Computers in analytical chemistry
Stephen R. Heller, Rudolph Potenzone Jr.,
US EPA, MIDSD, PM-218,
Washington DC 20460 USA
and
G. W. A. Milne, and Cherie Fisk,
NIH, NHLBI,
Bethesda, MD, 20205 U.S.A.
Various U.S. government agencies have been involved in the development of computerized databases of spectroscopic information which can now be linked to those on toxicology and chemical structure.
Over the past few years, as illustrated by the growth of computers in analytical instruments exhibited at the annual Pittsburgh Conference, the use of computers in analytical chemistry has reached the point where almost every piece of equipment, from balances to high-resolution mass spectrometers has a computer as an integral part. The need for such a tool to assist the chemist in everyday laboratory work is clear, for until now, scientists have spent most of their time in the laboratory recording and calculating experimental data. These activities are time consuming and create a very high degree of boredom and error. Computers can carry out these tasks repeatedly and without the need for coffee breaks or sleep, and because of this their introduction has improved productivity in most laboratories. However, there are other important tasks that computers can perform. I he computer is now being used in chemistry to compare and evaluate, both rapidly and accurately, large volumes of data and to find correlations which the human mind cannot readily perceive.
Searching the chemical literature
The use of computers for information handling and processing in chemistry is only starting to be exploited. Computer searching of the vast and growing chemical literature is now beginning to replace the classical manual approach. But this has met with some resistance. One reason for a reluctance towards computer searching is that manual searching is free to the individual scientist - once his library has paid the subscription fee to a journal. Where an on-line database is used, not only does a subscription have to
be paid, but an extra fee is charged each time a computer search is made (see Fig. 1). However, as budgets get tighter, the prices of printed publications, especially secondary abstracting services, are becoming too expensive for many people to subscribe. This has led to a slow, but marked, move towards the use of computers for bibliographic searching. Organizations such as Lockheed, SDC (System Development Corporation), BRS (Bibliographical Retrieval Services), Blaise in the U.K., and ESA (European Space Agency) m Italy, all have a number of databases storing chemical and related information available on their computers. Databases from Chemical Abstracts, the Institute for Scientific Information, and Excerpta Medico are readily searched by simply connecting a computer terminal to a telephone line or telecommunicattons network. It is believed that the largest of these compames has over 10,000 organizations, in countries throughout the world, regularly using over 100 databases. Other services, such as the ISI ASCA computer based search prof iles, provide weekly computer printouts of all the articles containing a given keyword or author name in the title or references. A service such as ASCA (Automatic Subject Citation Alert)* enables a researcher to keep track of hundreds of journals on a weekly basis. This job could not be done manually because of lack of time and the fact that library budgets are simply not large enough to enable so many journals to be purchased. Browsing through journals is fast becoming a lost art.
Development of spectroscopic databases
While the two areas mentioned above are of value to the analytical chemist, perhaps a more useful application of computers in analytical chemistry is in spectral identification. A number of simple organic molecules, or commonly occurring commercial chemicals (PCB, DDT, Valium, and so forth) may be easily identified using an MS, CNMR or IR spectrum, but most chemicals are not readily identifiable without reference to a spectral library. Since 1972, the US Government, led by the National Institutes of Health (NIH), Environmental Protection Agency (EPA) and the National Bureau of Standards (NBS), have been developing computerized databases of spectroscopic information and making these databases available in a number of ways. The result of this collaborative effort evolved into the NIH/EPA Chemical Information System (CIS). The CIS is a collection of computer databases which is available worldwide, via telecommunication networks (the same networks as are used for the bibliographic systems mentioned in Ref. 2). The CIS has, in addition to a number of spectroscopic databases listed below, databases containing toxicological and structural information as well as other data and modeling programs (2). The CIS spectroscopic databases of interest to the analytical chemist include mass spectra (both electron impact and chemical ionization), infra-red spectra, 13-C nuclear magnetic resonance spectra, X-rav powder diffraction spectra, X-ray single crystal data and organic X-ray structure data. The CIS, using the Chemical Abstracts System Registry Number, can link this information to produce one coherent result, thus the outline of the CIS, shown in Fig. I as a w heel with the various databases attached to the center hub. is appropriate.
An example of how the CIS works is shown in Fig. 2, where infrared (IR), C13 nuclear magnetic resonance (CNMR) and mass spectrometry (MS) data are used to positively confirm the identity of an unknown substance. Additional information (chemical structure diagram, toxicity data and a powder diffraction pattern) about the chemical, Estrone, is then printed out, as shown in Fig. 3.
A valuable tool - used wisely
The use of computers in analytical chemistry is growing in all areas, from data acquisition to data identification and analysis. Over the next few years the major advances in instrumentation, productivity and research are likely to come from those analytical chemists who understand and are well versed in the applications of this new tool. As the cost of computerization decreases (due to improved technology) its value to the analytical chemist increases. Its implementation by knowledgeable scientists will have a great impact in the 1980's.
References
1. Online Rev. (1981) 5, 61-83.
2. Heller, S. R. and Milne, G. W. A. (1980) J. Chem. Inf. Comput. Sci. 20, 204