160
INDUSTRIAL A N D ENGINEERING CHEMISTRY
this there should be a slow and gradual increase of temperature up to the optimum range. It is obvious that satisfactory results in the FriedelCrafts synthesis cannot be obtained when the delivery of the reacting materials takes place in a moist atmosphere, Maximum efficiency cannot be expected in a reactor or reaction medium that contains moisture. It is essential, therefore, to provide approximately anhydrous operating conditions. I n small-scale work this is simpler than in factory practice. It is feasible, however, to provide a separate dehumidified room for the reactor. All connections to the reaction vessel should be closed after each run to prevent the absorption of moisture by the hygroscopic residual material. It is furthermore desirable to sweep the reactor with hot, dry air or gases before each run. These precautions are simple and inexpensive steps in the right direction. Acknowledgment
VOl. 23, No. 2
chief of the Chemical Warfare Service, for information regarding the work done a t Edgewood Arsenal. Acknowledgment is also made of the courtesy shown him by the chemical personnel at Edgewood Arsenal, from whom much information of practical value in this field was obtained. Literature Cited (1) Beard, Dissertation Ohio State University, "Use of TiClc in the FriedelCrafts Reaction."
(2) Boswell and McLaughlin, Can. J . Research, 1, 400 (1929). (3) Groggins, IND. ENG. CHBM.,29, 620 (1930). (4) Groggins and Newton, Ibid., 21, 367 (1929). (5) Groggins and Newton, Ibid., 22, 157 (1930). (6) Heller and Schulke, Ber., 41, 3627 (1908). (7) Jacobson, U. S. Patent 1,445,082 (1921). (8) Lawrance and Oddy, J . A m . Chem. S O L ,44, 329 (1922). (9) McAfee, Trans. A m . Inst. Chem. Eng., 8 , 177 (1915). (IO) McAfee, I b i d . , 22, 209 (1929). (11) McMullen, J . A m . Chem. Soc., 48, 1965 (1921). (12) McMullen, I b i d . , 44, 2055 (1922). (13) Muller and Appenzeller, U. S. Patent 1,699,671 (1929).
Nitration of Diethylene Glycol' Wm. H. Rinkenbach and H. A. Aaronson PICATINNY ARSENAL,DOVER, N. J.
Diethylene glycol dinitrate has been prepared with a H E nitration of diethylIt was recognized that a yield of about 85 per cent of the theoretical by the use ene glycol (P,P'-dihycomplete study should evaluof a mixed acid differing radically in composition from ate the following variables droxyethyl ether) was those used in nitrating other compounds. first studied by Wurtz (3), with respect to maximum who showed that nitric acid The pure diethylene glycol dinitrate obtained is yield: caused the f o r m a t i o n of toxic when absorbed through the skin, or its vapors are Ratio of nitric to sulinhaled. The symptoms are similar to those produced glycolic, oxalic, and an un)(! func acid in nitrating acid named acid by oxidation, but by nitroglycerin. ( 2 ) . Percentage OF water in found no eviience of the' fornitratmg acid (3) Ratio of nitric acid to diethylene glycol mation of nitrate or nitro derivatives. He made no effort to (4) Solubility of dinitrate in spent acid usemixed acids for the purpose of bringing rtbout substitution. (5) Effect of temperature The preparation of the dinitric ester, CH2(NOs).CH2.0.(6) Loss by oxidation CH~.CHQ(NO,), was first reported by one of the writers ( W ) , (7) Losses during purification who obtained it by the use of several mixtures of nitric and In addition, it was considered desirable to study the stasulfuric acids. Little effort was made to determine the optibility of the purified product, the nature of the spent acid mum conditions of nitration with respect to the composition of the nitrating acid, as the yields obtained were considered from the nitration, and the toxicity of pure diethylene glycol sufficiently satisfactory for the purpose of preparation and dinitrate. Purification of Diethylene Glycol study. Diethylene glycol dinitrate was found to possess properties Commercial diethylene glycol was distilled under a pressure which suggested its application in the manufacture of ex- of 120 mm. and the fraction boiling a t 184-187' C. was sepaplosives, and an increasing interest in the compound has rated. This was further fractionated by freezing. The made i t desirable that a study be made of the mechanism of pure compound so obtained was found to have a freezing the nitration and the determination of the conditions giving point of -11.5' C. as compared with -11.3' C. reported prethe maximum yield. This paper gives the results of such a viously (2). study and adds to the existing data concerning the properties Water in Nitrating Acid of the compound. Commercial diethylene glycol usually contains about 5 The efficiency of a nitrating system, given the theoretical per cent of ethylene glycol, which gives a high yield of ethyl- requirement of nitric acid and constant temperature and ene glycol dinitrate when treated with mixtures of nitric pressure conditions, is largely dependent upon the dehydrating and sulfuric acids. Because of this, commercial diethylene value of the sulfuric acid present (DVS). This is best exglycol was purified by fractional distillation and freezing pressed as the quotient obtained by dividing the actual before nitration, A number of nitrations of the commercial sulfuric acid content by the total water present when the material were also made for purposes of comparison. reaction is completed (I). A number of nitrations were made by adding 25.0 grrtms 1 Received September 20, 1930. Published by permission of the Chief of pure diethylene glycol to 91.7 grams of mixed acid mainof Ordnance. U. S. War Department.
