrated flask. In order to ensure complete dissolution of the residue, repeat this procedure with three more 5-ml portions of acetone, adding them to the flask. Make the volume up to the mark with acetone and mix. Transfer 5.0ml of this solution into a 50-ml conical flask and add, by pipette, 0.40ml of sodium hydroxide solution. Read the absorbance, within 3 min of adding the sodium hydroxide solution, on a spectrophotometer at 576 nm in a 10-mm cell against water as reference solution.Prepare blanks by adding 0.40 ml of water to 5.0 ml of sample solution and 0.40 ml of sodium hydroxide solution to 5-0 ml of acetone. After taking into account these blank values, the amount of dinitolmide present in the samples can be calculated by reference to a standard graph. Preparation of Standard Graph Weigh accurately 40.0 mg of dinitolmide reference standard into a 100-ml calibrated flask. Dissolve the standard in and dilute to volume with acetone. Mix well, transfer 10.0 ml of the solution into a 200-ml calibrated flask, dilute to volume with acetone and mix the solution well. Transfer 5.0, 10.0, 20.0, 30.0 and 40.0-ml portions into separate 50-ml calibrated flasks, dilute to volume with acetone and mix.Transfer, from each flask, 5.0 ml into separate 50-ml conical flasks and proceed with the colour development and spectrophotometry as described above. Calculation from the following equation : w = 2-5 x p x Flm where p mg per 5 ml is the concentration of dinitolmide in the sample solution; F , the dilution factor; and mg, the amount of sample of feed or pre-mix taken for analysis. Shortened Method for Analysis of Pre-mixes Containing more than 1 g of Dinitolmide per 100 g Collect all dichloromethane extracts in the same evaporating flask. Swirl the flask carefully for 10s. Subtract both values from the sample value. Calculate the content, w (g per 100 g), of dinitolmide in the sample of feed or pre-mix Pre-mixes containing more than 1 g of dinitolmide per 100 g can be examined by the following method.Weigh accurately a sample containing 40-50 mg of dinitolmide into a 300-ml conical flask. Add 100.0ml of acetone and agitate the mixture mechanically for 15 min. Filter it through filter-paper, rejecting the first 15 ml. From this filtrate, prepare a solution with acetone so as to contain between 2 and 15 mg 1-1 of dinitolmide. Transfer 5-0 ml of this solution into a 50-ml conical flask and proceed with the colour development and spectrophotometry as described above. Calculate the amount, w (g per 100 g), of dinitolmide in the sample of pre-mix from the following equation : (symbols as used above). w = 2 x p x Flm Results and Discussion Contents of dinitolmide down to 10 mg kg-l can be determined by the above method.Other chemotherapeutic additives to feedstuffs (amprolium, ethopabate, nitrofurazone,484 SEVER1 JNEN AND BUIZER furazolidone, nitrovin, nicarbazin, acetyl enheptin, furnicozon, dimetridazole, meticlorpindol, buquinolate, decoquinate, methylbenzoquate, sulfaquinoxaline, sulfamezathine, sulfacet - a i d e , monensin, tetracycline, oxytetracycline, penicillin, streptomycin, zinc bacitracin, tylosin, oleandomycin, virginiamycjn and spiramycin) do not interfere when present in the usual amounts. For instance, a content of nitrovin of 100 mg kg-1 (usually 10 mg kg-1) gave a “dinitolmide recovery’’ of 10 mg kg--l. Dinitrobenzamide interfered, but could be distinguished from dinitolmide by using concentrated ammonia solution instead of sodium hydroxide solution.Dinitolmide did not react with ammonia to give a blue colour. The ratio of acetone to sodium hydroxide solution of 5.0: 0.44 gives a coloured complex, which is stable for about 3 min and the absorbance must therefore be read within this time. The above method has been in normal routine use in our laboratory for almost 2 years and has given full satisfaction. Results and 1. 2. 3. 4. 5. 6. 7. We incorporated dinitolmide into several feed ingredients : feathermeal, fishmeal, grassmeal, maize gluten feed and meatmeal. All these ingredients showed low blank values (less than 3 mg kg-1). When medicated with dinitolmide (125 mg kg-l) the method gave complete recoveries, except with meatmeal, when 95 mg kg-l was recovered. Results for a commercial poultry feed that was medicated with 125 mg kg-l of dinitolmide are given in Table I. The blank value of the feed was less than 1 mg k g l . The average value for the dinitolmide recovered was 124 mg k g l with a standard deviation of 4 mg kg-1, the limit of error for a 95 per cent. probability level was 2 mg k g l . TABLE I COMMERCIAL POULTRY FEED MEDICATED WITH 125 mg k g l RECOVERY OF 3,5-DINITRO-O-TOLUAMIDE IN 24 DETERMINATIONS ON A Dinitolmide recovered/mg kg-l r 128 126 127 12; 124 119 122 126 128 128 127 115 128 127 12 1 113 117 121 123 127 125 125 119 122 References Getzendaner, M. E., J . Ass. Off. Analyt. Chem., 1961, 44, 18. Smith, C. N., J . Agric. Fd Chem., 1960, 8 (3), 224. Analytical Methods Committee, Analyst, 1969, 94, 1159. Janovsky, J. V., Ber. Dt. Chem. Ges., 1891, 24, 971. Beckmann, H. F., J . Agric. Fd Chem., 1959, 7, 280. Smith, G. N., J . Agric. Fd Chem., 1961, 9, 197. Sax, N. I., “Dangerous Properties of Industrial Materials,” Van Nostrand Reinhold, New York, 1968. Received December 31st, 1974 Accepted January 28th, 1975

