[1]2 1 8 3 9 , 5 5 6 Uni'ted States Patent Office 2,8391556 IPRODUCTION,OF ALUMINUM HYDROCARBONS Karl Z.-egler and Roland Koester, Mulheim an der Rutir, Germany; said Koester assignor to said Ziegler No Draivit,,,g' Application July 27,1953 Serial No. 370,655 Claims priority, apl)lic-,ition Germany July 28, 1952 11) 29 Claims. (Cl. 260-448) This invention relates to improvement in the prqdue- 15 tion of altiminiim hydrocarbons, and more particularly relates to @the prodtiction of aluminum trialkyls or aluniintim triaryls along with cryolite and aluminum fluoride. Aluminum hydrocarbons have obtained iniportance L@is catalysts for the polymerization of olefiiis, and there ex- 20 ists a pronotiiiced need for commerically converiient methods for their prodliction. A ntiniber of proposals have been made iii receiit time, blit all of them were not fully satisfactory. In the two processes of the patent applications 354,- 626, filed May 12, 1953, now U. S. Patent No. 2,744,127, and 354,658, filed May 12, 1953, now U. S. Patent N6. 2,691,668, of the applicants, organic halogen compouiids are directly reacted with magnesium-aluminum alloys with the production of aluminurn trialkyls, or aluniinlini 30 is at first dissolved in alkyl halides to give the liquid socalled aluminum alkyl sesquihalides which are debalo.@e@nated by a subsequant treatment with magnesiUDI or magensium-aluminum alloys, thereby setting free a part of th-- aluminum charged in the metallic forn-i. In this ',3 5 process, the magnesium used is lost, at the end of the process, in the form of practically useless magensiurn halide. In the process of the patent applicati6n 354,624, filed PAay 12, 1953, now U@ S. Pttent No. 2,786,860, of the 40 -ipplicants, alkali metal hydrides are iiecessary as essential auxiliary agents for the production of the alliminuni alkyls. These alkali metal hydrides must be separately produced and have properties which render their comn-iercial use difficult. 45 One object of this invention is the prodliction of aluminum hydrocarbons without the above-tlientioned disadvantages. A further object of this inveiition is a process for the production of aluminum hydrocarbons which allows a 50 coinmercially simple mode of operation' A still further object of this invention is a process for the production of altiminum hydrocarbons which does not resL,.It in the formttion of useless waste products, but simtiltancously prodticas useful by-prodticts, such as purely ilior@anic fluorine-containing alumintim compounds, sticli as cryolite, Potassium alliminum fluoride, or aluminun-i fluoride in a form which is directly suitable for ftirther use, as, for example, for the production of alumiuum. Tliese, and still further objects will become apparent from the following descriptionin accordance with the invention, alumintim hydrocarbons, and particularly alun-iinum trialkyls or altiminum triaryls, are produced in addition to purely inor.-anic altililinum fltiorine compounds by reacting alliminiim 65 alkyl halides or aluminum aryl halides with alkali fluorides which may partially be present in the form of their complex compounds with aluminum fltioride, In the process of the invention, there are obtained as by-prcdtzcts either cryolite or potassium-alumintim fILio- 7( ride, respectively, depending on whether sodium fluoride or potassiii,-n fiiioride, respectively, have exciiisivcly be,-n Patented June 17, 1958 2 used as reagents. Aluminum fluoride is produced if a part of the alkali fluoride had been replaced by the complex compound thereof with altiminum fluoride. The pfocess may be represented by the folowina reactions: (1) AlumiiiLini alkyl fluorides or aluminum aryl fluorideg may readily be formed from the corresponding aluminum alkyl or aluminum aryl halides oth than fluori,,Ies by reaction with alkali fluorides or a fluoride alkali fluoride complex compounds. When iising the complex compouiids of the alkali fluoride for the reaction, aluminum fluoride is set free. AI(C2H5)2CI@+NaF--g-AI(C2H5)2F+NaCl (1) Al (C,H,) Cl,+2NaF->AI(C2H,)F2+2NaCl (2) 3AI (C2FI,),CI+Na,AIF,--> 3AI(C2H5),F+AIF3+,3NaCl (3) If desired, these new alkyl- and fluorine-containing aluminumcompounds inay easily be obtained in the pure form in this phase of the reaction, for example, by vacuum distillation. They are generally very viscolls, spontaneously inflammable liqtiids, which, in coldness, are often of a resin-likeviscosity. (2) The aluminum alkyl fluorides or alum'mum aryl fluorides disproportion in the presence of suitable amounts of alkali fluoride at elevated temperatures to give comple--, alk-aiialuminum fluorides, such as cryolite or "pgtassiu @m cryclite," on the one hand, and aluminum l@rialkyls or qluminum triaryls, respectively, on the otl-ier haiicl, accorcling to the following equations: 3AI(C2Hr,)2F+3NaF 3-2AI(C2HB) +Na3AIF6 (4) 3 3AI(@C,HG)F2+6NaF-->AI(C2,Hs)3+2NaBAIF6 (5) it is evident from the foregoing that the production of alliminum trialkyls or aluminum triaryls may at will, be coupled witb the production of cryolite or "potassilim cryolite," respectively, and/or of aluminum fluoride. Iii the course of a commercial production of aluminum trialkyls or aluminum triaryls, several or many charges are ordinarily reacted in succession. If in all batches alkali fluoride is exclusively used as one reactant, then cryolite or "potassitim cryolite". will be the by-p)roduct. It is possible in this oase to combine the reactions characterized by the Equations 1 and 2, one the one hand, and by the Equations 4 and 5, on the other hand, into one operation witl'O"t intermediately isolating the aluminum alkyl fluorides. Aiiother possibility is to use alkali fluoride only in the first charge, and then to react in the second batch, accordin.@ to Eqtiation 3, the aluminum alkyl halide with the cryolite obtained at the end of the flrst charge. In this case, if aluminum fluoride itself is to be obtained as byproduct, it is necessary, after the termination of Reaction 3, to first separate the, organic aluminum ffuorides. This is possible by dissolving out the same with indifferent solvents aiid filtration, or by vacuum distillation. Thereafter the Reactions 4 or 5 may be effected after the addition of alkali fluoride. When repeating this procedure, any sti ch charge obviously gives aluminum fluoride in the flrst stage and the cryolite at the end of the second stage, the cryojite being used for the beginning of the next batch. In total, only aluminum fluoride is thus obtained as the by-product, apart from the last charge at the end of the production. The total constimption of alkali fluoride in this case is naturally only half that of the prbduction of by-product cryolite. It is evident that it is possible by suitably guiding the process to obtain by-product aluminum fluoride and byproduct cryolite in any proportion desired. In these different process modifications, aluminum fluoride is always, and the cryolite is in certain embodiments, obtained in mixture with alkali lialides, such as sodiurn chloride, ilkali bromide, or alkili iodide. The process@
[2]2,839,556 3 in.- of sucl-i mixtures is effected in a very simple maiiner by treating ttiem witti water, preferably after a previous calcination, wliicli improves the filtering property. The difficultly soluble aluminum fluorine compounds are theii recovered by filtration. The s@-cond stage of the reaction as represented by Equa'Lious 4 and 5, respectively, i. e., the reaction of the aluminum alkyl fitiorides witli sodium fllioride, has been fotind to proceed via novel coi-nplex compounds of alkali fluoride aiid aluminum alkyl fluoride according to the following equations: 3AI(C,H,),F+3NaF--> 3NaAI(CH,),F, (6) 3NaAI(C2H5)2F2--> 2AI(C2H5)3'+Na3AIF, (7) Iii the last stage, the final products, alurfl-inlim tri-,iikyl@ 1 and cryolite, are formed with the cleavage of the ii),termediately formed so ditirn-aluininum-diethyl-difluor complex compound. The reaction of aluminum alkyl d;nuoride with sodii-im Quoi-ide according to EqLtatioii 5 proceeds likewise via a second stage with the