Carbon-Carbon Bond Formation by Reductive Coupling with Titanium(II) Chloride Bis(tetrahydrofuran)* John J. Eisch**, Xian Shi, Jacek Lasota Department of Chemistry, The State University of New York at Binghamton, Binghamton, New York 13902-6000, U.S.A. Dedicated to Professor Dr. Dr. h. c. mult. Günther Wilke on the occasion of his 70th birthday Z. Naturforsch. 50b, 342-350 (1995), received September 20, 1994 Carbon-Carbon Bond Formation, Reductive Coupling, Titanium(II) Chloride, Oxidative Addition, Carbonyl and Benzylic Halide Substrates Titanium(II) bis(tetrahydrofuran) 1, generated by the treatment of TiCl4 in THF with two equivalents of n-butyllithium at -78 °C, has been found to form carbon-carbon bonds with a variety of organic substrates by reductive coupling. Diphenylacetylene is dimerized to ex­ clusively (E,E)-1,2,3,4-tetraphenyl-l,3-butadiene; benzyl bromide and 9-bromofluorene give their coupled products, bibenzyl and 9,9'-bifluorenyl, as do benzal chloride and benzotrichlo- ride yield the l,2-dichloro-l,2-diphenylethanes and l,l,2,2-tetrachloro-l,2-diphenylethane, respectively. Styrene oxide and and ris-stilbene oxide undergo deoxygenation to styrene and fra/«-stilbene, while benzyl alcohol and benzopinacol are coupled to bibenzyl and to a mix­ ture of tetraphenylethylene and 1,1,2,2-tetraphenylethane. Both aliphatic and aromatic ke­ tones are smoothly reductively coupled to a mixture of pinacols and/or olefins in varying proportions. By a choice of experimental conditions either the pinacol or the olefin could be made the predominant product in certain cases. The reaction has been carried out with heptanal, cyclohexanone, benzonitrile, benzaldehyde, furfural, acetophenone, benzophenone and 9-fluorenone. In a remarkable, multiple reductive coupling, benzoyl chloride is converted into 2,3,4,5-tetraphenylfuran in almost 50% yield. The stereochemical course of two such couplings, that of diphenylacetylene to yield exclusively (E,E)-1,2,3,4-tetraphenyl-l,3-buta­ diene and that of acetophenone to produce only racem/c-2,3-diphenyl-2,3-butanediol, is inter­ preted to conclude that the couplings proceed via two electron transfer pathways (TET) involving titanium(IV) cyclic intermediates of the titanirene and the oxatitanacyclopropane type, respectively.

The monomolecular hydrodeoxygenation or bi- termining the reducing action of the resulting re­ molecular reductive coupling of a wide gamut of agent is uncertain. The ill-defined nature of such organic substrates has been found to occur by the reductants is readily evident from the numerous action of various reactive metals, metal hydrides titanium-based reagents reported to be formed or subvalent metal complexes [1,2]. Such reducing when TiCl4, TiCl3 or CpTiCl2 is treated with, agents often are employed in heterogeneous reac­ among others, RLi, RMgX, R3AI, LiAlH4, Li, K, tion media either as highly dispersed metal par­ Mg or Zn [2], Outstanding among these reducing ticles or as metals adsorbed on solid supports such combinations for its versatility in organic synthesis as graphite. In many other cases, the reducing is the McMurry Reagent, a black suspension of agent is generated, in situ, by treating a transition some form of titanium(O) generated when a 4:1 metal salt with a main group metal, metal hydride mixture of LiAlH4 and TiCl4 is added to THF [2], or metal alkyl. Although it is certain that the tran­ With this backdrop and in connection with our sition metal center is thereby reduced, the exact investigation of new routes to transition metal oxidation state formed is often uncertain and the borides [3], we recently found that titanium(II) role of the main group metal reductant in de­ chloride could be readily synthesized from ti- tanium(IV) chloride by simply adding two equiv­ alents of a metal alkyl to TiCl4 in toluene or tetra- * XIII Communication of the series, “Organic Chemis­ hydrofuran (eqs 1-3): try of Subvalent Transition Metal Complexes”; XII The titanium(II) chloride bis(tetrahydrofuran) 1 Communication: J. Am. Chem. Soc. 108, 7763 (1986). formed in eq. 1 could be obtained free of LiCl ** Reprint requests to Prof. J. J. Eisch. and analytically pure by evaporating the THF and

