ISSN:0144-557X    

DOI:10.1039/AP98926FX029

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANPRDI 26(8) 281-304 (1 989) Proceedings of the Analytical Division of The Royal Society of Chemistry 281 SUMMARIES OF PAPERS 281 Sampling to Determinations 281 283 284 285 289 290 'Concepts in Analytical Chemistry' by G. E. Baiulescu and J. D. R. Thomas 'Sampling and Standards' by 6. W. Woodget 'Strategies for Sampling Ambient Air for Pollutants of Industrial and Environm Interest' by Peter Clayton and Brian Davis 'Simplex Optimisation' by K. Burton 'New Strategies for Determining Trace Metals in Industrial Materials and Effl by A. M. Bond 'Modified Simplex Optimisation for Operating an Enzyme Electrode' by S. I< G. J. Moody and J. D. R. Thomas 293 First Prepare Your Sample 293 294 'Sample Preparation Procedures Used in Clinical Chemistry' by Andrew Taylor 'Automated Sample Handling for Drug Analysis from Biological Fluids Using by E. Doyle, R. D. McDowall, G. S. Murkitt, V. S. Picot and S. J. Rogers 296 Equipment News 299 New Deputy Director for Polytechnic 299 Publications Received 300 AOAC/Europe Regional Section 302 Ronald Belcher Memorial Lectureship (Rules) 302 1990 Merck Medal for Analytical Chemistry-Call for Papers 302 Conferences and Meetings 304 Courses iii Analytical Division Diary Typeset and printed by Black Bear Press Limited, Cambridge, England August 1989 Analytical Proceedings CONTENTS