T
.
INDUSTRIAL AND ENGINEERING CHEMISTRY
February, 1931
tained a t 5-10" C. As the amount of nitric acid, representing an excess of 54.4 per cent over that required by theory, was the same in each case, the nitric acid ratio was a constant in this series. The results are given in Table I. Table I-Effect
of Water in Nitratin$ B a t h on Yield of Diethylene Glycol Dinitrate
COMPOSITIOX OF MIXEDACID "0s
H?SOi
Ha0
%
%
% 50 50 50 50 50 50
0 0 0 0
0 0
30 35 40 42 45 50
0
20 1.5 10 7 5
0
0 35 0 0
1
DVS OF MIXED ACID
YIELD
%
0
1 02 1 44 2 07 2 50 3.15 5 39
0
0 65 0
0 0
0 5 21 5 78.3 81.0 83.9 75 2
161
whole fumed off in less than 2 hours. in Table 111. Table 111-Effect
COMPOSITION
OF
The results are given
of Temperature on Yield of Diethylene Glycol Dinitrate TIMEOF ADDITION
MIXEDACID TEMP.
HNOi
Has04
YIELD
HzO
%
%
%
50.26 50 26 50.26 50,26 50.26 50.26 50.26
44.74 44.74 44.74 44.74 44,74 44.74 44.74
5 0 2.0
=
c.
5-10 5-10 10-15 15-20 20-25 25-30 30-35
5.0 5.0 5.0 5.0 5.0
OF
DIETHYLENE GLYCOL
%
Min.
83.3" 84.6) 85.5b 83.6b 83.7b 84.lb
56
Noteb
43 32 16
a Pure diethylene glycol used; nitrogen content of product, 14.26 per
cent.
When these results are plotted (Figure l), it is seen that a nitrating acid having a DVS value of 3.15 represents an optimum with respect to yield of the dinitrate. It is believed that the reduction in yield when a mixed acid of higher DVS value is used is due to the tendency of strong sulfuric acid to dehydrate diethylene glycol. Such dehydration would result in the formation of polyethylene glycols (dihydroxy polyethers), which give lower yields of dinitrates on subsequent nitration.
b Commercial diethylene glycol used.
These results show that the presence of a small amount of other glycols does not interfere with the nitration of diethylene glycol and that the yield is not appreciably influenced by the use of temperatures as high as about 30" C. The effect of allowing the dinitrate oil to remain in contact with spent acid is shown by the fume-off noted above.
Nitric Acid Ratio
Sitrations carried out with all the variables held constant except the ratio of nitric acid used to that theoretically necessary gave the results shown in Table 11. Table 11-Effect
of Nitric Acid Ratio o n y i e l d of Diethylene Glycol Dinitrate
MIXED TEMP, ACID GLYCOL
OB MIXEDACID
Hh'Os
His04
M:&D
H20
Grams Grams
' C.
