LIQUID AMMONIA RESEARCH itz 1934-A REVIEW GEORGE W. WATT The Ohio State University, Columbus, Ohio
I
N ADDITION to the numerous publications dealing with liquid ammonia research which have appeared during the past year, this review also covers a number of papers published during the year 1933. A few of these were not available a t the time the first of these reviews (1)was prepared for publication. Others were inadvertently overlooked. In one of his earlier publications concerned with the chemistry of liquid ammonia solutions, Franklin (2) offered the term ammonolysis as "a suitable designation for a class of reactions in which ammonia plays the part analogous to the action of water in-ordinary hydrolytic reactions." Subsequent studies of ammouolytic reactions have shown that they are of quite general occurrence and that they may be utilized in the preparation of a rather wide variety of nitrogen pounds. Further, it is to be noted that reactions of this type may usually be carried out with a minimum of experimental difficulty. A number of typical arn-monolytic reactions will be discussed in the following pages. A method for the preparation of amino nitriles in high yields has been described by Menge (3). This method involves the ammonolysis of nitriles by means of liquid ammonia a t room temperature. In this manner, Menge has prepared aminoacetonitrile by the action of liquid ammonia on glycolonitrile and has stated that the method may be used for the preparation of other compounds of this type. Pinck and Hilbert (4) have shown that fluorenone imide is pwduced in high yields by the ammonolysis of fluorenone. Similarly, a liquid ammonia solution of ammonium chforide ammonolyzes fluorenone anil to fluorenone imide and aniline. The nature of the products formed when fatty oils are heated with liquid ammonia in an autoclave has been investigated by Oda and co-workers (5, 6 ) . Olive, spenn, linseed, coconut, castor, fish, walrus, and chiuese wood oils were completely ammonolyzed to solid products. It is believed that these reactions, which apparently proceed in a manner analogous to the hydrolysis of fats with water, give high yields of fatty acid amides. For example, Oda has shown conclusively that the principal product of the ammonolysis of coconut oil is the amide of lauric acid (7). In the course of his studies of the action of ammonia and amines on the esters of unsaturated acids, Morsch (8) has reported that methyl acrylate reacts with liquid ammonia a t room temperature. The products obtained varied with the length of time during which the reac1
tions were allowed to proceed. After 100 hours, the indicated yields of the following products were obtained: Methyl p-aminopropionate, HINCHnCHsC02CHz-2.5%
Dimethyl ester of &p'-aminobispropionic a c i d 4 2 % HN(CH2CH2CO9CHJz
Trimethyl ester of p,pl,p"-aminotrispropionic acid-
15.5% N(CHaCHL!O~CH&
After 14 days, p,pf-aminobispropiouamideand &B',pXaminotrispropionamide were found in approximately equal amounts, while a t the end of five months, the latter product was found to predominate, Morsch (9) has also referred to the preparation of ethyl j+aminobutyrate by the action of liquid ammonia a t room temperature on ethyl crotonate, CH,CH:CHCO&H6
+ NHa
-
CH~CH(NH~)CH~COIGH~
Govaert (10) has shown that acetyl and benzoyl chlorides are ammonolyzed readily by liquid ammonia with the production of the corresponding acid amides. The acid chlorides in dry ether solutions were added to anhydrous liquid ammonia, after which the ether and ammonia were evaporated and the acid amides extracted with a suitable solvent. In carrying out these reactions, care must be taken to eliminate water in order to prevent the hydrolysis of the acid chlorides. The reaction between hydrazotriphenylmethane and sodium in liquid ammonia at 0" has been studied by Pinck (11). Hydrolysis of the reaction products gave a 90% yield of triphenylmethane. Similarly, the hydrolysis of the product formed in the reaction of hydrazophenylfluorenene with sodium in liquid ammonia gave a high yield of phenylfluorene. Wooster and Morse (12) have shown that the reduction of r,r,r-triphenylpropyl iodide by means of a solution of sodium in liquid ammonia at its boiling point leads to the formation of a product which, upon hydrolysis, yields triphenylmethane. Of the possible mechanisms proposed to account for the above product, Wooster and Morse prefer the following:
+
--
+
(CnHi)rC-CH.CHd 2Na (CaHJrC-CHCH9Na NaI 2e ( C ; H ~ ) ~ C N CHz:CH1 ~ ('%~)~c-cH~cH~N~ NaOH. (CaHdzCNa HOH -r (GH&CH
+
+
+
+
Thus it appears that the sodium triphenylpropide molecule undergoes a molecular rearrangement involving a spontaneous cleavage which gives rise to a 1
neutral molecule of ethylene and sodium triphenylmethide. As an intermediate in the preparation of r,r,r-triphenylpropyl iodide, y,y,y-triphenylpropyl ethyl ether was prepared in liquid ammonia at its boiling point in accordance with the following equations:
They have also shown that y,r,r-triphenylpropyl iodide and j3,j3,@-triphenylethylchloride are not dehalogenated when treated with solutions of sodium nitrate in liquid ammonia a t its boiling point. Wooster and Ryan (13) have studied the reduction of a number of phenylated olefines with solutions of alkali metals in liquid ammonia at -33.4"C. While these reactions were often complicated by ammonolysis, polymerization, cleavage, etc., it was found possible under suitable conditions to saturate the double bonds in the side chains of every hydrocarbon investigated and to thus obtain the corresponding phenylated paraffins after decomposition of the intermediate organo-alkali compounds with ammonium chloride. The phenyl groups were not attacked by the alkali metals. Since it is possible to obtain highly concentrated solutions of the alkali metals in liquid ammonia, and since the solvent ammonia often enters into the reactions, this method of reducing phenylated