THE PREPARATION OF TRIMETHYLOX AMINE AND ISOMERIC ALKOXYL DERIVATIVES BY JAMES HERBERT HIBBEN B. S. University of Illinois, 1920 THESIS Submitted in Partial Fulfillment of the Requirements for the Degree of MASTER OF SCIENCE IN CHEMISTRY IN THE GRADUATE SCHOOL OF THE UNIVERSITY OF ILLINOIS 1922 * UNIVERSITY OF ILLINOIS THE GRADUATE SCHOOL January 21 j gg_ I HEREBY RECOMMEND THAT THE THESIS PREPARED UNDER MY SUPERVISION BY JAi/iES HERBERT BIBBER ENTITLED THE PREPARATION OP XRIkEIHYLOXAMINE ADD ISOkERIC ALKOXYL DERIVATIVES BE ACCEPTED AS FULFILLING THIS PART OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OE. SCIENCE KJn • y In Charge of Thesis nr, J. ^ Head of Department Recommendation concurred in* Committee Final Examination* ‘‘Required for doctor's degree but not for master's v Digitized by the Internet Archive in 2016 https://archive.org/details/preparationoftriOOhibb TABLE OF CONTENTS 1. Introduction. . . ...... 1 11. Historical 10 111. Experimental 12 IV. Conclusion 28 V. Bibliography 29 VI. Acknowledgments 30 1 . 1. INTRODUCTION. The ultimate purpose in the study of the amine oxide derivatives- more specif icially the isomers trimethylmethoxy- ammoniumhydroxide , (CH 3 ) 3 NtOCH 3 )OH, and trimethylhydroxy - ammonium methylate, (CH 3 ) 3 N(0H)0CH 3 , - is an investigation^ the phenomena involved with reference to the oxygen of the methoxy groups on the electrolysis of these compounds. If it can oe demonstrated that the methoxy group will function in one case as a negative ion, and in another as a positive ion this will contribute some evidence toward the explanation of the isomerism. Tho the experimental data thus far obtained deal only with the preparation of these isomers, a resume" of tne structural representations advanced is necessary for the consideration of any possible formulae of the resultant comp- ounds, as well as any isomeric interpretation. In 1899 Dunstan and Goulding 1 prepared trimethylmefch- oxyammoniumhydr oxide by the interaction of methyliodide and trimethylamineoxide and subsequent conversion of the ammonium iodide salt thus formed to the sulphate which was treated with barium hydroxide. Heisenhelmer 8 in 1913 by treating 2 . trime thylmethoxyammoniumiodide with silver oxide obtained the same compound. However, by treating the trime thy lhy dr oxy- chloride with sodium methylate he obtained not trime thy lmeth- oxyammoniumhy dr oxide but tri me thy lhy dr oxy ammonium methylate. The reactions and decomposition products are as indicated: ,0CH 3 1. (CH s ) 3 N~0 + CH 3 I — > (CH 3 ) 3 Nf V .OGH: 2. (CH 3 ) 3 N 7 T / ''OH + Ag s O— » (CH 3 ) 3 N 7 OCH. OH ,och 3 3. (KH 3 ) 3 n( + » (CH 3 ) 3 N + CHsO + H a 0 . b. (CH 3 ) 3 N=0 + HC1- OH (CH 3 ) 3 N. OH .OH OCH. 5. (CH 3 ) 3 N 7 + NaOCHa — > (CH^N 7 N C1 OH 6. (CH 3 ) 3 n( + (CH 3 ) 3 N=0 + CHaOH. \)CH. Thru these and similar reactions Meisenheimer came to the conclusion that, in order to obtain two sueh fundam- entally different isomers, one of the valences of the nitrogen atom was different from the other four. He rejected various possible explanations such as the Werner's ammonium formula 3 or oxonsznium formula discussed by Willstater , Hantsch and 4 . 6 G-ra.f , Yferner. In contrast to Werner, Meisenheimer assumed that all five radicals were bound by principal valences but '• '• ’ r '■ : ■ S : C . f ■ / . 3 that the fifth or different valence was in the outer zone. This was represented by: 1 . Fromrn however, maintained that Meisenheimer ' s ex- periments were not conclusive and capable of other interpre- tation. He said further that the decomposition of the two isomers into an aldenyde, amine, and water: amine oxide, and alcohol respectively was not sufficient to give absolute proof of the unique fifth valency. Fromm represents the struct- ure of the first compound as: /GH 3 (CH 3 ) 3 N=(X n oh. 7 Jones in discussing the electronic tautomer isn of hydroxylamine and its derivatives represents the structural and electronic tautomerism as follows: Structural: 1 > 2, 3 — > 4. Electronic: 1 — , 4~» 2. (1) + (Ii ) a -N-+0-H t > (H + ) 3 n±;o (2) jr jr (3) (H + ) 8 =N+-0- + H£— * ( H + ) a N+-0 (4) CH 3 ^ch 3 X N OH. CH. OCH. CH 3 ch 3 X N CH^ X 0H OCH- - . 