JACS volume 50 page 571 (1928)
A Preparation Method for Benzophenone
It is well known that simple carboxylic acids yield hydrocarbons to a greater or less extent when they are heated with finely divided metals, particularly copper and nickel.
In the case of aliphatic acids the reaction is usually difficult to effect. Acetic acid, for example, when heated with copper at 390-411°, gives mainly acetone and when nickel is used as the catalyst a considerable portion of the acid is decomposed into complex carbonaceous substances1. With aromatic acids the decomposition is usually easier but frequently, as with benzoic acid and nickel2, the reaction temperature is so high that the resulting hydrocarbons are largely transformed into other substances.
It has been found that this process of eliminating carbon dioxide from an acid can be used to advantage in the preparation of benzophenone from o-benzoylbenzoic acid.
The yields, 80-90% of the theoretical, exceed the average of those obtained by the Friedel and Crafts synthesis, in which benzene and benzoyl chloride react in the presence of aluminum chloride, by about 10 %. Furthermore o-benzoylbenzoic acid which is the intermediate in the synthetic anthraquinone process, is manufactured on a large scale and should be obtained without difficulty. It can also be prepared easily and inexpensively from phthalic anhydride, benzene and aluminum chloride3. This reaction goes nearly to completion, 95-97%, so that the yield of benzophenone based on the phthalic anhydride is still somewhat better than that obtained when benzoyl chloride and benzene are used, and the cost of the materials is much less.
When o-benzoylbenzoic acid is heated alone to about 300° some carbon dioxide is eliminated and benzophenone can be detected among the products. The yield, however is negligible. The catalysts which accelerate the reaction most efficiently are metals such as copper, nickel, iron and the o-benzoylbenzoic acid salts of these metals. It is not necessary that the metals be particularly finely divided as they go into solution in the acid, and the only apparent advantage in having a fine state of subdivision is to increase the rate of solution. The preferred catalysts are finely divided copper and the copper salt of o-benzoylbenzoic acid. The later is easily made by adding a copper sulfate solution to a solution of the sodium salt of the acid, washing, filtering and drying the precipitate at 110°. The velocity of the reaction increases with the amount of catalyst used, but too large a quantity causes inconvenience in getting out the product. It has been found that a satisfactory ratio is 1-3 g. of copper, or an amount of the copper salt of o-benzoylbenzoic acid containing an equivalent weight of copper, to 100 g. of the acid. In the presence of these catalysts carbon dioxide elimination begins at about 150° and increases as the temperature rises until 306° is reached, when the reaction mixture begins to boil. The speed of the reaction below 200° is too slow for practical purposes and above 280° there is decomposition, the formation of anthraquinone and unidentified tarry substances. The optimum temperature range is about 250-270°.
Preparation Procedure
Three hundred grams of pure o-benzoylbenzoic acid was placed in a distilling flask with a low side tube. To this was added 20 g. of the copper salt of the acid. The flask was fitted with a mechanical stirrer arranged through a mercury seal, and a thermometer the bulb of which was immersed in the liquid when the material melted. The mixture was heated with stirring; carbon dioxide evolution was appreciable at 200°, and it was allowed to escape through the side tube of the flask and to bubble through water, which gave an indication of the progress of the reaction. The temperature was allowed to rise and was maintained at 260° until carbon dioxide evolution ceased (about four hours). It is necessary to continue until the carbon dioxide evolution has completely stopped. If a distillation is attempted before this point, a considerable quantity of anthraquinone is formed, which contaminates the product. The stirrer was taken from the flask, a short air condenser attached to the side tube and the thermometer raised out of the liquid as for distillation. The crude benzophenone was distilled over until the drops of distillate became dark in color; the weight of this crude product was 209 g., 86.6 % of the theoretical. One crystallization from 95% alcohol gave pure benzophenone, m. p. 47-48°, in an amount corresponding to 82-84% of the theoretical based on the o-benzoylbenzoic acid. The crystallization may be replaced by distillation in a vacuum.
1 Sabatier and Senderens, Ann. chim. phys., [8] 4, 467 (1905).
2 Sabatier and Mailhe, Compt. rend., 159, 217 (1914).
3 Cain, “The Manufacture of Intermediate Products for Dyes,” Macmillan and Co., London, 1918, p. 240