olefines seldom results in reactions identical with those which occur in inert solvents. Govaert (14) has reported that the reduction of the tetrabromide of pentaerythritol, C(CH2Br),, by means of a solution of sodium in liquid ammonia leads to the formation of only very small quantities of tetraminotetramethylmethane, C(CH2NHp)r. I t is unfortunate that Govaert and others continue to formulate reactions in terms of the hypothetical compound, NaNHs, since the existence of such a compound was disproved more than a quarter of a century ago. The theory that sodium dissolves in liquid ammonia by virtue of the formation of the so-called '%odammonium" was long supported by Joannis (15). That such a view is untenable was iirst suggested by Seely (16) and definitely proved by Kraus and co-workers (17.18) and by Ruff and Geisel (19). It has since been quite generally recognized that sodium dissolves in liquid ammonia to form a true solution. Certain modifications of the method of Lebeau and Picon (20) for the preparation of but~ne-1have been described by Lai (21). He has reported that these modifications enable one to prepare this product in fairly large quantities with comparative ease. The above method involves the treatment of sodium acetylide with ethyl iodide in liquid ammonia a t low temperatures. While the first of the following equations is not in accord with the views of Lai, it probably represents the course of the primary reaction:
-
HCi CH + N a - H C 1 C N a +'/,H, HC i CNa IC2Hs HC i CCHZCHS NaI
+
+
Lai holds to the opinion that the monosodium acetylide is formed by the action of "sodammonium" on acetylene. Nieuwland and co-workers (22) have reported the preparation of methyl and ethylacetylenes by the method due to Bourguel (23). This method involves the treatment of sodium acetvlide in liauid ammonia with the appropriate dialkyl &ate. Vaughn (24) has pointed out that a solution of sodium in liquid ammonia may be used as an agent for the removal of hydrogen halides. He has investigated the action of this reagent, at -33.5'C., on a number of organic halides which might be expected to yield acetylenes upon dehydrohalogenation. Thus, 8bromostyrene, p-methyl-a-chlorostyrene,and styrene dibromide gave fair yields of the corresponding substituted acetylene. a-Chlorostyrene, stilbene dibromide, and 2-bromodecene-l were reduced to phenylethane, bibenzyl, and decene-1, respectively. Since all of the above reactions yielded some hydrogenation products, this method is not particularly well suited to the preparation of pure acetylenes. Miller and Roberts (25) have investigated the reactions of certain amino acids and related substances with sodium in liquid ammonia. They have attempted to correlate the volume of hydrogen evolved with the nitrogen content of the sample and the quantity of sodium used. Glycine and alanine were found to react as monobasic acids. Mixtures of amino acids and of amino acids and diketopiperazine yielded less hydrogen than would be expected on the basis of the amount evolved by the constituents of the mixtures taken separately. The peptide linkage in dipeptides was found to be non-acidic in liquid ammonia. The formation of 2:aminoquinoline by the action of alkali and alkaline earth amides on quinoline in liquid ammonia solution has been studied by Bergstrom (26). Yields of from 5 to 11% were obtained with alkali amides in the presence of metallic mercury, while with the amides alone, resinous materials were formed. Using barium amide, fairly high yields of 2-aminoquinoline were obtained in accordance with the equations :
-
++
+
2C&N Ba(NH& + (CoHSJNH)&a HS 2CpHINNH, Ba(OH),. (CpHaNNH)sBa 2HOH
+
.
Lithium, strontium, and barium nitrates and barium thiocyanate were found to act as positive catalysts for the first of the above reactions. A slight reaction was observed to occur between quinoline and strontium amide, while calcium amide was found to be uureactive, similarly, Berptrom (n) has shown that the reaction between iso-aninoline and certain metal amides s~~ in liquid ammonia solution a t room temperature results in the formation of amino-iso-quinoline. The highest yields were obtained with potassium amide, while the amides of lithium, sodium, strontium, and barium gave somewhat lower yields. Calcium and mamesium amides were found to be relativelv unreactive toward iso-auinoline. A study of thed action of a number of ammono-saltson iso-quinoline has shown that NaK%N,ZNHsand BaKN,2NH8 react to form ~~~
-
~~~
~
~~~
~~~
~
~
amino-iso-quinoline. The extent to which an amide or salt reacts with iso-quinoline is believed to be a function of its ability to furnish amide ions in liquid ammonia solution. In connection with their studies of the tautomerism of cyanourea with guanyl isocyanate, Blair and Smith (28) have shown that cyanourea, a mixed aquo-ammono-pyrocarbonic acid, behaves as a tribasic acid in liquid ammonia. They have prepared the mono-, di-, and tripotassium salts by treating cyanourea with potassium amide in liquid ammonia solution a t room temperature. The monopotassium salt is quite soluble in liquid ammonia while the di- and tripotassium salts are insoluble. An interesting study of the chemical behavior of benzopbenoneimine has shown that this compound exhibits the properties of an aromatic ammono ketonealcohol (29). This imine was prepared by the ammonolytic action of liquid ammonia at its boiling point on the addition product formed between benzonitrile and phenylmagnesium bromide. With hydroxylamine in liquid ammonia a t room temperature, henzophenoneimine reacts to form practically a quantitative yield of the corresponding oxime. A monosodium salt was formed by the action of sodium amide on benzophenoneimine in liquid ammonia a t room temperature, (W&C:NH
+ NaNH2
-
+ NHa
(GH&C:NNa
while a similar reaction with potassium amide resulted in the formation of a dipotassium salt, (CsH&C:NH
+ 2KNH2
-
(CsH&C(NHK)r
+ NHa.