4 . The OH group in formula (1) would act as a positive hydroxyl group giving tautomer (2) which would readily dissociate into active oxygen and ammonia, as is demonstrated by the oxidation of ferrous hydroxide to ferric hydroxide by hydroxylamine and the conversion of the latter into ammonia# Jones further supposed that if the hydroxyl group could f 'unction positively in this case then, provided that one hydroxyl group was positive and the other negative, in the compound: ~t~ ♦ /OH { R ) 3 =N. _ there should be two V 0H isomers ( electromers ) and compounds of the type: Rx -t- \ OK Rs + _iK ___ should exist in ,/ N 0H r 3 sterioisoraeric modifications. There exists the possibility that the optical isomerism might persist even to the corresponding amine oxides - (RiR s R 3 + ) K +-0. That the latter deduction is correct has been demonstrated by e Meisenheimer who resolved sueh an amine oxide into enantiomorph- ous modifications. Therefore, Jones asserted, the key to the disputed formulae of trimethylmethoxyammoniumhydroxide and trimethyl- hy dr oxy ammonium methylate lies in the electronic conception ’ 5 . by assuming that, in one case, the fourth valency is linked to a positive hydroxyl and the fifth to the negative meth- oxyl, and in the other, the linkage is to the positive methoxyl and negative hydroxyl groups respectively. This arrangement with the decomposition products is depicted thus: +, - + 0-+CH 3 C4) ( Cri 3 ) 3N+-O-+H k 5 ) “ — : ► (CH a + ) 3 N+±8-4.H + H 3 c"-0 8 TT (4) ( GH 3 + ) 3 N+i8-+§H a (5) —4 (ohJ) 3 n+±o + CH 3 +-0-+H. d Jones points out again that the isomers (1), (2 ), are alike structurally but different electronically and that the only difference in preparation lies in the order by which the groups are introduced. 10 Michael in 1920 stated that the acceptance of either Meiaenheimer ' s or Jones’ interpretation involved far reaching modifications in the present conception of valence, that isomeric trialky lhydroxylammonium salts and tri^alkyl- oxyammonium derivatives are not known, and there is, therefore, no experimental evidence in this field to sup >ort any of the foregoing hypotheses. The formula ascribed by Michael for the compound resulting from the reaction of methyl iodide and the trialkylamine oxide is obtained thus: 6 . ?/ H /O (CH 3 ) 3 N=0 + CH 3 I > (0H 3 ) 3 Nr-0 + — > (GH 3 ) 3 NQ +H s 0— 4 \/ GH 2 CH S (CH 3 ) 3 NHOH + CHgO > (GH 3 ) 3 N + (JHgO + H s O. In consideration of the Lewis -Langmuir theory of valency the following structures might be assigned to the compounds trimethylmethoxyammoniumhydroxide and trimethyl- hy dr oxy ammonium methylate, using the electronic conception of the isomerism. CW. h 3 c + j ( 2 ) h 3 g + ; H 3 C :N:0 :H ' 0:CH a H 3 G :N:OCHo 1 0 :H I I h 3 c * ! h 3 V ! Wherein it becomes apparent that, in the second case, the oxygen of the methoxyl group shares two electrons with the nitrogen , and the oxygen of the hydroxyl group shares two electrons only with the hydrogen, while the converse is true in case (lj , that is, the nitrogen shares two electrons with the oxygen og the hydroxyl group, and the oxygen of the methoxyl group shares electrons only with the CH 3 group. 