Extending motifs in lithiocuprate chemistry : unexpected structural diversity in thiocyanate complexes †

The new area of lithio(thiocyanato)cuprates has been developed. Using inexpensive, stable and safe CuSCN for their preparation, these complexes revealed Lipshutz-type dimeric motifs with solvent-dependent point group identities; planar, boat-shaped and chair shaped conformers are seen in the solid state. In solution, both Lipshutz-type and Gilman structures are clearly seen. Since the advent in 2007 of directed ortho cupration, effort has gone into understanding the structure-reactivity effects of amide ligand variation in and alkali metal salt abstraction from Lipshutz-type cuprates such as (TMP)2Cu(CN)Li2(THF) 1 (TMP = 2,2,6,6-tetramethylpiperidide). The replacement of CN with SCN is investigated presently as a means of improving the safety of lithium cuprates. The synthesis and solid state structural characterization of reference cuprate (TMP)2Cu(CN)Li2(THP) 8 (THP = tetrahydropyran) precedes that of the thiocyanate series (TMP)2Cu(SCN)Li2(L) (L = OEt2 9, THF 10, THP 11). For each of 9–11, preformed TMPLi was combined with CuSCN (2 : 1) in the presence of sub-stoichiometric Lewis base (0.5 eq. wrt Li). The avoidance of Lewis basic solvents incurs formation of the unsolvated Gilman cuprate (TMP)2CuLi 12, whilst multidimensional NMR spectroscopy has evidenced the abstraction of LiSCN from 9–11 in hydrocarbon solution and the in situ formation of Gilman reagents. The synthetic utility of 10 is established in the selective deprotometalation of chloropyridine substrates, including effecting transition metal-free homocoupling in 51–69% yield.


Introduction
Enhanced methods for the regioselective functionalization of aromatics that avoid the complications associated with the use of traditional main group organometallic bases 1 are of ongoing interest.This search led, in 1999, 2 to the development of the first of what have become known as 'synergic bases' 3 of the type R m M(NR′ 2 ) n AM (R = alkyl; m = 0, 2, 3; M = less polarizing metal; NR′ 2 = amide; n = 1, 2, 3; AM = (more polarizing) alkali metal).These have, for example, previously incorporated M = Al, 4 Cd, 5 Mg, 6 Mn 7 and Zn, 8 and have shown hitherto unachievable potential in anionic activation 9 and templated polymetalation. 10In a similar vein, new lithium cyanocuprates 11 have been central to the development of directed ortho cupration (DoC) 12 (Fig. 1).
Following the inception of lithium cuprate chemistry 13 through the development of R 2 CuLi, 14 attempts have focused on enhancing reactivity.Modifications have taken two major forms.Firstly, lithium amidocuprates have been developed, offering often unique reactivities as well as the potential of the amido group as a non-transferable ligand and as a chiral auxiliary. 15Secondly, there has been a focus on the incorporation of LiCN within lithium cuprates, 16 the presumption being that the cyano group would be transferred to Cu to give a higher order (tricoordinate) copper centre. 17The issue of the Cu-sequestering of cyanide has been discussed at length in the literature, though calculations, 18 spectroscopy 19 and X-ray diffraction 20 have increasingly pointed to the retention of lower order (dicoordinate) copper.This was noted too in the recently developed field of DoC transformations, with the 2 : 1 reaction of TMPLi (TMP = 2,2,6,6-tetramethylpiperidide) and CuCN giving complexes that could be characterized by X-ray diffraction.Results revealed that in the solid state such Lip-shutz formulation cuprates were dimers based on (TMP) 2 Cu-(CN)Li 2 (L) (L = OEt 2 1, THF 2) 12,21 monomers that clearly lacked Cu-CN interactions (Scheme 1).
Finally, the reaction of CuCN, RLi and TMPLi established that the inclusion of cyanide in the cuprate structure was by no means guaranteed, by furnishing the externally solvated amido(organyl) monomers RCu(TMP)Li(L) n (R = Ph, L = THF, n = 3 3; R = Me, L = TMEDA = N,N,N′,N′-tetramethylethylenediamine, n = 1 4). 22This observation took on added importance when recent DFT analysis suggested the in situ conversion of cyanide-incorporating cuprates to reactive Gilman reagents as the precursor to DoC reaction. 23ecent studies have sought to extend the principles established by the isolation of 1 and 2 to the use of general inorganic anions.This has led to the isolation and full characterization of (TMP) 2 Cu(X)Li 2 (L) (X = Cl 5, Br 6, I 7, L = OEt 2 , THF). 23,24These species have been viewed as being Lipshutz-type by virtue of their demonstrating essentially the same structural principles as 1 and 2. They have been successfully tested in the deprotonative metalation of halopyridines, 25 notably in the course of azafluorenone synthesis. 24Herein we extend this principle further, introducing the use of the thiocyanate anion as a non-toxic but potentially synthetically useful analogue of the cyanide components in 1 and 2. Preliminary results reveal novel variations in thiocyanatocuprate structure as a function of solvent both in the solid state and in solution.

