Metalation Studies on Titanocene Dithiolates

Titanocene bis-arylthiolates [(C5H4X)(C5H4Y)Ti(SC6H4R)2] (X,Y = H, Cl; R = H, Me) can be prepared from the corresponding titanocene dichlorides by reacting with the thiols in the presence of DABCO as a base. They react with n-butyl lithium to give unstable Ti(III) radical anions. While the unsubstituted thiolates (X = Y = R = H) react with lithium Di-isopropylamide by decomposing to dimeric fulvalene-bridged and thiolate-bridged Ti(III) compounds, the ring-chlorinated compounds can be deprotonated with LDA and give appropriate electrophiles di-substituted and tri-substituted titanocene dithiolates.


Introduction
Titanocene compounds, i.e., compounds of the type "Cp 2 TiX n " where "Cp" stands for a substituted or unsubstituted cyclopentadienyl ligand and "X" for any anionic or neutral ligand and n = 0-2 are arguably the second-most studied metallocenes after ferrocene.Soon after the first synthesis of (C 5 H 5 ) 2 TiCl 2 in 1954 [1], its potential for acting as a polymerization catalyst when combined with certain aluminum compounds was discovered [2].Further studies showed that a judicious choice of substituents on the cyclopentadienyl ring had a great influence on the stereochemistry of the produced polymers, which gave rise to an enormous number of publications and also many review articles [3][4][5][6][7][8].However, it was also found that the titanium catalysts were rather quickly deactivated and the corresponding zirconium compounds showed much higher stability and catalytic activity.Thus, the mainstream research on metallocene catalysts stopped for titanium-based metallocenes nearly completely until recently when the concepts of "Green Chemistry" and "Sustainable Catalysis" came into play [9][10][11][12].Another area of "applied research" opened up after the discovery that Cp 2 TiCl 2 showed antitumor activity [13][14][15].Here again, it turned out that the effectiveness of the titanocenes could be enhanced by introducing substituents on the cyclopentadienyl rings [16,17] as well as by modifying the "X"-ligands [17,18].Lastly, titanocenes were the subject of purely "academic" studies, e.g.studies devoted to synthesis, isolation, and characterization of the "true" titanocenes "Cp 2 Ti" [19] or of "chiral at titanium" complexes "CpCp'TiXY" [20].In addition, in these studies, the importance of cyclopentadienyl ring substituents was well established.All of the titanocenes with substituted cyclopentadienyl rings known so-far have been prepared by a reaction of the substituted cyclopentadiene with TiCl n (n = 2-4).This method fails for very electronegative substituents or substituents with ligating properties.Since our group focuses on the synthesis of Cyclopentadienyl complexes with such substituents for a long time, using an approach of performing halogen-metal exchange reactions followed by electrophilic substitutions on already coordinated perhalogenated cyclopentadienyl ligands [21], we wondered if our approach was also suitable for the titanocene system.Since our synthetic protocol always involves the use of lithium organyls or lithium amides, there was also the need for a proper titanocene starting material.It has been known for a long time that titanocene chlorides react with alkyl lithiums to thermally unstable titanocene alkyls Cp 2 TiR 2 , which easily decompose to titanium(III) products [22,23] and with lithium amides to mono-cyclopentadienyl titanium tris-amides CpTi(NR 2 ) 3 [24].In addition, with the often-used base KO t Bu, a mixture of products including those of splitting off the cyclopentadienyl ligands was observed [25].We, therefore, decided to look at the also long-known titanocene thiolates Cp 2 Ti(SR) 2 .These compounds are usually prepared either by a reaction of Cp 2 TiCl 2 with thiols in the presence of the base, which were sometimes contaminated with the mixed chloride-thiolates Cp 2 Ti(SR)Cl [26][27][28] or by oxidative addition of disulfides to in-situ prepared "Cp 2 Ti" [29].Some of these thiolates showed promising antitumor properties [30,31].