times difficult. Further, Muskat has shown that carhohydrates containingpotentially free aldehyde groups are oxidized bv iodine in liouid ammonia to the corresoonding acid akides. ~ ~ ~ a r a ttechnic, u s , and preca;tions to be observed in carrying out these reactions have been described. These reactions are usually effected a t the boiling point of liquid ammonia. The complete methylation of isosucrose (31) has been accomplished through the use of liquid ammonia as the solvent. A method for the preparation of aldonic and saccharinic acid amides by the ammonolysis of the lactones of the corresponding adds by means of liquid ammonia a t its boiling point has been desnibed (32). Thus, the ammonolysis of d-glucono-y-lactoue gave a high yield of d-gluconamide. This method is comparatively simple and is believed to have certain advantages over other known methods for the preparation of this type of compound. It has been found that narcotine methiodide and methochloride are converted into narceinamide hydriodide and hydrochloride, respectively, by treatment with liquid ammonia a t its boiling point (33). By a similar treatment, hydrastine methiodide was converted to methylbydrastamide hydriodide. Methylhydrastine and methylhydrastine hydriodide are not ammonolyzed by liquid ammonia a t its hoiling point. Coleman and Maxwell (34) have shown that tolane and substituted tolanes may be prepared in good yields by the action of potassium amide on lJ-diaryl-2chloroethenes in liquid ammonia solution: (GHa)sC:CHC1
+ KNH?
-
CsHrC i CCsHs
+ KC1 + NH,.
The reduction of benzophenoneimine by means of solutions of the alkali metals in liquid ammonia a t room temperature apparently led to apartial reduction of the imine to henzohydrylamine which reacted with excess imine to form benzophenonehenzohydrylimine,
Tolanes may also be prepared by the action of potassium amide upon diary1 substituted dichloroethanes in liquid ammonia:
+
Continuing their studies of the iodination of substituted acetylenes, Vaughn and Nieuwland (35) have iodinated xylyl, mesityl, o-, m-, and .fi-chlorophenyl, fi-hromophenyl, and a-methylvinyl dcetylenes with iodine in liquid ammonia a t its boiling point. They found that the halogen-substituted phenylacetylenes iodinated much more rapidly than did the alkyl substituted ones. This method for preparing iodoacetyl e n s is characterized by high yields and the absence of complicating side reactions. Fulton and Bergstrom (36) have shown that the electrolysis of concentrated liquid ammonia solutions of the potassium salts of acid amidines (the fatty acids of the ammonia system) at high current densities results in the formation of hvdrocarbons. The electrolvsis of potassium acetamidiie yielded methane in a manner analogous to the formation of hydrocarbons by the Kolbe synthesis in water:
(GHdC:NH
+ (GHdCHNHz NHz
-c
(Wr)rC:NCH(W&
By analogy to the reduction of an aquo-ketone, i t might be expected that the reduction of benzopknoneimine would result in the formation of an ammono benzopinacol or benzopinacolone, but no evidence for the formation of such compounds was obtained. The results of an investigation of certain reactions of carbohydrates in liquid ammonia have recently been published by Muskat (30). With the exception of the free sugars, which react with liquid ammonia to form amino derivatives, most carbohydrates are soluble in and unreactive with liquid ammonia. It has been shown that carbohydrates containing a free hydroxyl group react with alkali metals or their amides to form alkali salts. By allowing these salts to react with alkyl, aryl, or acyl halides, either in liquid ammonia or in an inert solvent, i t is possible to introduce various substituents into the carbohydrate molecule. It has also been found possible to introduce phosphorus, arsenic, antimony, bismuth, sulfur, silver, and mercury into the carbohvdrate molecule. althourh the isolation and identification of certain of these products is some-
(C;H&CHCHCIz
+ 2KNH2
2CHaC(:NH)NHK
-
CHrC i CCsHaf 2KCI
+ 2NH8
+ 2F2K (atCH8CH8 + N H C N (at anode) + cathode). -c
~
~.
Acid amidines derived from nitriles containing two or more carbon atoms yielded mixtures of methane and ethane.