ISSN:0003-2654    

DOI:10.1039/AN9750000482

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

Analyst, July, 1975, Vol. 100, p p . 485-488 485 An Ion-selective Electrode Method for the Determination of Nitrate in Grass and Clover A. W. M. Sweetsur and Miss A. G. Wilson The Hannah Research Institute, Ayr, Scotland, KA 6 6HL A method for determining the nitrate content of grass and clover involving the use of a nitrate-selective electrode is described. The method is rapid, accurate and precise and can be used for samples containing as little as 10 p.p.m. of nitrate-nitrogen. Methods currently exist for the determination of nitrate-nitrogen in plant materials by using a nitrate-selective electrode,14 but they have a lower limit of sensitivity of approximately 100 p.p.m. of nitrate-nitrogen. Hence they cannot be applied directly to grass and clover, which may contain much lower amounts of nitrate.This paper describes a procedure for the determination of nitrate in rye-grass, Blanca white clover and mixtures of them down to a lower sensitivity level of 10 p.p.m. of nitrate-nitrogen. Experimental Reagents All the reagents were of AnalaR quality except when stated otherwise. De-ionised water was used throughout. B u f e r extraction solution. The concentration of this solution was 2.5 times that used by Milham et u E . , ~ i.e., 0.025 M with respect to aluminium sulphate, 0-025 M to silver sulphate, 0.050 M to boric acid and 0.050 M to sulphamic acid (microanalytical grade). The pH was adjusted to 3.0 with 0.1 M sulphuric acid. Apparatus Electrodes A nitrate-selective membrane electrode (Corning Instruments Ltd.) was used in conjunction with a Corning silver - silver chloride glass double-junction reference electrode, the outer compartment of which was filled with 1 M sodium sulphate solution.Electrometer Nitrate-nitrogen concentrations of more than 5 p.p.m. in the test solution were read directly on the activity scale. Concentra- tions of between 1 and 5 p.p.m. were read from a calibration graph. Procedure Two procedures were employed, depending upon whether the sample contained more (A) or less (B) than approximately 500 p.p,m. of nitrate-nitrogen. (A). Weigh approximately 0-1 g of sample, add 5 ml of buffer extraction solution and 5 ml of water. Mix in order to moisten all of the sample and shake the mixture gently for 5 min. Add 5 ml of buffer solution to 5-ml portions of potassium nitrate standard solutions containing an appropriate range of nitrate-nitrogen.Calibrate the electrometer with the standard solutions at 20 "C according to the operating instructions. Measure the nitrate-nitrogen concentration of the sample. (B). Weigh approximately 1.0 g of sample, add 5 ml of buffer extraction solution and 5 ml of water. Mix, shake the mixture for 5 min and leave it for at least 2 h to complete the extraction. Calibrate the electrometer as for (A), and measure the nitrate-nitrogen con- centration of the sample. NOTE- but this was found not to shorten the response time. A Corning - EEL, Model 101, instrument was used. The test solution can be stirred mechanically in order to enhance the attainment of equilibrium,486 SWEETSWR AND WILSON : ION-SELECTIVE ELECTRODE METHOD FOR Analyst, VOZ.IOO Results and Discussion Fig. 1 shows the relationships obtained between millivolt readings and nitrate-nitrogen concentrations for standard potassium nitrate aqueous and buffered solutions containing between 0.14 and 7000 p.p.m. The difference between the curves was doubtless due to the difference in the ionic strengths of the solutions, but above concentrations of about 5 p.p.m. each approximates closely to that predicted by the Nernst equation. > E 00 Nitrate-nitrogen, p.p.m. Fig. 1. Nitrate-selective electrode calibration graphs for aqueous (A) and buffered (B) standard solutions. Attempts to use the buffer described by Milham et aL3 in conjunction with 1.0-g samples of grass resulted in erroneously high nitrate values being obtained, probably as a result of incomplete removal of interfering ions.Table I demonstrates that the use of the more concentrated buffer solution described above resulted in similar nitrate recoveries for different amounts of the same grass. Although good nitrate recoveries could be obtained from 1.5 g of grass per 10 ml of test