formati(ii of a novel complex compound of alkali fluoride ,Lnd alumiiium alkyl fluoride in accordance with the followiiig equations: AI(C-,H5)F2+Nar@NaAI(C,,,H5)F3 3NaAI(C,H,)F,+-)NaF--@,AI(C2H5)3+2N,,,A!F.l (9) Hov-/ever, this process is by far more unfz,.vorable, siiiec more alkali fluoride must be charged, aiid, as the final result, only half the quantity of aluminum hydrocarbons, but double the quantity of cryolite is obtained as coinpared with the yields of the aforenientior@ed process with the use of aluminum dialkyl monochloride. It has been found that when using aluminlim alkyl dichlorides and aluniinum aryl dichlorides in the process according to ttie inveiition, the quantity of solid salts loriiiing at first in the course of this process, as is obvious froin the eqiiations, is substantially larger than when starting with aluminiim dialkyl monolialides and aluminum diaryl inotiohalides, respectively, as the initial material. In additioii, the corresponding heats of reaction to be rCi,noved are much greater. Therefore, in the first case, a bigher expenditure with respect to the quantities of sblvent wbich may haye to be added, to the size of the reaction spaces, etc., is required for carryin@, out the process according to the invention. These difficulties, which will naturally also arise to a certain extent when using as starting materials the so-called aluminum alkyl sesquichlorides and alumini,-m aryl sesquichlorides, respectively (mixtures of RAIX2 aiid R2AIX, where R is all@yl or aryl, X is chlorine, bromine or iodine), iii@,ty be avoided if in ti-ie process according to the inver@tion the mixture of the aluminum alkyl sesquichlorides and q@lutminum aryl sesquichlorides, respectively, is at first completely converted witl-i a previously prepared portion of alumiiium trialkyl and aluminum triaryl, respectively, into three molecules of the aluminum di-,ilkyl rnoiiochloride, which is particularly suitable to the reaction. This conversion may proceed, for example, in accordalice with the following equation: AI(C,1-1,)CI2+AI(C,H5),Cl+ AI(C2H5)3-3A](C2H5),Cl (10) In the further processing, as is evident from eqiiation 7, double the quantity of the altiminum trialkyl or aluminti-,ii triaryl charged is recovered. The aluminum trialkyl and aluminum triaryl, respectively, formed by the process according to the iii,,ipntioil is extremely easily senarated from the reaction n-iixture ei'Lher by distillation or by dissolving in a su@itablesolvent. In detail, the process for the preparation of aluminum 1-iydrocarbons together with cryolite or aluminum fluoride may be effected in various different modes. For example, it may be preferable to operate in the preseiice of a so!-ve,.tl or sispend;n@, a,,eiit in order that the forming salt masses do not become too compact and remain stirrable 4 and easily movable, especially in the first and second stages. Suitable solvents are iii principle all materials to wliich t'@ie reactants used, and especially the aluminum trialkyls or aluminum triaryls formed are resistant ii,) to theoperati-@ig teniperature required, slich as hexaiie, benzetic, toluene, cl-ilorobeiizene, o-dichlorobeiizene or plleiiaiithrene. If all of the three stages are carried out iii the pres--iice @@) of a solvent, there sliould b-@ selected a solvent wh-ich has i s@,ift'i--iciit cliffereiiee in boiling point from the aluriiiii-Lim ti-ialkyls or aluminum triaryls to be produced. Alumiiiun-i triethyl, for example, boils at 196' C, Therefore, the@re should be selected a solvent, the boiling point of which is iiot substantially above 150' C. or iiot below about 300' C. The d;fference ii-i boiling point may be still smaller if a correspondiiig expenditure in equipment for separatiiig the solvents and the reaction product by distillation is allowable@ If low boiling solvents are selected, the third stage must be operated under pressure. If it is intended to use the aluminum trialkyl or aluminum triaryl only in solutioii, one will not be subjected to these limitatiors witt-i respect to the boilin- point of ',.he solveit. According to a special dmbodiment of the process of this inven'@ion, increased yields of aluminum liydrocarbons may be obtaiiied if tfie Lhermal decompos:ltioii, as described in Equation 7 and the separation of the aluminum hydrocarbons are coinbined in a sin.