II trends in analytical chcmisty, vol. 3, n6. 3,1984

Microcomputers, sohare and the analytical laboratory

Until a few years ago only specialized analytical chemical laboratories had their own computer and, because it was usually a central facility, the laboratory had a few specialists working on it. The other members of the laboratory often were not allowed to touch the computer and lived in considerable awe of the few privileged computer experts. The advent of the microcomputer has changed this. The computer is now not only in every laboratory but will soon also be on every desk. Many laboratories still have their specialist who knows a bit more than others about computers, but soon every scientist will be able to work with one. One may conclude that analytical chemistry laboratories have experienced a hardware revolution.

Software developments have not been so fast and many chemists cannot exploit the enormous possibilities of their micromputers to the fullest extent. Where the micromputers are not directly linked to an apparatus, and therefore function as dedicated computers running on software pro- vided by the instrument company, the micro is often used only as a textprocessor and for running a few programs written in the laboratory or obtained from friends or colleagues. Writing programs takes time and is a very costly business; after the first few months of enchantment, many laboratory directors are irritated by the efforts some of their younger associates put into informatics and the time taken from the primary goal of their laboratories, chemical analysis. The authors of the programs, on the other hand, often feel frustrated by the very limited distribution they can give to the product into which they have put so much time and thought.

Chemometrics These problems are felt even more

by the growing number of analytical chemists who call themselves chemometricians and those who are not chemometricians but are con- vinced that chemometrics can solve some of their problems. The

0165?3936/S4/rsO2.00.

Chemometrics Society has defmed chemometrics as:

‘the chemical discipline that uses mathematical and statistical methods (a) to design or select optimal meas- urement procedures and experiments and (b) to provide maximum chemi- cal information by analysing chemical data. In the field of analytical chemistry, chemometrics is the chemical discip- line which uses mathematical and statistical methods to achieve the aim of analytical chemistry, namely the obtention in an optimal way ofrelevant information about material systems’. This is not only a very nice definition of chemometrics but, if one leaves out the words ‘statistical and mathematical’, is also a very good definition of the fundamental aims of all analytical chemists. Even the words ‘statistical and mathematical’ .do not really discriminate the chemometrist from other analytical chemists. The differ- ence is simply a matter of degree. Where the general analytical chemist determines a calibration line with a simple linear regression, the chemo- metrician will determine whether some transformation of the data or weight- ing of the observations produces a better result and he may also try other than linear models.

One can conclude that what dif- ferentiates the general analytical chemist from the chemometrician is his larger knowledge of mathematics and statistics. This is not really true. Many of the chemometric tools are mathematically complex but concep- tually simple. What the general analytical chemist really needs to be able to use chemometrics is good software, which takes care of the mathematics. In fact, chemometrics cannot be employed without the right software. For those chemometricians who want to spread the use of chemometrics through the general analytical public, it is therefore logical to try to provide easy to use and cheap programs. For me personally this was the starting point of my involvement with the Elsevier Scientific Software project.

In this edition of TrAC you will find a Software Bulletin in which informa- tion is given on the chemometric programs which will be available from Elsevier. This is only a start (I hope), other programs will follow and chemometrics will become accessible to every analytical chemist.

D. L. MASSART

D. L. Massart is Professor of Analytical Ckemtitry in th Free University of Brussels, Laorbeeklaan 103, B-1090, Brussels, Belgium. He ir an advisory editor for TrAC. -

in forthcoming issues.. .

Applications of flow cytometry in clinical diagnosis

The measurement of enzyme catalysed rates of reaction by PD-NMR spectroscopy

Computerized interpretation of electrochemical data using deviation pattern recognition

Substoichiometry in trace-analysis

Enzyme immunoassays with electrochemical detection

Steroid analysis in the pharmaceutical industry

PIXE Possibilities in elemental micro and trace analysis

by Katharine A. Muirhead (USA)

by James A. Ferretti and Robert S. Balaban (USA)

by James F. Rusling (USA)

by Kiyoshi Kudo and Nobuo Suzuki (Japan)

by G. Sitta Sittampalam and George S. Wilson (USA)

by S. Gorog (Hungary)

by Rainer P. H. Garten (FRG)

@I 1984 Elscvier Science Publishers B.V.