0932-0776/95/0300-0342 $06.00 © 1995 Verlag der Zeitschrift für Naturforschung. All rights reserved. J. J. Eisch et al. ■ Carbon-Carbon Bound Formation by Reductive Coupling with TiCl2-2THF 343

TiC14 + 2 BunLi ----- — ------► TiCI2-2THF + 2 BunH (D - 2 LiCl 1

TiCl4 + 2 H2C=C H C H ,M gC l — _ u ■» TiCl2*2THF*2MgCl2*4rHF (2) * * L3H6 2

TiCL + 2 Me-iAl ------► TiCl2*Me2AlCl + 2 CH4 (3) - Me2AlCl extracting 1 into toluene. The titanium(II) chloride This observation is consistent with a 2:1 stoichi­ 2 formed in eq. 2 was weakly complexed with the ometry of reaction and the formation of tetrachlo- magnesium chloride by-product and that in equa­ rodititanoxane(III) 6. tion 3 formed a stable complex with Me2AlCl 3. Both 2 and 3, when admixed with an excess of b) Scope of organic substrates reducible by 1 R„A1C13_„, function as highly active hetero­ (Table I) geneous Ziegler catalysts for the polymerization of ethylene and higher olefins, as has been prelimi­ a) Hydrocarbons: Although titanium(II) chlo­ narily reported elsewhere [4]. ride in the form of complexes 1, 2 and 3 and ad­ With a well-defined, soluble subvalent titanium mixed with a six to eight-fold excess of Me2AlCl complex in hand, we were well-positioned to ex­ is able to catalyze the polymerization of ethylene plore the scope and the mechanism of reduction and other alpha-olefins [4], complex 1 in THF or of organic substrates by titanium(II) chloride bis- unsolvated TiCl2 suspended in toluene [6] caused (tetrahydrofuran) 1. We report here the results of neither reduction nor oligomerization of such ole­ our investigation thus far. fins as styrene and 1,1-diphenylethylene, even after 24 h in refluxing solution. Diphenylacetylene 7, Results however, underwent a slow bimolecular reduction to yield solely (E,E)-1,2,3,4-tetraphenyl-l,3-buta­ a) Reaction conditions and stoichiometry diene 9 (entry 1 in Table I) upon hydrolysis (eq. 5): Reductions with 1 were initially conducted with That the organotitanium precursor to 9 is most the lithium chloride-free reagent in refluxing tolu­ likely l,l-dichloro-2,3,4,5-tetraphenyltitanole 8 is ene or tetrahydrofuran solution. Since the pres­ supported by the photoreaction of ? 73-allyltitano- ence of the LiCl had no marked effect on the re­ cene 10 with 7, whereby titanole 13 is formed in ducing activity of 1 for most substrates, subsequent 60% yield [4]. The reaction mechanism leading to reductions were carried out directly with the THF 13 involves the photolytic loss of the allyl radical solutions of 1 still containing the suspended LiCl from 10 and the generation of titanocene(II) 11. (eq. 1). The ratio of 1 to the organic substrate This undergoes oxidative addition with 7 to pro­ ranged from 2:1 to 4:1. However, with diaryl ke­ duce titanirene 12, which inserts a further unit of tones, such as benzophenone, failure to remove 7 to produce 13 (Scheme 1). the LiCl prior to reduction led to a less active re­ ß) Halides: Aromatic halides and aliphatic ha­ agent [5]. lides, as typified by p-bromoanisole and 1-bromo- The stoichiometry of one reduction employing 3-phenylpropane, underwent no discernible re­ 1, which was free of LiCl, is significant: a 1:1 ratio duction by 1 during 24 h in refluxing THF. On the of 1 and benzophenone 4 gave a 48% yield of other hand, benzylic halides, such as benzyl bro­ tetraphenylethylene 5 (eq. (4)): mide 14 gave exclusively the bimolecular re-

2 Ph2C = 0 + 4 TiC l2 ------► Ph2C=CPh2 + 2Cl2Ti— O— TiCl2 (4) A 4 1 5 6 344 J. J. Eisch et al. • Carbon-Carbon Bound Formation by Reductive Coupling with TiCl2-2THF

Table I. Reduction of organic substrates with TiCl2-2THF 1.