ISSN:0144-557X    

DOI:10.1039/AP98926BX031

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, AUGUST 1989. VOL 26 28 1 Sampling to Determinations - The following are summaries of six of the papers presented at a Joint Meeting of the Western Region of the Analytical Division and the South East Wales Section of the Royal Society of Chemistry held on November 17th, 1988, in the Redwood Building of the University of Wales College of Cardiff. Concepts in Analytical Chemistry G. E. Baiulescu Department of Analytical Chemistry, National Institute of Chemistry, 202, Spl. Independentei, 77208- Bucharest, Romania J. D. R. Thomas School of Chemistry and Applied Chemistry, University of Wales College of Cardiff, P.O. Box 912, Cardiff CFI 3TB The process of sampling is integral with obtaining the required analytical information from the sample and of the outlook of the analytical chemist who obtains that information.In this way, and knowing the history of the sample to be analysed, the qualitative, quantitative and structural aspects of the complete analytical process of measuring and control can be properly related. These three concepts of quality, quantity and structure are interdependent and are representative of analytical chemistry as far as the total information is concerned in terms of the manner of producing the information, the associated calibration and of the adaptability of the analytical chemist. Producing the Information As previously stated,’ any analytical determination is based on the very simple relationship expressing a dependence between a physical property, P, and concentration, C. Hence, the widespread use of physical methods in analytical chemistry, a feature which, in Europe at least, and especially in Britain, has led to a dichotomy of views of the nature of analytical chemistry.The view of physicochem- ical analytical methods as belonging to physical chemistry concedes that the classical methods of gravimetry and ti- trimetry rightly belong to analytical chemistry, but con- veniently overlooks the fact that these methods are also based on physical properties. Analytical chemistry is broader and consideration of the analysis of the sample can be discussed in terms of the octahedron’ shown in Fig. 1. When (inorganic and organic) samples of relatively simple composition, especially in solution, need to be analysed, they are given to the analytical chemist as a performer of chemical P = f(C) .. . . . . . . (1) Mathematics Physics Fig. 1. The octahedron of analytical chemistry2 analysis. However, an altogether different line is frequently taken when the sample has a more complex composition and requires more advanced non-destructive methods, such as surface analysis approaches. The analytical chemist is then often excluded, and may even be replaced by a physicist. This is irrational thinking. The analytical chemist, in trying to make efficient use of the full arsenal of physical methods, non-destructive or destruc- tive, must know the nature of the sample in order to get the relevant information about it. The ability to have this knowledge implies being well informed in the parallelogram horizontal in three of its corners (Fig.l), namely, inorganic, organic and physical chemistry. The three branches have equal parts in the formation of a good analyst. Increasingly, the fourth comer is needed to complete the horizontal development, for the analytical chemist has to have rather more than an elementary knowledge of the basic nature of automatic instrumentation, and has to accept that it is microcomputer-equipped. Therefore, the parallelogram, con- necting the sample to “automatics” by means of the three fundamental chemical disciplines, has, as its centre, the fact that today the analytical chemist is no longer a simple performer, but also a thinker, a person who solves problems. The scientific base of the analytical chemist extends beyond the plane of the parallelogram by means of the two apices to yield an octahedron, that is, the development of the analytical chemist in the vertical direction to include physics and mathematics.Physics is the base for being able to use laboratory instrumentation, and to understand the physical significance of the various phenomena occurring in chemical interaction or reaction. A knowledge of mathematics (not only statistics) is required for the handling of data processing, that is, for converting and transmitting the analytical signal into a suitable form. Owing to the complexity of the data to be interpreted, there is a need for a mathematical background. Thus, for the complex mixture resolved by the GC - MS - computer system, or even by just MS analysis of some organic components, the analytical chemist has to reconstitute the original molecules knowing only the mass fragments.This problem cannot usually be solved by an analyst’s own unaided efforts, so that the computer has to be used. Although this is helpful, it is only the analyst who can know enough chemistry to be able to interpret the results. Calibration The need for computer pattern recognition, essentially a282 ANALYTICAL PROCEEDINGS. AUGUST 1989. VOL 26 branch of chemometrics defined2 as “the chemical discipline that uses mathematical and statistical methods to ( a ) design or select optimal measurement procedures and experiments, and (6) provide chemical information by analysing chemical data,” brings us to another point of importance in the analytical process, namely, calibration.The example selected here is that of the calibration of ion-selective electrodes (ISEs) used in body fluid analysis. This involves the conversion of measure- ments (voltages) to chemical information (concentration of analytes) which, in the ideal example, is the model R = kC . . . . . . . . (2) where R is the sensor response, C is the analyte concentration and k is the response constant (or slope of calibration line). Of course, for equation (2) to hold true, the analytical sensor must not be sensitive to interfering components and matrix effects must be absent. Interferences change the model to . . . . (3) where, if kB and CB are known and constant, they can be simply subtracted from R. In the absence of interfering ions, ISE response equations E = constant + Slog U A .. . . (4) can be directly modelled on the straight-line relationship: where y is the dependent variable (here E), x is the independent variable (here log aA), and 60 and 61 are the linear parameters of the model (intercept and slope). The quality of fit of the calibration is commonly judged by reporting the correlation coefficient but, because this does not take into account degrees of freedom, it is better to have a more general quality check based on an “analysis of variance” (ANOVA). Here, all the possible sums of squares are calculated, which add up to the total sum of squares, as illustrated here for the calibration of a sodium glass ISE3.4 (Fig. y = b o + b l x . . . . * - ( 5 ) 2). Corrected for the mean = 44.93 - \ Due to factors = 43.01 / / “-IJ pl = 1 Pure experimental uncertainty = 1.68 n-f = 25 Fig.2. ANOVA for glass sodium ISE demonstrating additivity of sum of squares and degrees of freedom334 for the straight-line calibration E = 210.9 + 56.84 log (“a] + 6.35 x 10-3 (K]), based on test solutions each containing sodium and potassium and each measured on 6 occasions. Key: n, number of measurements; p , number of parameters; f, number of independent factor combinations As is seen in Fig. 2, the “total sums of squares” (167.3) are used to evaluate the adequacy of the model, that is, the straight-line relation . . (6) by F-tests for “goodness of fit” and “lack of fit.” Thus 43.0111.92 28 = 628 for the F value in the goodness to fit test of the data of Fig. 2 compares with a 99% confidence level (1,28) of 7.0 10E’S = 10EdS([Na] + k c , , K [q) .. indicating a highly significant fit. In the lack of fit test, ’+/% = 1.19 for Fcrit (0.95,3,25) = 2.99, indicatingno statistical lack of fit, even at the 95% confidence level.3.4 Therefore, an ANOVA check helps to consider objectively the performance of the analytical device, here an ISE. The intermediately stored data can be used to compute confidence bands along the calibration lines, or to characterise the variance associated with particular estimates. Adaptability of the Analytical Chemist The chemicoanalytical control of industrial production is now generally carried out with automatic analysers with the objective of the optimisation of industrial output. Analytical chemists have adapted to this, but progress in analysis demands continued adaptability so that the analytical chemist of the future must be a cultured man.’ This will only be ensured by the continual improvement of the teaching of chemistry in general and analytical chemistry in particular.Central to this is the viewpoint of Abelson,s who wrote in an Editorial on “Science in the Twenty First Century”: “Whatever the changing shape of society, scientists and engineers will have essential r8les. The uncertainties, though, make it advisable to caution against excessive specialisation. In contrast, it seems desirable to adopt policies of maximum flexibility, of preserva- tion of options, of being prepared to pursue life-long learning.” It is certain that, in order to maintain form, a good specialist must continuously read.In this way the analyst has learnt to operate automatic analysers, but not to become part of them. Thus, the optimisation of quality control can be pictured as a pyramid2 with a “man - instrument - method” correlation at the base and “quality control” at the apex. In order for the correlation to lead to automatic quality control, the “man” factor has to ensure perfect matching of the operational parameters of the “method” to the most suitable “instrument.” In the case of on-line analysis, he would have to ensure that the system gets and processes the information continuously. Reproducibility of the analytical signal is a problem in on-line analysers; hence, instruments with a self-calibrating facility have to be designed. The analytical chemist is the only person in a position to judge or assess if the “method - instrument” couple is favourable, or when it needs replace- ment or adjustment.To be competitive in this and in a host of other respects, whether it be clinical monitoring or consumer protection, the analytical method must meet conditions of ( a ) the best possible accuracy, (6) the best selectivity, ( c ) good accuracy and ( d ) rapidity. Conclusion In order to correlate quality, quantity and structure in the analytical process and design it for obtaining the best informa- tion on the sample, it can be concluded that there is a clear r6le for analytical chemistry in education at undergraduate, post- graduate and post-experience levels. Improvements in the educational process and in teaching lead to better research, and better research improves products and industrial production. Man cannot be bettered, except within a continuing educa- tional process.’ References 1.2. 3. 4. 5 . Baiulescu, G. E.. Analyst. 1980, 105. 1045. Frank, I. E., and Kowalski. B. R.. Anal. Chern., 1982,54,232R. Otto, M., and Thomas, J. D. R., Anal. Chem.. 1985, 57.2647. Otto, M., and Thomas, J . D. R., Ion-Sel. Electrode Rev., 1986, 8,55. Abelson, P. H.. Science, 1979, 205. 1087.ANALYTICAL PROCEEDINGS, AUGUST 1989, VOL 26 283 Sampling and Standards B. W. Woodget Division of Chemistry, Hatfield Polytechnic, College Lane, Hatfield, Hertfordshire Sampling Philosophy Although the importance of correct sampling has been recognised for many years, it is only within the past decade or so that it appears to have become acceptable as a topic for scientific meetings.Of all of the stages within an analytical process, the sampling stage is the most important, by virtue of the fact that if this part of the process is not carried out correctly, the analytical result obtained eventually will be both inaccurate and misleading. No sampling procedure can be correctly designed, however, without extensive prior know- ledge of the system to be sampled. A sample taken correctly is termed a “representative sample”, which is defined as “a portion of a material taken from a consignment and selected in such a way that it possesses the essential characteristics of the bulk.” The sample taken must be representative, both chemic- ally and physically, of the bulk material from which it was obtained ( i .e . , none of the components has been lost, and with solid matrices, the particle size distribution and ratio must be the same as in the bulk material). Sampling terminology is well established and is illustrated in Fig. 1. Gross I sample 1 t Su b-sarn ple 1 I t Analysis I sample 1 Fig. 1. Sampling terminology and relationship between stages In Fig. 1, the increments taken from each sampling unit have been shown to be composited to produce a single gross sample, this being the scheme most often adopted. Once the initial sample has been taken, great care must be exercised to avoid losses of any of the components, or to avoid physical or chemical changes occurring within the sample. Single phase liquid samples and gas samples are fairly easy to homogenise in order that a representative analysis sample may be taken: solid samples may require several stages of homo- genisation, sieving and sub-sampling before a suitably sized analysis sample is available.However carefully the sample was taken, it is still impossible to know precisely the representative nature of the sample, and thus any sampling procedure relies heavily on statistical analysis and theory. Even after extensive investigation, the method actually chosen is that which appears to give the correct answer, the true answer only being available when all of the material has been reacted and analysed. Thus, a correct evaluation of error lies at the heart of any sampling procedure. Assessment of Sampling Error As with any other part of an analytical process, sampling error can be divided into that which is systematic and that which is random.Systematic errors are often caused by bias, due to poor method development or human weakness. Whilst it should be possible to detect and correct for these errors, random errors are impossible to eliminate completely. Ran- dom errors should follow a normal distribution pattern, but as the distribution is obtained on a finite amount of data, the true mean of the data (p) will never be known exactly, only the estimated mean (3 being obtainable. The true and estimated mean values are related together through the equation p = X L- tslnl where t is a statistical term obtained from the students t-test table, s is the standard deviation about the mean, and n the number of results.Thus, the margin of error within which we are confident that the true mean lies is given by the term ts/n*. Any value relating to the accuracy of a sampling procedure will inevitably be a combination of errors, brought about in both the sampling and determination stages of the analysis. In practice, the individual constituents within the population do not have an equal opportunity of being present in the sample taken for analysis, because the sample never contains one completely uniform species. Hence, an inherent variance (sampling error) is always present. It is usual to express this error as s2 = sg2 + s12 where s, so and s1 refer, respectively, to the standard deviations of the over-all, determination and sampling stages.The variance of the mean value X is given by s2/n, and this applies where one analysis is carried out on each of n sample increments. If the value is obtained by making N replicate measurements on the n sample increments, then the variance of the mean value becomes (so2lN + s12)/n For maximum precision, the variance of the mean is required to be as small as possible, and this can be achieved by either using a more precise method for analysis, or by increasing N. There is no real advantage, however, in reducing the value of so2 below about 10% of that of s12, as this will have little effect on the over-all value of s2. It is better to take a larger number of sample increments, as the confidence interval will decrease by increasing n. Thus, taking n increments, compositing them to produce one gross sample which is then analysed N times gives an over-all variance of s,VN + s12/n This over-all variance will be a little higher, but the analysis costs will have been considerably reduced.Outline of a Sampling Procedure When deciding upon a sampling procedure, the method t o be adopted will depend inevitably upon the population to be sampled. It is important also to realise that the properties of the population under consideration may vary from place to place284 ANALYTICAL PROCEEDINGS. AUGUST 1989. VOL 26 and from time to time. Obviously, there is little point in attempting to achieve a highly accurate method of sampling where sample removed from the same source shows consider- able variation with time. In these instances, it is best to take samples over a fixed period of time and bulk them together to obtain a time average analytical figure.When deciding upon a sampling procedure, the following questions must be answered: What size of error is acceptable? How many increments are required and what size? Where shall the increments be placed? How shall the increments be combined and further treated to provide a sample for analysis? Is there a variation in particle size? How homogeneous is the distribution of the analyte through- o u t the matrix? In choosing or placing increments, care must be taken to avoid bias whilst at the same time collecting a sample that contains the correct ratio of the components. The use of random number tables will help in overcoming bias, but may not produce a satisfactory sample where the population is highly segregated.This problem is best avoided by using imaginary sectioning,' increments being removed from each separately defined section. The process is, of course, easier said than done. In choosing an incremental size, a rough rule of thumb is to take a sample with a device which is 24 times that of the largest particle, bearing in mind that all increments should be of the same size. The incremental quantity may vary from mg to kg, dependent upon the parameters listed above. Attempts have been made to devise empirical formulae to define sample size for particular types of materials. However, these should be used only as a guide, the exact sample size being decided after considerable experimental evidence has been accumulated.Standardisation Standardisation can refer to a variety of procedures from the preparation of standard substances to their use in determina- tion, and to the development and validation of standard methods of analysis. One area which appears to receive little coverage, but which is applicable to many comparative analytical methods, is that of standard addition. The method is applicable whenever a linear relationship exists between instrumental response and concentration, and may be applied via a single or multiple addition procedure. Both methods offer the advantage over the use of calibration graphs of being able to account for some interference effects created by the sample matrix, but only the single standard addition allows an accurate quantitative answer to be obtained with the minimum of sample preparation and measurement. For an analytical method exhibiting no systematic error, the concentration of analyte may be found from two measurements, but three measurements (sample + 2 additions) are required when a systematic error is present.When used with ion selective electrodes, it is necessary, unfortunately, to know the value of the Nernstian response constant for the cell, and this minimises the potential usefulness of the technique in this context. The technique is most applicable where the standard can be added in situ to the solution being measured, thereby reducing still further the time required for sample preparation. Voltam- metric methods are perfect examples of this particular applica- tion.Many atomic absorption methods are prone to some interference, which can be a particular problem when analysing the one-off sample. Provided this is chemical interference, created either in the condensed or vapour phases, then standard addition should be capable of overcoming the effect, assuming, of course, that the calibration graph has a linear portion and that the total concentration of analyte in the sample plus the standard does not exceed this linearity. By making, say, triplicate measurements on each of the two solutions, the over-all precision should be similar to that obtainable by interpolation from a calibration plot. Reference Woodget, B. W.. and Cooper, D.. "Samples and Standards (ACOL Text No. 1):' Wilev. 