Wooster (37) has recently pointed out that liquid ammonia may be used to marked advantage in the study of the metal ketyls. He has shown that, in liauid ammonia a t its boilinp point, sodium benzop i k o l a t e is largely dissociated into the corresponding metal ketyl,
-
(CsHSnCONa
I
.-
Dennis and Work (41) have reported that mono- and dichlorogermanes react readily with liquid ammonia a t low temperatures. The following equations have been ~rooosed to describe the course of these reactions: * A
3xGeHaCl
-
+ 3xNHz 3xNHSI + xGeH4 + 2(GeH).. GeHE1. + 2NH8 G e + 2NH,CI. -r
A method for the preparation of pure monogermane, GeHa. in h i ~ hvields has been described bv Kraus and ~ a r n (427. e ~ ?his method involves the Leatment of It has also been shown that the intermediate formation magnesium germanide with ammonium bromide in of the corresponding disodium compound, (CaH&- liquid ammonia solution a t -33.4'C. The crude reCNaONa, is not a determining factor in the rapid action product, which consists of the monogermane, reactions of the metal ketyl. Wooster and Holland higher germanes, and hydrogen may be purified by (38) have shown that the true mechanism of the re- treatment with a liquid ammonia solution of sodium. action between a metal ketyl and an alkyl halide dif- This treatment results in the formation of sodium fers from that proposed by earlier workers. They have germauides, all of which yield monogermane when presented evidence supporting the view that a reaction treated with ammonium bromide. This method of of this type proceeds as follows: purification is somewhat unipue in that the higher germanes which are present as impurities are converted ClsHloONa G H a r NaBr (CUHUO) into monogermane in the course of the process. Kraus (C!~SH,~O) . . + CtaHzoONa + ( C B H E ) ~ C O (CsHs)nC(GHdONa . . . . and carney have shown that sodium trihydrogermanide the ketone being formed in the secondary reaction. ~ecomposeswith rising temperaturein accordauce This secondary reaction may be repressed by rapid with the equation: utilization of the ketyl in the primary reaction, while if NaGe% NaGe '/2H2. the alkyl halide is added slowly, high of ketone - yields . are obtained. With molecular oxygen, NaGeHa in liquid ammonia In connection with their studies on the chemistry of solution is apparently oxidized to an orthogemanate the triethylsilicyl group, Kraus and Nelson (39) have which loses water a t room temperature, carried out a number of reactions with triethylsilane and its derivatives in liquid ammonia solution: Using magnesium silicide as the starting material, Johnson and Hogness (43) have shown that the above method may be used in the preparatiou of monosilane, .. Tripheuylgermanyltriethylsilane, (CsHs)aGeSi(GHs)s, SiH,. Dennis and co-workers (44) have observed that inwas prepared by making the sodium salt of triphenylgermanium in liquid ammonia solution, evaporating dium trimethyl is soluble in liquid ammonia a t -35'C. the solvent ammonia, and adding a benzene solution They obtained no evidence of the formation of amof triethylsilicon bromide. Reactions of related silicon mines. McCleary and Fernelius (45) have shown that liquid compounds in solvents other than liquid ammonia have also been studied and appiyatus for carrying out ammonia solutions of the mouotellurides, -selenides, and -sulfides of sodium are oxidized bv molecular oxvthese readions has been described. Johnson and Wheatley (40) have reported solubility gen a t -33.4'C. to mixtures of the corresponding values for germanous and germanic sulfides, and for -ite and -ate oxygen salts. Sodium disulfide was oxidized directly to sodium thiosulfate, while more comsodium sulfide in liquid ammonia a t -33°C. plex reactions occurred in the case of the polytellurides, -selenides, and -sulfides. Kraus and Parmenter (46) have described certain improvements on a method for These sulfides are not ammonolyzed by liquid ammonia. the preparation of the oxides of potassium. This 7% sulfides of germanium were reduced sodium in method depends upon the oxidation of the metal in liquid ammonia in accordance with the following liouid ammonia solution bv means ,f omPen Pas. L~~ ~~~~~~-~ ~, ---equations: Ootimum conditions for the orenaratiou of relativelv pure specimens of K202and KzO, have been given. GeS 2Na NaB Ge Potassium trioxide, K203, has been prepared by the 4Na 2N%S Ge Ge8 oxidation of K202. Certain properties and reactions of 4Na Na&Ge. xGe these oxides have been described. When the sodium polygmauide was treated with amIt has been shown that titanium trichloride reacts monium bromide in liquid ammonia, monogermane was with liquid ammonia at low temperatures to form a produced: white hexammine, while under similar conditions, Na,Ge. GVH,Br 4NaBr 4NHa GeHs ( x - 1)Ge. titanium dichloride forms a pearl-gray tetrammine (CsHJ~CONa
+
~
A
-
~
2(CdHs).CONa
+ +
-
+ + +
+
-
~
---
+ +
+
+
~~
~
+
~~
2 0.-
A
+
0----
(47). When heated to 300°, TiCl&NHa loses four moles of ammonia and forms what is believed to be a diammine. Experimental procedures and apparatus used in this work are well described. G. Spacu and coworkers (48, 49, 50) have prepared the ammonates of a large number of double salts by treating the anhydrous double salts with liquid ammonia a t low temperatures. They have studied the behavior of these ammonates over a temperature range of from -80' to 120°C. In general, it was found that with increase in temperature, the higher ammonates which are formed a t low temperatures lose ammonia with the formation of more stable ammonates. Heats of formation for a number of these compounds have been calculated. Gmner (51) has prepared the solvates, BP04.3Ha0.NHz and BAsOa3HaO.NHs by extracting the hydrates with liquid ammonia. Bjerrum (52) has reported that he has been unable to obtain evidence for the existence of the complex ion, CU(NH~)~++, in liquid ammonia solutions. Schwarz and Giese (53) claim to have secured evidence for the formation of the nitrosamine, HsNNO, by the ammonolysis of nitrogen trioxide, nitric oxide, nitrosyl potassium, nitrosyl sulfuric acid, nitrosyl perchlorate, and nitrosyl chloride .at low temperatures.
Strecker and Schwarzkopf (56) have prepared selenium nitride by the ammonolysis of the methyl and ethyl esters of selenious acid,
+
-
+
+ +
OSe(OR)* 4NHs + Se(NHn)r HzO 2ROH 6Se(NH)x 6Se(NHs)d -r 12NHa 4NHa 2Na 2%. 6Se(NHI2 SelN,
+
+
+
Similarly, Strecker and Mahr (57) have prepared tellurium nitride by the ammonolysis of tellurium tetrabromide, Numerous properties and reactions of these nitrides have been described. The reactions of potassium amide with sulfur'in liquid ammonia solution have been studied by Bergstrom (58). The following reactions were found to occur under the conditions noted, KNHz in excess, a t -33'C.:
-+ + + -
+
+
+
6KNH2 3 s 2KB S(NK)pNH3 3NHs NHs S(NK)?NH1 heat + S(NK).