Entry Substrate3 Products13 Yield'

1 Diphenylacetylene3 (E,E)-1,2,3,4-Tetraphenylbutadienee 14 2 Benzyl bromide Bibenzyl 100 3 9-Bromofluorene Fluorene 18 9,9 '-Bifluorenyl 82 4 Benzal chloride 1,2-Dichloro-l ,2-diphenylethanesf 97 5 Benzotrichloride l,l,2,2-Tetrachloro-l,2-diphenylethane 92 6 Dichlorodiphenylmethane Tetraphenylethylene 96 7 Benzopinacol Tetraphenylethylene 39 Tetraphenylethane 51 Benzophenone 10 8 Styrene oxide Styrene 90 9 ris-Stilbene oxide rra«s-Stilbeneg 98 10 N,N-Diphenylaminomethyl phenyl sulfide Methyldiphenylamine 15 11 Benzyl alcohol Bibenzyl 70 12 Benzonitrile Benzyl phenyl ketone 10 13 Heptanal 7,8-Tetradecanediol 80 14 Cyclohexanone 1,1 '-Dihydroxydicyclohexyl dicyclohexylidene 60 15 Benzaldehyde rran5-Stilbeneh 98 16 Furfural (E)-l,2-Bis(2-furyl)ethane 95 17 Acetophenone (E)-2,3-Diphenyl-2-butene 88 Acetophenone rac-2,3-Diphenyl-2,3-butanediol 83 (E)-2,3-Diphenyl-2-butene 13 18 Benzophenone Tetraphenylethylene 58 19 9-Fluorenone 9,9 '-Bifluorenylidene 44 20 Benzoyl chloride 2,3,4,5-Tetraphenylfuran 47

a Unless otherwise specified, all reaction were conducted by allowing a 4:1 molar ratio of the LiCl-containing TiCl2 and the organic substrate to reflux in THF solution under an argon atmosphere for 24 h. The individual runs employed about 2.5 mmol of the substrate dissolved in 30 ml THF; b the product were isolated from the hydrolyzed reaction mixture by column chromatography and identified by comparing their TLC, GC, m.p. and JH and 13C NMR spectral properties with those of authentic samples;c the yields are those of the isolated components but are not yet optimized; d a 2:1 ratio of the acetylene to 1 were employed in a 60 h reaction; e a 40 h reaction time was employed; f a 2:1 ratio of racemic and meso isomers resulted; 8 the ds-isomer was present in 3% yield; h less than 1 % of the c/s-isomer was found.

Ph Ph Ph Ph TiCl, H ,0 2 Ph— C = C — Ph n V . (5) r " ^ T r 1 Ph'-'x /"-Ph H H c / V'C1

S ch em e 1 Ph Ph Ph N/ Ph Cy f T\ 'Cp Cff 'Cp

10 12 1 3 J. J. Eisch et al. • Carbon-Carbon Bound Formation by Reductive Coupling with TiCl2-2THF 345 duction product, bibenzyl 15 (entry 2 in Table I). Noteworthy is the inertness of the benzylic The more hindered 9-bromofluorene 16 provided C -C l bonds in 20 and 22 to further reductive elim­ 82% of 9,9'-bifluorenyl 18, but also 18% of fluo- ination and the formation of stilbenes and di- rene 17 (eq. 6) (entry 3 in Table I): phenylacetylene, respectively. By contrast, the pre-

16 17 Even polyhalobenzylic halides underwent fairly sumed intermediate in eq. 8, 1,2-dichloro-l,1,2,2- efficient coupling with 1: a) benzal chloride 19 tetraphenylethane, is readily dechlorinated by 1. provided a 2:1 mixture of racemic- and m eso- 1,2- y) Alcohols and ethers: Tetrahydrofuran itself dichloro-l,2-diphenylethanes 20 (entry 4 in Ta­ showed no sign of reductive cleavage to 1-butanol ble I); b) benzotrichloride 21 yielded 1,1,2,2-tetra- after 24 h reflux with 1. Ordinary alcohols were chloro-l,2-diphenylethane 22 (eq. 7) (entry 5 in likewise unreactive, although benzyl alcohol was Table I); and c) dichlorodiphenylmethane 23 pro­ slowly coupled to produce bibenzyl (entry 11 in duced tetraphenylethylene 5 (eq. 8) (entry 6 in Table I) (eq. 9). Benzpinacol 25 underwent a sig­ Table I): nificant amount of beta-bis(dehydroxylation) with the formation of 5, 25 and 4 (eq. 10) (entry 7 in Table I). R R The significance of the formation of 25 and the TiCl, V , * C1 Ph— C1 — C 1 — Ph m generation of benzophenone 4 will be analyzed Ph^ NC1 I I Cl Cl later in the discussion of the mechanisms of TiCl2- reductions (cf. infra). 19: R = H 20: R = H (rac+ meso) In this class of compounds, epoxides proved to 21: R = Cl 22: R = Cl be the most reactive: Both styrene oxide 26 and ds-stilbene oxide 27 were readily deoxygenated to the olefin in high yield (eq. 11) (entries 8 and 9