Solid state analysis
With this study aiming to probe new thiocyanatocuprate bases solvated by THF, OEt 2 and THP (= tetrahydropyran) it was first necessary to complete the series of cyanocuprates 1, 12 2 21 and (TMP) 2 Cu(CN)Li 2 (THP) 8. To do this a hexane solution of TMPLi containing also THP (0.5 eq.wrt Li) was added to a suspension of CuCN (0.5 eq.wrt Li) in toluene.Following the addition of hexane the mixture was heated to reflux and then filtered to give a pale-straw coloured solution from which block-like crystals could be obtained (Scheme 2).
The product was shown by 1 H NMR spectroscopy to incorporate TMP and THP in a 2 : 1 ratio, suggesting a formulation analogous to that previously reported for 1 and 2. Corroboration of this view came from X-ray diffraction, which established the product to be 8 and to be the analogue of 1 and 2 (Fig. 2).All three structures proved to be relatively flat.IR spectroscopy on 8 revealed a dominant CuN stretching mode at ν ¯= 2104.3 cm −1 , with a signal developing at 2138.3 cm −1 upon air exposure (see ESI, Fig. S1 †), which compared closely with prior work. 22oving to the employment of CuSCN in an attempt to render a safer analogue of OEt 2 -solvate 1, a low temperature solution of TMPH in Et 2 O/toluene was treated with n BuLi.The resulting solution was transferred to a suspension of CuSCN in toluene.The mixture was heated to reflux, turning from pale cream to grey-black, whereupon filtration gave a yellow solution.Storage at room temperature gave a low yield of needlelike crystals after 1 day, which dissolved with further standing, and after several days were replaced with crystals of pseudorhombic habit in low yield. 1 H NMR spectroscopy revealed these to comprise TMP and OEt 2 in a 2 : 1 ratio.Though 13 C NMR spectroscopy failed to suggest the presence of SCN, this ligand was obviously incorporated, with IR spectroscopy revealing peaks at 2064.6 (s) and 1996.6 (m) cm −1 (Fig. S2 †).Air exposure of the sample for 30 s. incurred the growth of a signal at 2107.1 cm −1 , which compared with 2066 and 2070 cm −1 for calculated free [SCN] − and LiSCN, respectively. 26,27Overall, these data suggest the presence of [S-CuN] − moieties.This spectroscopic suggestion of a (TMP) 2 Cu(SCN)-Li 2 (OEt 2 ) formulation was verified by X-ray diffraction, which established an essentially flat dimer 9 2 (Scheme 3 and Fig. 3).The thiocyanatocuprate was found to be based on an essentially planar 8-membered (LiNCS) 2 metallocyclic core (the maximal deviation of any atom from a mean plane defined by the four Li + ions and the two thiocyanate ligands being just 0.12 Å).The suggestion from vibrational spectroscopy that formally Li-S-bonded [S-CuN] − moieties are present 27 is perhaps most clearly reinforced crystallographically by the shortness of the N3-C19 distance (1.149(3) Å), though at 1.631(3) Å S1-C19 hints at some level of delocalization in the anionic ligand.Formal, S-anion behaviour is also suggested by the significantly inequivalent N3-Li1 and N3-Li2 distances (2.250(5) and 1.998(5) Å, respectively); these relative lengths are consistent with a single N-based lone pair bisecting the Li1-N3-Li2 angle but favouring interaction with Li2 on grounds of electrostatic directionality (Fig. 3a). 28aving obtained the OEt 2 -solvate of (TMP) 2 Cu(SCN)Li 2 and established its essential planarity, THF was introduced to probe whether the structure remained fundamentally unchanged (cf. 1 vs. 2).TMPLi in hexane and THF (0.5 eq.wrt Li) was added to CuSCN, allowing the isolation of colourless prismatic crystals (Scheme 3).NMR spectroscopy revealed that, as for 9, these crystals incorporated TMP and Lewis base in a 2 : 1 ratio and a SCN group (δ 141.5 ppm by 13 C NMR spectroscopy).Corroboration of the last point came from IR spectroscopy, with two peaks seen at 2050.4 (s) and 1998.0 (m) cm −1 (Fig. S3, † cf.2064.6 and 1996.6 cm −1 in 9 2 ), the signals being replaced by a peak at 2105.7 cm −1 after air exposure for 30 s. X-ray diffraction was undertaken, with data revealing a dimer based on (TMP) 2 Cu(SCN)Li 2 (THF) 10.However, in contrast to the structure of 9 2 , 10 2 exhibited a novel boat conformation (Fig. 4) based on the aggregation of two crystallographically independent monomers (of which one will be representatively discussed) and in which the geometry at   sulfur is fundamentally altered relative to that seen in 9 2 .The result is that the two sulfur centres in 10 2 constitute the hinge about which deviation from planarity of the dimer operates.In spite of this, diffraction fails to reveal any significant difference in the bond lengths associated with sulfur (cf.S1-C19 1.631(3), S1-Li2A 2.518(4) Å in 9 2 and S2-C24 1.625(3), S2-Li1 2.500(6) Å in 10 2 ).This suggested equivalent thiocyanate anion structure to that seen in 9 2 even if the angle at sulfur now dramatically constricted on account of the dimer folding along the S⋯S vector (Li2-S1A-C19A 103.92(13)°in 9 2 compares with Li1-S2-C24 91.72(17)°in 10 2 ).The two 6-membered Cu-containing CuLi 2 N 3 rings in 10 2 are essentially planar (angles at Li1 and N3 sum to 359.9°and 359.6°, respectively).
Having established significantly different geometries for 9 2 and 10 2 attention switched to use of the Lewis base THP.TMPLi in hexane/toluene and THP (0.5 eq.wrt Li) was reacted with CuSCN, leading to the isolation of a modest yield of colourless blocks (Scheme 3).NMR spectroscopy revealed the presence therein of TMP and THP in a 2 : 1 ratio.Meanwhile, IR spectroscopy demonstrated signals attributable to SCN at 2063.2 (s) and 2006.5 (w) cm −1 , these being replaced by peaks at 2173.7 and 2108.5 cm −1 upon air exposure for 30 s. (Fig. S4 †).X-ray diffraction confirmed the expected formulation of the product as the dimer of (TMP) 2 Cu(SCN)Li 2 (THP) 11 but revealed a structure (Fig. 5) that, in contrast to the structures of 9 2 and 10 2 , exhibited a clear chair conformation ostensibly by virtue of the significant displacement (by ±0.43 Å) of Li2 and its symmetry equivalent from the mean plane described by the Li1-N3-C19-S1 fragment and its symmetry analogue.In order to allow this to occur not only S1 but also N3 necessarily deviate from planarity (in contrast to the geometry of N3 in 10 2 ), with angles at nitrogen now summing to 351.6°.This small deviation from planarity ( presumably limited to maximize the electrostatic directionality of nitrogen) is reflected also in the geometry at S1 for which, at 96.09(12)°, the Li-S-C angle is intermediate between those seen in the two previous structures (103.92(13)°in 9 2 , 91.72(17)°in 10 2 ).This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Diffraction reveals bond lengths of S1-C19 1.632(3), S1-Li2A 2.464(4) Å associated with sulfur, which are essentially the same as those seen in 9 2 and 10 2 and point to a common thiocyanate ligand electronic structure.The solvent sensitivity of thiocyanatocuprate formation was also investigated by eliminating external Lewis bases.As described above, reaction of TMPLi (4 mmol) with CuSCN (2 mmol) in bulk hydrocarbon doped with L (= OEt 2 , THF, TMP; 0.5 eq.wrt Li) gave (TMP) 2 Cu(SCN)Li 2 (L) 9-11.However, crystallographic analysis of the product revealed that the avoidance of donor solvent afforded a convenient and clean route to material which demonstrated a single signal at δ 0.90 ppm by 7 Li NMR spectroscopy.Crystallography subsequently revealed previously reported (TMP) 2 CuLi 12, 23 which was a dimer in the solid state, with IR spectroscopy corroborating the absence of SCN ligands (see ESI Fig. S5 †).However, the suggestion from 7 Li NMR spectroscopy that omitting Lewis base avoided contamination of the product with minor impurities 23 led us to undertake further re-characterization, obtaining a simple 13 C NMR spectrum of the Gilman cuprate (see below) which served to aid our interpretation of the more complex behaviour of 9-11 in solution.

