U . S. Department of Commerce Research Paper RPl909 National Bureau of Standards Volume 41, August 1948 Part of the Journal of Research of the National Burea u of Standards

Aliphatic Halide-Carbonyl Condensations by Means of Sodium By Edgar A. Cadwallader, Abraham Fookson, Thomas W. Mears, and Frank 1. Howard

As a part of an in vestigation of the synthesis of highly branched ali phatic hydrocarbons that is being conducted at the National Bureau of Standa rds for the National Advisory Committee for Aeronautics, the l\'avy Bureau of Aeronautics, and t he Army Air Forces, se veral compounds have been prepared by interaction of alkyl halid es a nd various carbonyl compo unds in the presence of sod ium. This reaction makes possible the sy nthesis of ce rtain highly branched compound s not easily obtain able by other m eans.

L Introduction cooling of the reaction mixture was effected by application of hot or cold oil baths. In some The usc of the GrignaI'd reaction for synthesis cases, particularly with low-boiling solvents, both of highly branched compounds is limited by un­ small necks of the flask were fitted with reflux I desirable side reactions, which become progres­ condensers with the elimination of the thermom­ . sively more pronounced as the degree of branching eter. of the reactan ts is increased. These side reactions Sodium sand was prepared batch-wise by heat­ involve the l'eclu ction and enolization of the adduct ing a weighed amount of sodium in a purified ligh t rather than the desired addition of the alkyl group. mineral oil until melted, and then shaking the The recent work: of Bartlett and Schneider [1) 1 liquid sod ium vigorously in a closed E rlenm eyer and previous work by Morton and Stevens (13) fl ask until the des ired state of subdivision was ob­ indicate that the reaction between an alkyl halide tained. After the particles of sodium had cooled and an ester or ketone undel' the influence of sodium sufficiently, they were washed by decantation may be used to prepare certain compounds more with three or four portions of the solvent to be easily than by the GrignaI'd reaction, or may be used, transferred to the reaction vessel, and used in some cases in which the Grignard reaction immediately covered with solvent. fails completely as a synthetic tool. In this paper, A few of the first reactions tried in this series are described the )'Csults of a few reactions that were conducted in ?i-pentane (boiling range 34° to were designed to give information concerning the 38° C), but later ones were conducted in isooctane 2 scope of the reaction and the effect of variou s (boiling point, 99° C). It wa found that t he solvents an d temperatures of reaction . The work vigorous reactions were more easily con trolled of Bartlett and Schn eider has been extended to when the higher-boiling solvent was used. Also include a few additional reactants. the usc of isooctane as a solven t facilitated the use of higher temperatures for the reactions. II. Experimental Bartlett and Schneider carried out their reac­ tions by adding all of the alkyl halide to the sodium The apparatus u sed in making small-scale runs at once and then adding the carbonyl compound was the conventional three-neck flask of appro­ over a period of time. In attempting to usc this priate size, fitted with a stirrer, reflux condenser,

and thermometer. A separatory funnel was 2 rrhis is the commonly used name for 2,2,4-Lrimcthylpentao (}, which is a placed at the top of the condenser. H eating and pri mary standard for octane number mea surements. rrhe commercially obtained material is exceptionally pure and se rves admirably as all inert I Figures in brackets indicate the literature references at Lbe end of this solvent for reactions of tbis type. Tbe co mmon nallle is used throughout paper. this paper.