P h. JZ\ TiCl, Ph Ph in Table I): ^C=C( (8) With 27 it is noteworthy that the deoxygenation Ph Cl Ph Ph proceeded with high stereoselectivity (29 trans: 23 cis = 97:3).

TiCI2 2 PhCH2OH ------► PhCH2— CH2Ph (9)

p\ ,ph TiCl2 Ph— C — C — Ph ------► Ph2C=CPh2 + Ph2CH —CHPh2 + Ph2C = 0 (10) HO OH

24 5 (39%) 25(51%) 4(10%)

19 346 J. J. Eisch et al. • Carbon-Carbon Bound Formation by Reductive Coupling with TiCl2-2THF

P. H R Me Ph wp pL url/ / ,.,-.t___ \p . i t H i-i liv-l? n ‘ / /ii* Me JVIC^ TiCli 11V-IJ I I II \ Mev / JC ~ C v ------► y C — C (1 1 ) 2 ___C = 0 ------► Ph—C—C-Me + X=CV (15) p/ VR Ph' H ^ J I.. Ph' Me HO OH 34 35 36(92%) 26: R = H 28: R = H + 8 % Z -isom er 27: R = Ph 29: R = Ph

The pinacol 32 or the olefin 33 could in some d) Sulfides: As expected, sulfide linkages are cases each be made the predominant product by more readily cleaved than ether linkages. How­ varying the ratio of 1 to the amount of substrate. ever, their response to 1 is relatively slow. One Again illustrated with acetophenone, two equiva­ such cleavage observed thus far is the following lents of 1 to ketone gave an 83% yield of pinacol (eq. 12) (entry 10 in Table I): 35 while a 4:1 ratio of 1 to ketone produced 88% of olefin 36. Aliphatic aldehydes and ketones, such TiCl, as heptanal and cyclohexanone, gave principally Ph2N -C H 2-SPh Ph2N -CH 3 HS-Ph (12) the pinacol (entries 13 and 14 in Table I). With aromatic aldehydes, such as benzaldehyde 37 and e) Carbonyl derivatives: The great ease of reduc­ furfural 38, little or no pinacol was detected re­ ing carbonyl derivatives by 1 is undoubtedly con­ gardless of the proportion of 1 employed (eq. 16) nected with the high oxophilicity of the titanium (entries 15 and 16 in Table I): in any oxidation state. Because aldehydes, ketones H Ar and acid chlorides are readily reductively coupled TiCl, by 1, it should first be noted that carboxylic acids ; c = o Nc=c^ (16) A r' and esters have proved to be unreactive, and ni- a / n h triles only slow to reduce. Benzonitrile 30 is slowly 37: Ar = Ph converted to a product that yields benzyl phenyl 38: Ar = 2-furyl ketone 31 upon hydrolysis (eq. 13) (entry 12 in As already noted, the LiCl-containing 1 seemed Table I): to be less effective in coupling diaryl ketones. Even with a great excess of 1, neither benzo- O 1.T iC l2 phenone 4 nor 9-fluorenone gave more than a 40- 2 P h — C = N II (13) 2. H 20 Ph-C H 2—C — Ph 60% yield of the corresponding olefin 33 (entries 3 0 31 (10%) 18 and 19 in Table I). In fact, much more efficient coupling to form tetraphenylethylene can be ob­ Both aldehydes and ketones are reduced by 1 to tained by employing Ph2CCl2 with 1 (eq. 8). give mixtures of the pinacol 32 and the corre­ The most remarkable reductive coupling of a sponding olefin 33 (Table I) in varying pro­ carbonyl derivative by 1 is that observed with ben­ portions (eq. 14): zoyl chloride 39. The product of this reaction, ob­ tained in almost 50% yield, is 2,3,4,5-tetraphe- R' R T iC l, I I (14) nylfuran 40 (eq. 17) (entry 20 in Table I): 'c=o R — C — C — R' r = c I I H O OH Ph Ph o TiCl, R, R' = H, alkyl, aryl 3 2 33 II (17) 4 P h— C — Cl Ph Ph Where R and R' were groups of significantly O different steric demands, the preponderant con­ 3 9 4 0 figurations for the pinacols and olefins formed were the racemic (32) and the E-configurations (33), respectively. Thus, in the reduction of aceto- Discussion phenone 34, the pinacol 35 obtained was exclu­ The number and variety of publications con­ sively the dl-isomer and the olefin 36 was 92% of cerning the reduction of organic compounds by the E-configuration (eq. 15) (entry 17 in Table I): low-valent titanium reagents of ill-defined charac- J. J. Eisch et al. • Carbon-Carbon Bound Formation by Reductive Coupling with TiCl2-2THF 347 ter portray a lively and often bewildering field of tanium(II) reagent, 1, containing no