7 . It may be readily seen, therfore, that the degree of attach- ment of these groups to the nitrogen is different. In the second isomer the methoxyl group is bound by a polar and 11 in the first by a non-polar valence. Hence isomer ( 2 ) would be expected to ionize into negative (OH) and positive ( CH 3 ) 3 N(OCH 3 ) , and isomer (1) would furnish the ions negative (0CH 3 ) and positive (GH 3 ) 3 N(0H) . It is evident that the methoxy group functions, if this representation is correct, in one case negatively and in the other, positively, or in the last analysis, the oxygen behaves normally on one hand and on the other as if it were neutral or more positive than normal oxygen, and with consequent greater oxidation capacity analagous to the oxygen of hydrogen peroxide , such an analogy having been demonstrated in the case of the amine oxide by 12 Hantzsch and Hillard . The decomposition, therefore of the second isomer into the amine, aldehyde and water compared with the decomposition of the first isomer into the amine oxide and methyl alcohol might be expected. From this electronic configuration it may also be seen that the basisity of the trimethylhydroxylammonium methylate will be less than that of the trirne thy line thoxyl- ammoniumhydroxide . If comparison is made between tetra- methylammoniumhydroxide and trime thy lmethoxy ammonium methylate 8 . the Lewi 3 -Langmuir conception of valency phenomena leads to a possible explanation of the lesser basisity of trie latter compound, which might ue raised as an objection to the electronic representation. and monochloracetic acid, the influence of the increasingly electronegative groups , ( GH 3 — »H, H — * Cl ) is to make a stronger acid, or in other words, to cause the hydrogen of the hydroxyl group to ionize as a hydrogen ion to a greater and greater degree. Hence there is a seeming displacement of the electrons holding the hydrogen, toward the carbon atom. Similarly, therefore, the influence of the oxygen of the methoxyl group, which is more electronegative than the GH 3 group of the t e tr ame thy laurnoni urn compound as: would be to cause some electronic displacement between the carbon atom and one of the hydrogens of the methyl group as follows: TT and consequent tendency of the hydrogen thus less closely bound to neutralize the ionized hydroxyl group making this compound less basic than the tetramethylammoniumhydroxide. Considering such organic acids as acetic, formic, (CH 3 ) 3 N-CH a 'oh and H =N--0 :C £:H H . . 9 . The foregoing discussion reviews in brief various representations of the amine oxide derivatives. The motive has been several fold: first, a better understanding is obtained of the reaction phenomena involved in their prepara- tion, and second, as has been pointed out, the ultimate purpose in the preparation of these compounds is to obtain experimentally some evidence that may contribute tov/ard the substantiation or refutation of the electronic conception of the isomerism. Hence the previous discussion outlines to some extent the scope and point of vie?/ in such an attempt. HISTORICAL. 10 , 2 . In an attempt to prepare B- alkylhydroxylamine Lobry 13 de Bruyn in 1894 treated hydroxylamine with methyl iodide obtaining the salt CH 3 NH0H*HI according to his analysis. 14 Duns tan and Moulding during the same year while repeating de Bruyn* s work obtained trimethylhydroxylaninehydroiodide as a result of the same reaction. In addition the sulphate, hydrochloride and free trimethylhydroxylamine were prepared 1 5 from the original hydroiodide. Subsequently Duns tan and Goulding found that in addition to the trimethylhydroxylamine- hydoiodide, a mixture of the mono, di , and tri hydroiodide - hydroxylamine (NH a 0H*HI, (NH 8 0H) 3 HI, (NH s 0H) 3 HI ) was also is formed. This was substantiated by de Bruyn who obtained similar results in 1896 , admitting also the error of his first analysis. _ 17 In 1399 Dunstan and G-oulding investigated the prop- erties of the trialkyl salts and of the free trimethylhydroxyl- amine or, correctly, trimethylamine oxide as the derivative obtained in this