Aliphatic Condensations by Sodium III technique in our applications of the reaction, we were equipped with the type of distilling hear! often found that the reaction was so vigorous recently described [7]. that the sodium was completely consumed before Solvents used for t he Teactions were removed all of the carbonyl compound could be added. by distillation thTough these columns; the residues Apparently in these cases there is a self-sustaining were di stilled '(generally at reduced pTessures) in reaction between the sodium and the alkyl halide a Podbielniak H yper-cal column or in a Claisen that needs only a starting impetus given by the flask fitted with a short fraction ating column. carbonyl compound. In order to minimize this Yields were calculated on the basis of the eff ect only a small portion of the alkyl halide was amount of carbonyl compound that was added added at the start of the reaction, and the re­ to the reaction. Table 1 summarizes data ob­ mainder was mixed with the carbonyl compound tained on the sever al products, together with for gradual addition during the reaction. results reported in the literature. After the reaction had been carried to comple­ tion, the mixtuTe was decomposed eith er by pour­ 1. Diisopropyl Ketone and n-Propyl Chloride ing it into ice water or by direct addition of water Sodium sand (69 g , 3 g atoms) was added to 1 to the mixture in th e :flask. If any sodium was liter of isooctane in the reaction vessel. The visible in the flask, th e treatment with water was contents was heated to 80° C, and a solution of preceded by th e addition of small amounts of 114 g of diisopropyl ketone (1 mole) and 157 g of ethyl alcohol. After complete hydrolysis, th e two n-propyl chloride (2 moles) was gradually added. layers were separated and the organic layer After the reaction began, the temperature rose to washed with several portions of water. The or­ 95° C, which caused considerable re:fluxing. The ganic substance was finally dried over a suitable characteristic 3 greenish-black color appeared, drying agent. In those cases where the products and addition of the ketone-halide solution was of reaction were well known, only those physical continued at such a rate as to insure vigorous constants that were deemed necessary to identify re:fluxing. AdditiOll time was approximately 2 the product were determined. hours. After standing overnight the reaction Some of th ese reactions have been used for mixture was hydrolyzed, and the organic layer large-scale synthesis; the eq uipment used in these was washed and dTied. Isooctane and unre­ cases has been described previously [8]. acted n-propyl chloride were removed by distilla­ Methyl isobutyrate, t-butyl chloride, methyl tion , and the residue was distilled at reduced n-butyrate, isobutyl chloride, n-propyl chloride, pressure (Claisen flask). There was obtained and n-butylchloride were all made by standard 40 g of recovered diisopropyl ketone, an d 51 g of procedures. Ethyl trimethylacetate and hex a­ 2-methyl-3-isopropyl-3-hexanol (3 2% yield). This methylacetone (2,2,4,4-tetramethyl-3-pentanone) carbinol has been prepared by George [5] in 36- were prepared by methods previously described percent yield by the reaction of n-pl'opyl-mag­ [8]. Pentamethylacetone (2,2,4-trimethyl-3-pen­ nesium haJide and diisopropyl ketone. The resi­ tan one) was prepared by a large-scale run of the due (50 g) was a viscious material (boiling point reaction between ethyl trimethylaeetate and iso­ > 200 ° C, nn20 1.4590), and presumed to be a. propyl chloride by the method described herein. bimolecular product from condensation of diiso­ Pinacolone (3 ,3-dimethyl-2-butanone) [8] was pre­ propyl ketone. pared by the oxidation of 3,4,4-trimethyl-2-pen­ tene [8] with sulfuric acid-dichromate solution in a 2. Diisopropyl Ketone and Isobutyl Chloride manner analogous to that used previously for the A solution of 155 g of isobutyl chloride (1.68 preparation of 4,4-dimethyl-2-pentanone [8]. The moles) and 96 g of diisopropyl ketone (0.84 mole) other materials were ob tained commercially. was diluted with 200 ml of isooctane and added Only constant-boilillg materials that were col­ 3 'rhis reaction is visualized us starting with the formation of a sodium lected from distillation in a 180 by 2.5-cm Whit­ ketyl as the first step. 'rhis ketyl is colored all intense dark blue, blue-gl'crll. more and Lux [16] column were used in the purple, or black, depending, it seems, Oil one Or a combination of sP \'eral factors. It ha.c; bee n noted t h:lt in sc\'rra l reactions tripd, whero rraetion synthesis. Several of these columns were used ; failed to occur, n o chal'actcri:;t ic color was obscn ·cd. It is significant that some were packed with glass helices, the others this color appears fi rst on the surfa C4? 01 t he sodium. At this stasc, if or, ly a very small amo unt of ketone is prm~cn t in relation to t he amollnt DC halide', were packed with stainless steel helices, but all the co\oI' often oisappcal's as the ketonr I'racts.

112 Journal of Research TABLE 1. Reaction products

Yleld pel I Refracti ve I I I _____~~ct i o_n______~~o~=______ort ;~~~~ __~::~:: __ I_i_I_1(_le_X_' _"_10_ _r __D_ e~_, ~_,:t_.y_, _ Referen ce

°C/mm H g 90 to 95/22 ______} ' 178 to 182/760 ______1. 44 39 0.850'1 D ii ~ op r opy l ke t one and n -propy l 2-m eth y l-3-isoprO I)yl- 3- hexalloL ___ __ 32 1 3 13£/125 ______} chloride. 18£/760 ______1·4444 .8554 [5]