additional research. The applicability of such reagents in or­ Lewis acid (MgCl2, A1C13 or ZnCl2), to achieve ganic synthesis and their presumed reaction mech­ carbon-carbon bond formation with a wide variety anisms have been explored by many researchers of carbonyl derivatives, benzylic halides, acety­ and their findings have been ably reviewed and lenes and epoxides. The presence of exclusively ti- assessed by Fiirstner [1] and by McMurry [2], tanium(II) chloride in homogeneous solution What has impeded more definitive mechanistic in­ greatly facilitates any detailed investigation of re­ sight have been the heterogeneous nature of many action mechanism. such reductants and the uncertainty about the oxi­ One of the prime mechanistic questions to be dation states of the active titanium reagents in­ answered for the reactions of reagent 1 is whether volved. Often, the oxidation state is simply as­ single-electron transfers (SET) involving exclu­ sumed to be Ti(0) or Ti(II) without direct proof. sively Ti(III) intermediates are involved or whether The coupling of ketones into tetrasubstituted eth- concerted two-electron transfers (TET), oxidative ylenes by combinations of TiCl4 with zinc dust in additions leading to titanium(IV) intermediates, THF is a case in point: a titanium(II) chloride is are decisive for such reactions. These possible assumed to be the reagent, without assurance that pathways are depicted in Scheme 2 with ben- ZnCl2 or Ti(0) might play a role [7], zophenone. The previous use of well-defined subvalent ti­ tanium compounds for reductions of organic sub­ Schem e 2 strates has been rare. Corey and coworkers have Ph Ph employed the complex of TiCl2(AlCl3)2 with hexa- TiCl, I I 2 Ph2C = 0 2 Ph2C—O—TiCl2 Ph— C — C — Ph methylbenzene to form pinacols from ketones [8], 1 I Cl2TiO OTiCl2 The partly defined complex HTiC10.5THF has TE been implicated as an active reagent in the TiCl, Ph Ph McMurry reaction [9] and the complex, 1 I Ti(MgCl)2 xTHF has also been shown to serve as Ph,C=0 Ph— C — C — Ph a reagent for such ketone couplings [10]. Girolami OTi and coworkers have employed dimethylti- Cl tl tanium(II) bis(l,2-bisdimethylphosphinoethane) 4 6 for the catalytic dimerization of ethylene to 1-but- ene 43 [11]. It is thought that this titanium com­ Although an unqualified choice between SET plex 41, for which the crystal structure is known and TET pathways cannot now be made for all of [12], forms an intermediate titanacyclopentane by the organic substrates reacting with 1, stereochem­ oxidative addition (eq. 18): ical evidence for the reactions of 1 with diphenyl- acetylene (7, eq. 5) and with acetophenone (34, Me Me r P/'.. I ..>'p ~ \ h 2c = c h 2 ■P,. I entry 17, eq. 15) is more consistent with the oper­ (18) >-40 °C OKJ■P^ I - 20 °C ation of a TET or oxidative addition pathway than Me -dmpe Me - TiMe2 (dmpe) ligand = dmpe catalytic SET steps. Were SET operative, the intermediate radicals, 47 and 48, should couple for steric rea­ 41 42 43 sons and provide the Z,Z-isomer of 9, 49, and the Whitesides and coworkers demonstrated that ti- meso-isomer of 35, 50, respectively (eqs 20 and tanocene generated in situ could cyclodimerize 21). ethylene to produce substantial yields of 45 via 44 Since in fact, only the E,E-isomer 9 and the ra- [13] (eq. 19): cemic-isomer 35 are found exclusively in such couplings, SET processes are inadequate to ratio­ 2 e' (19) Cp2TiCl2 ■ 2cr [Cp2Ti] H2C=CHV Cp2T i ^ — O==<0 nalize the stereochemical course of reaction. On the other hand, intermediate 12 (Scheme 1) for the acetylene 7 and intermediate 51 for aceto­ In light of these reports, then, the present study phenone (34, Scheme 3) could be readily be appears to be the first to use a well-defined ti- formed by TET. The resulting rings could undergo 348 J. J. Eisch et a l ■ Carbon-Carbon Bound Formation by Reductive Coupling with TiCl2 -2THF