%
%
%
68.6 78.74 86.2 91.7 95.6
5-10 5-10 5-10 5-10 5-10
43.3 45.3 48.2 50.0 51.2
51.7 49.7 46.8 45.0 43.8
5.0 5.0 5.0
25.0 25.0 25.0 25.0 25.0
DVS
COMPOSITION
DI-
ETHYLENE
5.0
5.0
RATIO HNOa USED
ACID
TO
3.15 3.15 3.15 3.15 3.15
1.0 1.2 1.4 1.54 1.65
YIELD
% 66.6 74.8 80.4 83.9 85.4
When these results are plotted as shown by Figure 2, it is seen that the yield increases almost directly with the excess of nitric acid. The limit approached, however, is not 100 per cent, but about 89 per cent, because of the solubility of the dinitrate in spent acid and in water. The maximum yield recorded is therefore within 3.5 per cent of the practical maximum, and the curve indicates that this discrepancy could be reduced by increasing further the nitric acid ratio. In view of the nature of the curve given in Figure 2, it is evident that the choice of a nitric ratio to be used in a mixed acid which is to be economically optimum must depend upon the relative values of nitric acid and the dinitrate produced. For experimental purposes a ratio of 1.55 was considered satisfactory. Effect of Temperature
In order to determine the effect of temperature of nitration on the yield obtained, a number of nitrations were made with temperature as the only variable. In making these nitrations, 50 grams of diethylene glycol were added to 183.4 grams of mixed acid which was actually maintained within a narrower temperature range than indicated in each case. After the addition of the diethylene glycol was complete, the mixture was stirred for l/z hour, while cooling to about 5" C., before separating and washing the dinitrate. Nitration took place without difficulty. Allowed to stand a t 30-35" C., the temperature of the two-layer system gradually rose, and the
D,1! 5.
The temperature of nitration is of practical importance, however, if the time required is considered. Oxidation by the spent acid in the above experiments was prevented by cooling immediately after nitration, but on a large scale this would be less convenient. However, the time required for nitration might be reduced t o a minimum by carrying out the nitration a t 20-25" or 25-30' C. and then cooling. Composition of Spent Acid
As the composition of the mixed acid found best for the nitration of diethylene glycol differs radically from those used for the nitration of other polyhydroxy derivatives and the benzene homologs, the spent acids from two such nitrations were analyzed and the results are recorded in Table IV. Table IV-Composition
53 10 5s 10 93 a By difference.
of Spent Acid
R
'70
R
60 76
1 13 0 63
3 50 3 17
60 81
% 24 03 24 46
INDUSTRIAL AND ENGINEERING CHEMISTRY
162 Losses
It will be noted that the maximum yields of acid-free dinitrate are only from 82 to 85.4 per cent of the theoretical. It was thought that this might be due to solubility of the dinitrate in the spent acid, oxidation of the raw material or product, and solubility of the dinitrate in the wash waters; and efforts were made to evaluate the effect of each of these possible factors. In carrying out the nitrations the diethylene glycol was added slowly to the cooled and well-agitated mixed acid, the addition being made a t such a rate that the temperature of the mixture would be maintained within the desired 5" C. range. Agitation was continued for '/z hour after all the diethylene glycol had been added, the temperature of the mixture being lowered to about 5' C. by means of an ice bath. The acid and dinitrate were allowed to form two lagers and were separated by means of a separatory funnel. The acid was run into ice water and any dinitrate separating out was returned to the funnel.
RATIO
jtN+ USED To /+NO3 THEORETICAL
The dinitrate was first washed with about eight times its weight of water (300 grams), then with 200 grams of a 2 per cent sodium carbonate solution, and finally to neutrality to litmus with additional changes of water. Usually two water washes of about 250 grams each were sufficient. Any dinitrate adhering to the funnel was dissolved in ether and added to the bulk of the product. The ether was removed by air-drying and the dinitrate was placed in a vacuum desiccator for 3 or 4 days before weighing and calculating the yield. On analysis, spent acid from a nitration made under optimum conditions was found to contain 3.3 per cent of dissolved dinitrate. The solubility of the compound has been reported (2) as 4.1 grams per liter of water a t 24" C. On this basis it is possible to explain the losses encountered in nitration and purification. From 25.0 grams of diethylene glycol there should be obtained 46.2 grams of the dinitrate. With a yield of 85.4 per cent, 39.45 grams of purified material are obtained. The 1250 cc. of washes and spent acid dilution account for about 5.0 grams, or 11 per cent. This leaves 1.7 grams or 3.6 per cent, to be accounted for. From Figure 2 it would appear that this unexplained loss is due to the fact that the maximum excess of nitric acid was not used and that nitration was not complete. It was thought that oxidation of some of the diethylene glycol by nitric acid might be responsible for this loss, but tests of spent acid for oxalic, formic, lactic, malic, citric, and tartaric acids gave negative results. No increase in the yield was noted when the time of nitration was increased, and it is uncertain whether the loss of 3.6 per cent unaccounted for is due to incomplete nitration or to complete oxidation of a portion of the diethylene glycol to carbon dioxide and water. In view of the foregoing it does not appear that a yield