In this study, we report on the synthesis of several titanocene bis(aryl thiolates) (Cp)(Cp')Ti(SAr) 2 and their reactivity towards lithium alkyls and amides including functionalization of the cyclopentadienyl rings.It should be mentioned here that a related approach was used for deriving the so-called "Troticenes" CpTi(Cht) [32].
In this study, we report on the synthesis of several titanocene bis(aryl thiolates) (Cp)(Cp')Ti(SAr)2 and their reactivity towards lithium alkyls and amides including functionalization of the cyclopentadienyl rings.It should be mentioned here that a related approach was used for deriving the so-called "Troticenes" CpTi(Cht) [32].

Reaction of Dithiolates 2a and 3a with Butyllithium
Treatment of a suspension of 2a or 3a with 1.1 equivalents of n-BuLi in THF at −78 °C results in the formation of yellow solutions.Warming to room temperature leads to a color change from yellow to brown within five minutes.The resulting solutions appear to be paramagnetic since it turns out impossible to obtain any good-quality 1 H-or 13 C-NMR spectra.However, the solutions resulting from 2a show a 7 Li-NMR resonance at δ = 2.64 ppm while the solutions resulting from 3a exhibit a 7 Li-NMR signal at δ = 2.74 ppm (Figures S5 and S6).For comparison, the 7 Li-NMR signals of Compounds 2a and 3a had been reported before [26,27].All compounds were characterized by 1 H-NMR and 13

Reaction of Dithiolates 2a and 3a with Butyllithium
Treatment of a suspension of 2a or 3a with 1.1 equivalents of n-BuLi in THF at −78 • C results in the formation of yellow solutions.Warming to room temperature leads to a color change from yellow to brown within five minutes.The resulting solutions appear to be paramagnetic since it turns out impossible to obtain any good-quality 1 H-or 13 C-NMR spectra.However, the solutions resulting from 2a show a 7 Li-NMR resonance at δ = 2.64 ppm while the solutions resulting from 3a exhibit a 7 Li-NMR signal at δ = 2.74 ppm (Figures S5 and S6).For comparison, the 7 Li-NMR signals of THF solutions of Li(C 5 H 5 ), LiSC 6 H 5 , and LiSC 6 H 4 CH 3 can be found at δ = −8.37,1.98, and 1.90 ppm, respectively).Cooling these solutions down to −78 • C leads to a color change and it turns into a violet color.Warming up to room temperature reinstates the brown color and evaporation of the solvent in vacuum leaves a brown residue, 2a or 3a , respectively.When a frozen solution made up from 3a and four equivalents of butyl lithium in THF is gradually warmed up within a NMR spectrometer, a 1 H-NMR spectrum immediately taken at −80 • C shows no signals of the Li-CH 2 protons.However, a signal due to LiC 5 H 5 and several signals is presumably assigned to Tol-SBu.After warming to ambient, the signals due to 3a are re-installed and, while the LiCp signal has disappeared, the signals of Tol-SBu are still present (Figure S4).
Both solid compounds are extremely air-sensitive and yield within seconds in air violet solids, which can be identified as 2a or 3a by using 1 H-NMR spectroscopy.When the yellow solutions, obtained from 2a or 3a and n-BuLi at −78 • C, are treated with dimethyldisulfide at this temperature and the mixture is warmed to ambient, the color changes to red-violet.Evaporation of the solvent leaves violet oils that contain a mixture of various compounds where Cp 2 Ti(S-C 6 H 4 R)(SCH 3 ) (R = H, Me) and (C 6 H 4 R) 2 S 2 can be identified as unreacted starting materials (Scheme 2).