1987. 1. Strategies for Sampling Ambient Air for Pollutants of Industrial and Environmental Interest Peter Clayton and Brian Davis Department of Trade and Industry, Warren Spring laboratory, Gunnels Wood Road, Stevenage, Hertfordshire SG? 2BX Pollutants can arise in the environment from natural or man-made sources.They may be present as solids, liquids or gases and may affect the atmosphere, water or soil. For example, sulphur dioxide may affect plants as a gas or as acid rain. The effects of pollution may be in terms of health or nuisance or loss of amenity. Evaluation of the environment is complicated and difficult and is normally carried out on the basis of Air Quality Standards in the case of the atmosphere and Water Authority Standards in the case of water. The burden of assessment falls on the regulatory organisations such as HM Inspectorate of Pollution, the Hazardous Waste Inspectorate, the Department of the Environment and Local Authorities.Expansion of pollution control legislation and air quality standards will probably continue. Sampling Considerations Sampling for air pollutants may be by direct reading instru- ments, by filtration or scrubbing samples for subsequent analysis or by collecting samples of air for later analysis. The choice of a particular method depends on the portability and reliability of equipment in the field, the cost and the type of information required. Recent trends in manufacturing have been towards more direct reading instruments for specific substances. Any instrument must operate with the required level of sensitivity, accuracy and reproducibility under arduous field conditions.Often the contaminant under investigation is known as a result of a history of complaint, knowledge of the likely source or as a result of preliminary surveys. The problem then becomes one of quantification of a representative sample. It is therefore necessary to consider, where, how and when to sample, for how long to sample and how many samples to take. Where to sample depends on the circumstances; for a complaint investigation one site will not be enough since comparisons with background concentrations will be neces- sary. Alternatively, for air quality measurements it may be necessary to set up networks of samplers to measure temporal and spatial distributions. But the maximum amount of relevant information must be obtained from the minimum number ofANALYTICAL PROCEEDINGS, AUGUST 1989, VOL 26 285 sites.If it is necessary to determine the contribution of a particular source to ambient concentrations this would also affect site selection. How to sample depends on objectives and resources. Continuous monitors are expensive but cheap to operate; the reverse is true of manual methods. How long to sample and how many samples to take usually applies to manual methods and depends on the sensitivity of the method and the expected ambient concentration. Measure- ments for air quality standards must take account of specified averaging times and so the number of samples is stated. It is desirable to cover a range of meteorological conditions using suitable sampling periods together with measurements of wind speed and direction.A suitable sampling period depends on the method and its sampling time. If weekly sampling is to be carried out then not less than 3 months are required to obtain an over-all average concentration; 1 month may be sufficent for daily sampling unless an assessment of seasonal variations is to be undertaken. If an event or complaint occurs only under particular conditions it will be necessary to design the programme in order to obtain data under those conditions. When assessing the contribution of a source to ambient concentrations it is necessary to measure with the wind in the correct direction and this is the basis of the WSL directional instrumentation. The Siting of Samplers The siting of samplers depends on the nature of the pollutant, gaseous or particulate, and the objectives of the campaign.It is usual to site the apparatus or continuous gas monitor within buildings, using an external sampling probe with the end about 1 m from the wall with a downward facing funnel to avoid entry of rainwater. The inlet should be well away from possible sources of the pollutant to be measured and far enough above ground level to minimise vandalism, The EC Directive for smoke and SOz suggests between 1.5 and 5 m above the ground. Black smoke can be measured at similar sites because the particle size sampled is much less than 5pm. For other types of particulate sampling consideration should be given to local topography, other buildings and the distance of the source or other sources of similar material.It is possible, by using suitable siting, to differentiate between stack and close to ground level fugitive emissions. When sampling particulate material samplers should be more than 3 m from buildings on the windward side. Preferably there should be a clear line of site to any source being investigated. On the leeward side of buildings the sampler should be at least five building heights away in order to avoid enhanced measurements in the recirculation region. Whilst sites should be representative of the area under review they should not be in enclosed situations or subject to local emissions such as bonfires. The position of major roads in the vicinity should be considered, as when monitoring a lead works close to a motorway.The Warren Spring Laboratory directional samplers will be described and the strategy involved in their use explained. Interpretation of Results After collecting the requisite number of samples from the ambient air it is necessary to interpret the results. If there is an air quality standard for the pollutant being considered then interpretation is relatively simple, assuming that the averaging time for sampling is compatible with that referred to in the AQS. For example, in an extreme case, it is worthless to compare a concentration obtained from a 1-h sample with an annual average. Air Quality Standards are based on epidemio- logical evidence and are intended as health protection stan- dards. It may be necessary to use other criteria in assessing the results, such as amenity effects, soiling and reduction in visibility.There are only three AQSs applicable to air pollutants in the UK and the averaging times involved are long term as opposed to the more short term practical pollution problems. Thus, reconciliation of short-term results with long-term AQSs is difficult. Some Directives give guide values for short-term measurements with which comparison may be attempted. In the absence of suitable AQSs comparisons must be made with other measurements made in similar situations elsewhere or, as a last resort, by comparison with the published Occupational Exposure Limits (OEL). These limits are for the normal 8-hour day, 40 hour working week for fit people. They do not take into account the perhaps continuous exposure of people living close to a source.Some workers have therefore tried to compensate for continuous exposure of the general populace, which includes the more vulnerable very young, the old, or the sick, by using fractions of the OEL such as 1/30,1/40 or 1/10. This is not to be recommended because the OEL was devised for particular circumstances and the list specifically states that the OEL should not be extrapolated for long term, non-occupational exposure. Simplex Optimisation K. Burton Department of Science, Polytechnic of Wales, Pontypridd, Mid-Glamorgan CF37 I DL The optimisation of chemical systems will be discussed in this paper, the emphasis being upon Simplex optimisation methods. The increasing availability of cheaper yet more powerful computers has led to an instrumental revolution in analytical chemistry, with the end result that many analyses can nowadays be carried out almost automatically on a routine basis.Even if full automation is not possible, the “determina- tion” step of the analysis can be considerably simplified by the use of modern instrumentation. However, in all these instances optimisation of instrumental performance can be very impor- tant. Similarly, the underlying “wet” chemistry of the analy- tical method prior to the determination step must not be neglected: rubbish in means rubbish out. A chemist may wish to develop a mathematical model to describe an experimental system; here again, optimisation can be used to improve the model so that it gives a better prediction of the experimental system.Up to a few years ago the application of optimisation procedures to analytical systems was uncommon and often rudimentary. This is in contrast to the industrial area, where control and optimisation of systems has been practised for many years, for instance, the optimisation of a process to give better product quality yield, etc. There are two major problems to be addressed: how to optimise several responses simultaneously; how to optimise a system where there are several controlling variables. With any system the actual response to be optimised has to be defined. Considering the analyte determination step of an analytical system, typical parameters might be: selectivity, sensitivity, accuracy, precision, analysis time, reliability (stab- ility). Operational Variables or Factors The response of a system will be defined by the conditions under which the system operates, in other words by the286 ANALYTICAL PROCEEDINGS.AUGUST 1989. VOL 26 controlling factors of the system. Typical examples of these are given below. Flameless AAS: drying, ashing, atomisation temperatures and times; heating rates; hollow cathode lamp current; slit width. GLC: column temperature; types of stationary phase; stationary phase loading; carrier gas flow- rate; support mesh size. HPLC: column temperature; type of stationary phase; varying composition of mobile phase with time; mobile phase flow-rate. Optimisation Methods The actual approach adopted depends upon the information that is available. There are basically three situations. (i) There is a reasonably full knowledge of the theoretical principles (physicochemical concepts) underlying the system to be optimised.Thus, mathematical models are available or can be derived, and mathematical and graphical techniques can be employed to optimise the particular system. Examples of such optimisation procedures include the “Window” method of Purnell as applied to gas - liquid chromatography, high- performance liquid chromatography, nuclear magnetic reson- ance and electrochemical methods, and the solvent selectivity triangle in chromatography.’ (ii) The information base is more limited, but data is available either from records over a period of time or generated from an experimental design. Thus, it is possible to generate a model of the response surface, which could take the form of a polynomial equation relating response to levels of the various controlling variables. This equation could then be optimised for maximum or minimum response by using mathematical search techniques.hl 0 (D U L c 3 Possible search dir Equilateral triangle .ections Factor 1 4 Possible search directions Tetrahedron / Factor 1 Fig. 1. Simplex figures for two- (top) and three-dimensional systems (iii) There is effectively no database available and optimisa- tion is carried out by reference to measured responses under different operating conditions of the system, whereby a move is made towards the optimum conditions by means of a “learn- ing” method approach. This is of more general applicability than (i) or (ii) above as it can be used with more complex systems that cannot easily be rigorously described in theor- etical terms, such as optimisation of the performance of an AAS or optimisation of systems where interactions between controlling factors are present.Attention is focused here on situations of type (iii), which is the so-called “learning” approach. This latter approach is of more general applicability and can lead to efficient optimisa- tion quite rapidly and with quite trivial calculations. Software based upon this “learning” approach can also be incorporated into instrumental systems. Initial Choice of Controlling Factors or Variables For any system, certain factors may be more significant than others in influencing the response. It is therefore desirable to run experiments beforehand using a factorial design so that the major controlling factors can be identified.It is better, if possible, initially to simplify the system to be optimised. The influence of secondary factors in the “optimised” system can always be examined later. The factorial design will also evaluate the effects of interactions between factors. Optimisation Methods of Type (iii) (a) One factor at a time This is the “classical” approach adopted by many chemists, where only one factor is changed at a time, the others being held at specified levels. The optimum response under these conditions is located, and then another factor is selected as a variable. Such an approach can require a lot of experiments and is therefore inefficient. I t does not take into account interactions between factors.(b) Method of steepest ascent The method is based upon developing a linear model which relates response to the controlling factors: where xl, x2, x3 . . . are factors, a,b,c . . . are constants which define the influence of the individual factors upon the response, and co is a constant relating to the response not explained by xI, x2, x3, etc. This model is then used to identify the path of steepest ascent up the response surface to the maximum (optimum). This path of steepest ascent is deter- mined by the constants a, 6, c, . . (slope terms), which are associated with the factors. The drawbacks of the method are the number of experiments required and the fact that the method may converge on a local optimum, rather than a global one. Also. of course, a linear model may not fit the response surface, in which case a higher order model, possibly with interactions, might have to be used.2 R = ax1 + 6x2 + C X ~ + . . . + c() . (c) Simplex methods The approach used in simplex optimisation is to vary all n factors simultaneously in order to search the n-dimensional response space. A simplex figure with n + 1 vertices (n + 1 experiments) is used, which allows a choice of movement in n + 1 directions. Fig. 1 shows simplex figures for two- and three-dimensional systems. The basis of the simplex methods is a sequential search, where the next move of the search is defined by the responses of the current simplex. There are several types of simplex method and these will be looked at in turn, starting with the original method developed by Spendley,3 which was initially applied to analytical chemistry by Long.4 Rules of the Basic Simplex 1.Identify the major factors (using factorial design?). 2. Decide upon the working range for each factor and identify any constraits to the system (factor levels which are not allowed or not attainable). 3. Decide upon the step size for each factor (the change in a factor which gives a reasonable change in response). 4. Define the initial experimental conditions, which might come from previous knowledge of the system. This is vertex I of the simplex. 5. Hence create the initial simplex. For the two-dimensional simplex the vertices of the simplex, which is an equilateral triangle, (Fig. 2) are given by:ANALYTICAL PROCEEDINGS, AUGUST 1989. VOL 26 287 Vertex 1 (O,O), Vertex 2 ( l , O ) , Vertex 3 (0.5,0.866) i.e., vertex n ( x , y ) where x refers to the step fraction for the first factor and y refers to the step fraction for the second factor.Thus, experiment 1 (vertex 1) is taken as the origin (0,O) and the other vertices related to that. This is shown below: 0.5, 0.866 3 0.0 1 to Fig. 2. An equilateral triangle representing the two-dimensional simplex Note that the side of the simplex does not have to be parallel to a factor axis. An example of the calculation of the experimental con- ditions for the initial simplex of a GLC optimisation is summarised below: Factors: Temperature ( x ) , Flow-rate (y) Range Initial step size Tempera t urePC 8O-140 20 Flow-rate units 5-50 20 Vertex Coords. Temp.F.-rate 1 0-0 80 20 2 1 ,o 100 20 3 0.5 .O. 866 90 37.5 For higher dimensional simplex figures the vertices can be calculated by using basic geometrical principles. A summary of vertex coordinates is given in the paper by Long.4 6. Run experiments 1, 2 and 3 and obtain the responses. 7. Carry out a reflection step by rejecting the vertex which has the worst response and reflecting through the centroid of the remaining vertices to obtain a new vertex (experiment 4). See Fig. 3 below. cv 0 m u. L 4- 5 R - 2- Reflection ..!& ~ ~~ Lett Response then vertex ..!& (0.38): If 0.38 use next worst Factor 1 Fig. 3. Representation of a higher dimensional simplex For higher dimensional simplexes the coordinates of the new vertex may be calculated following the procedure of Long,4 or from geometrical principles.8. Run experiment 4. Compare the responses in the new simplex (triangle 234 in the two-dimensional example) and reject the worst vertex in order to reflect. If the new vertex gives the worst response (triangle 234) use the new simplex but reject the next worst vertex (in this case vertex 2 of triangle 234). 9. Near the optimum the simplex will start circling (in a two-dimensional system). In this instance the step size can be reduced and the contracted simplex used for a more detailed search. With higher dimensional simplexes this circling or “close-packing” may not occur. 10. If a simplex has been retained for n + 1 vertices it is advisable to re-run this experiment to check that it was not a false high result. If a result was false (re-check if necessary) then adopt the new response and continue.11. Optimisation criteria, i . e . , convergence criteria: (i), for the final simplex the variance of the n + 1 responses is less than a pre-set value; (ii), the variance of the n + 1 responses from the centroid response is less than a pre-set value. This could be violated, for instance, if the simplex were too large, and might therefore indicate that a contraction is required. Note that after a move on the response surface during this optimisation procedure only one new experiment has to be run. Thus, effective use is made of the previous experiments. Disadvantages of the Basic Simplex Method (i) It is not always obvious when the optimum has been reached; tetrahedra and higher dimensional simplexes (n > 2) will not close-pack.