Vol. 23, No. 2
much greater than 80.5 per cent of that theoretically possible can be obtained on a practical production basis. Those reported above include the dinitrate recovered by drowning the spent acids, and this is not economical if the spent acid is to be recovered and concentrated. Furthermore, the dinitrate dissolved in spent acid is decomposed when the latter is heated or allowed to stand a t ordinary temperatures. The purification by washing is necessary, as diethylene glycol dinitrate is unstable in the presence of free acid; and this purification cannot be abbreviated if an acid-free product is to be obtained. Stability
As stated above, diethylene glycol dinitrate is unstable in the presence of free acid. This is reflected by the fact that when spent acid having the composition 10.7 per cent nitric acid, 60.8 per cent sulfuric acid, 0.9 per cent oxides of nitrogen, 3.3 per cent diethylene glycol dinitrate, and 24.3 per cent water is allowed to stand a t room temperature for about 2 hours, the dissolved dinitrate is oxidized and a fumeoff takes place. The diethylene glycol dinitrate prepared from pure diethylene glycol in the course of this work was of high purity, as shown by the nitrogen content of 14.26 per cent noted under Table 111, while the theoretical value is 14.29 per cent. This material was used in studying the inherent stability of pure diethylene glycol dinitrate. This compound is easily decomposed by the application of heat ( 2 ) , and the empirical heat tests for stability are no more applicable to this than to nitroglycerin. Storage tests were made by placing a sample of pure, anhydrous neutral dinitrate in an unheated magazine for 1 year and another portion of the same sample was covered with an equal weight of water and placed in a small magazine for 3 months during the hottest part of the summer season. Each bottle was stoppered tightly so as to permit the acceleration of decomposition by the development of gas pressure incident to decomposition. When tested a t the end of the times noted, the dry sample was found to contain less than 0.005 per cent, and the wet sample not more than 0.004 per cent, of free acid calculated as nitric acid. These results indicate that pure diethylene glycol dinitrate is a t least relatively stable when kept under storage conditions which are normal with respect to temperature. Toxicity
It had previously been reported ( 2 ) that diethylene glycol dinitrate did not appear to be toxic. I n the course of the first experimenbs made under the program reported in this paper the occurrence of headaches was noted. It was thought that these might be due to the formation of glycol dinitrate in the nitration of commercial diethylene glycol, but it was considered advisable to redetermine its toxicity with a larger number of subjects. In making the test, 1 cc. of alcohol containing 0.02 gram of pure diethylene glycol dinitrate was placed on the wrist and forearm of each of twenty-seven people of both sexes. The results were definitely positive. Ten of the subjects were unaffected, five were only slightly affected, and twelve reported headaches during the next 12 hours. Those having nervous temperaments were much more affected than those classified as phlegmatic. The symptoms, similar to those produced by nitroglycerin, varied from a feeling of tenseness of the head, with some discomfort, to headache of extreme intensity, and considerable duration. I n the case of nitroglycerin many workers have developed
February, 1931
INDUSTRIAL AND ENGINEERING CHEMISTRY
an immunity after experiencing the usual headaches, and the question might be raised as to possibility of their doing this when handling diethylene glycol dinitrate. Although data of sufficient scope are not available, it might be noted that severe headaches lasting from 2 to 4 hours were experienced by the worker after making the first two nitrations. In subsequent work no headaches were caused, but if the worker was feeling physically subnormal, a feeling of slight heaviness and tenseness about the head with general discomfort was experienced after making a nitration. It would appear that diethylene glycol dinitrate is definitely toxic and that, since its structure and the symptoms produced
163
are similar to those of nitroglycerin, it affects the heart action and blood pressure in approximately the same ways. Acknowledgment
The authors desire to express their appreciation of their associates who volunteered to submit to tests to determine the toxicity of diethylene glycol dinitrate. Literature Cited (1) Groggins, Chem. Met. Eng.,36, 466 (1928). (2) Rinkenbach, IND. ENG.CHEM.,19, 925 (1927). (3) Wurtz, A n n . c h i n . , [31 69, 317 (1863).