Inorganics 2018, 6, x FOR PEER REVIEW 3 of 12 THF solutions of Li(C5H5), LiSC6H5, and LiSC6H4CH3 can be found at δ = −8.37,1.98, and 1.90 ppm, respectively).Cooling these solutions down to −78 °C leads to a color change and it turns into a violet color.Warming up to room temperature reinstates the brown color and evaporation of the solvent in vacuum leaves a brown residue, 2a′ or 3a′, respectively.When a frozen solution made up from 3a and four equivalents of butyl lithium in THF is gradually warmed up within a NMR spectrometer, a 1 H-NMR spectrum immediately taken at −80 °C shows no signals of the Li-CH2 protons.However, a signal due to LiC5H5 and several signals is presumably assigned to Tol-SBu.After warming to ambient, the signals due to 3a are re-installed and, while the LiCp signal has disappeared, the signals of Tol-SBu are still present (Figure S4).Both solid compounds are extremely air-sensitive and yield within seconds in air violet solids, which can be identified as 2a or 3a by using 1 H-NMR spectroscopy.When the yellow solutions, obtained from 2a or 3a and n-BuLi at −78 °C, are treated with dimethyldisulfide at this temperature and the mixture is warmed to ambient, the color changes to red-violet.Evaporation of the solvent leaves violet oils that contain a mixture of various compounds where Cp2Ti(S-C6H4R)(SCH3) (R = H, Me) and (C6H4R)2S2 can be identified as unreacted starting materials (Scheme 2).Scheme 2. Reaction of 2a or 3a with n-BuLi and air or MeSSMe.

Ring Metalation of 2a-c with Lithium Diisopropylamide and Reaction with Electrophiles
When a THF solution of 2a is treated at −78 °C with a freshly prepared THF solution of LDA (1.2 equivalents), which is followed by the addition of an equimolar amount of SiMe3Cl and warming up to room temperature, the product isolated after work-up contains mostly apparently unreacted 2a together with small amounts of the silylated compounds (C5H4SiMe3)(C5H4Z)Ti(SC6H5)2 (Z = H, 4a; SiMe3, 4b) and a di-nuclear complex of formula C32H28S2Ti2 (5).When dimethyldisulfide was used as an electrophile, the only identifiable product was the dimer 5. Attempts to purify the compounds by column chromatography lead only to its decomposition.However, compounds 4 and 5 could be unambiguously characterized by HRMS.

Ring Metalation of 2a-c with Lithium Diisopropylamide and Reaction with Electrophiles
When a THF solution of 2a is treated at −78 • C with a freshly prepared THF solution of LDA (1.2 equivalents), which is followed by the addition of an equimolar amount of SiMe 3 Cl and warming up to room temperature, the product isolated after work-up contains mostly apparently unreacted 2a together with small amounts of the silylated compounds (C 5 H 4 SiMe 3 )(C 5 H 4 Z)Ti(SC 6 H 5 ) 2 (Z = H, 4a; SiMe 3 , 4b) and a di-nuclear complex of formula C 32 H 28 S 2 Ti 2 (5).When dimethyldisulfide was used as an electrophile, the only identifiable product was the dimer 5. Attempts to purify the compounds by column chromatography lead only to its decomposition.However, compounds 4 and 5 could be unambiguously characterized by HRMS.
Treatment of a THF solution of 2b with a solution of LDA at −78 • C followed by the addition of hexachloroethane leads, after a chromatographic workup, to a 72% yield of the desired 1,2-dichlorocyclopentadienyl complex (C 5 H 3 Cl 2 )(C 5 H 5 )Ti(SC 6 H 5 ) 2 (6a).Similar treatment of 2c with LDA and C 2 Cl 6 led to an inseparable mixture of (C 5 H 3 Cl 2 )(C 5 H 3 ClY)Ti(SC 6 H 5 ) 2 (Y = H, 7a, Cl, 7b) in a combined yield of ca.40% together with a 47% yield of the olefin C 2 (SC 6 H 5 ) 4 .Similar treatment of 2b with LDA and chlorotrimethylsilane gave the chiral 1-chloro-2-trimethylsilyl-cyclopentadienyl complex 6b.While 6a,b could be characterized by NMR and mass spectrometry, compounds 7a,b could only be identified by HRMS.The NMR spectra of 6a show weak signals that might be assigned to the 1,3-regio-isomer of 6a (ca.6%).The NMR spectra of 6b show weak signals that might be due to PhSH or PhSSPh.The mass spectra of both compounds show peaks due to (C 5 H 3 ClY)(C 5 H 5 )Ti(SC 6 H 5 )Cl (Y = Cl, 6a , SiMe 3 , 6b ).
