Aug. 20, 1953 NOTES TABLE I 4103 APPARENT IONIZATION CONSTANTS, DIPHENYLDIAZOMETHANE REACTIVITIES AND ETHYL ESTER SAPONIFICATION RATES OF SUBSTITUTED BENZOIC ACIDS Substituent Av. half-point\"*b KA X lOeC Concn., moles/l.d Half:life,d mtn. Av. krb,? I./mole min. kr X 10',?,* l./mole min. kr X 101 bJ I./mole min. Sone 5.80f0.01 (5.75)' 3.80f0.01 (5.85)' 1.59 0.0487 .OS74 .0967 13.21 7.29 6.22 7.41 7.35 13.71 7.72 8.69 13.22 1.08f0.01 (1.04)' 1.15fO.01 (1.05)' 28.7f0.50 8.68f0.50 p-( CH&Si- m-( CH8)sSi- 1.59 ,0800 ,0823 .0440 23.3f0.6 9.98f0.44 6.00f0.02 1.00 1.10f0.01 12.6f0.3 6.64f0.08 (6.01)' (1.07)' P-CHI- 6.00&0.01 1.00 .... ... ......... ......... ......... (5.94)O P-CHsO- G.12fO.02 0.76 ... ... ........ ........ ......... (6.07)' a Reading on PH scale of PH meter calibrated for aqueous buffer solutions at half-neutralization point using glass and volume) solu- saturated calomel electrodes without correction for liquid junction potentials in 50% water-50% ethanol (by tions at 25\". Standard deviations are appended to the average values. Calculated assuming unit activities and readings of pH meter scale equal to logarithm of reciprocal of hydrogen ion concentrations. For reaction of benzoic acids with dipheriyldiazomethane in absolute alcohol solution at 30'. e Rate constants for saponification of ethyl esters in 56% acetone (by weight) at 24.9\5b. f Rate constants for saponification of ethyl esters in 87.83'% ethanol (by volume) at 30.0\see ref. 5a. 0 Values obtained previously' with different apparatus and materials. The procedures have been described earlier in detail.',* and p-trimethylsilylbenzoic acids The ethyl esters of m- were prepared in 50-57% yields by heating the silver salts of the acids' with a slight excess of ethvl iodide in ether: ethyl m-trimethylsilylb&zoate, b.p. 96-91' (1.5 mm.): Anal. Calcd. for C12HlaOZSi: C, 64.54; H, 8.12. Found: C, 64.y; H, 8.10. Ethyl P-trimethylsilylbenzoate, b.p. 105-106 (2.7 mm.). Anal. Calcd. for C12H1802Si: C,64.54; H,8.12. Found: C,64.70; H,8.11. The experimental results for the trimethylsilyl derivatives along with those for some reference compounds are given in Table I. A summary of available u-constant data is pre- sented in Table 11. Reasonable experimental agreement was found with the earlier investigation' and from all of the results it is clear that the net electrical effect of the tri- methylsilyl group on the reactivities of benzoic acids or esters is essentially negligible in the para-position but rather electron-donating in the meta-position. The spread of the meta-a-values is large compared with what is customarily founda and may be related to the bulkiness of the group.4~8 .OS10 .0737 ,0474 Direct Synthesis of Organotin Halides. I. Prepa- ration of Dimethyltin Dichloride' BY A. C. SMITH, JR.,* AND EUGENE G. ROCHOW RECEIVED MAY 4, 1953 TABLE I1 a-CONSTANTS FOR THE TRIMETHYLSILYL GROL-P Reaction\" (Metal (Para) -0 020 (-0 OOOP +0.002 (-0 022)b ionization constants in 50% -0,155 ethanol at 25' ( 0. 208Ih Diphenyldiazomethane rates f0.025 (-0 018)h in absolute ethanol at 30 0\" Alkaline hydrolysis in 87.83% ethanol at 30.0' -0 041-1 Alkaline hydrolysis in .56? acetone at 24.9\" -0.150 Average -0.080 =t 0.042' ( -0.113) - to 02i -0.040 -0.008 f 0.014c ( 0. O4ob Organotin halides have been known for a hun- dred years,a and many compounds of the type R,- SnX+, have been synthesized. With the excep- tion of two sealed-tube syntheses of iodide^,^^^ all preparations have been by indirect and often diffi- cult methods.6 The direct synthesis of organosilicon and organo- germanium halides' led to a study of the direct syn- thesis of organotin halides8 The effects of several metals as catalysts was studied by placing each metal over a thin film of evaporated tin on a micro- scope slide and heating the slide in an atmosphere of methyl chloride for one-half hour at 300\". Any unusual reactivity of the metallic couple was indi- cated by reaction and removal of the tin at the junc- tion of the two metals. Copper proved to be the most promising catalyst, with some activity also ex- hibited by silver and gold. Mercury, iron, sele- nium, arsenic, titanium, antimony, tellurium, cal- cium, magnesium, zirconium, aluminum, chro- (1) From a thesis submitted by A. C. Smith, Jr., to the Graduate School of Arts and Sciences of Harvard University. (2) Metal and Thermit Fellow at Harvard University 1949-1050. (3) C. Lowig. Ann., 84, 309, 313 (1852). (4) A. Cahows, ibid., 114, 373 (1860). (5) Karantassis and Basileiados, Compt. vend., 206, 460 (1937). (6) An extensive discussion is given in E. Krause and A. v. Grosse, \"Die Chemie der metall-organischen Verbindungen,\" Chapt. V, Borntraeger, Berlin, 1937, pp. 311-372. JOURNAL, 67, 963 (194.5); 69. 