105 to 11 0/38 _____ .. ___ .. _ f I. 44 30 0. 8486 180 to 185/i00 ______._ . . __ ) Diisopropyl ketone a nd isobutyl 2,5-dimethyl-3-iwpropy l-3 -hexanoL­ 14 1 11,0/125 ______•••• } chlorid e. .8470 [5] 190/760_. ___ •.••••• ______1.4423

III to 11 2/128 ____ • ______} I. 4397 0.8550 175/i60 ____ . ______.. --.- Penta melh ylacetone a nd met h y l 2,2,3,4-telramethyl-3- penLanoL ______43 bromide. 94/50 ____ ••••• _____ .•••• _ } 1.4405 .8565 [8] 179/760 _. ____ ••• _____ •.••

SO to 90/55 ____ ... __ . __ . __ J. 44 10 0.8555 29 { P enta mctll,'lacctonc a nd II1c tll yl 2,2,3 ,4-lelralllctby l-3 -pcnla noL - ----­ 94/50 ___ •••• ____ .. _. ____ _ 1.4405 .8565 (8) iodide. 93 to 97/35 ..... __ ._. ___ .. l. 4479 0.862 Pentamelh)'lacclonc and ethy l 2,2,4-l rimetb yl-3-cth yl-3- penlanoL -­ 44 {185 to I 9/76[J. . ______... I. 4481 .8628 [14) bromide . 188 to 19//760 __ ••• ______}

2,2,4 ·trimelhy l -3- isopropyl- 3 pen - I05 to 107/35 ___ .. ______._ 1. 4550 Peutamethy lacetoJ] c and i ~ op r opy l 1 42 { chloride. tanol. 103 to 104 /27. ____ • __ ._ •. _ I. 4569 [1)

H exem eth y lr.cetone 2G d l1 -propyl 2,2-dimcth y l-3-I·butyl-3-hcxa nol ' __ _ 38 117/22. __ .. _... __ • __ . __ .. 1. 4563 0.8595 cbloride.

]'inacololJ e and n-but yl c hloride {2, 2,3,6,6. penta meth,-I-a-hcPtcnC-5'} {I03 to 107/30 _ .. ---.----- 44 87 .8439 .. onc. --.----.. ' 83/10.5 _.. ------.-.. - .. - , 1.4498 ' .8432

Hexamcth-y lacetone and l-b lll y l No prod ucL ______------cbloride. f 8 io 52/20 ___ .. ______1. 4115 ]V[ethy l ,,-b utyratc a nd I·buiyl 2,2-d in: eth yl-3-hexa nonc 6 __ •• ___ •• _. 148 to 152/760 ___ ... ______} I. 40[5 (0) chloride ("n ail run). 146 to 148/760 ____ ..•. _•••

{'6.2 to 77.7/86 to 89 ____ .. I. 409 2,2-d imethy!-3- hcxa none ______. 6. 6 146 to 148/760 •.. ______L4095 [6, 12, (0)

M cthy l-n-bu tyral e a nd I· bulyl 156/760 ____ ... _. ______} chloride (large run) . 90 to 91/87 1.089 __ .. _____ t. 4262 70 to 78/28 to 30 ___ ...... 2,2-d imcthyl-3-hexanoL . ____ ... ____ _ 4.8 155 io 157/760 ______..•. _ } 1.4280 (a t [6, 12, 10) 153 to 156/755 ______17° C) .

HCXEmcth yleth ane _____ ... _.. _____ ...... _. __ 99 to 102 __ _. ______. __ _ Solid J. 4067 Pcn trmcthyIHeelone - -- -. ------.-. 50 {~~~~~60 ::: ::::: ::::::::: L4059 (8) \ Met h yl i;;o butyratc a ncl I·butyl I 55/760 _ . ______... _. ___ _ t. 4280 2,2,4-trimethyl-3-pcnta nol .. ______. 7.3 {83/82 ______. ____ .. } chlor ide. , 1. 4280 152/700 , __ ... ______.. __ _

11 3/37 ___ . ______2, 2,4,4' tctramc thyl -3- iso prop yl -3 - 19 { I. 4655 pcntanol. 120 to 122/30_. _. ______1.4623 [2)

I. 3912 [8] iEt h yl t llmeth ylaeeta!e - -- -•••.. --.. If: : ~ ::::: ::::-:::: :::: :::: 1. 3906 Diethy lcarbonatc a nd I-buiyl chlo­ ride. J.4198 ) R cxa methylaectonc ____ •••• ___ .. _. _. 9 {: ~~ ::: :::::::: ::: ::: : : ::: I. 4192 l Plva IOllL ______. ___ Meltl ll g POlIlt , 78° C ___ _ Soli d See f ootn otes at e nd of tuble.