-le process iii such a manner that the aluminuni hydrocarbons formed @,0 remain only as short as possible a time in th-- heated reactioii zone. Iii carryipg out this modificatior, of the process, it is even possible to i:tse a higl-ier decoiiiposition temperature than wo,,ild generally be allowable by the ttiermal resistaiice of the aluminum hydrocarboiis. @@n @35 this way, ihe process can be greatly shortened in itiaiay cases. The execution of the process mty niost cf--;iiveniently be realized by heating the products of the reaction between aluminum alkyl halides of altiminu-,i-i aryl halides and alkali fluorides obtained at moderate 40 temp.,ratures, after previously bavin.- driven off the solvents wbich may have been used, under vacuum it temperatures of between 200' and 300' C. and tal@iiig care, by the use of pumps of proper capacity, tiiat as high as possible a vacum is continuously maintained iii 15 the apparatus. The requirements with respect to plimpiiig capacity will increase witl-i the size of tl-ie -,ilkyl or aryl radicals combincd with the aluminum. It should be noted, for example, that as low -,t vaclium as may be produced by means of ordinary water-jet 50 vacuum pumps will be sufficient in the production of tl-le aluminum trimethyl. The splitting teniperature m,,ty be increased in this case to as high as 2700-300' C. In spite of these high temperatures, aluminum trim--thyl is obtaitted iii i very high and prlictically q@,.iaitititive2 yield. 55 This specific embodiment of the process of the ;nvention may be carried out without the use of @,,-,cLium or with the use of only a slight vacuum if heated in different gas currents, preferably superheated varlors of liquids which are indifferent to the aluminum @lkyls or alumi60 -ium aryis are used as the heat transferriiig media wl@iil@takin- care, by properly selectiiig the boiling point of these liouids, that the aluminum hydrocarbons obta'- @ieci may easily be separated from the heat-transferring i-.nedia. For example, superheated v,,ipors of benzene, pentane, 65 hexane or also butane have been found to be A7ery siiitable. To effect these process modificatiolis, it is @,irssiblc, for example, to simply pass the vapors of the abrvementioned liquids after previotis superheatin-, to 200'- 70 300' C. into the reaction vessels in which the solidsz,,Itlike products of the process, est)ecially the coiiiplex coripounds of the formulq, NaAlR2F2 where R is alkyl or aryl, are contained, while the reaction vessels are likewise heated to a temperature sufficient to prevent conden75 sation of the vapors and allow the reacticii producis to
[3]5 . actually reach the splitting temperatures required. The escaping superheated vapors are condensed, preferably with the iise of heat exchangers for the evaporation of the freshly charged portions, and freed from the aluminum hydrocarbons carried along by distillation. The auxil5 iary liqtiids are returned into the process. It is also convenient to condense the aluminum hydrocarbons, which boil at a higher temperature than the heat-transferring media, by partial condensation as liquid. This modification of the process may with excellent 10 success be carried out in the so-called fltiid bed, thus vet-@l easily permitting continuous operation. The advantage of the process described over the known processes described at the beginning of the specification consists in that in addition to the desirable aluminum 15 trialkyls or alurninum triaryls the comniercially knowii fluorine con-ipolmds of alumintim, i. e., cryolite or alt, minum fluoride, are obtained as by-product in the forni of ufflizable commercial products which are, of importance in the electrolytic prodtiction of aluminum. 20 To obtain bi.