Cobalt Driers’ E. Gebauer-Fuelnegg and Gottfried Konopatsch U N I V E R S I T Y MEDICAL SCHOOL, CHICAGO, UXIVERSITYOF VIENXA,VIENNA,AUSTRIA
DEPARTMENT O F RESEARCH BACTERIOLOGY, h-ORTHWESTERN
I L L , , A N D LABOR.4TORY OF C H E M I C A L
TECHNOLOGY
I T H respect to the Several cobalt compounds of known structure and Experimental action of the siccapurity have been tested as siccatives. Their solutions tives on drying oils in organic solvents have been found to be efficient driers I n the writers’ experiments only the bare facts are known. in most cases. Some such siccative solution caused cobalt compounds were used The metals salts apparently the formation of glass-clear films of China wood oil. as driers and China wood oil have to be c o n s i d e r e d as Simple salts such as the anhydrous cobalt halides have asthedryingoil. The former c a t a l y s t s , since comparaalso been tested and found to be satisfactory driers. were chosen, because a larger tively small quantities speed variety of compounds could up the drying process and film formation. Our presclnt knowl- be prepared and tested than with the other metal salts so far edge of the process indicates that the drying of the oil is a used as siccatives, while the latter was selected.to save time durcomplex reaction, involving colloidal as well as chemical ing the drying process and on account of the relative uniformity processes, polymerization, oxidation, and even loss of sub- of its chemical constitution (80 per cent eleostearic acid). stance have been observed. Catalysts can be positive or The preparation of cobalt compounds of definite structure negative, which in the case of drying oils means speeding and their use as siccatives were a t first tackled from the viewup and retarding the process, respectively. It is probable point of their solubility or dispersion in the oil. Most of that in the course of the present work both types have been the simpler compounds of known structure were found t o be encountered. either insoluble or soluble only with difficulty. To facilitate With the hope of getting an insight into the action of their solubility the use of an organic solvent,2miscible with siccatives, an attempt was first made to simplify the driers, the oil, was thought to be advantageous. The requirements both as to the purity of the chemicals used and as to their of such a solvent and solution may be summarized thus: constitution. (1) It must dissolve the siccative readily; (2) it must be easily The metal resinates or oleates that have been used have miscible with the drying oil; (3) it should be non-aqueous several obvious disadvantages. Since the metal part only is and not miscible with water; (4) it must not appreciably the carrier of the drying (catalytic) activity, the organic retard the drying action of the siccative, and (5) it must have remainder, of high molecular weight, is largely ballast. The no bad effect on the film to be formed. organic part, however, is considered indispensable for furBenzene and its homologs were first used, but since some of nishing dispersibility of the siccative in the oils. To de- the cobalt compounds were found to be insoluble, benzonicrease the molecular weight of the organic part should there- trile, benzaldehyde, benzyl chloride, benzyl alcohol, benzyl fore prove advantageous. It is difficult to obtain metal cyanide, furfural, aniline, nitrobenzene, and amyl alcohol resinates and oleates of standard activity, no matter by what were also tested. Intensely colored solutions were obtained. method they have been prepared, since they are chemically Work on other solvents is in progress. not uniform. As a result the quality of the products depends With the exception of furfural such siccative solutions were largely on the skill and experience of the manufacturer. found to be well miscible with China wood oil. Furfural The preparation of siccatives of well-defined chemical con- as such was likewise found to be miscible with the oil, but stitution vas therefore attempted. in presence of the cobalt salts it separated out again. The literature shows that attempts have already been Aniline and amyl alcohol were found to retard appreciably made in that direction. For example, the use of metal oxa- the catalytic action of the substances tested and the filnis lates, benzoates, succinates, palmitates, and napthenates as formed were inadequate. Benzyl chloride had to be dissiccatives has been patented. -411 but the last two are of carded on account of the possible formation of hydrochloric comparatively small molecular weight and, except for the acid and its insufficient solvent action. last, it should be easy t o prepare them in sufficient purity to Benzonitrile, benzaldehyde, benzyl alcohol, benzyl cyanide, insure products of definite standard. The napthenates and nitrobenzene, therefore, must be considered the best alone, however, arc sufficiently soluble in the oils t o make 2 T h e terms “solvent” a n d “solution” as used in this paper are not an efficient drier. 1
Received October 15 1930.
meant t o imply definite statements as t o whether t h e cobalt compounds are in t r u e or (ultra) colloidal solution