I Aliphatic Condensations by Sodium 113 TABLE 1. Reaction products- Continued

Y ield per­ Refracti\"e Density, Reaction Product centage Boiling poin t index, '1lbo Reference of theory I D" ------o Gj mm H g ( P entamethylacetone ______1. 4059 g;~ ::::::::::: : :::: :::::: 1.4059 [8]

Ethyl trimethylacetate and iso­ 2,2,4-trimeth yl-3-pentanol ______1. 4287 propyl chloride. n~~ ::::::::: ::::::::: :::: 1.428 [4] II 0;33 ______[ 2, 2, 4 -trimethyl-3-i,opro pyl-3 -pen - 1 { 1. 4576 tanol. r 103 to 104/27 ______1.4559 [I] Paraldehyde and I-b utyl chloride __ No producL ______

Para formaldehyde and I- butyl chlo- ___ __ do ______ride. Acetamide an n I-butyl chlori de ______do ______

Acetic anhydride and I-butyl chlo- _____ do ______ride.

I Y ields based u pon the carbonyl or ester compound. ' T he tem peraturc con ve rsion to 760 mm H g was done by a method described by Lippencott [II]. a T he data in italics are from the litera ture. ' A freezing pOin t of -10.10 C fo r this compound was made by E . H. R icb . • T he values for a known sam ple of 2,2,3,6,6-pcn tamcLhyl-3-hep Lene-5-onc, as measured in this laboratory. 5 Obviously, a m ixLure of ketone and ca rbinol. Oonstan ts from th e carbinol previously prepared in this laboratory. over a period of 3 hours to a suspension of 69 g had been added, and the contents of the flask was of sodium sand (3 g atoms) dispersed in 1 liter of allowed to stand overnight_ The reaction mix­ isoo ctane at 80° C. The reaction mixture turned ture was· worked up in th e usual manner, and the purplish blue as soon as the addition was begun, isooctane and umeacted methyl bromide were and the temperature rose quickly to 98° C. Stir­ removed by distillation. The residue, which ring was con tinued for 1 hour after the addition was distilled at reduced pressure (Podbielniak was completed. The reaction mixture was hy­ still), yielded 62 g of 2,2,3, 4-tetramethyl-3-pen­ drolyzed after standing overnight, and the organic tanol, (0.43 mole, 43% yield). layer washed several times with water and dried. 4. Pentamethylacetone a n d Methyl Iodide The solvent and unreacted isobutyl chloride were distilled off, and the residue was distilled Pen tamethylacetone (128 g, 1 mole) and 284 under reduced pressure (Claisen flask) t o give g methyl iodide (2 moles) were diluted with 200 20 g (14 % yield) of 2,5-dimethyl-3-isopropyl- ml of isoo ctane and add ed over a period of 2 3-hexanol. hours to a suspension of 69 g of sodium sand (3 I g atoms) in 300 ml of isooctane a t 80° C. A rapid

3. Pentamethylacetone and Methyl Bromide rise in temperature occurred , followed by the I appearance of the characteristic blue color. After A suspension of 69 g of sodium sand (3 g atoms) standing overnight the reaction mixture was hydro­ in a solution of 128 g of pentamethylacetone (1 lyzed, and the organic layer washed several times mole) and 1 liter of isooctane was heated to 80° C. with water and dried. The solven t and unreactecl An excess of methyl bromide was added from a methyl iodide were removed by distilla tion, and methyl bromide dispenser 4 through a tube th e r esidue, which was distilled under r educed extending below the surface of the liquid. The pressure, (Claisen flask ) gave 41.5 g (29 % yield) of I reaction began in the usual manner, wi th the 2,2,3,4-tetramethyl-3-pentanol. temperature rising quickly to 98° C, and the appearance of the blue color. Stirring was con­ 5. Pentamethylacetone and Ethyl Bromide tinued for 1 hour after all of the methyl bromide A solution of 163 g of ethyl bromide (1. 5 moles) 'Arrow Products 00., Oarlstadt, N. J. and 64 g of pentamethylacetone (0.5 mole) was

