For a truly complete literature search, at least for a compound that has long been known, one must do the following things.


1. Search all of Chemical Abstracts, beginning with Volume 1 from 1907. The collective indices save a lot of tedium. These were published every 10 years originally, but now appear every five years. You can search by name or formula, or by a compound's registry number if you know it. One can also do subject and author searches.

One must be insightful with regard to searching by name, since IUPAC conventions change over the years. It's often a good idea to do a preliminary search by empirical formula so as to find out what the compound was named in Chemical Abstracts during a particular period. Often one can't do the entire search easily by the empirical formula, if there are several compounds that have the same composition. One can, however, use an initial formula search to fish the correct name out from the many that will be listed.

Searching by formula is highly useful if the compound under consideration has few or no constitutional isomers.

Searches of earlier volumes of Chemical Abstracts presently have to be done manually. Online searching of the later volumes can be done via SciFinder Scholar. SciFinder is a very powerful search engine, and allows searches by structure, substructure, chemical name (traditional or IUPAC), registry number, formula, author, journal, or keyword. Sorting of the references in several different ways can also be done.

Collective indices of Chemical Abstracts up to ca. 1964 (I don't remember the exact final year) are available at Gustavus. Individual volumes past that point up to the time where online access begins have to be searched manually.


2. It is sometimes the case that simple derivatives of compounds have been made and reported without the parent compound having been reported. For compounds or substructures where this is deemed to be a reasonable possibility, likely derivatives, analogs, and homologs should be searched in addition to the parent. A general example might be esters of acids. It would be highly possible in such a case that both ethyl and methyl, and possibly phenyl esters could have been reported, so those compounds should be searched.

An example from our research is the phthaloylphosphino ring system (see Ryan and Jeff's poster in the hall outside my office). In my original preparation of the anion, I searched for the both anion and the protonated form, and in addition searched for the P-methyl, P-ethyl, and P-phenyl derivatives. I found that, though I was the first person to make the phthaloylphosphanide anion, the charge-neutral P-ethyl and P-phenyl derivatives had been made by a different route. Thus, the anion was new, but the ring system was not. Searching only for the anion would have caused me to omit important references that were highly relevant to the research.

Searches for ionic species should be made for all likely salts (e.g. sodium, potassium, cesium, and lithium would be likely ions for pairing with phthaloylphosphide). A search by formula would be likely to work well for these cases.


3. For organic compounds, a search of Beilsteins Handbuch der organischen Chemie must be done (click the highlighted name for links to Beilstein-related web sites). This goes back in time much farther than Chemical Abstracts (to 1771), and much organic chemistry was done during the latter part of the 1800's. "Beilstein", as it's usually known, is intended to be a compilation of all known organic compounds, organized by ring types (also including acyclic compounds, of course) and functional groups. The more recent volumes of Beilstein are available online (but not at Gustavus), though at a steep fee.

In searching out preparations of compounds, Beilstein is often more useful than Chemical Abstracts in that it provides much more detailed abstracts, often listing physical properties and details of preparations. One can thus sometimes avoid the necessity of acquiring the entire original article in which the compound was reported.

We have hard copies of Beilstein up to 1968 in the Chemistry Department. A decision, very poor in retrospect, to stop acquiring Beilstein was made at that time. I came to Gustavus in 1985, and so am not complicit in that decision.


4. A search of an early German analog of Chemical Abstracts called Chemisches Zentralblatt should be done, at least for the years of World War II. Chemical Abstracts was inefficient at abstracting German chemistry during that period. Most large university libraries have Chemisches Zentralblatt.


Note: One can search the Web for leading references and other information, but a Web search does not constitute a literature search.