-h yields and to secure a smooth cotirse of reaction,.c,,ir-- s'@iould be taken for the followin.@: If the operation is directed at cryolite or "potassiu-rn cryo@ lite" as the by-product, the qtiantity of alkali fluoride ntaiiied - '-> 5 must be exactly equivalent to tw@ce the lialogen co in the aluminiiin a@lkyl or alliniinum aryl halide used. An excess of alkali fluoride shotild be avoided, if possible, or kept within very narrow limits, since altiiiiinum trialkyls and aluminum triaryls have been found to form very stable non-distillable complex compounds with alkali 30 fluorides. The composition of such complex compotinds being, for e-@,,imple, NaAIX3F where X is alkyl or aryl. However, in the process pha,,,e qs Ciescribed, for exaniple, by Equation 31 one is not botind to this conciiiion of equivalents, ailcl excess complex litioi,ide may be Lis,,@d; @,5 In this case, liowever, the condition of equivifents iyitist again be p.-eserved in Eq-,iatioii 4 or 5. Men prodlicing aliimintim tr,.'eilkyl,@ or @ a]Liininum triaryls in the manner clescribed above with the recovery of aluini.@ium fluoride as the by-product, tlip. qtiin40 tity of cryolite obtained iii accordatice with, for ex,,iniple, Equation 4 corresporids eyactly to the - o-Liantity reqtiired in accordance witli Eqtiation 3 if the sequoiice of reactions is cotiti-iued. Wheri operatiig witli exi,,@tly th,- molar ratios corresi)oiidiig to Equatio@i 3, tlie concentrcl4--1 tion of the two reactlilts, for exaniple - altiminuiii dietliyl chloride and cryoli'ce,, will dim-iiiisli very nilich tov,/,-@rci the end of the react;oi2. This restilts in a decrei,@,e of the xeaction r,,iie vvi)ich is the greiter si,@ice ,i re@ictio@i bet@,vecii a liquid (possibly di,-sol-@eei) ani ,t @;ol;d mat,-ri,@l is inC)o volved. It is obvious, ifierefore, that the conil,)It-te co-,iversion of the cryolite ivill offer certain diffielilties - ,ind will take a very long t@:Me uiider certain cir eti@i-istance-q. However, these diffictilties may easily be o-ver,,@ome by a further modification o,4L the process, namely @f @he conversion between cryolit-- and altiminum allyl halide or alumin-Lim aryl halide is cirried o,,it in two or sei,ei-al stages in stich a rnanner that the reaction, with conversion being stiil incomplete, is interrupted if - reaction rate beconics too slow.@ Iii this pb@).se ttic orga ... c CO alumin-am comi3oL,,nds wheti-ier by dissoltitio-ii or by distillation, are s-.Parated from the solid P',lase. Theii, ti . o solid phase whicii is not yet completely convor@ted is treated with fyesh aiumiiitim alkyl halide and ali-iminum aryl halide, respectii,ely, thereby obl,,tining rai)idly tl-i-- chloride, wlii'le tl@e iinconv,,rted or.-ariie alumizilim coiilpound,s are ftirther con,,,erted with fresh ci'yol ite. It is obvious tnat also stich a process may L)c carried out in an Liiinterri-tpted sequence of i-ndividual oper- ations, which, witli the sole use of, for example, -,Iumi!@ num dibthyl nionochloride and the correspo-tidi,.i@ quantity of sodium fl-aoyide, permits th-@ following moleciilar equation to be realizod: 3A](C,2H@)2CI+3Nal,-->AIF3+3.NaCI+2AI(q2H,)@ (11) 74 complete con-ve-ision iitc) fl@,ioride aiid soditim 6 The aluminum alkyl or aryl fluorides, as, for exampid, are produced in accordance with Reactions 1, 2, and 3, and the alkali metal aluminiii-n alkyl or aryl:ffuorides are novel compounds and constittit.e valuable intermediates for the production of the alumintim trialkyls or triaryls and the cryolites, etc. The new aluminum alkyl or aryl fluorides have general formula: AIR@,F, in wbicn P, is a radical, selected from the group consistifi@- of alkyl and aryl radicals, and x is one and y the other of the numbers I and 2. Preferable among these novel intermediates are aluminum diethyl fluoride AI(C2HB)2F and alumintim monoethyl difluoride Al (C2H5) F2 The novel complex compounds with the sodium or potassiiim alkali metals have the geiieral formula Me)klR,,Fb, in which Me is a@-i P-Ikali metal selected frori-i the gro,,tp consisting of