1729 (1947). (7) E. G. Rnchnw. THIS (8) The experiments reported herein were conducted in the period 1949 to 1951. After a full account had been prepared, U. S. Patent 2,625,559 in the same field, (Frederick A. Smith, assigned to Union present Carbide and Carbon Corp.) appeared on Jan. 13, 1953. The paper is a condensed version dealing almost entirely with those aspects of our work not touched upon by the F. A. Smith patent. Full de- tails of our own werk are available in the thesis of reference (1). - Data for log k, and p are given elsewhere.1J Standard deviations. DEPARTMENT OF CHEMISTRY AND LABORATORY FOR NUCLEAR SCIENCE AND ENGINEERING MASSACHUSETTS INSTITUTE OF TECHNOLOGY CAMBRIDGE 39, MASSACHUSETTS Ref. 1. (5) (a) K. Kindler, Ann., 460, 1 (1926); (b) E. Tommila and C. N. Hinshelwood, J. Chcm. SOC., 1801 (1936). (6) Cf. ref. 5 of J. D. Roberts, R. A. Clement and J. J. Dryadale, THIS JOUINAL, 7S, 2191 (1951); C. C. Price and D. C. Lincoln, ibid., 73, 5136 (1951). 4104 NOTES VOl. 75 Gas Outlet Gas Inlet Fig. 1.-Dimensions: reaction chamber, 65 X 200 mm., condenser (body), 80 X 200 mm. mium, silicon, manganese, cobalt, nickel, zinc, cad- mium, germanium, lead, bismuth and iodine gave no indications of being good catalysts, although the mercury dissolved the coating of tin. passed through molten ting Following this survey, methyl chloride was with various proportions of copper added to the tin at various temperatures, as catalyst. Reaction occurred point of the tin at the melting range was practically pure dimethyltin dichloride,1° to- of 300-350\" (232\") being optimum. and up to 450°, The product with the gether with very small amounts of trimethyltin chloride and methyltin trichloride. conditions, a yield At optimum per hour could be obtained from an input of of 0.3 g. of dimethyltin dichloride per minute the excess methyl chloride being recoverable at the (1 g. per hr.) of methyl chloride, most 10 cc. of output. could be expected that copper(I1) chhide, cop- Under the operating conditions employed, it per(I1) oxide, copper(1) oxide, tin(I1) chloride, tin- impurities at the start (11) oxide, and tin(IV) oxides might be present as during operation. tested separately for its effect both Each or be formed to some extent of these substances was period on the induction tion. of the reaction and on its continued opera- marked effect, and that was to increase the induc- Of these, only tin(I1) oxide showed any tion period, an effect also noticed when the molten tin stood in contact with air before admitting methyl chloride. gen before admitting methyl chloride niaterially Treatment of the tin with hydro- reduced the induction period. The effects of zinc, lead, sodiuni, iiiercury water tion then were studied. on the starting .tnd and the showed any pronounced Of progr thesc., only cause an increased output effect, And that sotliuin WAS to and dium &loride a corresponding retention of trimethyltin chloride of chlorine as so- The useful dgratiaq iq the reactiaq vggsel, af ture, by the aewm~latian the reaetiart was shorterred particularly at the surface af reddues in of the reaction the tin. When- mix- (9) Anat 001% Fe, 001% Pb, 0.03% Zn, no As or Cu An- other source prod&ed 99 981% pure tie seutainrng kXf304% Bb, 0 001 4;: Cu, 0 007% Sb, and 0 008% Fe, but this did uot read readrlv with OHa- til 0 2 to 1 0% of Zn was added 16) White erystais, easitj. wbWat Bu to e. 50?, m.p 106\" ipef b SOo) Asai Cald fos e2lf&nC% 11 34; SX 384 CI. 17 2R t ounn c, 1 I li H. 2 00. ('1 R2 21 ever considerable residue appeared, the yield of dimethyltin dichloride decreased and the product became increasingly contaminated with by-prod- ucts and with tarry material. found to be composed chiefly of copper-tin inter- The residues were metallic compounds and porous carbon. ating the reaction between pure tin and methyl By oper- chloride at the optimum conditions determined with copper as catalyst, it was found that dimethyl- tin dichloride could be produced less