Inorganics 2018, 6, x FOR PEER REVIEW 4 of 12 could only be identified by HRMS.The NMR spectra of 6a show weak signals that might be assigned to the 1,3-regio-isomer of 6a (ca.6%).The NMR spectra of 6b show weak signals that might be due to PhSH or PhSSPh.The mass spectra of both compounds show peaks due to (C5H3ClY)(C5H5)Ti(SC6H5)Cl (Y = Cl, 6a′, SiMe3, 6b′).

Discussion
Our previous approach towards functionalization of metal-coordinated cyclopentadienyl rings needs perhalogenated cyclopentadienyl complexes as starting materials, which are not known for the titanocene system.To obtain such titanocenes with perhalogenated Cp rings, stepwise introduction of halogens via alternate metalation-electrophilic halogenation sequences might be a useful strategy, which was successfully applied by us [35][36][37] and others [38,39] in the ferrocene system and also in the cymantrene system [40].For the metalation step, we used lithium bases like butyl lithium or lithium amides.As outlined in the introduction, these reagents cannot be used with titanocene chlorides but may work with the corresponding aryl thiolates.Application of the known synthetic protocol "titanocene dichloride + aromatic thiol + base" with replacement of the usual NEt3 Scheme 3. Reactions of 2a-c with LDA and electrophiles.

Discussion
Our previous approach towards functionalization of metal-coordinated cyclopentadienyl rings needs perhalogenated cyclopentadienyl complexes as starting materials, which are not known for the titanocene system.To obtain such titanocenes with perhalogenated Cp rings, stepwise introduction of halogens via alternate metalation-electrophilic halogenation sequences might be a useful strategy, which was successfully applied by us [35][36][37] and others [38,39] in the ferrocene system and also in the cymantrene system [40].For the metalation step, we used lithium bases like butyl lithium or lithium amides.As outlined in the introduction, these reagents cannot be used with titanocene chlorides but may work with the corresponding aryl thiolates.Application of the known synthetic protocol "titanocene dichloride + aromatic thiol + base" with replacement of the usual NEt 3 by DABCO gives not only slightly better yields for the already known bis-arylthiolates 2a and 3a but also allows the synthesis of the new chlorocyclopentadienyl thiolates 2b,c and 3b,c.All chlorocyclopentadienyl compounds are air-stable and are highly viscous violet oils that withstand all attempts of crystallization.Since the used starting materials [(C 5 H 4 Cl)(C 5 H 4 X)TiCl 2 ] are always obtained as mixtures with the unsubstituted Cp 2 TiCl 2 , which are extremely difficult to separate, the corresponding dithiolates were also obtained as mixtures that were easily separated by using chromatography.The lower yields of 2b,c and 3b,c in comparison with their unsubstituted analogs 2a,3a are, therefore, probably due to losses in the purification step and not a consequence of an intrinsic instability.The particularly low yield of 3b is probably due to an adventitious presence of moisture in the reaction mixture.