sodium and potassitini, R is a radical selected from the group consisting of alkyl and aryl radicals, a is one of I and 2, and b is one of 2 and 3, which, added to a will equal 4. Preferable among these novel complexes is sodium aluminum diethyl diffuoride NaAI(C2H5)2F2 and sodium aluminum ethyl trifluoride NaAI(C2H5)Fg. The following exaniples are given by way of illustraticii and not limitation: Evample I ,20.5 arams aluiiiinum diethyl efiloride ai-c dissolvc,,d tinder nitro-,en in 200 cc. of dry hexaiie and then 42 gr-,Ials of finel3i pulverized and previous]3( well dried sodiijm fluoride are odded to the soltition while stirring the same. The mixti-ire will heat tip and the hex-,vne boil if tb , c addition is effected rapidly. The stirring is contintied for ibout I hotir and the hexane is then distd]@@d off. The residue is further heated under high vacutim and the highly viscotis aluminuni diethyl fluoride distils off at a pressure of 1-2 rnm. and a temperature of b,@t I ween 90' and 100' C. The yield is nearly qiiantitative. 104 -rams of the monofluoride obtained are intimately m'@xed under nitrogen with 42 grams of finely pulverized iiid dried soditim fluoride and heated i'Or 3 hotirs at 180'190' C. Then the aluminum triethyl formed is distilled off Linder vactium. The yield is 65-70 grams of alumintini triethyl having a boiling point of 128'-130' C. at So nl,,,. PLire cryolite remains as residue in the distillin@apparattig. Instead of tluminum diethyl chloride, aluminum diethyl bromide may be used with the same success for this exLmple. Exaniple 2 92 grams of alunlintim dimethyl chloride are dissolved in 200 cc, of benzene, mixed with 84 grams of sodii-ii-n fluoride (both steps under nitroacn) and the inixture is then heated for 5 hourg at 180' . C. in a horizontal atitoclave which is arranged so that it may be rotated. In addition, some sharp-edged pieces of metal, such as some pieces of hexagonal steel cut to a size at mthich their length equals their diameter are placed iii the autoclave for the purpose of @comniinuting the sait masses slispended in the @benzene during the rotating motion of the autoclave. The al-itoclave is cooled, the liquid po-rti,(n of the co-ntents of the alitoclave is forced ',hrough a filtering device Linder nitrogen and the salt n-iasses substantially reMIining in the autoclave are washed two times with benzene with slibseqtient pressing off of the liquid. All of the combined beiizene solutions are freed under nitro,@en from beizene in a column, thereby obtaining as residue 80-90% of the theoretical yield of aluminum trimethyl. 7 wific"lti @ i oil s expediently egected by means of vacti-
[4]21839,556 7 um disiillatioii@ The solid salt masses remaining in the atitoclave and in tl-ie filtering device are decomposed with water and freed 'Lrom residlial beiizene by injecting ste,@,m. In this way there results cryolite susp.-nded in a sodiuni chloride solution from whicti the cryolite is easily recov- 5 ered by centrifugin- apd washiii.- vvith water alC, sLibs-@- quent dryiiig. Exai,nple 3 100 kilograms of coi-iimercial phenanthre;ie v,,hich ha@@i -i ' previously been piirified by pro'@on.@ed melllii.- togetfici7 with sodium with heating aiid by distillation were placed in a suitable slirring vessel followed by the additicil f)f 100 liters of toluene and 25.2 kilograms of fi@iely pulverized and dried sodium fluoride. The free sdice of the i reaction vessel is filled with nitrog@-n and the-ri 2,@.7 kilograms of aluminum ethyl sesquichloride are careftil-ty allowed to drop in while preventing excessive heati-,ig by properly coitrolling the additioii and externally cool@ng. Theii the solutioti is boiled while stirring tl-ie saine, ?@nd 20 the toluene is distilled down by ineans of a short columii until the contents of tl-te vessel show a teniperature o@-' 180'-190' C. At ,his moi-nent the deseendin,- cooler is: replaced by a reflux cgndenser tnd the heatiig is contiiiiied for furvn-.r 5-6 hours at 180'-200' C. while stir- 25 rin,@- This is followed by distillp@tiori in a column under vacuti,n. AL frst, certain portions cf tolucite escape, theii a@boi-it I 0 kilo-rams of aiiiiiiinun-i triethyl pags over.. which 21 are iftnally mi@ed w@th scir@e pl-@enajithrene. 