114 Journal of Research added to 72 g of sod ium sand (3 .1 g atoms) yield of 2,2-dimethyl-3-t-butyl-3-hexanol. Two suspended in 900 ml of isooctane. The r eaction analyses 5 on this carbinol gave (1) C, 77.6 percent· began in the usunl manner with the appearance H , 14. 3 percent and (2) C, 77.6 percent; H , 14.7 of the blue color and a r ise in the temperature from percent. The calculated values arc: C, 77.42 per­ 80° to 98 ° C. After standing overnight the reac­ cent; H , 13.98 percent. The molecular refrac­ tion mixture was hydrolyzed, and the organic tions were determined: lv[R (observed), 58.97. portion was washed and dried. J sooctane was MR (calculated ),6 58 .86 . removed by distillation, and the residue was dis­ It is presumed that the carbon and hydrogen tilled at a reduced pressure (Claisen flask) to analyses are high due to the presence of som e give 35 g (44% yield) of 2,2,4-trimethyl-3 -ethyl- olefinic material formed by the dehydration ot 3-pentanol. N asarow [14] prepared the same the carbinol during distillation. carbinol in 7-percent yield by using penta­ methylacetone and ethylmagnesium bromide. 8 . Hexamethylacetone and i-Butyl Chloride A large run of this reaction has been carried A solution of 71 g of hexamethylacetone (0.5 out as a part of a synthesis. More precise data on mole) and 138 g of t-butyl-chloride (l.5 moles) the physical properties of the carbinol, its dehy­ was added to 69 g of sodium sand (3 g atoms). dration, and iden tmcation of 4,4-dimethyl-3-i 0- The reaction began at 78 ° to 0° C, but it did not propyl-2-pentene as the dehydration product will proceed well. Considerable gas, probably 1S0- I be reported when further work, which is now in bu tylene, wa given oft' during the reaction. The progress, has been completed. ketone-halide solution was added to the reaction 6 . Pentamethylacetone and Isopropyl Chloride m ixture over a period of 2 hours. The contents of the flask was worked up in the usual manner, A solution of 64 g of pentamethylacetone (0.5 and the solvent was removed at atmospheric I mole) and 98 g of isopropyl chloride (1.25 mole) pressure. The residue (85 ml) was distilled from was diluted with 250 ml of isooctane and added a Clai sen flask at reduced pressure to give: 25 to a reaction flask containing 72 g (3. 1 g atoms) ml of residual isooctane, 36 g of recovered hexa­ of sodium and in 300 ml of isoo ctane at 80° C. methylacetone, boiling point 65° to 72° C at 32 The usual procedure was carried out, and the mm Hg, n~o = l.4133 to l.4195 , and 3 ml of organic layer was worked up in the usual manner. heavy viscous material of unknown composition. After removing the solvent, the r esidue was dis­ An attempt to crystalize tri-t-butyl carbinol [1] from tilled at reduced pressure (Claisen flask) to give a pentane solution of the res idue by using first I 36 g (42 % yield) of 2,2,4-trimethyl-3-isopropyl- an icc-salt mixture and secondly a dry icc and 3-pentanol. This carbinol was prepared by Bart­ acetone m ixture, fa il ed to yield any solid material. lett and Schneider [1] in 6-percent yield from the reaction of diisopropyl ketone with t-butyl chloride 9 . Pinacolone and n-Butyl Chloride by means of sodium. A solution of 100 g of pinocolone (lmole) and 7 . Hexamethylacetone an,d n-Propyl Chloride 185 g of n-butyl chloride (2 moles) in 700 ml of isooctane was added to 69 g of sodium sand (3 A solution of 142 g of hexamethylacetone (1 g atoms) in 300 ml of isooctane at 80° C over a \ mole) and 157 g of n-propyl chloride (2 moles) period of 2 hours. Th e reaction that ensued was I was diluted with 200 ml of isooctane. This solu­ similar to those already d escribed. After stand­ : tion was added to 69 g of sodium sand (3 g atoms) ing overnight, the reaction mixture was worked i in 600 ml of isooctane at 80° C over a period of 2 up in the previously described manner. The iso­ hours. A sudden rise in temperaure occurred and octane and unreacted n-butyl chloride were re­ I was followed by the appearance of the usual blue­ moved at atmosph eric pr essure, and the residue I black color in the r eaction mixture. After stand- was di stilled at reduced pressure in a Claisen ing overnight the contents of the flask was hydro­ flask. There was obtained 43 ml of material l lyzecl, washed, and dried. Isooctane and un­ having these properties: boiling point, 103° to I reacted n-propyl chloride were r emoved, and the I We are indebted to the Anal yti cal Chem istry and Standard Samples I residue was distilled at reduced pressure in a Sect ions, NBS, for t. he ca rbon and hydrogen determination. Claisen flask. There was obtained a 38-percent 6 Lorentz-Lorenz equation .