(ii) There is no provision for acceleration of the search. (iii) It is possible to attain a false optimum. This can depend upon the initial starting position of the simplex and its orientation. It is advisable to try several starting positions. Thus, there have appeared modifications to the original simplex method which attempt to improve on its performance. The Modified Simplex Method This modification, which was developed by Nelder and Mead,s is more flexible than the original simplex method as it allows expansion or contraction and therefore better adaptation to the response surface. Fig. 4 summarises possible moves of the simplex. N 0 m LL L c 1 1 2 3 Expansion factor Y * Factor 1 Key: W, N, B: worst, next worst and best vertex; R: reflected vertex through the centroid M; E: expanded vertex: CR, CW: contracted vertices. Note the expansion factor axis (Y) which defines the vertices W.M. R and E. Fig. 4. Summary of possible curves for the modified simplex methods The possible options, depending upon relative responses at the different vertices, are to use the simplexes BNE, BNR, BCRN, BCWN, BNW. The rules of the modified simplex are conveniently sum-288 ANALYTICAL PROCEEDINGS, AUGUST 1989. VOL 26 marised in the flow diagram from the paper by Aberg and Gustavsson .6 The advantages of the modified simplex are that it follows the response surface more readily, adapts to changes in gradient, is more efficient and more likely to converge. Obviously, with dimensions higher than two, graphical methods of defining vertex coordinates cannot be used.The coordinates can readily be calculated by using the following equations: Centroid M vM,j = [ n i l v j , j - v , , , j ] / n 0‘ = 1 . . . n) i = 1 j variables (factors); i vertices. Vertices R and E V, = Y + VM,j + (1 - Y)Vw,, 0’= 1 . . . n ) V,,,: “worst vertex” For VR,,, Y = 2 For Y = 3. The articles by Aberg and Gustavsson6 and by Deming7 both give clear descriptions of the modified simplex. Further developments of simplex methods have concen- trated upon. ( i ) , Search direction, where it is possible to match more closely the local gradient of the response surface by modifying the position of the reflection point (which was originally the centroid of the remaining vertices) by weighting the coordi- nates in relation to the responses of the remaining vertices.This is known as the weighted centroid method (WCM).8 ( i i ) , Expansion, where by using the modified simplex a fixed expansion is possible ( Y = 3). An alternative is to use a variable value for Y, depending upon the gradient of the response surface. This has led to the development of the super modified simplex (SMS) .9710 Convex 1 I YOPT a C 0 a n a / I I / 1 I I Vertex W 0 1 2 M Y Concave I r 0 1 2 Convex I / - 0 1 2 Fig. 5. Typical response contours Super Modified Simplex In this method an equation is set up which relates the response (R) to the expansion factor (Y). A second order polynomial equation may be used: R = A + B * Y + C * Y Z where A , B and C are constants. Below are summarised the responses for the different Y values: Y: 0 1 2 R: R(W) R(M) R(R) corresponding to the worst, centroid and reflected vertices.On the basis of these values A, B and C can be calculated. Differentiation of this equation gives dRld Y , which equals zero for a maximum or minimum value for R. Hence, Yo, can be calculated, i.e., the expansion vertex E. Typical response contours are given in Fig. 5. In the convex case the Yo,, value may be calculated. By using this method the coordinates of the expansion vertex E may also be calculated and this experiment then run. This new simplex figure can then be used as the basis for the next move, which may end up as a contraction if the optimum is, in fact, fairly close. In the concave or convex cases, where no optimum is in fact present, a fixed expansion may take place, followed by the generation of a new second-order polynomial to describe the newly spanned response surface.Recommended operating limits for Yare - 1 d Y d 3 (Van der Wielll), in order to prevent excessive movement of the simplex. There are also various modifications to super modi- fied simplex.10-11 Applications of Simplex Methods12 Some examples of applications in analytical chemistry are as follows. Emission spectrometric analysis The factors involved in this technique are arc current, slit widthheight and arc optical path. The response is the SINratio of the emission intensity of 10 pg (3171-3 Ca2+ solution.13 H PL C: Phospholipid separation The factors here are the volumes of methanol and ammonia in chloroform - methanol - ammonia mobile phase.The response is by peak resolution and shape.14 Finally, Berridgels has discussed an automated, multi- parameter optimisation of HPLC separations. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. References Berridge, J. C . , “Techniques for the Automated Optimization of HPLC Separations,” Wiley, Chichester, 1985. Sharaf, M. A., Illman, D. L., and Kowalski, B. R., “Chemometrics,” Wiley, Chichester , 1986. Spendley, W., Hext, G. R., and Himsworth, F. R.. Techno- metrics, 1962, 4, 441. Long, D. E., Anal. Chim. Acta, 1969,461. 193. Nelder, J. A., and Mead, R., Comput. J.. 1965, 7, 308. Aberg, E. R., and Gustavsson, A. G. T., Anal. Chim. Acta, 1982, 144, 39. Deming, S. N., and Parker, L. R., Crit. Rev. Anal. Chem., 1978 (Sept.), 187.Ryan, P. B., Barr, R. L., and Todd, H. D.. Anal. Chem., 1980, 52, 1460. Routh, M. W.. Swartz. P. A., and Denton, M. B.. Anal. Chem., 1977,49, 1422. Van der Wiel. P. F. A., Anal. Chirn. Acta, 1980, 122.421. Van der Wiel, P. F. A., Maassen, R., and Kateman. G.. Anal. Chim. Acta, 1982, 153, 83. Deming. S. N.. and Morgan, S. L., Anal. Chim. Acta, 1983, 150, 183. Rippetoe, W. E. R., Johnson, E. R.. and Vickers, T. J., Anal. Chem., 1975.47, 437. Rainey, L., and Purdy, W. C., Anal. Chim. Acta, 1977, 93. 211. Berridge, J. C., Analyst, 1984, 109, 291.ANALYTICAL PROCEEDINGS. AUGUST 1989, VOL 26 289 New Strategies for Determining Trace Metals in Industrial Materials and Effluents A. M. Bond Division of Chemical and Physical Sciences, Deakin University, Geelong 321 7, Victoria, Australia The determination of metals in industrial materials and effluents at the trace level presents a severe challenge to the analytical chemist.Not only must the metal of interest be identified near to the detection limit of many techniques, but a knowledge as to whether total metal or a particular metal species, in a given oxidation state, etc., are being detected, may be required. Inherently present at the same time are notorious problems of sampling and contamination and the question as to whether the analytical procedure used modifies the sample of interest always needs to be addressed. All of the above considerations are highlighted in developing strategies for determining trace metals in effluents and industrial materials.Recent developments include battery operated detection systems, which amongst other features allow the instrument to be taken to the sample, rather than the more usual method of bringing the sample to the laboratory. Additionally, there is the combination of automated chromatographic methods with spectrophotometric or electrochemical detection of metals and automated procedures enabling the continual monitoring of industrial effluents and plant operation. The computer has a role to play in the above procedures, but there has to be an awareness of the common sources of error related to the various relevant examples under consideration. Battery-operated Instrumentation Many factors are stimulating the development of sophisticated battery operated analytical instrumentation.For example: continuous monitoring of industrial effluent and chemical processes in industrial plants must take place, even in the event of power failure; battery operated devices are inherently low noise instruments and may have improved detection limits compared with mains powered devices; monitoring of chemical species in the environment is frequently required in remote areas where mains-powered instrumentation cannot be used. Initial developments in battery-operated field instruments were predominantly based on very simple analogue electronic devices. For example, hand-held portable pH instruments are well known to environmental chemists. In standard laboratory instruments, major advances in performance have taken place which are commonly attributable to the widespread availability of mains powered digital microprocessors.Field based instru- ments have yet to receive the full impact of access to computer operated components. Present technologies in low powered electronic devices have expanded so rapidly that the sophistica- tion of battery operated instruments can now approach that of the mains operated instrumentation used in conventional analytical laboratories. That is, the future generation of battery operated field based analytical instruments will be “intelligent” analytical systems with the capability of revolutionising an important and rapidly expanding field of analytical chemistry. There are interesting developments in battery operated, computer based electroanalytical instruments for detecting and monitoring metals in effluents and the environment and they represent the considerable advances taking place in this field of chemical instrumentation.Electroanalytical techniques are ideally suited to the applications of battery-operated monitor- ing instruments. Voltages and currents are easy to generate and easy to measure. Instrumentation is relatively cheap and electrochemical techniques are uncomplicated and frequently complementary. In addition, electrochemical instrumentation can be compact, miniaturised and operated with very low energy requirements. Automated Metal Detection with HPLC High-performance liquid chromatography, HPLC, with elec- trochemical, spectrophotometric or other forms of detection has become well accepted as a sensitive method for determin- ing a wide variety of organic and inorganic compounds.In these laboratories, interest has been focused on develop- ing analytical methods for trace metal determinations of metals such as copper, nickel, chromium( HI), chromium(VI), manganese, iron, cobalt, mercury and lead by the HPLC method in industrial effluents. The design principles of a new, completely automated HPLC system for on line monitoring offers: (i) 7 day continuous operation without maintenance; (ii) complete automation of all procedures including sampling from the industrial effluent. This has required the development of a newly designed, low pressure mixing system, a string bead reactor coil to ensure complete mixing, a bubble chamber to remove bubbles formed during the mixing procedure and automatic injection of metal complexes; (iii) the ex situ method of metal complex formation has been implemented in an automated version.Previously, in situ complex formation was used, which is inherently simpler from an instrumental point of view, but does not necessarily provide optimal sensitivity; (iv) a heating unit is provided as an option to increase the rate of formation of metal complexes when reaction rates are slow at ambient temperatures; (v) the flow-rate has been decreased to well below 1 ml min-1 in order to provide substantial savings of expensive and toxic organic solvent. At the same time, microbore chromatography has been introduced so that higher resolution is retained. Bibliography Further details and information on the above subjects are available in recently published papers.Bond, A. M., Hudson, H. A., Tan, S. N., and Walter, F. L., Trends Anal. Chem., 1988, 7 , 159. Bond, A. M., Garrard. W. N. C., Heritage, I. D., Majewski, T. P., Wallace, G. G., McBurney, M. J. P., Crosher, E. T., and McLachlan, L. S . , Anal. Chem., 1988. 60, 1357.290 ANALYTICAL PROCEEDINGS. AUGUST 1989. VOL 26 nplex Modified Simplex Optimisation for Operating an Enzyme Electrode S. K. Beh, G. J. Moody and J. D. R. Thomas School of Chemistry and Applied Chemistry, University of Wales College of Cardiff, P.O. Box 912, Cardiff CF13TB In many chemical activities the output or response of a system is often required as a function of experimental conditions. Optimisation is employed to attain the best response from a system for certain desired criteria; for example, in a flow injection analysis (FIA) system the response time and response peak height, or a combination of both.A common approach used by chemists to optimise, is the “one factor at a time” method of optimisation. An obstacle to optimisation of real systems is their multi-dimensional nature, and in most instances many experimental parameters are to be optimised and response variables monitored simultaneously. A common, but better, method is the simplex approach,’ the basis of which is a sequential search depending on the responses obtained. For example, when all factors (n) are varied simultaneously in the n-dimensional response space an initial simplex of n + 1 vertices is used; this allows for a choice of n + 1 search directions.In a two-variable system the possible search direction is three with an initial simplex of three vertices. However, the search step in the simplex approach is fixed, making it less efficient and difficult to converge. Another problem is the possibility of a false or localised optimum. In order to overcome these drawbacks a modified approach by Nelder and Mead2 allows the simplex search to expand or contract, thus enabling better adaptation to the response surface. Fig. 1 summarises the possible moves of the simplex and a description of the algorithm. Table 1. Initial data for optimisation Parameters- Labelled as Limits Flow Volume 10 d Flow Q 90% 0.3 S Vol. < 1.8 cm3 Variables- Labelled as Peak height Response time Objective Maxim i s e Minimise Estimated limits 0 s peak < 120A.1 0 s time G 200s Relative weights* S1 0.9 0.1 s2 0.5 0.5 s3 0.1 0.9 * Criteria for convergence <O.1. The initial steps are as follows. Firstly, identify all major experimental parameters and response variables; secondly, decide on constraints of parameters and response variables; R3 > R4 > R 2 R3 > R2 > R1 reflect ion MI ( 3 3 Initial simplex R1 > R4 E,R5 If R5 > R4 ____) If R4 > R5 - Substitute R1 for R5 Substitute R1 for R4 \ New s If R5 > R4 _____) If R4 > R5 _____) If R5 > R4 If R4 > R5 _____) Get new initial simplex 1 Check for optimum * Fig. 1. Modified simplex algorithmANALYTICAL PROCEEDINGS, AUGUST 1989. VOL 26 Buffer 29 1 *Waste lniection thirdly, determine the relative importance of the response variables (weighting); and fourthly, establish the criteria for convergence.80 60 &? - 2 40 LL 20 Parameters Variables Flow-rate Sample volume Peak height n Resoonse time - - - - U valve Enzyme electrode Peristaltic Pump Fig. 2. FIA system to be optimised The procedure is illustrated by the optimisation of a flow injection analysis system (Fig. 2) for monitoring glucose using an enzyme electrode. The responses such as peak height and 100 22 (b) 2o I 0 10 20 Iteration 100 n 0 1 Volume/cm3 2 Fig. 3. (a). Total response function (TRF) with respect to flow and volume for S1; (6). total response function (TRF) with respect to iteration for Sl; (c). simplex optimisation of flow and volume for S , response time are important because a large peak height offers better resolution of the system and quick response times enable greater analysis turnover.The parameters that affect these variables are sample volume and flow-rate. In order to illustrate the ability of the modified simplex algorithm to design experiments where certain criteria have to be met. the initial conditions and relative importance of each variable are defined in Table 1. The initial simplex chosen for each of the three experimental criteria (S, to S3) is the same, that is, high flow-rate and small sample volume. Three initial set of points were taken (Table 2). Table 2. Initial simplex Flow. Yo Volume/cm3 1 90 0.3 2 85 0.3 3 85 0.4 Results In order to obtain a better view of the modified simplex approach three-dimensional plots are shown (Figs.3-5) in which the planar axes consist of flow and volume and the vertical axis is the total response function, where the two response variables are normalised. At the optimum, based on the pre-determined criteria, the standard deviation of the response function of each of the vertices must be less than for the criteria for convergence (Table 3). This is usually dependent on the resolution of the ex pen men ts. ~~~~~~ ~ Table 3. Final simplex for S1, S2 and S3. TRF is total response function S1 s2 s3 Flow/ Volume/ Flow/ Volume/ Flow/ Volume/ YO cm-7 YO cm3 Y O cm3 Vertex 1 18 1.5 32 1.2 82 0.45 Vertex 2 26 1.45 42 1.125 85 0.4 Vertex 3 18 1.65 26.5 1.425 80 0.4 Peak/ Time/ Peak/ Time/ PeaW Time/ nA s TRF nA s TRF nA s TRF Vertex 1 116.8 57.33 19.94 110.4 34.16 53.81 37 10.18 89.45 Vertex2 116.8 47.52 20.04 100 23.71 53.81 34 8.84 89.45 Vertex3 118.0 58.00 20.04 120 40.76 53.96 37 9.0 89.56 S.D. 5.77 x lo-’ 8.62 x 10-2 5.77 x 10-2 Discussion From Table 4, experiments can be designed by specifying the relative importance (weighting) of the individual response function, in this case peak height and response time. As is illustrated (Table 4) S , has the largest peak height but the longest response time and S3 has the shortest response time but the smallest peak height. S2 is a compromise of S1 and S3. However, peak height affects the optimisation more than the response time, even though the response is in between S , and Sz, but the bias is towards peak height. ~ ~ ~ ~ ~ ~~ ~ ~~~~~~~~~~ Table 4. Results obtained from the optimisation. Wt = weighting System S1 s2 s3 Peak Wt. 0.9 0.5 0.1 Time Wt. 0.1 0.5 0.9 Flow. Yo 20.7 k 15.9 33.5 k 27.0 82.3 2 8.6 Volume/cm3 1.5 k 0.35 1.25 k 0.5 0.42 k 0.10 Peak (A/)/nA 117.2 k 2.4 110.1 2 34 36 ? 5.96 Time/s 54 k 20 32.8 k 29.5 9.34 2 2.5292 89.5 LL c I- ANALYTICAL PROCEEDINGS, AUGUST 1989, VOL 26 - 54 53 LL 52 51 50 0 89.4 0 0 U n o - 0 0 0 0 0 80 60 8 - $ LL 40- 0 - - 0 0 0 10 Iteration 100 I 1 0' 1 J 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Vo I u rneIcrn3 Fig. 4. ( a ) . Total response function (TRF) with respect to flow and volume for S2; ( b ) , total response function (TRF) with respect to iteration for S2; (c), simplex optimisation of flow and volume for SZ 0 0 0 89.3 I I 0 2 4 6 8 Iteration 70 ' I 0.25 0.35 0.45 0.55 Volu me/crn3 Fig. 5. ( a ) , Total response function (TRF) with respect to flow and volume for S 3 ; ( b ) , total response function (TRF) with respect to iteration for S3; ( c ) , simplex optimisation of flow and volume for S3 Conclusion The modified simplex approach can be used as an efficient method of obtaining the optimum for any given experimental criteria. Also, it has the capability of coping with multi- response functions and the algorithm can be computerised. The authors thank the Trustees of the Analytical Chemistry Trust Fund of the Royal Society of Chemistry for the award of an SAC Research Studentship (to S. K. Beh). Professor A. A. S. C. Machado and Mr. L. M. Ribeiro of the Universidade do Porto, Portugal, are also thanked for helpful discussions made possible by NATO grant No. oo69,84. References 1. 2. Spendley. W.. Hext, G. R.. and Himsworth, F. R., Technome- trics, 1962, 4. 441. Nelder, J. A., and Mead. R., Compur. J.. 1965, 7. 308.