S i n excess, a t -33'C.: 6KNH2 6KNH2
12s + 2K& 10s 2K&
+
+ S4(NK)*+ 4NHs
+ S4(NK)a+ 4NHa
Nieuwland and co-workers (59) have described a method for the preparation of sodium amide in liquid ammonia a t its boiling point, and have used this amide in the preparation of a number of acetylenes from organic halides by dehydrohalogenation. They believe They believe that the nitrosamine, which they were unable to isolate, is unstable and decomposes in ac- that the conversion of sodium to sodium amide is catalyzed by ferric nitrate. Lome (60) has shown that cordance with one of the following schemes: femc nitrate is reduced by sodium in liquid ammonia a t 2H*NNO NH4NO% N% its boiling point to very finely divided and highly reH2NN0 + Nz + H1O active metallic iron. Since metallic iidn is known to be In connection with this work, the following reactions an active catalyst for this reaction (61), it is highly probable that the catalytic influence observed by were also shown to occur: Nieuwland and co-workers is due, not to ferric nitrate, KNH2 + ONOSOIH NHJ KNOB H ~ N S O ~ N H I ) but to the freshly precipitated iron. In fact, the above 2NaC1 + N a N 4 + 2NH$ NI 3NaNHI + 2NOC1 N2 Hs0. writers state that "at the end of the reaction, most of KNH, + (CsHn),O KNO1 2CaHnNOz the catalytic material is in the form of a-linely divided It has been pointed out previously by Feraelius and black precipitate. . Loane has also reduced nickel Watt (54) that the nitrosamine, H2NN0, is an aquo- nitrate, cobalt nitrate, manganous iodide, and a ammononitrous acid. Their studies on the saponifica- copper salt to the free metals by means of sodium in tion of diarylnitrosamines in liquid ammonia a t room liquid ammonia. The pyrophoric metals so protemperature have shown that the decomposition of the duced were oxidized a t low temperatures and the reunstable salt, KHNNO, occurs in accordance with the sulting oxides used as catalysts fur the oxidation of latter of the above decomposition schemes. carbon monoxide. The efficiency of oxide catalysts preThe ammonolysis of phosphorus pentachloride by pared in this manner compares favorably with that liquid ammonia a t -50°C. has been studied by Moureu of similar catalysts prepared by other methods. and Rocquet (55). The ammonolysis was found to Additional evidence has been given by Franklin (62) proceed as follows, in support of the view that hydrazoic acid is an ammononitric acid. Various methods of preparing hydraPels 8NHa P(NH)%NH* 5NHdCI. wic acid and its salts, and the reactions of this acid When heated, the phosphorus compound so formed with certain metals, halogen acids, inorganic, and orloses ammonia, forming phospham, and, finally, phos- ganic compounds have been described. The formaphorus nitride, tion of potassium azide by the action of potassium amide on potassium nitrate in liquid ammonia is particularly worthy of note,
-
+
-+ - +
+
+
+
+
."
+
-
+
KONO*
+ 3KNH.
-
KN:N i N
+ 3KOH 4-NHI
The highest yield of potassium azide was obtained by heating the reactants in a sealed tube to 130' to 140°C. in the presence of liquid ammonia. The above reaction involves the ammonolysis of potassium aquonitrate to potassium ammononitrate. Cupric hydroxide has been found to be insoluble in liquid ammonia (63) and it has been suggested that this may be due to the dimerization of ammonia in the liquid state. The vapor pressures of systems consisting of kaolins or clays in liquid ammonia have been shown to undergo a gradual decrease (64). This is indicative of an absorptive retention of the ammonia by the solid phase. I t has been found that liquid ammoniaremoves water only from the allophane portion of kaolin. It is interesting to note that Taylor and Jungers (65) have prepared an ammonia containing 99% deuterium. This ammonia was found to have a boiling point of -30.9"C. (-33.45"), a freezing point of -74°C. (-78'), a vapor pressure of 63 mm. at -70.9"C. (77 mm.). and a latent heat of vaporization of 5990 cal./mol (5797 cal./mol). The values given in parentheses are those given by Taylor and Jungers for ordinary ammonia. Kameyama and Mori (66) have reported that the end result of the electrolysis of sodium nitrate in liquid ammonia at -40°C. may be expressed by the equation, 3NaNOs
+ 2NHa
-
3NaNOz
'+ N2 + 3H20
While the formation of sodium nitrite was found to be the principal cathodic reaction, small quantities of alkali and hydrogen were also formed. The above workers evidently observed no further reactions involving the sodium nitrite. Earlier work by Maxted (67) has shown that "disodium nitrite," NanNOa, is deposited on the cathode during the electrolysis of sodium nitrite in anhydrous liquid ammonia. The same product was obtained by treating sodium nitrite in liquid ammonia with a liquid ammonia solution of sodium. Similar observations have been recorded by Zintl and Kohn (68). A liquid ammonia calorimeter, an%its application to the determination of heats of solution and heats of reaction in liquid ammonia have been described by Kraus and Ridderhof (69). They have determined the heats of solution of a number of inorganic salts and the heat of solution of metallic sodium. The heat effects accompanying certain reactions have been measured and from these data the heats of formation of a number of compounds have been calculated. Kraus and Prescott (70) have made improvements on the above method and bave determined the beats of solution of a number of salts not investigated by Kraus and Ridderhof. In addition, they have checked some of the data given by the latter and offer values which they believe to be more nearly correct. In general, heats of solution in liquid ammonia are positive, while in water the corresponding values are negative. Kraus and Schmidt (71) bave extended this work and have given values for the heats of solution of lithium, sodium,
potassium, sodium bromide, ice, acetamide, urea, and phthalimide, and for solutions of sodium in the presence of potassium bromide, and of potassium in the presence of sodium bromide. The energy changes involved in the reduction of trimethyltin bromide to trimethyltin and to sodium trimethylstannide have been determined. From these values, they have calculated the energy of bromination of trimethyltin and the energy of reduction of trimethyltin to the corresponding negative ion. Pressure-temperature-concentration relations in the binary system, hydrazine trinitrideammonia, and in parts of the ternary system, hydrazine-hydrogen trinitrideammonia, over ranges of from 0 to 3200 mm.. -50" to 80°C., and 0 to 100% ammonia, bave been investigated by Howard and Browne (72). Only one solvate of the hydrazine salt was found, namely, hydrazine trinitride hemiammonate, 2N2HsNa.NHs. At temperatures greater than -g°C., this substance is ammonolyzed in accordance with the equation, 2NzH&kNHs
+ NH, S 2NHaNs + 2N2H4.