'the distillation is continued until 5 kilo-rams of the briginally prese-.it ,0 100 kilorains of pheiiaiithre@iil, h-qvc disiilledDi/er. These 5 kilograms of. phenailthrene 8till contaiii ali,31nigiim trielhyl and ,ire, after beiii,- combiiied witli the rnain fractioii of alumiiiiim triethyl, agaia fractionated iii vacuo, tliereby coiiipletely sedarating the phenanthrepe froiii the @;5 aluminum triethyl. - The residi-ic reniainin.- iii t,ic slirrin.- vessel is allclwed to cool somiawl-iat and tiicn the toluene d@stilleci off is agaiii added while stirrin,-. Th-, toli:,cnc at'this temperatul,e (70'-80' C.) d;ssolves )If of the and it "L 0 is now possible, by 'Lr,-,-@isferriiig the total cont-,i-,,ts of the vessel into a suitable filtering (!.cvice, to easily separat,. the salts from the solvents ,ind to subsequeiatly wash the salts free from phenanthrene witli toluene. The motlier 1.5 liquor is directly passed to 'LI-ic iicxt batch. It is preferable, prior to the addition of the aluminum ethyl sesq-Liichlor@de, to distil of,' at first sufficient toluene as to obtain the initial state of the first experiment. In doing so, certaiii po@-tions of atmospheric oxygoi-i and ntoisture oosorbed dtiring the filtration of the salt masses are remove Theii the phenanthrene recovered iii the secoiid distill@itioii of the ,ilumin,,iiii triethyl is added, aiid the rext experiment is ready to start. The salt masses are worked up in exactly the san-icmaiiner as described iii Exam-ole 1, aiid processed iiito pure cryalite, Nvhicli reslilts in a qiliantity of 20-21 kilo.- rams. The phenanthrene has beeii selected -in this case only as aii exampIL - of a high-boili-@i.- solvent whicli is completely indifferent tc, aluminutii triethyl, it may b.@ i-c.. (30 placed by any other solveiit having the s-,tme proaerties. Aromatic halogen conipounds, slich as chlorinated d-iphenyl, are also suitable as high-boiling solvents of this kiiid. Exai??,ple 4 65 61 grams of aluminum-bi-ityl-sesqui-iodide as is easily obtained fron-i aluiiiintim chips ,iiid butyl iodide are dissolved iii 100 cc. of o-dichicrobenzeiie atid mixed wilii 25.2 grams of sodiliii-i fitioride. Followiiig this, the SOILL- 70 tioii is boiled for 6 hours. Theeafter, it is @'iltzred tinder nitrogen from the separated salts, these salts are wastieci with some dichlorobenzeiie, the diChICTobenzene is distilled o@ff under vacuum by means of@ a column, a-iid the residue is rectified under high vacui.,m. About 15 grams 75 of aluminum tributyl are obtained. Exaniple 5 29.5 grams of alliniinum-pheiiyl-sesqui-iodide produced iccording to "J<)tirnal Organic Ch@,,mistry," 5, 114-118 (1940), are ref-luxed tinder i3itro,,@,eii for 5 hours iii I(jO cc. of ,i completely liydro,-enated Fischer-TrGpseli diesel oil fraction boilin.- between 190' and 200' C. together with 12.6 -,rams of so@dium fluoride, or are shaker@ in a closed vessel under the s,,ime conditions. Thp. mixc-Lire is filtered i,,7itli t'le c-,"-Iusion of air and ttie diesel oil 's d-still,-,d vtclitim, thereby obt,,iinin.- as resi,,i-tie - abr,,,it '-O g,a,-,-is of a'iuniini-im triphenyl, whicli at -oiic-- crysiallizes @Lid ma@t be recrystallized with Vn-, exclus@on of air ifrjm hot clilorobenzene with the addition of so@o-e pf,Lrolet-,m 0, t b 'Th-- scparalio-@i c-," the cryolite in the Exnn-iples 4 -,tnd 5 is l@ossibl-, iti the s,,ime in,,lnner E,,s in 'LI-ic other ccampl--S. Example 6 In a vessel of 15 liters capacity provided with a doublewall jacl@et aiid the other moiintings required, a suspeiision 2.8 Idio,@rai-ns of clj:y aiid finely palveriztd comsodium fiuo,@@ide, -ii 6 liters of chlorobcizene is heated under reflux until it boils, and 4.2 kilograms of aluminum diethylchloride are allowed to flow in witliin Tialf an liour. During this phase of the experiment, the external heating is shut