The reaction of 2a/3a with one equivalent n-BuLi yields reversibly paramagnetic solutions, which are extremely air-sensitive and produce the starting materials quantitatively upon a deliberate addition of air.Treatment of the reaction solutions with dimethyldisulfide gives the mixed titanocene aryl-alkyl-thiolates 2d/3d together with the symmetric and asymmetric disulfides ArSSAr and ArSSMe.We think that this behavior is due to a temperature dependent redox-equilibrium with the room temperature solution containing a Ti(III) radical anion, according to Scheme 4. by DABCO gives not only slightly better yields for the already known bis-arylthiolates 2a and 3a but also allows the synthesis of the new chlorocyclopentadienyl thiolates 2b,c and 3b,c.All chlorocyclopentadienyl compounds are air-stable and are highly viscous violet oils that withstand all attempts of crystallization.Since the used starting materials [(C5H4Cl)(C5H4X)TiCl2] are always obtained as mixtures with the unsubstituted Cp2TiCl2, which are extremely difficult to separate, the corresponding dithiolates were also obtained as mixtures that were easily separated by using chromatography.The lower yields of 2b,c and 3b,c in comparison with their unsubstituted analogs 2a,3a are, therefore, probably due to losses in the purification step and not a consequence of an intrinsic instability.The particularly low yield of 3b is probably due to an adventitious presence of moisture in the reaction mixture.
The reaction of 2a/3a with one equivalent n-BuLi yields reversibly paramagnetic solutions, which are extremely air-sensitive and produce the starting materials quantitatively upon a deliberate addition of air.Treatment of the reaction solutions with dimethyldisulfide gives the mixed titanocene aryl-alkyl-thiolates 2d/3d together with the symmetric and asymmetric disulfides ArSSAr and ArSSMe.We think that this behavior is due to a temperature dependent redox-equilibrium with the room temperature solution containing a Ti(III) radical anion, according to Scheme 4. Such a redox equilibria are known from electrochemical studies of Cp2TiCl2 [41,42].It was found that dependent on the cyclopentadienyl ring substituents, the intermediate radical anion might either reversibly split off a chloride anion yielding Cp2TiCl (the so-called "Nugent-Rajanbabu-reagent" [43]) or irreversibly one of the cyclopentadienyl ligands to give CpTiCl2.It was also reported that treatment of the dialkyltitanocenes Cp2TiR2 with organolithium compounds first produced unstable "back-onium complexes" [Cp2TiR3] − Li + , which decomposed at 20 °C to CpTiR2, RH, and LiCp [23].At least in the observed time and temperature frame used by us, the presumed Ti(III) thiolates 2a′/3a′ are more stable since we could not observe any LiSAr or LiCp in the reaction solution.Quite interestingly, when a fourfold excess of butyl lithium was used in an NMR experiment, LiCp can be detected in the cold solution, but it disappears when warming-up.At the same time, Tol-SBu can be detected.Therefore, the first reactions in this case are outlined below.
On warming, apparently the LiCp and the "CpTi(SAr)" fragment change back to the starting dithiolate and unidentified decomposition products.In the case of the 1:1 stoichiometric reaction, the first step might be the same, but a further reaction is different due to the absence of excessive butyl lithium.
Clearly, for the butyl lithium, the redox reaction is preferred over ring deprotonation.We, therefore, turned towards lithium diisopropylamide as a possible deprotonation agent.When 2a was used, with SiMe3Cl as quenching reagent, small amounts of the deprotonation products could be identified together with an unexpected dinuclear compound of formula C32H28S2Ti2 (5). 5 was the only identifiable product when MeSSMe was used as a quenching reagent.We believe that this Such a redox equilibria are known from electrochemical studies of Cp 2 TiCl 2 [41,42].It was found that dependent on the cyclopentadienyl ring substituents, the intermediate radical anion might either reversibly split off a chloride anion yielding Cp 2 TiCl (the so-called "Nugent-Rajanbabu-reagent" [43]) or irreversibly one of the cyclopentadienyl ligands to give CpTiCl 2 .It was also reported that treatment of the dialkyltitanocenes Cp 2 TiR 2 with organolithium compounds first produced unstable "back-onium complexes" [Cp 2 TiR 3 ] − Li + , which decomposed at 20 • C to CpTiR 2 , RH, and LiCp [23].At least in the observed time and temperature frame used by us, the presumed Ti(III) thiolates 2a /3a are more stable since we could not observe any LiSAr or LiCp in the reaction solution.Quite interestingly, when a fourfold excess of butyl lithium was used in an NMR experiment, LiCp can be detected in the cold solution, but it disappears when warming-up.At the same time, Tol-SBu can be detected.Therefore, the first reactions in this case are outlined below.