ISSN:0144-557X    

DOI:10.1039/AP9892600281

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, AUGUST 1089. VOL 26 293 First Prepare Your Sample The following are summaries of two of the papers presented a t a Meeting of the East Anglia Region held on November 15th, 1988, at Unilever Research, Sharnbrook, Bedfordshire. Sample Preparation Procedures Used in Clinical Chemistry Andrew Taylor Department of Clinical Biochemistry and Clinical Nutrition, St. Luke’s Hospital, Guildford, Surrey GU 1 3NT Clinical chemistry grew from the early work of physiologists, who applied chemical procedures to the study of normal body function. Analytical methods, developed 60 or more years ago for the measurement of metabolites such as glucose, urea, etc., were later applied to the wider investigation and care of sick subjects. Analytical chemists such as Van Slyke, Fohn and Henry, who were responsible for this pioneering activity, not only developed many new chemical tests but also designed and built ingenious pieces of equipment that proved to be invaluable for the growth of clinical chemistry. Clinical chemistry is the study of the relationship between the chemistry of body fluids and tissues, and disease.Pursuit of this objective is particularly concerned with analysis, method development and advice. Analysis involves the determination of a wide range of parameters to diagnose disease, to monitor clinical progress and to help understand the development of disease or the response to treatment. New and improved analytical and investigative procedures increase the contribu- tion made by the laboratory, as does the advice given to doctors as to the most suitable tests to be carried out and to assist with the interpretation of results.This is especially important with specialist investigations, such as those for endocrinological disorders. In practice, much of clinical chemistry involves the measurement of analytes in appropriate specimens using a variety of different analytical techniques. Sample preparation is fundamental to this task. The repertoire of tests carried out by a busy department can extend into several hundreds and this number is continually being enlarged. There is, however, a group of tests (referred to as the “top twenty” or by some other trivial descriptive term) which are especially important for the following reasons: they are required in large numbers every day, need to be measured in every District Hospital laboratory, provide information on the status of vital organs in the body and may be needed very quickly to permit decisions relating to patient management to be made.Other tests are performed less frequently and with a lower degree of urgency. Most measurements are carried out on specimens of plasma or serum. A small number of tests are performed on urine and a very few with other types of samples. In order to obtain serum and plasma, these fluids must be separated from the blood cells, a step that is cumbersome, time consuming and cannot be satisfactorily accomplished by automated equipment. However, with liquid specimens there are no problems relating to lack of homogeneity within the sample and it is simple to dispense, transfer and mix them with reagents.Other features associated with these specimens, and that are relevant to the analysis, are the complex matrix (proteins, lipids, sugars, minerals, etc.) and the very small changes in concentration of many analytes that need to be accurately measured. Thus, it is essential that methods are very reproducible. Although many analytical techniques have a place in clinical chemistry, colorimetric procedures have always been the most important and account for the large majority of tests carried out each day. The methods introduced by Van Slyke and others had, in general, the following features. They required 0.5-1.0 ml of sample per test (this volume could be a limitation if several tests were requested or investigations were on young children), they had a slow throughput (each batch of tests took many minutes or hours to complete), they required great technical skill to achieve good results, they involved a colorimetric measurement and protein precipitation was necessary in order to prevent flocculation and interference in the colour reaction.Consequently, investigations were restricted to a minority of patients until the development of the flame photometer. It then became possible to measure the concentrations of sodium and potassium in dozens of samples within a few minutes and it was recognised that an expansion of the numbers of “top twenty” tests would be desirable. However, protein precipitation proved to be extremely difficult to automate and the solutions to this challenge represent the real developments in sample preparation in clinical chemistry during the last 20 years. Table 1 shows the automated approaches to sample prepara- tion that have been developed.Continuous flow analysis was the first successful technique and employed dialysis to remove the protein. Although expensive to purchase and run the AutoAnalyzer has been very widely used. It is simple to operate and has led to staff changes with less technically skilled personnel able to achieve very good results. As an alternative to removal there have been attempts to develop chemical methods that tolerate the high concentrations of protein. Table 1. Automated approaches to sample preparation Continuous flow analysers ( AutoAnalyzer) Discrete analysers syringdpump technology centrifugal analysers Solid phase chemistry analysers Discrete analysers have been designed that operate with very efficient and precise syringe and pump technology to provide for rapid dispensing of specimens and reagents, or with centrifugal forces to transfer and mix fluids.The instruments have improved photometry to allow large sample dilutions, or are used either with very sensitive colour reactions (for the same purpose) or with solubilising agents that prevent protein flocculation. A completely different analytical technique is that of dry film chemistry. Based on technology developed from the photographic chemistry these methods involve reagents incor- porated into gels mounted on an inert strip. Samples applied to the surface diffuse through the gel, mix with the reagents and the colour can be measured with a specially designed col- orimeter.Discrete analysers and solid-phase instruments are able to analyse a large number of specimens very quickly, with several different tests carried out simultaneously on each sample. Generally, their use is restricted to a limited range of “top294 ANALYTICAL PROCEEDTNGS, AUGUST 1989. VOL 26 twenty” tests and the chemical methods are those developed by the instrument manufacturers or by reagent kit producers. The instruments need a computer both to control the operation of the syringes, valves, dispensers, etc., and to process the large number of readings that are made (300 per hour is typical) and have to be reported. Thus, the persons responsible for much of the laboratory work have little influence on the methodologies and no longer need the skills and expertise associated with an analytical chemist.For clinical chemistry, therefore, sample preparation involves two quite separate attitudes. With the “specialist” analyses, drugs, hormones, trace elements, etc., there are few differences from the methods employed by other workers having similar interests. However, for a limited range of work, preparation cannot be separated from the measurement and the presentation of a result. In this context, high precision engineering and computer control of instrumentation have been applied to the methods of the analytical chemists of up to 60 years ago to give large numbers of results very quickly, from small volumes of specimen, with good reproducibility, with a small number of relatively junior staff, at a low cost per test and with high levels of microbiological safety.You just have to load up the instrument: sample preparation takes care of itself! Automated Sample Handling for Drug Analysis from Biological Fluids Using HPLC E. Doyle*, R. D. McDowall, G. S. Murkitt, V. S. Picot and S. J. Rogers Department of Drug Analysis, Smith Kline and French Research Limited, The Frythe, Welwyn, Hertfordshire AL6 9AR The analysis of drugs from biological fluids involves their separation from a biological matrix ( e . g . , blood, plasma, urine or bile). Solid phase extraction columns are particularly well suited to this purpose.’ Small columns packed with chemically bonded silica derived from HPLC column technology provide a convenient means of extracting the analyte and introducing the total extract on to the HPLC analytical column. Solid-phase extraction is simple, and sensitivity is attained because there is no sample loss through transfer.In fact, the sample handling and HPLC separation stages of an assay are now so similar in their operation that they can be built into a single analytical system, providing the analyst with a fully automated assay. When sample preparation and HPLC are performed together, the limitation on the number of samples that can be assayed in a day shifts from human constraints placed on it by a technician, to the reliability of the equipment and the HPLC run-time. Today, advances in sample handling methods and the now routine use of fully automated assays have achieved this shift.Two systems that provide full automation to sample handling are the Advanced Automatic Sample Processor (AASP, Varian Associates, Walton-on-Thames, Surrey) in combina- tion with a Gilson 222/401 autosampler with dilutor (Anachem Ltd., Luton, Bedfordshire),z and HPLC column switching, which was first described by Roth et af.3 Both of these systems are dedicated sample preparation units developed specifically for routine assays; they are relatively inexpensive and highly productive, and can also be adapted easily from method to method. In the authors’ laboratory the Gilson - AASP system has been used for the analysis of M and B 22948, a cGMP specific phosphodiesterase inhibitor,4 and SKF G6022, an H+/K+ ATPase inhibitor.5 Automated Sample Extraction Using the GilsodAASP The Gilson 222/401 provides a means of automatically adjust- ing the pH while adding internal standard solution to plasma in the Gilson sample rack.The Gilson was connected to the purge port of the AASP switching valve (port 5, Fig. 1) so that the solid phase was activated on-line, sample added and washed, and then automatically eluted on to the analytical column. The Gilson 401 dilutor, in conjunction with the 222, was used to draw the following liquids into a holding loop in reverse order; methanol (1 ml) and water (1 ml) required to activate the solid * To whom correspondence should be addressed. phase, plasma and internal standard mixture and a buffer solution used to wash the cartridge. The liquids were each separated by a segment of air (Fig.1). The liquid “train” was then passed via the injection port of the Gilson 222, through the AASP cartridge in situ in the AASP. After sample extraction, the mobile phase was switched through the cartridge, eluting the drug, internal standard and any other retained compounds directly on to the analytical column. \ - Loading Elution Injection4 1 I I phase AASP 1 AASP Plasma IS. Buffer HzO MeOH Gilson Fig. 1. Schematic diagram of Gilson - AASP automated system After the elution of each sample from the AASP cartridge, the AASP was programmed to re-set to the load position at 0.6 min. Direct Injection, Column Switching The sample was injected on to the pre-column in a flow of extraction eluent. After 2 min the flow of solvent was switchedANALYTlCAL PROCEEDINGS.AUGUST 1989, VOL 26 295 using a high pressure switching valve so that the gradient passed through the pre-column and the analytical column (Fig. 2). The pre-column was dry packed and meshes rather than frits were used in order to avoid blockage. The column was “primed” two or three times with control sample before use. The “priming” may serve to de-activate the silica backbone of the column and avoid a dual mechanism of retention in which the analyte interacts with the residual silanols as well as being retained by interaction with bonded alkyl chains on the stationary phase surface.“7 With this system, the pre-column was used for approximately 150 injections of plasma (100 pl). The pre-column was flushed separately from the analytical column with a suitable solvent during the equilibrium time between assays because a gradient was used.Column Switching Flow Diagram Rheodyne 7010 I I LJ A A Detector Fig. 2. Schematic diagram of direct injection, column switching Discussions The only sample preparation required for most biological fluids prior to introducing the sample to a Gilson-AASP or direct injection system is centrifugation. However, dilution is per- formed off-line when assaying samples that have high drug concentrations, for example, following intravenous adminis- tration. It may also be necessary to simplify the matrix, for example, by removing bile salts with ammonium sulphate,8 or proteins by precipitation with organic solvent.9 Further manip- ulation, such as the addition of internal standard, an organic modifier to free drug bound to proteins, or buffer to adjust the pH, can be performed automatically.9 However, when process- ing the sample automatically it is important not to allow any precipitation to occur, because this will block the extraction column.The Extraction Column The sorbents used for extraction columns have been derived from HPLC technology and there are many different phases available. The particle size is relatively large (2540 pm) in comparison with that used for an analytical column. This is to allow large amounts of biological fluid to pass through without blocking the column. The average length of a column for on-line extraction is 10 mm and it contains 50-100 mg of sorbent. By using this large pore size material and a 10-mm column, the sample has to pass slowly through the column to achieve good recovery of the analyte.A flow-rate of approxi- mately 1 ml min-1 is sufficient for most compounds. Of the wide range of sorbents available, the reversed phase C2, c8 and cl8 are most popular.’ However, for some applications a CN phase is preferred,lO and very selective extraction of diols has been achieved by using a chemically bonded borate phase.11.*2 The authors suggest that if the analyte is lipophilic and has no potential to interact with selective phases a C2 phase should be tried first. The advantage of this approach is that a C2 phase will often retain the analyte but the background from endogenous compounds in plasma can be more easily removed than with the more strongly retaining c8 and c18 phases.The analyte can also be eluted on to an analytical column more easily from a C2 phase than from higher phases, which facilitates coupling of the extraction and analytical columns. This assumes that there is no interaction between the analyte and the residual silanol groups. Conclusion There are sufficient examples of fully automated assays to show that automation is a reality. Routine systems, such as the Gilson-AASP and direct injection, have slight differences, but both are highly productive and leave the analyst free to develop the science rather than be the rate limiting factor in sample throughput. They may seem complicated, but in practice the time taken to learn to use them is only a few days. All of the equipment is commercially available and is now developed to a point that it is reliable. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. References Huber, R.. and Zech, K., in Frei, R. W., and Zech, K., Edirors. “Selective Sample Handling and Detection in High- performance Liquid Chromatography,” Chapter 2, Elsevier, Amsterdam, 1988. Murkitt, G. S., Peake, J. C., and McDowall, R. D., Chromatographia, 1987, 2 4 4 1 1. Roth, W., Beschke, K., Jauch, R., Zimmer, R., and Koss, F. W., J. Chromatogr., 1981, 222, 13. Rudd, R. M., Gellert, A. R., Studdy, P. R., and Geddes. D. M., Br. J . Dis. Chesr, 1983, 77, 78. Huber, R., Mueller, W., Banks, M. C., Rogers, S. J., Norwood, P. C., and Doyle, E.. J. Chromarogr., in the press. Doyle, E., Pearce, J. C., Picot, V. S., and Lee, R. M., J. Chromarogr., 1987, 411. 325. Nahum, A.. and Horvath, Cs., J . Chromarogr., 1981.203.53. Picot, V. S., Doyle, E., Read, L. J., and Lee, R. M., Chromatographia, 1987, 24,282. Hux, R. A., Mahammed, H. Y., and Cantwell, F. F., Anal. Chem., 1982,54, 113. Werkhoven-Goewie, C. E., de Ruiter, C., Brinkma’n, U. A. Th.. Frei, R. W., de Jong, G. J., Little, C. J., and Stahel, O . , J. Chromatogr.. 1983, 255, 79. Hagemeier, E., Kemper, K., Boos, K.-S., and Schlimme, E., J. Chromatogr., 1983,282, 663. Schlimme, E., Boos, K.-S., Hagemeier, E., Kernper, K., Meyer, U., Hobler, H., Schnelle, T., and Weise, M., J. Chromatogr., 1986. 378, 349.