reaction is not a true hydroxylamine der- ivative, but an isomer of it. They criticised the work of is Hantzsch and Hillard who by a similar reaction obtained a compound which they designated as a carbonate instead of the 11 . amine oxide. In addition Duns tan and Moulding made the tri- methylmethoxyammoniumiodide salt and trime thy lmethoxy ammonium - hydroxide at this time. The next important advancement was the preparation 18 of the trialkylamine oxides by the simple oxidation of trialkylated amines by these same investigators a few months 19 later. This was repeated by Meisenheimer in 1913 who syn- thesized in addition several trimethylalkyloxyammoniuniiodide salts in addition to the two isomers, trime thy line thoxy ammonium- hydroxide and trimethylhydroxyammonium methylate and other similar alkyl substituted compounds. . 12 . 3. EXPERIMENTAL. The procedures employed in the preparation of these compounds which have been chronologically considered, may be briefly summarized by the following equations: 1. H s NOH +3CH 3 I V (GH 3 ) 3 N^ + S-WI. .OH OH , s OH (GH 3 ) 3 N( 2. 2 (CH 3 ) 3 N/ + Ag 8 S0 4 » V S0* + 2AgI. N I (ch 3 ) 3 n/ OH 3. (CH 3 ) 3 N(OH) ) 3 S0 4 + Ba(OH) a — > 2(0H 3 ) 3 N=0*2H S 0 +BaSO. 4. ( CH 3 ) 3 N=0 • 2HgO + (GH 3 ) 3 N=0 + 2HgO. 5. (CH 3 ) 3 N + H 8 0 s +H 8 0 — l> ( SH 3 ) 3 N=0* 2H 8 0 OH 6. (CH 3 ) 3 N=0 + HC1 ( CH 3 ) 3 N^ XJl 7. (CH 3 ) 3 N=0 + ch 3 i — s> (gh 3 ) 3 n/ ,OCH. Vi OCH 3 och 3 8.2(CH 3 ) 3 N' + Ag 8 0-^2(CH 3 ) 3 N( + 2AgI. N I '' 9* (GH 3 ) 3 N^ + NaOCH 3 — > ( CH 3 ) + NaCl. OH OH OH \ Cl bGH. 10. (CH 3 ) 3 N(0H)0CH 3 + — > ch 3 oh + (ch 3 ) 3 n=0 13 . 11. (CH 3 )3N(0CH 3 )0H 4 . -i-* (CH 3 ) 3 N +H a 0 + GH s O. For the purpose of clarity in the consideration of the preparation and properties of these compounds, each individual will be considered seperately, the procedure i? being as outlined by Dunstan and Moulding and by 19 Meisenheimer . / 0H (ch 3 ) 3 n; x i OH HgNOH +3CH 3 I — > ( GH 3 ) 3 Ir + 2HI . N I Methyl iodide was added to a dilute solution of hydroxyl amine in methyl alcohfol prepared from the interaction of sodium methylate and hydroxylaminehydrochloride . The hydrochloride salt was obtained from treating hydroxylamine sulphate with barium chloride. After the addition of the methyliodide a white precipitate separated, which on reflux- ing the mixture for a short time went into solution. The refluxing, while preventing the formation of (NH 8 0H*HI, (NHgOH)gHl, and (NH 8 OH) 3 HI ) in as large a quantity as would otherwise occur, also results in the separation of free hydriotic acid due to the partial decomposition of the salts even at thi3 temperature. 14 . The excess of methyl alcohol was distilled under 18 mins pressure and the trimethylhydroxyainmoniumiodide separated from the remaining hydroxylaminehydroiodide salts by means of fractional recrystallization with ether. The salt thus obtained was purified by recrystallization from hot alcohol and ether. The theoretical yield from 15 gms. of hydroxylamine was 91 gms. A yield of 8.8 percent was obtained. Decomposition began at 95* and became total at 125*uc. Noticeable decomposition was a,parant on exposure to sunlight. On a second attempt a yield of 17.2 percent was obtained. The higher yield in the second case being attributed to less decomposition during a shorter period of refluxing. A pur- ified sample of the salt almost totally decomposed on ten days standing. .OH (0H 3 ) 3 N( \so 4 ( CH 3 ) 3 v/ \)H 2 (GH 3 ) 3 N^ + Ag 8 S0 4 ->( ( CH S ) 3 N0H ) 3 S0 4 +2Ag£. N I Trimethylhydroxylammoniurnhydroiodide was added to a solution of silver sulphate in water, the precipitate of silver iodide filtered, and the aqueous solution evaporated under I 15 . diminished pressure. The residue was dissolved in absolute methyl alcohol, separating any excess of silver