Cp 2 Ti(SAr) 2 + BuLi → "[Cp 2 Ti(SAr) 2 Bu]Li"→ "CpTi(SAr)" + ArSBu + LiCp (1) On warming, apparently the LiCp and the "CpTi(SAr)" fragment change back to the starting dithiolate and unidentified decomposition products.In the case of the 1:1 stoichiometric reaction, the first step might be the same, but a further reaction is different due to the absence of excessive butyl lithium.
Clearly, for the butyl lithium, the redox reaction is preferred over ring deprotonation.We, therefore, turned towards lithium diisopropylamide as a possible deprotonation agent.When 2a was used, with SiMe 3 Cl as quenching reagent, small amounts of the deprotonation products could be identified together with an unexpected dinuclear compound of formula C 32 H 28 S 2 Ti 2 (5). 5 was the only identifiable product when MeSSMe was used as a quenching reagent.We believe that this compound is a di-titanium(III) compound triply bridged by a fulvalene-diide and two phenylthiolate ligands (Figure 1).compound is a di-titanium(III) compound triply bridged by a fulvalene-diide and two phenylthiolate ligands (Figure 1).A zirconium compound of an identical structure (5, M = Zr) was obtained by an oxidation reaction of a Zr(II) complex and characterized by NMR spectra and crystal structure determination [44].Similar fulvalene-bridged Ti(III) complexes with chloride, hydride, or sulfide bridges instead of the aryl thiolate bridges were obtained by a sodium reduction of Cp2TiCl2 [45] or thermolysis of the Ti(II) complex Cp2Ti(Me3Si-C≡C-SiMe3) [46].A related mono-cyclopentadienyl Ti(III) complex without a fulvalene bridge known as [CpTiCl]2[µ-SAr]2 was obtained from the corresponding mononuclear CpTiCl2(SAr) upon a reduction with sodium amalgam [47].We assume that the ring-lithiated primary product of the 2a-LDA reaction is unstable under the reaction conditions and decomposes after splitting off LiSPh first to a Ti(III) radical centered on the Cp ring, which dimerizes to form the finally observed dinuclear compound 5.
However, when the chlorocyclopentadienyl dithiolates 2b,c were treated with LDA followed by quenching with hexachloroethane or chlorotrimethylsilane, the desired di-substituted complexes 6a,b and 7a,b could be isolated in moderate yields.Starting from 6b,b, repeating the treatment with LDA and C2Cl6 or SiMe3Cl gave the corresponding tri-substituted compounds 8a,b in yields of 24% and 75%, respectively.An attempt of a "one-pot" synthesis of pentachloro-titanocene thiolates using alternating additions of LDA and C2Cl6 gave a 52% yield of 6a (in the case of 2b as starting material) but no higher substituted products.Apart from the observations in the ferrocene system [35], either the reactivity towards deprotonation or the stability of the formed chlorinated products decreases with an increasing number of chlorine substituents.According to the 1 H-NMR spectra, both 6a and 8a are formed as regio-isomers with one largely dominating.By comparing results from the ferrocene system, we conclude that the major isomers are the 1,2 and 1,2,3-substituted ones.There are also weak signals in the NMR and mass spectra of 3c, 6a, 6b and 8a that might be assigned to chloride-mono(phenylthiolate) complexes, which are always accompanied by signals attributable to PhSH.Since chloroform was used as a solvent both for NMR and mass spectra, it seems possible that the complexes are unstable towards this solvent according to the equation below.
An alternative explanation would be that the starting materials were contaminated with the chloride-mono(thiolate) complexes.However, no signs of this can be seen in the spectra of 2b, 2c, 3a.