ISSN:0144-557X    

DOI:10.1039/AP9892600293

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

296 ANALYTICAL PROCEEDINGS, AUGUST 1989, VOL 26 Equipment News Automated Injection for HPLC The Shimadzu SIL-9A is a stand-alone auto-injector with a reagent dispensing function. It features a sample rack which accepts six types of sample vial and a sampling needle which moves in X , Y and Z directions, permitting random access to different vials. This allows various types of pre-treatment, including sample dilu- tion, addition of internal standard and pre-column derivatisation, before sam- ples are automatically injected on to the column. The system is computer con- trolled and program files are provided for various types of sample pre-treatment. The injection volume is highly reprodu- cible (coefficient of variation less than 0.5% for a 10-pl sample) and cross con- tamination is effectively eliminated through rinsing of the sampling needle after each run.The SIL-9A may be connected to an HPLC of any brand. Dyson Instruments Ltd., Hetton Lyons Industrial Estate, Hetton, Houghton-le- Spring, Tyne and Wear DH5 ORH. Automated Injection for HPLC The Model 728 is a microcomputer con- trolled autosampler which is compatible with any HPLC and any computer or integrator. Its design, based on a unique positive displacement technique, elimi- nates the need for accessory gases, pro- vides viscosity independent sample trans- fer and contains diagnostics which can detect problems in the sample transfer. The autosampler can perform up to three injections each on 64 different HPLC samples, and it can rinse the system, start and stop integrators and other external components, or be controlled itself by a laboratory computer. Severn Analytical Ltd., Unit 2B, St.Francis Way, Shefford Industrial Park, Shefford, Bedfordshire SG17 5DZ. HPLC System The SA 6000 isocratic HPLC system consists of a pump, an injector, a column, an ultraviolet - visible detector and an integrator. Easy to use, the system is ideal for routine work. It features a dual piston designed pump, and delivers an accurate and pulse-free solvent flow from 10 pl to 5 ml min-1 up to 490 bar operating pressure, as well as offering sensitivity down to 0.0005 AUFS from the variable wavelength ultraviolet detector. Data reduction is achieved by a compact in- tegrator featuring an integral, full-width printer - plotter, extensive HELP mes- sages and chromatogram storage and replay with base-line construction and naming of peaks.A Rheodyne 7125 injec- tor and a 5 pm x 25 cm column complete the system. Severn Analytical Ltd., Unit 2B, St. Francis Way, Shefford Industrial Park, Shefford, Bedfordshire SG17 5DZ. Finger-Tight Fittings for HPLC The Slip Free HPLC connector requires no tightening tools and is capable of withstanding pressures up to 700 bar. Manufactured by Keystone Scientific in the US, the Slip Free comes with various pre-cut lengths of 1/16 in outer diameter tubing. Single or double versions of the connector are available for column - detector or column - column connections. Shandon Scientific Ltd., Chadwick Road, Astmoor, Runcorn, Cheshire WA7 1PR. Extraction Column for Drug Testing The Bond Elut Certify solid phase extrac- tion column features a sorbent optimised for the extraction of acidic, neutral and basic drugs from urine prior to confirma- tional analysis. It enables laboratories to use single-column extraction for more than twenty major drugs of abuse.Bond Elut Certify columns feature an integral 10-ml reservoir and are available with laminated method sheets describing proven extraction procedures. Analytichem International, P.O. Box 234, Cambridge CB2 1PE. Isotope Ratio Detector The Isomass bench-top continuous flow isotope ratio detector is suitable for pre- cise, routine analysis of 13C or 15N in 15N with total N, down to natural abun- dance levels. This facility enables fully automated elemental and isotope ratio analyses to be made on a series of samples.Operation is simplified through the use of specially developed VG Isogas system software, which allows the system to be used for routine daily runs with minimum user intervention. VG Isogas Ltd., Aston Way, Middle- wich, Cheshire CWlO OHT. Refractive Index Detector The Philips PU4026 refractive index detector is now available for rental. It features a built-in Peltier effect thermo- regulator, providing a high degree of temperature stability and rapid warm-up times. Capable of being used in many applications, such as separating sugars in solution and analysing organic materials in river water, it is suitable, the makers claim, for specific property detection as well as for the measurement of the bulk properties of materials. Livingston Hire, Livingston House, 2-6 Queen’s Road, Teddington, Middlesex TWll OLB.Flue Gas Analysis System The Model 9150 two-in-one flue gas analyser monitors both oxygen and car- bon monoxide. It comes standard with three percentage oxygen ranges (0-5%, &lo% and 0-25%) and two carbon monoxide ranges (0-500 p.p.m. and VG Isomass isotope ratio detector organic or inorganic materials where a 0-1000 p.p.m.). The system is housed in large number of measurements are an enclosure rated NEMA 4, 12 and 13 required in the minimum time. Designed and IP 55. to operate with the world’s best elemental Teledyne Analytical Instruments, analysers and running on-line, the Iso- The Harlequin Centre, Southall Lane, mass can analyse for 13C with total C or Southall, Middlesex UB2 5NH.ANALYTICAL PROCEEDINGS, AUGUST 1989.VOL 26 297 Personal Gas Detectors A range of pocket sized gas monitors capable of giving audible and visual warn- ings in the event of dangerous gas levels are offered in two basic versions: the first, for toxic gases, is available for sensing such gases as CO, H2S and SO,; the second is specific to oxygen. Both operate from either dry cells or rechargeable NiCad Batteries to give 600 and 150 h operation, respectively. The bleeper sounds and an LED flashes every 15 s to indicate that the unit is functioning cor- equipped with the new electrode will produce results in less than 5 min and within a standard deviation of less than 1%. Over nine specific methods for differ- ent sample types have been developed and the list is being expanded.Orion Research UK, Freshfield House, Lewes Road, Forest Row, East Sussex RH18 5ES. Conductivity Meter The SC82 compact, lightweight, micro- Servomex personal gas detectors rectly. The units are approved by CENELEC as intrinsically safe for use in hazardous areas (approval code EEx ia IIC T6). Servomex Ltd., Crowborough, Sussex TN6 3DU. Eftluent Monitor Alarm System A fail-safe alarm unit has been introduced to draw attention to any failure in supplies to the makers' 6800 Series of TClT'OC process analysis monitors. The new alarm unit checks sample flows, carrier gas pressures, air flows, furnace temperatures and, for TOC monitors, acid inputs. Ionics (Instruments Division) Inc., 10 Statham Avenue, Lymm, Cheshire WA13 9NH. Surfactant Electrode A surfactant electrode (Model 9342) is designed for use with automatic titrators such as the makers' 960 autochemistry system.It provides both anionic and cationic surfactant analysis. Virtually no sample preparation is required and the use of toxic chemicals like chloroform is eliminated. An automatic titrator processor based, hand-held meter features auto-ranging, automatic temper- ature compensation and auto power-off. All settings are retained in the power-off mode, The large LCD display shows conductivity, temperature and coeffi- cient. Special conductivity cells for pure water and high conductivity measure- ments are available. Yokogawa Electrofact Ltd., Unit 6, Ashville Close, Gloucester GL2 6HY. Thermal Analysis Instrumentation New product developments include a low-temperature DSC with integral cool- ing and the TMA 1500, which provides accurate measurement of dimensional changes as a function of temperature and time under tension, load or zero-load operation.An encapsulation - crimping tool prevents the loss of volatile com- ponents in a material when it is heated during differential scanning calorimetry, for example, which allows mixed solid, liquid and vapour phases to be measured. Tests have demonstrated that volatiles can typically be contained up to pressures of 10 bar. All the standard thermal analysis techniques are supported by the TRACE 3 data acquisition system, IBM compatible software, which operates on either PS series or AT compatible machines. The QTM-D3 quick thermal conductivity meter provides direct spot measurement of thermal conductivity of a range of materials, including brick, glass, wood, foods and plastics as well as pow- dered, viscous and filmy substances.Stanton Redcroft Ltd., Copper Mill Lane, London SW17 OBN. Cryostat The CC4060 offers a heat removal capac- ity of 4000 W at - 10 "C, a pressure pump with a maximum pressure of 3 bar and a maximum flow-rate of 30 1 min-1, a bath opening with a depth of 300 mm, an analog interface for various functions, such as connection for external control sensor (Pt100), external measuring sensor (PtlW), remote alarm, etc., and a digital interface RS232C upon request. Haake Mess-Technik GmbH u. Co., Dieselstrasse 4, 7500 Karlsruhe 41, West Germany. Viscometer The VT500 viscometer for quality control is equipped with intelligent micro-elec- tronics and performs all calculations.Rheological data are digitally displayed and results can be transferred to record- ers, computers or printers. The VT500 gives high accuracy and a viscosity range of 1-107 mPa s. It features cylinder sensors according to DIN 53019 for medium viscosity substances, double-gap and cone and plate sensors according to DIN standards for low and high viscosity substances, and spindles corresponding to ASTM, I S 0 and other standards. Haake Mess-Technik GmbH u. Co., Dieselstrasse 4, 7500 Karlsruhe 41, West Germany. Centrifuges The Cryofuge 8500 6-1 centrifuge can generate a g-force of 8360 with its maxi- mum load. It is extremely quiet in opera- tion and offers a spin speed of 5000 rev min-1. It features Sepacontrol, which enables users to pre-select and store 32 centrifugation programs. Also incorpor- ated is an air-cooled refrigeration unit programmable from -20 to +40 "C.Nine acceleration profiles and nine braking profiles are programmed into the memory. Three models are available, the 8500,6000 and 5000. Heraeus Equipment Ltd., 9 Wales Way, Brentwood, Essex CM15 9TB. Diaphragm Pumps Available in five sizes ($, 3 , 1, 19 and 2 in) with capacities to 27.2 m3 h-1 and operat- ing pressures to 758 kPa, American pumps are fabricated in 100% pure PTFE or polypropylene and are capable of298 ANALYTICAL PROCEEDINGS, AUGUST 1989, VOL 26 pumping any chemical listed in the Chem- ical Dictionary, including fuming nitric acid, concentrated sulphuric acid, 100% hydrochloric acid and all aromatic sol- vents.Features include an air operated double diaphragm design with a modular air assembly, which is non-stall, ex- ternally serviceable and oil free. All the pumps are sealless and leak-proof, self- priming and offer infinitely variable flow over the full range of the pump perfor- mance curve. American Pump, Osmonics Inc., 5951 Clearwater Drive, Minnetonka, Minne- sota 55343, USA. Calibration System for Flow Meters The Cal-Bench system allows calibration of flow control and measurement devices within sight of the line operations in only 10 min. Engineered to calibrate gas mass flow controllers, gas mass flow meters and most types of flow devices, the system accommodates most gases, has a range of 3-30000 SCCM, and measurement is based on the US National Bureau of Standards’ primary measurements of length and time.Acal Auriema Ltd., 442 Bath Road, Slough, SL1 6BB. Level - Material Interface System The Model 4160 non-contacting level - material interface measurement system features a sensor that emits neutron energy to measure process level or material interface (gas - foam or liquid - solid) through several inches of steel without contacting the process. Also featured is the Model GLT indicating transmitter suitable for installation in a hazardous environment. Acal Auriema Ltd., 442 Bath Road, Slough SL16BB. Automatic Sieve System The Labcon P114 automatic sieve system for particle size analysis features simple menu-driven software, fully automatic operation of balance and sieve shaker, stored values of sieve weights and sizes for future use, archiving of results for later analysis, on-screen graphical representa- tions, the facility for screen contents to be copied to the in-built printer, and the availability of bespoke software service for special applications.Labcon Ltd., 24 Northfield Way, Aycliffe Industrial Estate, Newton Aycliffe, County Durham DL5 6EJ. Filtration System The Bio 2000 liquid filtration and separa- tion system enables the filtration of con- taminant particles down to molecular size (0.01 pm). This is achieved with a com- pact low-maintenance system. Designed to get closer to totally toxin-free process- ing of blood, it is also suitable for cell separation and the processing of blood products such as Factor 8. Small enough to fit a laboratory bench, it can process high value liquid production runs from as low as 1 1 h-1 to 120 1 h-1.Its monitoring and control system enables staff to control and monitor sensitive or critical toler- ances such as flow pressures, pH values and conductivity. Bio-Flo Ltd., Glasgow . Disposable Syringe Filters By the incorporation of a pre-filter the new Anotop 25 Plus range of disposable syringe filters gives improved perfor- mance for sample preparation. Anotec Separations Ltd. , Banbury, Oxfordshire. Electrophoresis System in Veterinary Diagnosis The Paragon agarose gel electrophoresis system has been proved to be a cost effective and efficient aid in the diagnosis of veterinary disorders. By comparing an animal’s protein pattern with a normal pattern for the species and, if necessary, quantitating each component, much use- ful information can be obtained.For example, alpha-1 globulins are increased in haemoconcentration, alpha-2 in acute inflammation and beta-1 (in horses) in strongyle damage. Beckman, Progress Road, Sands Industrial Estate, High Wycombe, Buckinghamshire. Specialised Diagnostics Newly available in the makers’ range of systems and instrumentation for screen- ing, identification and quantification of specific proteins are the Rheumatoid Fac- tor test kit, which eliminates time con- suming heat activation, the improved C-Reactive Protein kit, which eliminates off-line sample preparation and improves sensitivity, and the KappaLambda test kits, which allow for quantification of both free and bound light chains.An information pack details the components of the range and describes such instrumentation as the Paragon agarose gel electrophoresis system, the Appraise densitometer with Data Network, the Array protein system and ICS protein analysers. Beckman, Progress Road, Sands Industrial Estate, High Wycombe, Buckinghamshire. Molecular Modelling Software The January ’89 release of Chem-X mol- ecular modelling software contains a range of new features. The molecular fitting capabilities have been enhanced to allow chemical structures to be fitted together by matching vectors as well as by superimposing atoms. Also featured is the automatic identification of the local max- ima and minima of any function plotted out as a contour map, including wave function, electrostatic potential and elec- tron density maps. The January release of the ChemQM module of Chem-X now provides a transparent interface to Ver- sion 4.0 of the quantum mechanical pro- gram, MOPAC, which has also been implemented on the makers’ transputer based MITIE superworkstation. Other programs already implemented on the MITIE include the molecular dynamics program, AMBER and the ab initio pro- gram, GAMES, as well as the rule- based Chem-X conformational analysis routines.Chemical Design Ltd. , Unit 12,7 West Way, Oxford OX2 OJB. Prints from Video Virtually instant high definition prints of video images are now possible through a complete print-out package, central to which is the IEEE 488 interface available as an option in the makers’ KBFS video framestore, which links a video camera to the Alcatel Visor D laser printer.The latter offers a resolution capacity of 2048 pixels per line, each of which can be one of 256 grey levels. The King KSBS-1 video framestore permits detailed storage and display of video images. To enable users to purchase a complete operating system a laser printer, camera and moni- tor plus framestore with interface are offered. Oggitronics Ltd., Poole House, 37 High Street, Maldon, Essex. Literature The September issue of ARL News, a Spectrographer’s Newsletter, is available in English, German and French and includes items on ARL service, Spectra- Span VB , a new analytical tool for cement makers, new literature and an improved direct current plasma source.ANALYTICAL PROCEEDINGS, AUGUST 1989, VOL 26 299 ARL SA, En Vallaire, CH-1024 Ecublens, Switzerland.A brochure provides details of control and computation facilities for Beckman’s range of ultraviolet - visible spectropho- tometers. Featuring DU Data Leader and DU Data Capture software, it explains that both programs were designed to enhance data acquisition, storage, man- ipulation, formatting and reporting to make the DU Series 60 more powerful and flexible. Special Soft-Pac modules are mentioned; they suit a wide variety of laboratory needs. Beckman, Progress Road, Sands Industrial Estate, High Wycombe, Buckinghamshire. Catalogue 27 covers capillary and packed gas chromatography, HPLC, solid phase extraction, industrial hygiene and chem- ical standards. Supelchem UK Ltd., Shire Hill, Saf- fron Walden, Essex CBll 3AZ. The Howe - Sigma range of centrifuges are described in a brochure. Featuring brushless drive, they are silent and main- tenance free. Microprocessor controlled with automatic rotor identification, they feature free programming of the centri- fuge parameters, allowing up to 100 user programs to be stored. They also incor- porate numerous safety features. A wide range of rotors is available. V. A. Howe and Co. Ltd., 12-14 St. Ann’s Crescent, London SW18 2LS. A guide, “Chromoscan-Instrument Design and Applications,” describes the application of scanning densitometry to the quantification of electrophoretically separated proteins and the way certain design choices affect the ability of a densitometer to generate these data. Joyce-Loebl Ltd., Dukesway, Team Valley, Gateshead NEll OPZ.