sulphate, clarified with Eastman's special bone black and reprecipit- ated on concentration of the solution, by the addition of ether. The resulting sulphate crystals were white, very sol- uble in water, less soluble in ethyl alcohol, and insoluble in ether. The crystals decomposed at 155 * uc. and react with barium carbonate to yield carbon dioxide, the free base, and barium sulphate. After several purifications to get a constant decomposition point, the yield was 40 percent of the theoret- ical. Later this compound was prepared from the hydrochloric of tri methyl amine oxide in a similar manner but with an unsatisfactory yield. The hydrochloride lends itself to pur- ification with greater facility than the hydroiodide and for that reason was selected. However, on evaporation even under diminished pressure of the aqueous solution, after filtering off the silver chloride the sulphate suffered no little dec- omposition. The excess silver sulphate might plausibly act as a catalyst in such a decomposition. An analysis of the compound resulted in the following: 1 6 . ( (CH 3 ) 3 N0H) 8 S0 4 . 1 . Calculated Percent S0 4 in substance . . . 38.67. Weight substance. . .0.2282 gms. Weight BaS0 4 . . . 0. 2453 gms. Percent S0 4 found a 44.2. 2 . Calculated Percent S0 4 in substance 38.67. Weight substance ..0.2925 gms. Weight BaS0 4 .. 0.3187 gms. Percent S0 4 found 44.75. Calculated molecular weight 248. Molecular weight from analyds.. 216 . A third method of preparation of the sulphate is by the direct action of sulphuric acid on trimethylamine oxide. This, despite the difficulty in obtaining the pure amineoxide* hydrate, has resulted in a purer compound than the foregoing methods. The sulphate does not lend itself easily to analysis. ( CHo ) 3 N=0 , ( (CH 3 ) 3 N0H) 3 S0 4 + Ba( OH) 2 — > ( CH a ) 3 N=0- 2H S 0 + BaS0 4 ( CH 3 ) 3 N + H§0 + HgOg } ( CH 3 ) 3 H=0 * 2HgO . 17 . This compound may he prepared in two ways. The first method adopted was hy the interaction of the previously des- cribed sulphate with barium hydroxide, the excess barium hydroxide being precipitated by passing carbon dioxide thru the solution. The aqueous solution was distilled under 10 6ms pressure. The residue was then dissolved in absolute methyl alcohol and recrystallized with ether or from hot ethylic al- cohol solution. The resultant product is x the amine oxide with two molecules of water of crystallisation. The best yield obtained was 31.2 percent. The hydrated form crystallized in radiating needles with a melting point of 96 ' . The substance is very deliquescent, very soluble in water and also in methyl alcohol, less soluble in ethyl alcohol and insoluble in ether. It will not reduce Fehlings solution, reacts alkaline to litmus and methyl orange, will not liberate I from KI and quickly oxidizes a solution of ferrous sulphate to ferric hydroxide. Dunstan and Moulding found, in addition, that this compound would react with benzyl chloride to form benz- aldehyde and trimethylamine. 13 Hantzsch and Hilland, on the contrary, obtained a compound which would reduce Fehlings solution, liberate I from KI and absorb carbon dioxide. They were led to the 13 . conclusion that the methoxyammoniumhydrate could exist only in an ionized state. The results obtained substantiate Hantzsch and Hi Hand in only one particular, that is, the disposition of the hydroiodide and me thoxy iodide salts of trimethylamine oxide to liberate iodine. A small quantity of the free trimethylamine oxide was obtained by subliming the hydrated form at .5mm pressure. The original hydrate was heated in a tube ( Plate 1 ) at 12^* until all bubbling had ceased. After the temperature then was raised to 150* for ten minutes, the hydrate was caied and subsequently reheated at 190-200* for one and one half hours. Subliminal ion began at 180* . Only after