2 T iC l, T iC l, Me Ph 2 Ph— C=C — Ph 2 P h — C = C . (20) Me. l.Cp2TiC,H5 I I ^ c= o Ph— C — C -M e (2 2 ) 7 Ph 2. HiO I I HO OH 34 35 Ph Ph \ / (Z,Z)-1,2,3,4-TPB 0, C _ C ❖ v Cl2Ti— C C —TiCl2 4 9 Ph Ph One final observation can be offered in support of intermediates like 51 in reactions of ketones with 1. When benzophenone is treated with 1 and Me M?. O TiCl, an alcohol or with Ti(BH4)2, significant amount of 2 /C=0 2 ^C—O—TiCl2 ► Ph-C—C-iPh (2D / V Ph Ph' Cl2TiO Me 1,1,2,2-tetraphenylethane are produced in addition

3 4 4 8 to tetraphenylethylene [16], We suggest that inter­ mediate 46 (Scheme 2) undergoes cleavage of its

meso-glvcol C-Ti bond and the resulting Ph2CH-OTiCl2 un­ SO dergoes reductive coupling to Ph2CHCHPh2. These and other mechanistic aspects of the reac­ tions of 1 are receiving our continuing attention.

Experimental Section Scheme 3 General procedures Me Ph ph,ve ^ePh All procedures involving the purification of re­ Me TiCl, o—c :".pA c «»Me ' c —C :c=o - y // . / \ action solvents, the preparation of titanium(II) Ph' y v o chloride bis(tetrahydrofuran) 1 and the reactions cr ci c r 'c i of 1 with the various organic substrates were con­ 34 51 ducted under an atmosphere of anhydrous, oxy­ gen-free argon. The drying and deoxygenating of argon, as well as of the tetrahydrofuran and the insertion of a second unit of acetylene or ketone toluene used in reactions of 1, were carried out with the ring and the substituents controlling the according to established procedures [17]. stereochemical course of adding the C-Ti bond Instrumentation and analyses (Schemes 1 and 3). The second acetylene is thereby compelled to undergo C-Ti bond ad­ All melting points were measured with a dition in a syn-fashion. Similarly, by an approach Thomas-Hoover capillary melting point apparatus and are uncorrected. Infrared spectra (IR) were of the Si pi-face of 34 to the C -T i bond of 51, recorded on Perkin-Elmer spectrophotometers. steric repulsion of the approaching Ph and Me Models 457 and 283 B, which were equipped with groups is minimized. sodium chloride optics. Nuclear magnetic reso­ Experimental evidence for such TET processes nance spectra (!H and 13C NMR) were obtained and insertions for acetylenes is reported by Alt with a Bruker spectrometer, Model AM-360, on and coworkers, who found that Cp2Ti(PMe3)2 re­ pure samples or as 10% solutions in pure deuteri- acts with acetylene itself to form successively a ti- ated solvents. The 'H NMR data were reported on tanocene like 12 and a titanole like 13 [14], Anal­ the <5 scale in parts per million with reference to ogous evidence for acetophenone was reported by internal tetramethylsilane. Mass spectral data our group, when we found that titanocene, gener­ were collected with a Hewlett-Packard gas chro­ matograph-mass spectrometer. Model 5882 B. ated by the thermal decomposition of ?/3-allyltitan- Gas-liquid phase chromatographic analyses (GC) ocene in THF, effected the bimolecular reduction were carried out with an F&M temperature-pro­ of acetophenone into exclusively the racemic-pin- grammed chromatograph. Model 720, equipped acol 35 (eq. 22). A TET-pathway analogous to that with dual 12-ft column of a 10% UC-298 phase on in Scheme 3 can be deduced from this obser­ a Chromosorb W support and with an electronic vation [15]. peak-area integrator. Thin-layer chromatographic J. J. Eisch et al. • Carbon-Carbon Bound Formation by Reductive Coupling with TiCl2~2THF 349 analyses (TLC) were done on Eastman Chromag- filtered from the LiCl. The black filtrate was ram Sheets, no. 13181, consisting of silica gel with evaporated to form the TiCl2 residue. After drying fluorescent indicator. Analyses of 1 were for 3.5 h in vacuo the black solid was found to con­ conducted on vacuum-dried samples in the follow­ tain 23.53% Ti and 34.76% Cl; ratio of Ti:Cl, ing ways: 1) the dihydrogen evolved from hy­ 1.0:2.00. These values correspond to drolyzed samples was collected and measured; 2) TiCl2-1.2 THF. From these analytical values we the tetrahydrofuran liberated by such hydrolysis conclude that titanium(II) chloride in THF solu­ was extracted and analyzed by GC for any content tion exists as the bis(tetrahydrofuran) but that in of 