ISSN:0144-557X    

DOI:10.1039/AP9892600296

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, AUGUST 1989, VOL 26 299 I New Deputy Director for Polytechnic Polytechnic South West has named Professor Les Ebdon as its new Deputy Director responsible for academic plan- ning and control. He is currently Head of the Polytechnic’s Department of Environ- mental Science in Plymouth. Professor Ebdon takes up his new post on September 1st. He was selected from a strong international field of over 70 candi- dates. In 1986 Professor Ebdon was awarded the Royal Society of Chemistry’s SAC Silver Medal, the first polytechnic lec- turer to be honoured in this way. He has publishedover 100scientific papers, super- vised nearly 30 PhDs and MScs and is Chairman of the Editorial Board of the Journal of Analytical Atomic Spec- trometry. He is also a member of the Publications and Information Board of the Royal Society of Chemistry. Professor Ebdon will maintain his professional activity through the poly- technic’s Analytical Chemistry Research Unit.

ISSN:0144-557X    

DOI:10.1039/AP989260299a

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

302 ANALYTICAL PROCEEDINGS, AUGUST 1989, VOL 26 Conferences and Meetings Grove Anniversary s ~ P O s i u m To commemorate the 150th Anniversary William Grove, in 1839, a Fuel Cell September 18-21, 1989, London of the discovery of the fuel cell by Sir Symposium will be held at the RoyalANALYTICAL PROCEEDINGS, AUGUST 2989. VOL 26 303 Institution. The Symposium will begin on the evening of September 18th with the Sir William Grove Anniversary Lecture, delivered by Professor J . M. Thomas, and will conclude with lunch on September 21st. The aim of the symposium is to review the current status and prospects for fuel cell development in various countries. The principal types of fuel cell and their aRplications will be covered. Attention will be drawn to the potential of fuel cell technology to make a major impact in the spheres of energy efficient use, conserva- tion and environmental control, for both stationary and mobile applications.The speakers are international experts in this field. Their brief will be to provide an overview of the subject, rather than to detail technical problems, and also to anticipate likely developments in the exploitation of fuel cell technology. Attendance will be limited to 300 and early registration is recommended. For further details please contact Miss Geor- gina Mason, IBC Technical Services Limited, 56 Holborn Viaduct, London EClA 2EX. Recent Developments in the Assessment of Electrostatic Hazards in Industry September 28, 1989, London This IBC Technical Services seminar will be at the Tower Thistle Hotel, London E l .The speakers will be Dr. N. Gibson, Dr. P. Cartwright, Professor A. G. Bailey, Dr. N. Wilson, Dr. M. Glor, Dr. P. Tolson, Mr. G. Whitaker and Dr. S. Singh. For further information contact: Sara Mountford, IBS Technical Services Ltd., Bath House (3rd Floor), 56 Holborn Viaduct, London EClA 2EX. Second Annual Winter Conference on Flow Injection Analysis January 3-5, 1990, Orlando, FL, USA The second in this series of conferences will be held in the Compri Resort, Orlando. Both papers and posters are being sought. For information write to WCFIA, Department of Chemistry, Miami University, Oxford, OH 45056, USA. High Performance Capillary Electro- phoresis '90 January 29-31, 1990, San Francisco, CA, USA Following upon the highly successful First International Symposium on Capillary Electrophoresis held in Boston, The Second International Symposium (HPCE '90) will be held at the San Francisco Hilton on Hilton Square, San Francisco, California.The three-day programme will include lectures, poster presentations and discussion sessions. Topics will include Zone Electrophoresis, Isoelectric Focus- ing, Micellar Separations, CEMass Spec- trometry, Gel Columns, Isotachophor- esis, Detector Design, Instrumentation and Analytical and Micropreparative Applications for Pharmaceuticals, Pep- tides, Proteins, Carbohydrates, Oligo- nucleotides, Sub-cellular Structures and Whole Cells. Invited lectures will be presented by, among others, Aharon S. Cohen, Northeastern University, Franz M. Everaerts, Eindhoven University of Technology, Andrew G.Ewing, Pennsyl- vania State University, Eli Grushka, Heb- rew University, Jack Henion, Cornell University, Stellan Hjerten, University of Uppsala, James W. Jorgenson, Univer- sity of North Carolina, Barry L. Karger, Northeastern University, Milos Novotny, Indiana University, Tsuneo Okuyama, Tokyo Metropolitan University, Fred E. Regnier, Purdue University, Pier Giorgio Righetti, University of Milan, Richard D. Smith, Batelle Northwest Labs, Shigeru Terabe, Kyoto University, Edward S. Yeung, Iowa State University and Richard N. Zare, Stanford University. The deadline for submission of abstracts is September 1, 1989. For further information and abstract forms, please contact Shirley Schlessinger, Symposium Manager, HPCE '90, 400 East Randolph Street, Suite 1015, Chi- cago, Illinois 60601, USA.Characterisation of Macromolecules Used as Pharmaceutical Excipients March 7-9, 1990, Gothenburg, Sweden The advances within modern drug forma- tion have, to a large extent, been based on the use of macromolecular or polymeric materials as excipients. Such components often serve as regulators of the diffusion or permeation of the drug from the formulation to the surrounding fluids in the body. Many macromolecular materials have been described in the pharmacopoeias, but the tests therein have been rather vague when it comes to the characterisa- tion of their composition. Viscosity is usually the only parameter that is measured. Sometimes tests which indi- cate differences in degree of substitutions have been described.Technical proper- ties are considered to be related to the construction of the formulation and are therefore of less interest to include in an official monograph for excipients. Those are more the property of the manufac- turer. On the other hand, many new physico- chemical techniques are now available, which can give useful information on the composition of, or show differences between, various macromolecular materials. Chromatographic techniques, which today are becoming more and more powerful, can give information on the oligomer distribution of many polymeric materials. Spectroscopic and thermal methods can also alone give additional interesting data. The combination of chromatography with detectors giving information on molecular weight is now available and adds great power to the instrumental arsenal.All data collected from such sophisti- cated instruments are, however, of little value if they cannot be correlated with the behaviour of the polymeric and macro- molecular material in the formulation. The aim of this Symposium is therefore to bring together experts from both the analytical chemical field and those work- ing in the fields of pharmaceutics and polymer chemistry, to see how applica- tion of the physico-chemical techniques has advanced the know-how and helped in trouble shooting. This symposium is to be organised by the Swedish Academy of Pharmaceutical Sciences and held in the Scandic Crown Hotel, Gothenburg. For information on the symposium, and details of the invited speakers, contact the Symposium on Cha- racterisation of Macromolecules Used as Pharmaceutical Excipients, The Swedish Academy of Pharmaceutical Sciences, P.O.Box 1136, S-111 81 Stockholm, Sweden. XIIth World Congress on Occupational Safety and Health May 6-11, 1990, Hamburg, FRG The professional world of occupational safety and health will gather in Hamburg for the XIIth World Congress. Partici- pants are expected from about 130 coun- tries. Congress participants will have the opportunity to submit communications relating to one of the subject areas until October 1, 1989. The Congress, an event arranged by the International Social Security Association (ISSA) and the International Labour Office (ILO), usu- ally takes place in a three-year cycle. The XIIth World Congress will be held in Hamburg, organised jointly by the National Federation of Industrial Employment Accident Insurance Funds, the National Federation of Accident Insurance Institutes of the Public Sector and the National Federation of Agricul- tural Employment Accident Insurance Funds.The objective of the Congress will be a comprehensive and intensive exchange of experiences on new develop- ments in occupational safety and health. Participation of the following groups of304 ANALYTICAL PROCEEDINGS, AUGUST 1989, VOL 26 persons will be particularly welcome: safety engineers; occupational safety staff; occupational physicians and health staff in enterprises; technical inspectors, factory inspectors and similar officers responsible for the monitoring of occupa- tional safety and health; scientists work- ing in the fields of occupational safety and health , occupational medicine and ergo- nomics; management from all economic sectors and administration; trade union representatives; personnel officers and works councils; management of accident insurance institutions; and other persons concerned with occupational safety and health.The following three subject areas are foreseen: general subjects intended to provide a comprehensive survey of cur- rent aspects of occupational safety and health, particularly of recent develop- ments and the harmonisation of safety regulations in supranational areas; specific subjects dealing with chosen top- ics in more detail, particularly with the protection against harmful substances at work, safe working with automatised systems, transfer of technology, contribu- tion of research in developing occupa- tional safety and health policies in enter- prises; and special meetings which will bring together experts in common fields, particularly in agriculture, the construc- tion industry and the mining industry. For details contact XIIth World Con- gress on Occupational Safety and Health, Hamburg Messe und Congress GmbH, Postfach 30 24 80, D-2000 Hamburg 36, FRG.Fourteenth International Symposium on Column Liquid Chromatography May 20-25, 1990, Boston, M A , USA This symposium will be held at the Boston Park Plaza Hotel, Boston, Massachusetts, USA. The symposium, which alternates between the United States and Europe, is an important scientific meeting for the presentation of the most recent advances in the rapidly growing fields of liquid chromatography and related separation methods.Various sessions will address advances in high-performance liquid chromatography, analytical biotechnol- ogy, novel detectors and detection methods, high-performance capillary electrophoresis, large molecule separa- tions, preparative LC, mechanisms of biopolymer separations, sample pretreat- ment and derivatisations, MS coupling, ultra-trace analysis, chiral recognition, system optimisations, newhovel station- ary phases, emerging application areas, fast LC, microbore and capillary LC, inorganichon chromatography, biorecog- nition chromatography, supercritical fluid chromatography, field flow fractionation and related techniques. General and invited lectures will be devoted to specific topics; both fun- damentals and the application of HPLC and related techniques will be empha- sised. Poster sessions, informal discussion sessions and an exhibition of the latest instrumentation offer the scientist for- mats of his choice. Expert scientists will be invited to present lectures on topics of high interest and additional speakers and poster presenters will offer a wide variety of topics. The deadline for submission of abstracts is October 1, 1989. For further information and abstract forms, please contact: Shirley Schlessinger, Symposium Manager HPLC '90, 400 East Randolph Street, Suite 1015, Chicago, Illinois 60601, USA.