I Aliphatic Condensations by Sodium 115 107° C at 30 mm Hg, n~ = 1.4487 D20= 0. 8439. probably isobutylene. The reaction mixture was The reported constan ts for the expected product, stirred overnight at a temperature of 56° C. 2,2,3-trimethyl-3-heptanol, are boiling point, 84° Five gallons of 95-percent ethanol was then to 87° C at 13 mm H g, n~ 1.4409 , D 20 0.8487 added over a half-hoUl' period, and the mixture [3]. Since the constants of the fraction obtained was kept at reflux temperature all day in order to do not agree with the reported values, it was remove as much as possible of the excess sodium. presumed that pinacolone had dimerized in the This was followed by 10 gal of water, which was presence of sodium to yield, after dehydration, added cautiously to the cooled reaction mixture 2,2,3,6,6-p en t am e thy 1-3-heptene-5-one, boiling under a nitrogen atmosphere during 6 hours. point, 83° C at 10.5 mm Hg, n;'= 1.4498, D 21 = After th e last traces of sodium had disappeared, 0.8432. A carbon and hydrogen determination the mixture was washed three times with a large gave C, 79 .1 percent; H , 12.6 percent. The cal­ amount of water. About 8 to 10 gal of organic culated values for the dimer arc: C, 79.12 per­ material was obtained, which was dried over cent; H , 12.08 percent, and for 2,2,3-trimethyl-3- calcium chloride prior to distillation. heptanol, C, 76 percen t; H , 13.92 percen t. The The isooctane was removed at atmospheric formation of the dimer was expected rather than presSUl'e, and the r esidue, 2,500 ml, was submitted the addition reaction, because pinacolone, being to vacuum distillation (Podbielniak). There were a methyl ketone, easily aldolizes in the presence obtained 552 g of 2,2-dimethyl-3-hexanone (4.3 of sodium. moles, 6.6 %) and 408 g of 2,2-dimethyl-3-hexanol A 2,4-dinitrophenylhydrazone was prepared (3 .1 moles, ,4.8%) [10]. There were no fractions from some of this material, which melted 146.5° that could be identified as the desired product, to 147° C. The same derivative was prepared 2 ,2-dime thyl-3-t-buty1- 3-hexanol. from an authentic sample of 2,2, 3,6,6-penta­ The poor recovery of starting materials is in­ methyl-3-heptene-5-one and melted at 147° to dicative of the fact that little coupling occurred ; 147.5° C. A mixture of th e two derivatives the boiling points of the coupling products are all melted at 146.0° to 146 .5° C. H ence the main well above that of isooctane, and they would produce from this r eaction was the dehydrated certainly have appeared in the residue left upon bimolecular product from pinacolone. stripping the isooctane. H ence their absence can only be accounted for as a consequence of side I 10. Methyl n-Butyrate and i-Butyl Chloride reactions. The t-butyl chloride was lost partly A solution of 102 g of m ethyl n-butyrate (1 as isobutylene during the course of the reaction, mole) and 368 g of t-butyl chloride (4 moles) was and the remainder distilled with the isooctane added over a period of 3 hours to 46 g (2 g atoms) during the stripping operation. The ester must of sodium sand in 200 ml of isooctane at 80° C. have been lost as a consequence of alkaline hy­ After standing overnight the mixture was hydro­ drolysis when the reaction mixtUl' e was worked up, lyzed, and th e organic layer was washed several the hydrolysis products being discarded with th e times with water and dried over anhydrous sodium raffinates. carbonate. The isooctane was removed, and the 11 . Methyl Isobutyrate and i-Butyl Chloride residue was distilled at reduced pressure in a Claisen flask. There was obtained thereby 11 g In a 50-gal stainless steel reactor, 6.9 kg of I (9% yield) of 2,2-dimethyl-3-hexanone. sodium sand (3 00 g atoms) was prepared in 10 Since more information was desired on this gal of isooctane. To this was added 17.6 kg of reaction than was afforded by the small-E\cale run, t-butyl chloride (191 moles), and the mixture was I a large run was made. A solution of 17.6 kg of heated to 45 ° C, after which 6.4 kg of methyl t-butyl chloride (190 moles) and 7.54 kg of ethyl isobutyrate (62.8 moles) was added over a period n-butyrate (65 moles) was added over a period of of 6 hours. The reaction mL'{ture was stirred 7 hours to 6.9 kg of sodium sand (3 00 g atoms) in overnigh t and hydrolyzed . The organic layer 10 gal of isooctane at a temperature of 55° to 60° C was separated, washed, and dried. I sooctan e and under a nitrogen atmosphere. The reaction unreacted t-butyl chloride were distilled at atmos­ mixture turned greenish-black during the course pheric pressure. The following compounds were of the addition and evolved large quantities of gas, isolated from the residue by distillation on a