repeated attempts: however, were these results obtained. A small amount of impur- ity will cause the decomposition of the dehydrated form to begin at 17°'-l8 Q * instead of 205* (Meisenheimer used a pressure of 10-12 mm. His product decomposed at 2^8* ) The pure trimethylamine oxide is extremely hygroscopic and virt- ually impossible to obtain in any practical quantity. Phos- phorous pentoxide and calcium chloride were used as drying agents in the neck of the sublimination tube. Hence it may be seen that the preparation of the VLBTE J — 19 . anhydrous form by means of the foregoing procedure is not very satisfactory. The method outlined for the preparation of the hydrated form yields a net result of approximately two percent. For this reason, and because of greater facility 18 in the preparation, the hydrogen peroxide method was adopted. A three percent solution of hydrogen peroxide was added to a twenty nine percent solution of trimethylamine and the solution allowed to stand twenty four hours when, if the trimethylamine odor had not yet disappeared, more hydrogen peroxide was added. The large volume of water was evaporated off under l3 cms pressure until a viscious brown nonvolatile residue remained. This was redissolved in methyl alcohol and crystallized with ether. Several recrystallizatiors were necessary to obtain a corstant melting point of 95*. The compound ansv/ers the description previously given for trimethylamine oxidehydrate . The yield was thirty percent. It was noticed* that the heating effect of the reaction drove off the trimethylamine so that by connecting the mouth of the vessel in which the solutions were allowed to stand to a curved mercury column thus allowing for expansion and prevent- ing the escape of the trimethylamine . The maximum yield was eighty percent. ( Meisenheimer 90-95 ) . 20 . (CH 3 ) 3 N .OCHs i (GHs) s N=0 + CH 3 I » (OH 3 ) 3 N v / V ,OCH« The above equation represents the use of the dehydrated, base. In the actual experiments, however, on account of the difficulty in obtaining and handling the anhydrous forin^ the hydrated base was used throughout. There was no appreciable derogatory effect. A solution of the tri methyl amine oxide in water is not used in these reactions because of the difficulty encountered in recovering the soluble products without decomposition. Hence an alcoholic solution of trimethylamine oxide is added to methyliodide and the mixture allowed to stand in a sealed jar for forty eight hours. The methoxy- iodide , which is soluble in water, hot methyl alcohol and insoluble in ether, was crystallized from an alcohol ether mixture in the form of white plates. As a by-product tetramethylammoniumiodide is formed only to the extent of from five to ten percent with proper precautions. It is separated from the methoxy compound by dissolving the latter in the minimum amount of methyl alcohol. Tne slightest trace of impurity will cause the partial decomposition of !i 21 . the methoxyiodide even at room temperature. The decomposition ranged from 155'’ to 162**, ( Meisenheimer 162“) which by no means represents its stability as a whole. The maximum yield obtained was 53-6 percent. (Meisenheimer 40 percent ). The following analytical results were obtained: (0H 3 ) 3 N(OCHa)l. 1 . Calculated percent I in substance 58.58 Weight substance.. 0.9644 gms. Wt Agl ....1.0365 gms. Percent iodine found 58.10 2. Calculated percent I in substance 58.58. Weight substance 1.2032 gms. Weight Agl... 1.2951 gms. Percent iodine found 58.20 A sample of the precipitate insoluble in a small amount of methyl alcohol, presumably tetramethylainmoniumiodide, analyses as follows: (CH 3 ) 3 N-I. 1 Calculated percent I in substance 63 . 2 . Weight substance. 