Below -go, the ammonolysis proceeds as follows, Griengl and co-workers (73) have made an extensive study of the conductivities and solubility relationships in the ternary systems, sodium-potassiumammonia and sodium-lithiumammonia, between -40" and -70% The conductivities of dilute solutions of sodium and potassium and of sodium and lithium in liquid ammonia are practically additive. No evidence was obtained, from conductivity data, for compound formation between the alkali .metals, but the solubility curve for the system, sodium-potassiumammonia (liquid) indicated the formation of the compound, NanK. The solubility of urea in anhydrous liquid ammonia has been determined by Scholl and Davis (74). The solubility was found to increase from 25.1 grams urea/100 grams NHa at -26.4', to'1024 grams urea/100 grams NHa at 10l°C. Evidence was found for the existence of the ammonate, CO(NHa)z.NHs. Solubility values and vapor pressure data were obtained at a number of points over the above temperature range. The solubility of hydrogen in liquid ammonia at 25", 50°, 75'. and 100°C., and-at pressures up to 1000 atmospheres, has been determined by Wiebe and Tremearne (75). Values ranging from 1.695 cc. Hz/gram NHa a t 25°C. and a pressure of 25 atmospheres, and 388.2 cc. H*/gram NH, at 100°C. and 1000 atmospheres, have been found. The solubility of hydrogen in liquid ammonia was found to increase with increase in temperature. The work of Guyer, Bider, and Schmid (76) on the solubility of the chlorides and nitrates of sodium and potassium in ammonia-water mixtures is to be noted. Their data, presented graphically, cover the temperature range of -40" to 25'C. Similarly, they have determined the solubility of calcium chloride octam-
monate (77) in ammonia-water mixtures and the solubility of sodium chloride in liquid ammonia in the presence of sodium nitrate. The metathetic reactions of calcium nitrate with sodium and potassium chlorides in ammonia and ammonia-water mixtures have also been described. The potentials of the cells, Zn (amalgam Nz), ZnClr6NH3(,,, NH4C1 (in NH3(1), f), TlCI(,), T1 (amalgam N,) a t 25'C. have been measured by Elliot and Yost (78). The cells are reversible and, with the zinc and thallium in the amalgams a t unit activity, give a standard potential, Eozss = 0.9016 volt. With zinc and thallium present as the pure metals, the standard potential is Enzos= 0.8293 volt. Using these and other data, they have calculated the free energies, heat contents, and virtual entropies for zinc chloride, its mono-, di-, tetra-, hexa-, and decammonates. They have reported that thallous chloride is soluble in liquid ammonia a t 25'C. to the extent of 0.0259 mdle/1000 grams NH3. Costeanu (79) has studied the cells, Cd, Cd(N03)2.4H20; NH4N03; AgN03, Ag, and Zn, Zn(N0&6Hz0; NH4N03;AgN03, Ag, in liquid ammonia and in ammonia containing 26.5%, 45.2%, and 80% water and a t temperatures ranging from -75' to 1 8 T . The addition of water produced only slight changes in the potentials of the above cells and but little change in potential was produced by rather large temperature changes. The electrode potentials of a number of metals in liquid ammonia a t -50°C. have been determined by Pleskov and Monoszon (80). Particular attention was given to the problem of obtaining pure anhydrous materials. Where anhydrous salts could not be obtained for use as electrolytes, the corresponding ammonates were used. The following values, referred to the half-cell, Pb, 0.1 N Pb(NO& in NH&) as 0 a t -50°, were found: Zo 0.848
Cd 0.522
Cu 0.103
Ag 0.472
.