ISSN:0144-557X    

DOI:10.1039/AP989260302c

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

304 ANALYTICAL PROCEEDINGS, AUGUST 1989, VOL 26 Statistical Analysis of Laboratory Data October 9-13, 1989, Amsterdam This course, which will be held at the Carensa Crest Hotel, will cover Proce- dures for Summarising Data, Reference Distributions, Internal Estimates and Hypothesis Testing for the Mean and Variance of a Population, Diagnostic Checking of the Data, Experiments for Comparing Two Treatments, Predicting One Variable from Another, Simple Linear Regression Models, Model Build- ing, Comparing More Than Two Treat- ments, Planning the Experiment and Control Charts for Process Control. In addition, there will be personal computer workshops. For further information contact The Center for Professional Advancement, Palestrinastraat 1, 1071 LC Amsterdam, The Netherlands. High-performance Liquid Chromato- graphy: Fundamentals, Equipment and Operation October 23-25, 1989, Amsterdam The basic course covers an introduction to HPLC, The Stationary Phase and HPLC Columns, The Mobile Phase and Modes of HPLC, Equipment, Special Detection, Chromatography Theory, Quantitation and Detectors, Qualitative Analysis, Trace Analysis, Sources of Information, Size Exclusion Chromatography, Column Maintenance, Adsorption and Normal Phase, Reverse Phase, Ion Pair and Ion Exchange and a Panel Discussion.For information contact The Center for Professional Advancement , Palestrina- straat 1, 1071 LC Amsterdam, The Netherlands. High-performance Liquid Chromato- graphy: Beyond the Basics October 26-27, 1989, Amsterdam The course follows on from the basic course described above.It will go on to discuss other (HPLC related) techniques, followed by Bioseparation Methods, Bioseparation Applications, Other Forms of Chromatography (particularly when they are more appropriate), Pre-chromat- ography Clean-up, Preparative Chromat- ography, Ion Chromatography and Opti- misation of LC Separations. Like the course above, it will be held in the Novotel Alpha, and enquiries should be addressed to the Center for Professional Advancement. Thermoanalytical Methods October 23-26, 1989, Amsterdam This course, which will be held in the Crest Hotel, will consist of an introduc- tory day followed by sessions on Thermo- gravimetry, Introduction to the Applica- tion of Thermal Analysis to Polymer Systems and Computer Applications of Thermal Analysis. For additional details contact The Center for Professional Advancement , Palestrinastraat 1, 1071 LC Amsterdam, The Netherlands. Fourier Transform Infrared Spectroscopy November 6-8, 1989, Amsterdam The course will take place in the Crest Hotel and will consist of an Introduction followed by sessions on Instrumentation/ Hardware, Sampling and Data Manipula- tion, Qualitative Identification - Com- puterised Search, Quantitative Analyses, FTIR Techniques and Applications, How to Buy an Infrared Instrument and Prob- lem Solving. For further details write to The Center for Professional Advancement, Pales- trinastraat 1, 10701 LC Amsterdam, The Netherlands. Laboratory Information Management Systems November 6-9, 1989, Amsterdam This course will cover computerised LIMS systems and the Application of Bar Codes. For details contact The Center for Professional Advancement, Palestrina- straat 1, 1071 LC Amsterdam, The Netherlands.

ISSN:0144-557X    

DOI:10.1039/AP9892600304

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, AUGUST 1989, VOL 26 Analytical Division Diary SEPTEMBER Monday to Friday, 4th to 8th: Loughborough Radiochemical Methods Group. Workshop in Liquid Scintillation Counting. The Workshop will consist of Lectures, Practicals, Tutorials and Problem Solving Clinics. University of Technology, Loughborough. Registration in necessary, and numbers are strictly limited. Cost $395 including accommodation. Contact: Dr. P. Warwick, Nuclear Chemistry Laboratories, University of Technology, Loughborough LEll 3TU. (Tel. 0509-222585). Wednesday and Thursday, 6th and 7th: Salford North West Region, jointly with the Manchester Liaison Committee. Quality Assurance in the Chemical Industry. The University, Salford. Registration is necessary. Contact: Dr.G. Davison, 34 Beechfields, Doctors Lane, Eccleston, Chorley, Lancashire PR7 5RE. (Tel. 0257- 452537). Monday to Wednesday, 11th to 13th: Manchester North West Region, jointly with the Manchester Liaison Committee and the Institute of Physics. Surface Analysis-Techniques and Applications. UMIST, Manchester. Registration is necessary. Contact: Dr. G. Davison, 34 Beechfields, Doctors Lane, Eccleston, Chorley , Lancashire PR7 SRE. (Tel. 0257- 452537). Tuesday, 19th, 10 a.m.: London Particle Characterisation Group. Particle Characterisation in Forensic Science. School of Pharmacy, London. Registration is necessary. Contact: Dr. N. A. Orr, Formulation and Analytical Development Department, Beecham Pharmaceuticals, Research Division, Clarendon Road, Worthing, West Sussex BN14 8QH.(Tel. 0903-39900, Ex. 388). Tuesday, 19th, 10 a.m.: Edinburgh Scottish Region and Automatic Methods Group. Automatic Analysis in Health and the Environment. The automation of chemical analyses never has been simply a problem of automating the analytical process. The analyst has to be concerned with the specimen collection proce- dures, the fast and accurate identification of the specimen, the automation of sample processing and the automated conversion of the data to information through the use of techniques like Chemometric and Expert Systems. This Meeting will explore these facets of automated analysis. “Towards an Automated Specimen Reception System,” by L. B.Roberts. I l l ... “Robotic Sample Preparation-From Concept to Implementation .’* by M. Crookes. “Automated Environmental Monitoring,” by J . Webster. “Advances in Flow Injection Analysis.” by P. J . Worsfold. “Biosensors into the ~ O S , ” by J. Cooper. “Automated Thermal Desorption Analysis.” by K. Saunders. “Quantum-a LIMS and Spectral Database,” by N. Cook. “Some Uses of Chemometrics in Analysing Instrumental Data.” by J. Thompson. Heriot-Watt University, Riccarton, Edinburgh. Registration is necessary. Cost f50 to RSC members, 275 to non-members and &20 for retired and student members. Accommodation is extra. Contact: Mr. R. I. Aylott, United Distillers PLC, Group Central Laboratory, Menstrie, Clackmannanshire F K l l 7ES.(Tel. 0259-61701). Tuesday to Thursday, 26th to 28th: Loughborough Analytical Division. RSC Autumn Meeting. Sample Preparation and Presentation. “Digestion of Heterogeneous Matrices to Produce a Sample for Tuesday 24th-Session 1 : Environmental “Sample Preparation for Environmental Chemistry,” by M. Cresser. Multi-component Metals Analysis,” by M. Kibblewhite. “Problems Associated with the Extraction of an Organic Analyte from an Inorganic Matrix,” speaker to be announced. “Understanding the Importance of Contamination in Atomic “Letting the Punishment Fit the Crime,” by T. E. Edmonds. Wednesday 27th-Session 2: Atomic Spectroscopy “Sample Preparation for Atomic Spectroscopy,” by M. Thompson. “New Approaches to Sample Preparation/Presentation in ICP Spectrometry,” by C.W. McLeod. Spectroscopy,” by C. W. Fuller. “An Appraisal of Direct Introduction of Solids into ICP Sources.” by C. Pickford. Wednesday 27th-Session 3: Radiochemistry “Liquid Scintillation Counting,” by P. Warwick. “a-Counting,” by A . Lalley. “a-Counting,” by G. Sutton. “Activation Analysis.” by D. Green. Thursday 28th-Session 4: Chromatography “On-Line Sample Preparation for Enhanced Sensitivity and Sensi- tivity in HPLC,” by U. Brinkman. “Supercritical Fluid Extraction of Herbal Medicines,” by R. M. Smith. “Combustion in Ion Chromatography,” by J . P. Senior. “Speciation Studies of Metals in Foodstuffs,” by R . Massey. University of Technology, Loughborough. Registration is necessary. Contact: Dr. J . F. Gibson, Royal Society of Chemistry, Burlington House, Piccadilly , London W 1V OBN. (Tel. 01-437-8656). Friday, 29th: Uxbridge Micro & Chemical Methods Group. Elemental Analysis User Forum. Meeting to include: What’s New in Elemental Analysis?; Determi- nation of Phosphorus; Determination of Fluorine. Brunel University, Uxbridge. Contact: Mr. P. R. W. Baker, 55 Braemar Gardens, West Wickham, Kent BR4 OJN. (Tel. 01-777-1225).

ISSN:0144-557X    

DOI:10.1039/AP989260iiib

出版商:RSC    

年代:1989

数据来源: RSC