116 Journal of Research ------

Podbielniak column: 59 g of hexamethylethane, layer was separated, washed several time with 4,036 g of pentamethylacetone (50 % yield), 599 water, and dried over anhydrous calcium chloride. g of 2,4 ,4-trimethyl-3-pentanol, and 2,229 g of Isoo ctane was distilled off, and the residue was 2,2,4,4-tetramethyl-3-isopropyl-3-pentanol (19 % distilled at reduced pressure in a Podbielniak yield). The latter carbinol was prepared in 15- column. The following products were isolated percent yield by Bartlett and Sclmeider [IJ by the and identified from their physical constants: 2.5 reaction of pentamethylacetone, t-butyl chloride, moles (4% yield) pentamethylacetone, 2.73 moles and sodium. (4% yield) 2,2,4-trimethyl-3-pentanol, and 32.6 moles (54 % yield) 2,2,4-trimethyl-3-isopropyl-3- 12. Diethylcarbonate and i-Butyl Chloride pentanol. Distillation data are shown in figure 1. Ethyl trim ethylacetate (9% yield) and hexa­ methylace tone (9% yield) were obtained from the reaction of 277 g of t-butyl chloride (3 moles) and ~ ______------11. 46 236 g of diethylcarbonate (2 moles) in the presence U 1.44 N of 92 g of sodium sand (4 g atoms) using petroleum ." ether as the solvent. There was considerable . - t 4 2 ! difficulty in starting the reaction, but once started

the reaction became very vigorous. The sodium P 10 0 z .32 101M HG turned black, then bronze, and the reaction mix­ o 78 MM HG z ture thickened considerably so that additional ~ 75 - z solvent was required to permit stiJ'ring. After :; o hydrolyzing the reaetion mL'(ture, the organic • ' 0 layer was worked up in the usual manner. After ,, - removing the petroleum ether, the residue was 1000 2000 3000 4000 distilled in a Podbielniak column to give 0.185 RE COVERY IN lol L mole of ethyl trimethylacctate, and 0.176 mole FIGUHE 1. R esults of distillation of the producls from of hexamethylacetone. i nteraction of ethyl trimethylacetate and iSOp1·0pyl A second run was made by using the same ch loride. quantity of materials, but isooetane was used as '['he s m~1I plateau in the boilin g point curve, bCi[inning at about 1.000 m l. distilled, represents prntumct bylacetonc; Ow JlC 'X L small platrau represents the solvent. From the residue 25 g of solid 2,2,4-trimcthyl-3-pelltanol. ']' ho lo ng 1)!aLC'3 u represents 2,2,4.... trim cthy l-3- (melting point, 78 ° C), which was identified as isopropyl-3- pe ntallol. 'rhcsc cu[\'es arp similar LO t hos(' obtained in the t) Lh er pivaloin (melting point, 80° C), and a very high experiments. boiling material were obtained. There were no fractions isolated · havin g the properties of either 14. Paraldehyde, Paraformaldehyde, and i-Butyl ethyl-trimethylacetate or hexamethylacetone. Chloride

13. Ethyl Trimethylacetate and Isopropyl Chloride An attempt was made to extend the Bartlett­ Schneider reaction to include the reaction of an In a 50-gal stainless steel reaetor, 15.4 Ib of aldehyde with an alkyl halide. Because monomeric sodium sand (3 00 g atoms) was prepared in 10 aldehydes aldolize under alkaline conditions, it was gal of isooctane. · To this there was added, over decided to try paraldehyde and paraformaldehyde. a period of 6 hours, a solution of 180 moles of A mL'{ture of 132 g of paraldehyde (0.5 mole) and isopropyl chloride and 60 moles of ethyl trimethyl­ 138 g of t-butyl chloride (1.5 mole) was added in acetate. The reaction mL'(ture exhibited the small portions to 39 g of sodium sand in 400 ml characteristic appearance of a blue color accom­ of isooctane. The temperature of the reaction panied by a rise in the temperature to 54 °C. mixture was maintained at 80 ° C until all of the The reaction was kept at 50° to 70° C by means mL'cture had been added. There was no evidence of warm v.rater in the kettle jacket. After stand­ of reaction at this temperatme; consequently the ing overnight, excess sodium was destroyed by 5 mL'(ture was heated to 90° C for 1 hour. After : gal of 190-proof ethyl alcohol, followed by ample standing overnight, no reaction had taken place. quantities of ,vater for hydrolysis. The orgame The experiment was discontinued.