Hg 0.414
HI -0.351
In contrast with the results obtained b y a number of earlier workers, Pleskov and Monoszon have apparently succeeded in obtaining a reproducible potential with a hydrogen electrode in liquid ammonia. They have found it necessary to allow sufficient time to elapse for the system to come to equilibrium before reproducible values can be obtained and that the length of time required is a function of the preliminary treatment of the electrodes. They have demonstrated that variation in the potential during this initial period is due to the passivising of the electrode by oxygen, and have been able to reduce this effect to a minimum. Due to the lack of certain essential data, the above workers were unable to calculate accurately the activity coefficients for the ions in liquid ammonia solution. However, in the following table, values for a (the approximate ion activity coefficient in 0.1 N solutions), the (the normal electrode potential when Pb=O) and for comparison EH,O (the normal potential in aqueous solutions when Pb =0) are given:
Vasil'ev and co-workers (81) have reported values for certain optical constants of the crystalline ammonates of calcium chloride and cyanide and sodium chloride and cyanide. The ammonates of NaCN were found to be stable only a t low temperatures. They have determined the mutual solubility of sodium chloride and sodium cyanide in liquid ammonia between -41' and 10°C. The solubility of sodium cyanide in liquid ammonia was found to increase with rising temperature. Similarly, Portnov and Vasil'ev (82) have determined the solubility of calcium nitrate and of barium nitrate in liquid ammonia. The solubility of both salts was found to increase with rising temperature and the tetrammonates of both salts have been identified. In connection with their studies on the physical relationships among the hydrides of the elements of the fifth group of the periodic arrangement, Robinson and co-workers (83) have collected numerous data of interest to those working with liquid ammonia. They have, for example, given values for the surface tension and parachor of ammonia a t temperatures ranging from -56' to -33°C. While all of the values given cannot be reproduced a t this time, it is of interest to note that they have obtained considerable evidence which indicates that ammonia is largely bimolecular a t its boiling point and still more hixhly .~. associated a t lower temperatures. Hunt and Larseu (84) have described a method for measnrine the vaoor Dressures of liauid ammonia solutions and have presented vapor preSsure data for liquid ammonia solutions of ammonihm nitrate, chloride, bromide, and iodide. These measurements, made a t 25'C., cover a wide range of concentrations. The heat conductivitv of liauid ammonia over the temperature range, -15' to 30°C., has b ~ e ndetermined by Kardos (85) and Oldham (86) has published a new chart embodying the latest data on the thermodynamic properties of ammonia. In a preliminary report on his studies of acid catalysis in non-aqueous solvents, Shatenstein (87) has presented data which show that the reaction between santonine and liquid ammonia is markedly catalyzed by either ammonium nitrate or ammonium chloride (acids in liquid ammonia). This reaction, which involves the opening of the 7-lactone ring, proceeds as follows,
. .
% ,
CHI I H C C-O--C:O
/\/\
O:C
C
I
CH-CH-CH.
CH8H OH
/\/\
0:C
C
CONH?
I
CH-CH-CHI
The course of the reaction was followed by the polarimetric method. Apparatus and technic applied in the
above work have been described. Shatenstein has indicated his intention to continue the study of a d d and basic catalysis in liquid ammonia and other nonaqueous solvents, using as catalysts substances which should, on the basis of modern theories of acidity, function as adds and bases in these solvents. A method for effecting reactions in liquid ammonia, in cases where one or more of the reactants is only sparingly soluble, has recently been patented (88). Details of methods and apparatus for the manufacture of NHeCOOK from KC1 and NH2COONH4, and of (NH&S04 and Ca(NO& from CaS04and NHINO~are described. Two patents (89, 90) covering methods for the purification of sodium chloride have been issued. The pure chloride is extracted from the crude salt with liquid ammonia a t or near the temperature a t which sodium chloride exhibits its maximum solubility, namely, -9.5'C. Another recent patent (91) covers a method for the preparation of pure sodium nitrate. This preparation is accomplished by treating sodium chloride in liquid ammonia with an alkaline earth nitrate. A patent bas been issued to Pranke (92) on a method for the preparation of sodium cyanide from sodium and crude calcium cyanamide or from sodium calcium cyanide. In either case, the sodium cyanide formed is separated from the other reaction products by extraction with liquid ammonia. Still another recent patent (93) is concerned with a method for the prepara-
tion of concentrated solutions of aqueous ammonia by the addition of liquid ammonia to water, or preferably to water containing not less than 25% NHs. The use of liquid ammonia in the extraction (94) of substances such as sugar, nicotine, caffeine, etc., from animal or vegetable materials and in the preparation of amino nitriles (95) has been patented. An example of the latter is the preparation of 8-aminopropionitrile by heating acrylic nitrile with liquid ammonia in an autoclave. A review of new methods for the use of liquid ammonia in industrial processes has recently appeared (96). It has been found that urea may be prepared in a 50-60% yield by the reaction between carbon dioxide and liquid ammonia (97). This reaction has been effected by heating the reactants in an autoclave a t 175-85' for approximately one hour. The use of liquid ammonia solutions of sodium in the research laboratory has been discussed briefly and an apparatus in which reactions involving such solutions may be carried out has been described (98). Numerous interesting analogies between the corresponding members of the nitrogen and oxygen systems of compounds have been pointed out by Franklin (99), and certain reactions in liquid ammonia have been discussed by Audrieth (100) in his review of the chemistry of hydrazoic add and its inorganic derivatives. An elaborate compilation of solubilities of inorganic substances in non-aqueous solvents, particularly liquid ammonia, has been made by Shatenstein (101).
LITERATURE CITED
(49) G. SFACUAND P. SPACU,ibid., 217,804 (Feb., 1934). (50), G. SPACU.P. SPACU.AND P. VOICAESCU, ibid., 217. 33-45 (May. 1934): (51) E. GRUNER,zbzd., 2 (52) J. BJERRUM.Kgl. Medd., 12,67 pp. (1934). AND H. GIESE,Ber., 67,110&15 (June, 1934). (53) R. SCHWARZ (54) W14Cd_Fym:ys AND G. W. WATT,J. A n . Chem. Soc., 55,
.
z-0 ( , 3 . , . , , .