solution (5 ml of buffer solution plus 5 ml of water), the electrode response was very slow (about 5 min) and therefore this level of grass was not used for routine measurements. TABLE I EFFICIENCY OF BUFFER SOLUTION FOR NITRATE EXTRACTION Nitrate-nitrogen, p.p,m. Ratio of sample (g) Grass to buffer solution (ml) Test -&SS 1 0.1 : 10 8 800 0.3 : 10 25 833 1.0 : 10 82 820 1.5 : 10 122 814 2 0.1 : 10 35 3500 0-3 : 10 104 3463 1.0 : 10 384 3480 1.6 : 10 618 3466July, I 9 75 487 It was found that complete nitrate recoveries could be obtained for 0.1-g samples using procedure A outlined above. However, Table I1 demonstrates that a 2-h incubation period after shaking is required for complete recovery from the 1.0-g samples, and that an increase in the shaking period had little effect on the recoveries.TABLE I1 EFFECT OF SHAKING AND INCUBATION TIMES ON NITRATE RECOVERIES THE DETERMINATION OF NITRATE IN GRASS AND CLOVER 2 Y .- f 4d c F .- 2 0 !i l:: 1000 2000 3000 4000 Amount Shaking Sample analysed/g time/min Grass . . . . 1.0 5 20 Grass . . . . 0.1 5 20 Clover . . . . 1.0 6 20 Clover . . , . 0.1 5 20 a zk 3500 L 3 B e 2 2 2000 u 3000 Q 2500- c, a2 - a2 Q - E 1500 a C .- 7- 0 0.5 - - - - Incubation time/h 1 2 4 6 24 A \ Nitrate-nitrogen, p.p.m.r 370 365 360 372 718 705 716 721 10 11 10 10 2950 2940 2930 2930 421 435 705 718 14 13 2920 2950 482 473 714 716 16 16 2950 2920 468 470 720 720 15 16 2910 2920 459 468 716 722 16 15 2930 2950 1 47 1 476 722 715 16 16 2910 2930 Duplicate determinations were made of nitrate-nitrogen concentration on 41 dried grass, grass plus clover or clover samples by the electrode method and the standard AOAC pro- ~ e d u r e . ~ The relationship obtained is shown in Fig. 2. The regression equation was [nitrate (electrode)] = -2.92 + 0-97 [nitrate (AOAC)] and the correlation coefficient of 0-9981 was highly significant. The slightly lower values obtained with the electrode method may have been caused by salt interference^.^ Duplicate sub-samples were taken from 27 grasses or clovers and paired nitrate-nitrogen determinations made on each by the electrode method.Eight of the 54 paired determinations differed by more than 2 per cent. and three by more than 5 per cent. The means of three of the 27 duplicate determinations differed by more than 2 per cent. and one by more than 5 per cent. The accuracy, precision and repeatability of the method are therefore good. 5 4 0 0 0 5 Nitrate-nitrogen in sample (AOAC procedure), p.p.m. Fig. 2. Relationship between nitrate - nitrogen concen- tration of grass and clover determined by the electrode and the AOAC procedures. Table I11 demonstrates that the recoveries of added nitrate to low, intermediate and high nitrat e-cont aining samples are complete.488 SWEETSUR AND WILSON It can be concluded that the procedure described provides a simple, accurate and rapid method for the analysis of grasses and clovers containing as little as 10 p.p.m.of nitrate- nitrogen. TABLE I11 RECOVERY OF ADDED NITRATE BY THE ION-SELECTIVE ELECTRODE METHOD Sample Grass . . .. Grass + clover . . Clover . . .. Nitrate-nitrogen added, p.p.m. 0 36 70 140 350 700 0 36 70 140 350 700 0 35 70 140 350 700 Total nitrate-nitrogen in test solution, p.p.m. 2 37 72 142 352 702 29 64 99 169 379 729 310 345 380 450 660 1010 Nitrate-nitrogen determined in test solution r- p.p.m. Recovery, per cent. - 2 37 100.0 73 101.4 141 99.3 350 99.4 706 100.6 29 64 100.0 100 101.4 167 98.6 382 100.9 734 100.7 310 1 344 97-1 379 98-6 452 101.4 657 99.1 1005 99.3 - The authors thank Dr. D. Reid for his statistical analysis and Miss N. McGregor and Mrs. J. Cuthbertson for their technical assistance. References 1. 2. 3. 4. 6. Paul, J. L., and Carlson, R. M., J . Agric. Fd Chem., 1968, 16, 766. Baker, A. S., and Smith, R., J . Agric. Fd Chem., 1969, 17, 1284. Milham, P. J., Awad, A. S., Paull, R. E., and Bull, J. H., Analyst, 1970, 95, 751. Barker, A. V., Peck, N. H., and MacDonald, G. E., Agron. J., 1971, 63, 126, Horwitz, W., Editor, “Official Methods of Analysis of the Association of Official Agricultural Chemists,” Tenth Edition, Association of Official Agricultural Chemists, Washington, D.C., 1965, p. 347. Received December 4th, 1974 Accepted January 16th, 1976

ISSN:0003-2654    

DOI:10.1039/AN9750000485

出版商:RSC    

年代:1975

数据来源: RSC