dependably but in favorable yields without a catalyst. The accumulation of harmful residues was reduced that the reaction vessels did not become so greatly choked nor the reaction inhibited, but there was usually an induction period. inhibited by tip(I1) oxide, but relatively uninflu- The reaction was enced by tin(I1) chloride. be catalyzed or aided by the product, which, be- The reaction appears to cause reaction zone. of its volatility, escapes readily from the small vert the methyl chloride into other reactions. amounts Before any product appears, very of inhibiting substances may di- Such substances may very well be those which ac- celerate the pyrolysis of methyl chloride,\" for if the reaction with tin is delayed in getting started the \"right\" way, the accumulation of carbon irom pyrolysis rolysis and keeps the reaction going in the \"wrong\" of methyl groups accelerates further py- way. areas carbon, particulate residues, etc.) interfered in this (of Experiments indicated that extended surface fritted glqss distributors, finely-divided way. those using molten tin were preferred because they Many different reaction vessels were tried, but presented a fresh surface of the metal to the incom- ing methyl chloride. in this direction involved spraying a fine jet The most extreme attempt of mol teii tin into an atmosphere of methyl chloride at 300-350\". inethyltin dichloride was produced, but the yields The reaction was successful in that di- were extremely low because a stream liters of two to four the liquid tin. of gas per minute was necessary to atomize reaction time, and most of the tin agglomerated on This large volume gave a very short the walls, which acted as baffles. diameter and considerable length probably would A tower of large be necessary to get high conversions. generally adaptable to laboratory The type of reaction vessel found to be most consisted which were sealed a gas of a vertical cylinder well. delivery tube 6.5 experiments aX 20 cm., into nectetl cooltd condenser through X horizontal outlet near the top con- nd a thermom- eter a 24/40 joint to tals collected. 8 X 20 cni. in which the crys- an inclined air- ing the crystals and collecting the liquid The product was removed by melt- livery tube at held approximately the lower end. The reaction chamber from a de- about 1 kg. ments were 100 of 8. ner indicated e&& af produet. tin; the condenser below. out in Nearly a hundred experi- such vessels, in the man- Experimental Copper-eatatyzed Reaction of Methy1 with In one example, - 1400 g. of tin a vessel and 130 of the type described was Cl@@de wMi g. of copper powder, and charged Tin.- was (I 11 R Wpislrr. I nrw 7 , 63. I82 (IHYW) Aug. 20, 1953 heated to 305\" under an atmosphere of methyl chloride. The temperature was raised to 370' to start the reaction, and then dropped to 315' after 1.5 hours. Methyl chloride was introduced at a rate of 30 cc./min. at room temperature. Crystals of dimethyltin dichloride formed in 12 minutes. Over an interval of 862 hours the rate of formation of (CH3)a- SnC12 increased slowly to 3.3 g./hr., then declined after 250 hr. to 0.71 g./hr. The addition of fresh tin raised the rate to 2 1 g /hr. but after 400 hr. it declined to 0.81 g./hr. again. The average yield was 1.8 g./hr. Uncatalyzed Reaction of Methyl Chloride with Tin.-In one example, a small reaction vessel of the same general type was charged with 170 g. of molten tin at 460° under an atmosphere of methyl chloride. The gas input was ad- justed to 30 cc./min. Crystals of dimethyltin dichloride formed within 20 minutes, but the gas jet plugged inter- mittently for the first 120 hours and production during that period was low. Following this period operation was satis- factory at an average yield of 0.58 g./hr. for 774 hr. NOTES 4105 halide with metallic tin,' some new methods for preparing organotin halides were found which represent improvements over the previous meth- ods.