(56) (57) (58) (59) (60) (6ij (62)
.o..m s u AND P. ROCQUET,Compt. rend., 197, 1643-5 [1833). AND H. E. SCHWARZKOPP, 2. anorg. allgem. W. STRECKER Chem., 221, 193-8 (Dec., 1934). W,, STRFCKERAND C. M m , ibid., 221, 199-208 (Dec., (1934). ibid., 221, 113-23 (Dee., 1934). F. W. BERGSTROM, T. H. VAUGHN, R. R. VOGT,AND J. A. N I E ~ A N DJ., Am. Chem. Soc., 56, 2120-2 (Oct., 1934). C. M. LOANE.3.Phw. Chem.. 37.615-22 (1933). H. L. KAHLEU, JR., h i v a t e com&unicati&. E. C. FRANKLIN,J. Am. Chem. Soc., 56, 568-71 (Mar., ~
~
TR34)
( 6 2 ) G. Errrscn. E. HELLEEIGEL. AND D. KRfkER.. Ber... 67.224 .
(Jan., 1934).
(71) (72) (73) (74) (75)
Z. anorg. allgem. Chem., 215, 1-18 (1933). E. GRUNER, AND J. C. JUNGERS,3.Am. Chem. Soc., 55, H. S. TAYLOR 5057-8 (1933). N. KAMEYAMA AND H. MORI, J. Soc. Chem. Ind. (Japan), Suppl. Biding, 37, 167-8 (Apr., 1934). E. B. M ~ T E DJ., Chem. Soc., 111,1016-9 (1917). E. ZINTLAND 0. KOHN,Ber., 6lB, 189-99 (1928). C. A. KRAUSAND 1. A. RIDDERROP. 3.Am. Chem. Soc.. 56, 79-86 (Jan.. 193k). C. A. KRAUSAND R. F. PRESCOTT,ibid., 56, 8 G 8 (Jan.. 1934). C. A. KRAUSrn F. C. Scwarnr, ibid., 56, 2297-300 (Nov., 1934). D. H. HOWARD, JR., AND A. W. BROW, ibid., 56,2348-57 (Nov., 1934). AND K. STEYSKAL, Monatsh.. 63, F. GRIENGL, F. STEYSKAL, 394-426 (Jan., 1934). AND R. 0. E. DAVIS,Ind. Eng. Chem., 26,1299W. SCROLL 301 (Dee.. 1934). J. Am. Chem. Soc., 56, R. WIEBE AND T. H. TREMEARNE, 23.57-60 (Nov., 1934).
\----,.
(78) N. ELLIOTAND D. M. YOST,J. Am. Chem. Soc., 56, 1057+0 (May. 1934): for corrections see ibid.. 56. 2797-8 (DC;., 1934): Compt. rend., 197, 11134 (1933). (79) G. I. COSTEANU, (80) V. A. PLESKOVAND A. M. MONOSZON, 3. Phys. Chem. (U. S. S. R.), 4,696-702 (1933). (81) B. B. VASIL'EV,J. L. ETTINGER, AND M. P. GOLO'N~OW, Z. anorg. allgem. Chem., 219,341-7 (Sept., 1934). AND B. B. VASIL'EV,ibid., 221, 149-53 (82) M..-A. PORTNOV (Dee., 19J4). J. (83) A. A. DURRANT,T. G. PEARSON.AND P. L. ROBINSON, Chem. Soc., 1934, 730-5; see also T. G. P E ~ S O AND N P. L. ROBINSON, ibid., 73643. (84) H . HUNTAND W. E. LARSEN.J. P h ~ s .Chem... 38.. 801-7 (June, 1934). 2. ges. K d t e I d . , 41, 1-6.2935 (1934). (85) A. KARDOS, (86) B. C. OLDHAM, Ice and Cold Storage. 3 7 , 9 6 7 (1934). J. Phys Chem. (U. S. S. R.), 4, 703-5 (87) A. I. SRATENSTEIN,
. .
(88) (89) (90) (91) (92) (93) (94) (95) (96)
11022> ,A"--,. I. G. FARBENIND. A,-G., Ger. Pat. 600,485, July 24, 1934. R. S. HARA,Brit. Pat. 401,612, Nov. 16, 1933. A s m GARASU K. K., Fr. Pat. 757,080, Dec. 20, 1933.
A. GUYERAND A. BIELER,FT.PQ1.763,769,May 7,1934. E. J. PRANK&U. S. Pat. 1,947,570, Feb. 30, 1934. Dutch Pat. 32.748, May 15. 1934. I. G. FARBENIND. A,-G., Ger. Pat. 596,091, Apr. 26, 1934. I. G. FARBENIND.A.-G., Brit. Pat. 404,744, Jan. 25, 1934. B. B. VASIL'EV,AND M. A. PORTNOV, G. I. VoiN~~ovrcn, J. Chem. Ind. (Moscow), 1934, 43-5. AND A. N. POPOVA, J. Chem. Ind. (Moscow), (97) B. A. BOLOTOY 1934, No. 9, 1934, No. 9.32-8; see also B. I. LEVI,iw., 3-3. (98) H. N. GILBERT,N. D. SCOTT,W. F. Z ~ E R L IAND , V. L. HANSLEY, Ind. Eng. Chem., 25, 73-1 (1933). (991 E. C. FRANKLIN. Chem. Rev.. 14. 219-50 AD^.. 1934).
Division, 1934.
.