off since the reaction heat issi-ifficient to maintaiii ttie reaction mixture in a lively boiling state. Following this, the mixttire is stirred for 10 hours at ttie boililig temperature of the mixture. Thereafter, the chlorobenzene is distilled off- as compl,-tely as possible with the use of a short coliimn under a vacu,,im of 10-20 mm. at a temperature of the bath beginning with 70' C., and finally increased to 130' C. (temperature of the heat-transferring medium in the double-wall jacket). Now the column is replaced by as short as possible a still head. During this operation, the admittance of air is to be exeltided by suitable means, as, for example, by introducing nitrogen. The still head is connected via a cooler and a receiver to an oil-air pumr) which stirs as vigorously as possible. The followill-f,' measilires comprise evacuating to a pressure of belo@v 1 1-im., if possible, and gradually heating to a final temperature of 230' C. During this time, 2.6 kilograins of - a.Liminuiii trietj--iyl corresponding to 97% of tl-ic theory distil off. Example 7 Subst n 6uosTa tially the same experiment as described in E),ample 6 is carried out, using 3.25 kilo.-rams of alu@-ninum diinethylehloride instead of the ethyl compound. At the end of the experiinent, a vacuum of 10-200 mrii. is applied, but the heating is effected to 270'-300' C. 1.6 kilograms of aluminum trimetltyl, corresponding to 95% of the theory, are obtained. Exainple 8 This experinient is at first carried out exactly as described in Example 6 until to that point where in Example I the heating under vacuum wotild start. 'Eheii the reaction vessel is connected via a steam superheater to a distillin.- vessel in m7hich 10 kilograms of benzene are contained. The temperature in the jacket of the reaction vessel is adjusted to 200'-250' C., and the temperature of the benzene vapors introduced into the reaction vessel to 250'-300' C. The reaction vessel is connected to a suitable descending cooler which condenses the benzene vapors togetlier with the aluminum triethyl formed. The two products are separated by distillation and the benzene recovered, if required, is repeatedly used for fne sanie ptirpose until the decomposition of the complex aluminum-organic intermediate compound is a complete one. This is normally the case after 2 to 3 cycles.
[5]9 Example 9 i.2 kilograins of the finely ground solid salt-like comPlex compound NaAI(CH3)2F2 are, filled with the exclusion of air into a tube of 15 crn. diameter and I m. 5 length, which is provided at the base with a suitable sieve plate. The tube is externally heated over the whole length to 250'-300' C. and peiitane vapors from a small vessel fihod with pentane under pressure are passed in from below. The pentane vapors are at fir@t 10 passed through a superheater, ivhere they are superheated to 250'-300' C. The rate of introduction is adjusted in such a manner that a good fluidized bed develops within the salt mass. The effluent vapors are cooled and the pentane is subsequently separated from i5 the aluminum trimethyl by distillation. The pentane is returned into the process. Aluminum trimethyl is obtained in an almost quantitative yield. By connecting several of such tubes in series, the process may easily be made a continuous one. In do- o ing so, the complex compound is continuously charged to the first tube, while the finished decomposed residtie of cryolite is withdrawn from the last tube. The complex compound NaAI(CH3)2F2 required for dimethylchloride with the equivalent quantity ofsodium ', fluoride in boiling chlorobenzene, separating the sodium chloride formed by filtration, and once again heating with the same quantity of sodium ftuoride, and finally distilling off the chlorobenzene. It is also possible, however, to use with the same success for the process @O described in this example the solid salt mixture which is obtained by boilin- aluminum dimethyl chloride in chlorobenzene with 2 mols of sodium fluoride and siibsequently distilling off the chlorobenzene. @5 We