1-butanol; 3) the titanium(III) ion generated by the solid state one THF unit is labile to dis­ such hydrolysis was oxidized by H 20 2 to ti- sociation. tanium(IV) ion and the latter ion was determined To assure ourselves that the coordinated THF by a complexometric titration with the monosod­ was not actually an n-butoxy group bonded to Ti ium salt of ethylenediaminetetraacetic acid (Com- and possibly formed by reductive cleavage, a plexon III); and 4) the chloride ion in hydrolyzed sample of 1 was hydrolyzed and the evolved THF samples was determined by the Volhard method. analyzed by GC. The THF was found to contain <2% of 1-butanol. Preparation of titanium(II) chloride bis (tetrahydrofuran) 1 Typical procedures for the reactions of lithium chloride-containing titanium(II) chloride To 250 ml of anhydrous THF cooled to -78 °C bis (tetrahydrofuran) 1: trans-Stilbene (solid C 0 2-acetone bath) were slowly added 40 ml from benzaldehyde of a 1.0 M solution of TiCl4 in toluene. After 1 h of stirring at -78 °C a bright yellow solid suspen­ Since all the reactions were conducted in an sion had formed. Then over 2 h 32 ml of a 2.5 M analogous manner and formed products which are solution of «-butyllithium in hexane was gradually known, well-characterized compounds, the follow­ introduced as the suspension successively turned ing procedure should suffice to illustrate the ap­ yellow-green and then light brown. The reaction propriate experimental operations. mixture was thereafter brought to room tempera­ To a solution of TiCl2-2THF 1 (10 mmol) in ture and stirred for 18 h. At this point a black solu­ THF (30 ml) at 25 °C was added freshly distilled tion with suspended LiCl had formed. The reagent benzaldehyde (265 mg, 0.25 ml, 2.5 mmol). The 1 formed at this point was employed directly for resulting reaction mixture was heated under reflux the reactions reported in this article. for 24 h, quenched with H20 (100 ml), and filtered Analytical samples of 1 could be obtained by through a Celite cake. The Celite cake was washed removing much of the THF under reduced pres­ twice with Et20 (2x25 ml). The aqueous layer was sure at 25 °C. Filtration of the black suspension then extracted with Et20 (2x25 ml) and the com­ under argon and washing the filter residue with bined organic extracts dried over anhydrous toluene gave >95% of gray LiCl. The filtrate was MgS04. The extracts were then freed of solvents then evaporated to dryness in vacuo and solid resi­ on a rotatory evaporator and the crude product due washed slowly on the filter with portions of was purified by flash column chromatography a 1:1 (v/v) THF-toluene mixture. The black filter (eluent hexanes/THF 50:1) to give 220 mg of residue was dried in vacuo to yield 95% of essen­ frans-stilbene as a white crystal. M.p. 120-121 °C tially pure 1. (lit. 122-124 °C). 'H NMR (CDC13) (3 (in ppm): Anal. Calcd for C8Cl2H160 2Ti: Ti, 18.21; Cl, 7.48 (d, 4H), 7.32 (t, 4H), 7.22 (t, 2H), 7.08 (s, 26.96; evolved H 2, 0.5 mol; ratio of T :C1, 1.0:2.0; 2H); 13C NMR (CDC13) (3 (in ppm): 137.36, Calcd for C4Cl2H8OTi: Ti, 25.09; Cl, 37.14. 128.71, 128.65, 127.58, 126.51. This product con­ Samples of 1 prepared in the following manner tained < 1% of ds-stilbene. gave the following analyses: Ti, 19.8, Cl, 30.0, ev­ olved H2, 0.45 mol; ratio of Ti:Cl, 1:2.05. These Acknowledgements values correspond to TiCl2-1.75 THF. Alterna­ tively, samples of 1 could be obtained from the This research was initiated under a material sci­ original reaction mixture by complete removal of ence project sponsored by Akzo Corporate Re­ all the THF and toluene at 30 °C in vacuo to pro­ search America Inc., continued under support by duce a gray-black solid residue. The residue was the U.S. National Science Foundation Grant CHE- then stirred with 150 ml of a 1:1 (v/v) mixture of 87-14911 and brought to fruition with funding THF and toluene for 1 h and the suspension then from Solway & Cie, Brussels, Belgium. We are in- 350 J. J. Eisch et al. ■ Carbon-Carbon Bound Formation by Reductive Coupling with TiCl2-2THF debted to Dr. S. L. Pombrik for assistance in im­ proving the preparation of titanium(II) chloride as the THF complex 1 and as the 1:1 complex with Me2AlCl 3.