Aliphatic Condensations by Sodium. 117

704~50- 48-- 3 A similar attempt ,vas made with paraformalde­ the aliphatic ketone-aliphatic halide type, four hyde and t-butyl chloridc as reactants, but no were of the aliphatic ester-aliphatic halide type, reaction was observed. and four were a miscellaneous type. In those reactions that produced desirable compounds, the 15. Acetamide and I-Butyl Chloride yields were as good or better than where the Small portions of 59 g of acetamide (1 mole) and Grignal'd reagent was employed. 277 g of t-butyl chloride (3 moles) were added to This reaction is apparently, applicable to com­ a reaction vessel containing 69 g of sodium sand pounds whose carbonyl group is sufficiently re­ (3 g atoms) in 1 liter of isooctane. The reaction moved from influences that cause enolization, was very sluggish with evolution of a gas (probably aldolization, or hydrolysis. Table 1 is a resume of isobutylene), and the temperature did not rise results obtained in this work. above the initial temperature of 80° C. After IV. References hydrolysis with ethyl alcohol and water, a strong odor of ammonia developed indicating hydrolysis [1] P. D. Bartlett and A. Schneider, J. Am. Chem. Eoc. of the acetamide to acetic acid. Fractionation 67, 141 (1 945). gave a cut (boiling point 78 ° C n~= 1.3862) [2] G. Calingaert, H . Soross, V. Hnizda, and H. Shapiro, identified as ethyl acetate [9]. This is not sur­ J. Am. Chem. Soc. 66, 1389 (1944). [3] J . B. Connant and A. H . Blatt, J. Am. Chem. Soc. prising, since acetic acid and ethyl alcohol could 51, 1227 (1929) . esterify under conditions of the experiment. No [4] A. Faworsky, J . Prakt. Chim. 88,651 (1913). other reaction product was obtained. [5] Richard Stanley George, PhD Thes is, Pennsylvania State College (1943). 16. Acetic Anhydride and i-Butyl Chloride [6] A. Haller and E. Bauer, Compt. rend. 150, 586 (1910). In a large Erlenmeyer flask, 94 ml of acetic [7] F. L. Howard, Anal. Chem. 19, 144 (1947). anhydride was added to 23 g of sodium sand in 100 [8] F. L. Howard, T. W. Mears, A. Fookson, P. Pomer­ ml of isooctane. The mixture was heated to 75 ° antz, and D. B. Brooks, J. Research NBS 38, 365 (1947) RP1779. C, and 217 ml of t-butyl chloride was added in [9] E. H . Huntress and S. P. Mulliken, Identificat ion of small amounts. From time to time the flask was pure organic compounds, p. 378 (John Wiley & heated on a hot plate. There were no visible signs Sons, Inc., N'ew York, N. Y., 1941) . of reaction, nor did any reaction occur when the 110] M. J . Lerodie, Ann. Chim. Phys. [9]16, 369 (1921). contents of the flask was heated to the boiling [1l] s. B. Lippencott and Margaret M. Lymen, Ind. & Eng. Chem. 38, 320 (1946). point of isooctane (boiling point, 99° C). The [12J R. Locquin and L. Leers, Compt. rend. 178, 2097 addition of t-butyl chloride was discontinued, and (1924); Bul. soc. chim. [4] 39, 434 (1926). the reaction mL'{ture was allowed to stand over­ [13] A. Morton and J . R. Stevens, J . Am. Chern. Soc. 53, night. No change had taken place during the 2244 (1931) . night, and the experiment was therefore dis­ [14] I. N . Nasarow, Bel'. deut. chem. Ges. 69B, 21 (1936). [15] F. C. Whitmore, J . M. Church, and R. U. McGrew, continued. J . Am. Chem. Soc. 56,176 (1934). III. Summary 116J F. C. Whitmore and A. R. Lux, J. Am. Chern. Soc. 54,3451 (1932). In the study of these condensation reactions, seventeen runs were made, of which nine were of Washington, January 26, 1943.

118 Journal of Research