One interesting aspect in the mass spectra of nearly all compounds is the presence of peaks assignable to [C10H8Ti2(SAr)2] 2+ , which corresponds to the suggested structure of 5 (without the terminal C5H5-ligands).

Experimental Part
All solvents were of analytical grade and were distilled over the Na or Na/K alloy and stored over the Na wire.All reagents (n-BuLi: 1.6M in Hexane, thiophenol, thiocresol, Di-isopropylamine, hexachloroethane, dimethyldisulfide, and DABCO) and Cp2TiCl2 were obtained from commercial suppliers and were used as such.(C5H4Cl)(C5H4X)TiCl2 (X = H, Cl) were prepared according to literature procedures [33,34].Fresh solutions of LDA were prepared from Di-isopropylamine and n-BuLi in THF.Chromatographic separations were performed in glass columns (30 × 7 cm 2 ) filled A zirconium compound of an identical structure (5, M = Zr) was obtained by an oxidation reaction of a Zr(II) complex and characterized by NMR spectra and crystal structure determination [44].Similar fulvalene-bridged Ti(III) complexes with chloride, hydride, or sulfide bridges instead of the aryl thiolate bridges were obtained by a sodium reduction of Cp 2 TiCl 2 [45] or thermolysis of the Ti(II) complex Cp 2 Ti(Me 3 Si-C≡C-SiMe 3 ) [46].A related mono-cyclopentadienyl Ti(III) complex without a fulvalene bridge known as [CpTiCl] 2 [µ-SAr] 2 was obtained from the corresponding mononuclear CpTiCl 2 (SAr) upon a reduction with sodium amalgam [47].We assume that the ring-lithiated primary product of the 2a-LDA reaction is unstable under the reaction conditions and decomposes after splitting off LiSPh first to a Ti(III) radical centered on the Cp ring, which dimerizes to form the finally observed dinuclear compound 5.
However, when the chlorocyclopentadienyl dithiolates 2b,c were treated with LDA followed by quenching with hexachloroethane or chlorotrimethylsilane, the desired di-substituted complexes 6a,b and 7a,b could be isolated in moderate yields.Starting from 6b,b, repeating the treatment with LDA and C 2 Cl 6 or SiMe 3 Cl gave the corresponding tri-substituted compounds 8a,b in yields of 24% and 75%, respectively.An attempt of a "one-pot" synthesis of pentachloro-titanocene thiolates using alternating additions of LDA and C 2 Cl 6 gave a 52% yield of 6a (in the case of 2b as starting material) but no higher substituted products.Apart from the observations in the ferrocene system [35], either the reactivity towards deprotonation or the stability of the formed chlorinated products decreases with an increasing number of chlorine substituents.According to the 1 H-NMR spectra, both 6a and 8a are formed as regio-isomers with one largely dominating.By comparing results from the ferrocene system, we conclude that the major isomers are the 1,2 and 1,2,3-substituted ones.There are also weak signals in the NMR and mass spectra of 3c, 6a, 6b and 8a that might be assigned to chloride-mono(phenylthiolate) complexes, which are always accompanied by signals attributable to PhSH.Since chloroform was used as a solvent both for NMR and mass spectra, it seems possible that the complexes are unstable towards this solvent according to the equation below.
An alternative explanation would be that the starting materials were contaminated with the chloride-mono(thiolate) complexes.However, no signs of this can be seen in the spectra of 2b, 2c, 3a.
One interesting aspect in the mass spectra of nearly all compounds is the presence of peaks assignable to [C 10 H 8 Ti 2 (SAr) 2 ] 2+ , which corresponds to the suggested structure of 5 (without the terminal C 5 H 5 -ligands).