0.4152 gms. Weight Agl. 0.4744 gms. Percent iodine found 61.75 22 . The slightly too low a percent iodine from the theoretical might be due to the presence of some of the methoxy iodide s al t • (CH 3 ) 3 < Til (GH 3 ) 3 N=0 + HC1 » (GH 3 ) 3 n( X C1 The trimethyloxaminehydro chloride is prepared by the direct action of hydrochloric acid on the amine oxide. It is the most stable of all the salts obtained. The aqueous solution was evaporated over a steam bath to dryness, the residue was then crystallized from hot methyl alcohol by cooling and adding ether, or by repeated cooling and filter- ing with subsequent evaporation. The yield was 93 percent. The compound is soluble in water, hot methyl alcohol, and insoluble in ether. It melts without decomposition at 212* uc. (Dunstan and Moulding 205p-210‘ ) ( CH 3 ) 3 N (OCH 3 ) OH. ,och 3 2 (CH 3 ) 3 N' + Ag 8 0 .0CH 3 Ndh ►2 ( CH 3 ) 3 N ^ + 2AgI 23 . This compound has been prepared only in solution. The trimethylmethoxy iodide salt in dilute aqueous solution is added to a molecular proportion of silver oxide, the silver iodide formed filtered off, and, for identification purposes, the solution acidified with hydrochloric acid, wherein, on subsequent evaporation, there remains behind! the hygroscopic crystalline trimethylmethoxyarnmoniumchloride. The chloride can be determined by conversion to the chlor- 19 platinate. Meisenheimer found that the base decomposed completely o# evaporation of the solution and he quantatively determined the trimethylamine and formaldehyde formed on decomposition. As has been pointed out in the introduction, this compound and trimethylammoniumhydroxyl methylate are, if possible, to be electrolysed. For that purpose it is extrem- ely essential that they be obtained in a pure state. On account of the instability of the methoxy iodide, and the fact that the sulphates have uniformly been more stable than the iodides, an attempt was made to prepare the tri- methylmethoxy lammonium sulphate by converting the iodide with silver sulphate. This was not successful for after stirring for several hours and subsequent evaporation of the water solution under diminished pressure, noticeable decomposition 24 . ensued. The trimethylamine oxide was then treated with dimethyl sulphate to obtain similar results according to the following reaction: OCH s (gh 3 ) 3 n' 2 (gh 3 ) 3 n=o + (ch 3 ) 8 so 4 — b ;so 4 (ck 3 ) 3 n( n 0CH 3 The excess dimethyl sulphate was distilled in the apparatus (Plate 2 ) which had been previously ?/eighed. Phe resultant residue was a yellowish, viscious, hygroscopic , non crystalline mass. The procedure was repeated with similar results. Quantatively represented the results were: 1 . Calculated weight of substance from 10 gms ( CH 3 ) 3 N=:0 . 2H 2 C 12.4000gms. Weight Substance Found 11.7964 gms . final weighing was made after the complete removal of any excess dimethylsulphate and solvent by evacuating to a pressure of two millimeters. However, the residue being non crystalline would not, therefore, be a3 good a medium a3 the trimethylmethoxy- ammonium iodide for the prej^aration of trimethylmethoxyammonium- hydroxide. Hence this method was abandoned. ' . \ PLRTE 25 . By repeated recrystallisation of the (CH S ) 3N(0CH s )l- filtering the alcoholic solution directly into a large quan- tity of ether the quick precipitation preventing the occlusion of inpurities- a sufficiently pure sample of the methoxy iodide salt was obtained. However, it is only stable under ether and will decompose under warming or standing for several days. To an alcoholic solution of this sample at a temperature of -10 sufficient potassium hydroxide was added. The odor of _ ition 3 Loticeable even at this temperature. The solution was then distilled at .5mm pressure by means of the apparatus in. Plate 3. The distillate was , o xept at -15 and the aldehyde odor was only faintly apparant. Hydrochloric acid gas was passed thru the alcoholic solution of the distillate which was then evaporated leaving a chloride residue. The samples used were small and the yield not suffic- ient for analysis. The experiment will be repeated in a like manner, and also with the use of silver oxide in aqueous solution. is hoped that, with sufficiently reduced pressure and great cooling, the trimethylmethoxyammoniumhydroxide may be dis- ^j-llsd without decomposition and purified in this manner. 