It should be remembered that the IR spectrometers used in most of the '40s were single beam instruments where the standard NaCl prism spectrometers gave the 4000 - 700 cm^-1 region in six sections, each with a sloping black-body radiation background with carbon dioxide and water vapor bands superimposed.  In order to get a percent transmission spectrum of a chemical, one had to replot this, point by point, by manual division with a background spectrum with nothing in the beam. 

The IR spectrometers at the Stamford Labs had relatively long focal length spherical mirror optics.  It had become clear that infrared spectroscopy was a valuable and a practical tool for chemical analyses, but for certain plant applications, more compact instruments were needed.  Unfortunately, when short focal length instruments were designed and built at Stamford, it was found that the resolution was unacceptably poor.  This was traced to the spherical optics.  When these were replaced with appropriate off-axis parabaloids and ellipsoids, the short focal length instruments equaled or surpassed the older larger instruments in resolution and general quality.

It was during this time that the Perkin-Elmer Corporation, which was then an optical company, was asked to polish an NaCl prism for Cyanamid's new short focal length spectrometer.  Perkin-Elmer suggested that instead, they would like to manufacture the whole spectrometer using Cyanamid's design, which was agreed upon.  This was the Model 12 single beam spectrometer to be used with the Cyanamid double galvanometer recorder system, (and later with the DC amplifier system).  Cyanamid got the first spectrometer, and since I was far and away the least important member of the IR group, I was given the job of testing the instrument out.  In particular, the instrument had exchangeable prisms, and I especially checked out the region below 800 cm ^-1 using the KBr prism where the NaCl prism had poor transmission.

I decided to run the KBr spectra of about 150 aromatic compounds in the region below 900 cm^-1 where aromatics were known to absorb.  (See the chart illustration).  Only a few aromatics had actually been run down there.  I made purely empirical correlations of ortho, meta, para, and other isomers in this region which is where out-of-plane CH wag bands were thought to occur.  When I tried to extend the correlations to substituted pyridines and condensed aromatic rings like naphthalenes, I found that the only general correlation that worked was that for the number of adjacent hydrogens between substituents on the ring.  Meta substitution, for example, had three adjacent hydrogens and one isolated hydrogen between the substituents.  In pyridines, the ring nitrogen has no hydrogen on it, and acts like a substituted ring carbon.  One can easily distinguish alpha and beta substituted naphthalenes from the adjacent hydrogen correlation; alpha substitution has four and three adjacent hydrogens and beta substitution has four, one, and two adjacent hydrogens etc.  This was my first correlation (around 1945).

During the late '40s, the Stamford Labs of American Cyanamid was doing a lot of varied chemical research.  The infrared lab was involved in solving problems of chemical structure and composition for the various chemical groups, making use of Cyanamid's early developed spectrometers and skills in spectral analyses.  During the course of this work, use was made of the known group frequencies and we were able to establish many new group frequencies because of the variety of chemical compounds we had been analyzing.  By 1949, I had assembled all our known group frequencies from my own work and from other contributors into a new chart of spectra-structure correlations.  This new chart made use of the higher resolution in the C-H stretch region made possible by the use of LiF and CaF₂ prisms and the low frequency region down to 400 cm^-1 using a KBr prism.  In addition, it contained many correlations below
1375 cm^-1.  In the 1944 Cyanamid book, the statement had been made that correlations below 1375 cm^-1 were less reliable because more of the molecular bonds were involved.  For example, the C-O stretch in alcohols interact with the other C-C bonds in the alcohol.  While this is true, it did not invalidate certain correlations.  The important feature is not the lack of interaction with the rest of the molecule, but that there is a constant interaction with the rest of the molecule.

I had started this project to teach myself but after I finished the chart in 1949, my group leader R. C. Gore sent a copy of it to Wallace Brode at The Bureau of Standards.  Brode was also Editor of the Journal of the Optical Society of America, and asked me to write a paper to go with the chart that he published in 1950.  This work was successful because Cyanamid was one of the first industrial companies to make and use their own infrared spectrometers, so I was in right place at the right time.  The 1950 chart, which was developed using almost entirely single-beam infrared spectrometers, turned out to be a big hit.  It came out just about the same time when easy-to-use commercial double beam spectrometers became available.  Many new users needed these correlations for spectral interpretation.  A large number of reprints were distributed and the chart was reproduced in several books including The CRC Handbook of Chemistry and Physics.

I continued with my correlation work and collected a lot of published work by others in the field.  I made a much more extensive six-page chart which was published in 1964 in a book I co-authored with Daly and Wiberley for Academic Press, which is now in its third edition.  This chart is still in current use. 

Return to first page