Experimental Part
All solvents were of analytical grade and were distilled over the Na or Na/K alloy and stored over the Na wire.All reagents (n-BuLi: 1.6M in Hexane, thiophenol, thiocresol, Di-isopropylamine, hexachloroethane, dimethyldisulfide, and DABCO) and Cp 2 TiCl 2 were obtained from commercial suppliers and were used as such.(C 5 H 4 Cl)(C 5 H 4 X)TiCl 2 (X = H, Cl) were prepared according to literature procedures [33,34].Fresh solutions of LDA were prepared from Di-isopropylamine and n-BuLi in THF.Chromatographic separations were performed in glass columns (30 × 7 cm 2 ) filled with silica 60 (Merck, 0.063 to 0.2 mm).All reactions were run under N 2 atmosphere using standard Schlenk equipment.Work-up and chromatographic purifications were performed in the air.All peaks found in the mass spectra, according to the fragmentation pattern, are included in Table S1 of the Supplementing Information.

Reaction of 2a with LDA and MeSSMe
A violet solution of 2a (0.199 g, 0.50 mmol) in THF (15 mL) is treated at −78 • C with a freshly prepared THF solution of LDA (from HN(i-Pr) 2 (0.085 mL, 0.60 mmol) and n-BuLi (0.38 mL of a 1.6 M solution in hexane) in THF (15 mL) at 0 • C).After stirring for 10 min, MeSSMe (5 drops) is added and continuously stirred at this temperature for 15 min.Then the cooling bath is removed.After the solution has reached r.t., the solvent is evaporated in vacuo.The residue is extracted with benzene (20 mL).Evaporation of the violet extract in vacuo yields a violet oil.MS analysis of this oil shows the presence of 5 as the only identifiable product.Attempts of chromatographic purification lead only to complete decomposition.

Reaction of 2c with LDA and Hexachloroethane
A deep violet solution of 2c (0.200 g, 0.43 mmol) in THF (20 mL) is treated at −78 • C with a freshly prepared THF solution of LDA (from HN(i-Pr) 2 (0.13 mL, 0.90 mmol) and n-BuLi (0.56 mL of a 1.6 M solution in hexane) in THF (20 mL) at 0 • C).After stirring for 5 min, hexachloroethane (0.407 g, 1.72 mmol) is added and the cooling bath is removed.When the solution has reached r.t., the solvent is evaporated in vacuo.The residue is placed on top of a silica gel column.Toluene elutes two fractions.Evaporation of the first brown fraction leaves tetrakis(phenylthio)ethene as a brown oil (47 mg).The second violet fraction yields an inseparable mixture of (C 5 H 4 Cl)(C 5 H 3 Cl 2 )Ti(SPh) 2 (7a) and (C 5 H 3 Cl 2 ) 2 Ti(SPh) 2 (7b) as a violet oil (0.114 g).Due to strongly overlapping signals of the two compounds, no NMR data can be attributed to the single components.However, identification is possible via HRMS: 7a: C 22 H 17 35

Conclusions
The unsubstituted titanocene arylthiolates Cp 2 Ti(SAr) 2 cannot be ring-functionalized either by a n-BuLi/electrophile or an LDA/electrophile sequence.In both cases, unstable Ti(III) radicals are formed, which either form the starting materials, dimerize to fulvalene bridged dititanocenes, or decompose completely.With chloro-substituted ( Cl Cp)( X Cp)Ti(SPh) 2 (X = H, Cl), α-deprotonation can be achieved with LDA and Di-substituted and tri-substituted-cyclopentadienyl complexes can be obtained.However, the stability of the chloro-substituted titanocene thiolates decreases with an increasing degree of ring-substitution and, thus, the desired perchlorotitanocenes could not be obtained.
C-NMR and mass spectroscopy including HRMS.The NMR spectra of 3b/c show the presence of HS-C 6 H 4 CH 3 or [S-C 6 H 4 -CH 3 ] 2 and very weak signals (<2%) derived from 1b/c and (C 5 H 4 X)(C 5 H 4 Y)Ti(S-C 6 H 4 R)Cl.