[1] A. Fiirstner, Angew. Chem., Int. Ed. Engl. 32, 164 [9] L. E. Aleandri, B. Bogdanovic, A. Gaidies, D. J. (1993). Jones, S. Liao, A. Michalowicz, J. Roziere, A. Schott, [2] J. E. McMurry, Chem. Rev. 89, 1513 (1989). J. Organomet. Chem. 459, 87 (1993). [3] J. J. Eisch, J. Lasota, S. L. Pombrik, “Novel Route to [10] L. E. Aleandri, S. Becke, B. Bogdanovic, D. J. Jones, Titanium Bis(borohydride) and its use as a Precur­ J. Roziere, J. Organomet. Chem. 472, 97 (1994). sor to Titanium Dichloride”, Invention Disclosure [11] M. D. Spenser, P. M. Morse, S. P. Wilson, G. S. Girol- No. R-920, Dec 3, 1991, The Research Foundation ami, J. Am. Chem. Soc. 115, 2057 (1993). of the State University of New York, U.S.A. [12] J. A. Jensen, S. R. Wilson, A. J. Schultz, G. S. Girol- [4] J. J. Eisch, S. L. Pombrik, X. Shi, S.-C. Wu, Macro- ami, J. Am. Chem. Soc. 109, 8094 (1987). molecular Symposia, in press. [13] J. X. McDermott, M. E. Wilson, G. M. Whitesides, J. [5] The retarding effect of LiCl on the reducing action Am. Chem. Soc. 98, 6529 (1976). of 1 with diaryl ketones may stem from com- [14] H. G. Alt, H. E. Engelhardt, M. D. Rausch, L. B. plexation to form LiTiCl3 intermediates, but this as­ Kool, J. Am. Chem. Soc. 107, 3716 (1985). pect merits further study. [15] J. J. Eisch, M. E. Boleslawski, J. Organomet. Chem. [6] We are currently investigating the reducing ef­ 334, C l (1987). ficiency of TiCl2 generated by treatment of TiCl4 in [16] J. J. Eisch, S. L. Pombrik, X. Shi, unpublished toluene directly with «-butyllithium in hexane. studies. [7] D. Lenoir, Synthesis 8, 553 (1977). [17] J. J. Eisch, Organometallic Synthesis, Vol. 2, pp. 7 - [8] E. J. Corey, R. L. Dansheiser, S. Chandrasekaran. J. 25, Academic Press, New York (1981). Org. Chem. 41, 260 (1976).