* Tin oxide was found to impart an inhibiting ef- fect to the direct reactian of methyl chloride with metallic tin.' In an attempt to determine the rea- sons for the inhibition, methyl chloride was passed through a Pyrex tube containing powdered tin ox- ide (either with or without copper powder) at 300\". In either case, at the start of the reaction a narrow yellow band formed at the gas-inlet end of the tube and progressed slowly down the tube to the outlet end. During the passage of this band along the tube, trimethyltin chloride was produced. From its color and the known reaction of diethyltin with Discussion ethyl chloride to produce triethyltin chlorideI3 we The observed very high proportion of symmetri- conclude that the material was dimethyltin. Its cal dimethyltin dichloride follows a trend noted in formation can be postulated as the corresponding reactions with silicon and germa- 3SnO + 2CH,Cl+ (CH&Sn + SnOCll + SnOa ni~m~,'~ and may be explained in the same way as those found in the case of ~ilicon.'~ A speculative The insoluble, non-volatile dimethyltin then ab- mechanism for the catalyzed reaction is similar to sorbs more methyl chloride to form volatile tri- methyltin chloride the one proposed for silicon.14 ~CU CH3C1 CUCl CUCH~ (1) Sn+ CuCl + (SnCl) Cu (2) CUCH~ + CU CH3 (3) (SnCl) CH3 +( CHaSnCl), or (4) (SnCl) CuCH3 + (CH3SnCl) Cu (5) + ca. 300' + + + + + CHsCl + (CH&Sn + (CHs)sSnC1 + and so on, until tetra-substitution has occurred. Rearrangement of methyl and chlorine groups at the reaction temperature then results in dimethyltin dichloride almost exclusively. In the uncatalyzed reaction it could be assumed that tin can split methyl chloride in much the same manner as copper does. Alternatively, the \"in- electrons of tin allow divalent intermediates ert\" 5s2 which are covalently unsaturated and may absorb more methyl chloride. CH3Cl + Sn + CH3SnCl (which rearranges to (CH3)nSn and SnC12) (CH3)2Sn + CH3C1 +( CHI)3SnC1 SnClz + CH&1 + CH3SnC13 (6) (7) (8) followed by rearrangement as before. In support of the latter mechanism, yellow polymeric di- methyltin and molten colorless tin(I1) chloride have been observed in \"uncatalyzed\" reactions, and the feasibility of reaction (8) was demonstrated by passing methyl chloride into melted tin(I1) chloride, yielding methyltin trichloride. (12) E G Kochow, THIS JOURNAL, TO, 436 (1948) (13) P D Zemany and F P Price, *bid, 70, 4222 (1948) (14) D. T HurdandE G Rochow,tb MALLINCKRODT LA~O~ATORJI AP\\\"AfrD PNJVFkSITY' 38: %A$ '' ' * Only one-third of the tin thereby is converted to volatile product, leaving a powdered mixture of tin(1V) compounds and causing the reaction zone to move along the tube. In a similar way trimethyltin bromide was pre- pared from methyl bromide and tin oxide, and al- though no pure trimethyltin iodide was isolated from the analogous reaction of methyl iodide, a small amount of liquid with the expected boiling and freezing ranges was obtained. Pfeiff er and Heller* previously have reported the preparation of compounds of the type RSn13 by the reaction of RI with tin diiodide. We have ex- tended this method in the present investigation to the combination of methyl chloride and tin di- chloride at 300\" to give methyltin trichl~ride.~ Similarly, in testing the effect of tin dioxide on the reaction between methyl chloride and tin, some tin dioxide was mixed with 8% by weight of copper(I1) oxide and treated with methyl chloride at 300\". Within a few haurs a considerable quantity of di- methyltin dichloride containing small amounts of tin dichloride was obtained. It seems probable that the tin dioxide is reduced by the methyl chlo- ride to tin oxide and tin dichloride, both of which are capable of reacting with methyl chloride to produce methyltin chlorides. Such a mixture of methyltin chlorides is known to rearrange at the re- action temperature to produce dimethyltin dichlo- ride as the principal product. using a reaction vessel of the type described ig F&f€%rt$ Q) BF hay fsttsd %a* mgtjtyl brsmids 4 2. $fntths qqd R 9. BoShoV, Tzjq JOBtT'Lt.3 18, 4193 grause and A v Grosse, \"Die Chemie der metall-organischen Verbindungen,\" Chapt. V, Borntraeger, Berlin, 1937, pp 31 1-372 (3) P. Pfeiffcr, Ber , 44, 1269 (1911). (4) P Pfeiffer and I Heller, ibid , 87, 461% (1904) (5) This reaction was carried out on a half-kilogram scale by Mr John W. Farnham while associated with us in this investigation, and he found that the addition of 10% by weight of copper powder to the melted SnClt enahlcrl the reaction to proceed more smoothly and rap- idly New Preparative Methods for Organotin Halides BY ALBERT C. SMITH, JR., AND EUGENE G. ROCHOW RECEIVED MAY 9, 1953 '#?'& Sf.3 During the investigation of the preparation of or- ganotin halides by the direct reaction of an alkyl