26. (CH 3 ) 3 N (OH) OCH a . (CH 3 ) 3 ( GH 3 ) 3 OH N \ >S0 4 \0H OH + 2 NaOCH 3 — > 2 (CH 3 ) 3 + NagSO^ v 0CH 3 The use of the trimethyloxamine sulphate in place 19 of the hydrochloride used by Meisenheimer seems more advisable because of the greater insolubility of the resultant sodium sulphate, consequently a larger portion may be filtered off before attempting the distillation of the trimethylhydroxyammonium methylate in the manner just outlined. The influence of impurities as catalysers toward the decom- position of the amine oxide derivatives has been pointed out. An attempt to distil this compound was made as ment- ioned. The distillate showed no aldehyde reaction with phenyl hydrazine. The work on the trimethylmethoxyammoniumhydroxide and trimethylhydroxyammonium methylate will be continued. Tnere is a possibility that if the distillation method of purification should prove unsuccessful, this might be ac- complished in the same manner by which Knorr isolated the 27 . enol fora of acetoaceticester— co .'ling with carbon dioxide snow. It seems proper to say that the foregoing results were obtained only after considerable familiarity with the compounds and the requisite manipulation had been acquired. 28. 4. CONCLUSION. The work of Duns tan and Moulding and Jacob Meisen- he&mer on the preparation of trimethylamine oxide and alk- oxyl derivatives has been repeated and substantiated. Slight variations in procedure have, in some cases, yielded improve- ment, while in others such attempts have been without pos- itive results. Sufficient quantaties of the derivatives have been prepared to allow further experimentation along the lines indicated. It is hoped that, with the possible purif ication of the trialkylmethoxyammoniumhydroxide and trialkylhydroxy- ammonium methylate, that these isomers may be subsequently electrolysed and, as has been pointed out, the results may throw some light upon the electronic conception of their isomerism. 29 . 5 BIBLIOGRAPHY. 1. Dunstan and Moulding. J.C.S., 75, 792, 139$. 2. Jacob Meisenheimer. Ann. ,397, 273, 1913. 3. Willstatter . B er . , 33, 1638, 1900. 4. Hantzsch and Graf. Ber. , 33, 2154, 1915. 5. Werner. (Neuere Anschauug en) 204-210, 19°9. 6. Emil ^romm. Ann., 399, 366, 1913. 7. Wm.L. Jones. J. Am. O.S. , 36, 1269, 1914. 8. Meisenheimer. Ber., 4l, 3973, 19°8. Ann., 335, 117,1911. 9. Wm.L. Jones. Science, 46 (69) 493, 1917. 10. Michael. J. Am. C. S. , 42,1232, 1920. 11. Bray and branch. J. Am. G. S. , 35, 1441, 1913. 12. Hantzsch and Hilland. Ber., 31, 2°53, 1898. 13*LQ&bry de Bruyn. Receuil Trav. , 13, 46, 1894. 14. Dunstan and Goulding. Proc., 10, 133, 1894. Chem. News. 69,308 15- Dunstan and Goulding. J.C.S. , 69, 839, 1896. 16. Lobry de Bruyn. Rec.Trav., 15, 135, 1896. 17. Dunstan and Goulding. J.C.S., 75, 793, 1399 18 . Dunstan and Goulding. J.C.S., 75, 1004, 1899. 19. Jacob Meisenheimer. Ann. 397, 285, 1913. 30 . ACKNOWLEDGMENTS. In concluding this thesis, the author wishes to express his appreciation of the assistance given by Dr. W. A. Noyes, under whose guidance this work has been carried out. By criticisms, suggestions, and interest taken in the problem, Dr. Noyes has contributed much toward the successful preparation of the compounds. In addition the author is indebted to Dr. Roger Adams, Dr. J. H. Reedy, and Dr. W. H. Rodebush for many kind suggestions and criticisms.