NMR spectroscopy
The improved synthesis of 12 made available clean NMR spectra of the Gilman cuprate (TMP) 2 CuLi, with 13 C and 7 Li NMR spectra obtained in d 6 -benzene that could then be deployed in order to deconvolute the solution behaviour of thiocyanates 9-11.The 13 C NMR spectrum revealed a simple set of signals attributable to [TMP] − with singlets due to the 2,6-, 3,5-and 4-positions of the rings seen at δ 54.2, 42.1 and 19.2 ppm, respectively, and two Me signals located at δ 40.1 and 34.5 ppm (Fig. 6d).Comparison with the analogous spectrum obtained for diethyl ether complex 9 (at a concentration of ∼20 mg/0.7 mL d 6 -benzene, Fig. 6a) revealed it to be dominated by essentially identical signals, with a small amount of decomposition to give TMPH peaks, with minor traces of Lipshutz-type cuprate (see below) also manifest.These data strongly suggest ostensible conversion from the Lipshutz-type structure seen crystallographically for this system (Fig. 3a) to a Gilman formulation in solution.
Moving to the 13 C NMR spectroscopic data for THF-solvate 10 at the same concentration, a more complicated system is revealed.(The SCN component itself can be located at δ 141.5 ppm for this system in a 50 mg/0.7 mL sample, Fig. S11.†) It is immediately obvious that whilst signals attributable to the amide ligands in the Gilman cuprate remain, they no longer represent the dominant species in solution (Fig. 6b).Instead, major signals are now observed at the high-field side of each Gilman resonance (at δ 53.5, 40.6, 38.3, 34.4 ppm) and, in one case (δ 19.2 ppm), coincident with it.These are attributed to the corresponding carbon atoms in the Lipshutz-type structure, which appears to be substantially retained in solution, with an approximate Gilman : Lipshutz-type ratio of 1 : 3 suggested.Equivalently concentrated THP complex 11 (20 mg/ 0.7 mL in d 6 -benzene) revealed a similar picture to that demonstrated by 10, albeit the distribution of Gilman and Lipshutz-type species is approaching equivalence (Fig. 6c).Moreover, similar to 10, the analysis of a more concentrated sample (50 mg/0.7 mL in d 6 -benzene) located the thiocyanate resonance at δ 141.7 ppm (Fig. S11 †).
Coincident with the near-quantitative in situ conversion of Lipshutz-type crystalline 9 to a Gilman cuprate in d 6 -benzene (Fig. 6a) the deposition of a white powder was observed in the NMR tube to which 9 (20 mg/0.7 mL) had been added.Based on a control experiment in which pre-isolated Lipshutz-type cuprate was dissolved in benzene and the resulting white deposit analyzed by IR spectroscopy, we attribute this observation to LiSCN precipitation (Fig. S6 †). 7Li NMR spectroscopy on this sample therefore accorded no signal attributable to LiSCN.Rather, it revealed a dominant peak precisely matching the δ 0.90 ppm Gilman cuprate peak in 12 (Fig. 7d) accompanied by the development of a minor high-field signal at δ 0.65 ppm (Fig. 7a).Consistent with the 13 C NMR spectroscopic data, the high-field signal can be attributed to the retention of a small amount of Lipshutz-type cuprate in hydrocarbon solution, with the relative integrations (of 1 and 0.4) suggesting an approximate Gilman : Lipshutz-type ratio of 5 : 1.
Moving to 10 in solution it is immediately apparent that the 7 Li NMR spectrum revealed a significant change in behaviour, with the high-field signal at δ 0.71 ppm now dominant (Fig. 7b).Taken together with 13 C NMR spectroscopic data, this confirms the attribution of the high-field signal as retained Lipshutz-type cuprate and emphasizes the solvent dependence of LiSCN abstraction, with a Gilman : Lipshutztype ratio of 0.2 : 1 calculated based on the 7 Li NMR datawhich substantiates the 13 C NMR spectrum in Fig. 6b.Fig. 7c reinforces this view, with the Gilman : Lipshutz-type ratio of 0.6 : 1 calculated from 7 Li NMR data for 11 correlating with the appearance of Fig. 6c.

Chloropyridine derivatization
Based upon recent work on the use of halide-containing Lipshutz-type cuprates to facilitate the synthesis of azafluorenone frameworks 24 it was decided to test the reactivity of new thiocyanatocuprates in the selective elaboration of halopyridines.The in situ preparation of 10 was therefore undertaken using THF solvent and the resulting mixture was tested in the selective deprotometalation of both 2-chloropyridine 13 and 2,3-dichloropyridine 14.Under the conditions employed it was expected that reaction of 13 would occur at the aromatic 3-position, while the employment of two adjacent acidifying halogens would promote attack of the 4-position in the dichlorinated substrate.
Results of the use of in situ-generated 10 in the selective metalation of 13 are reported in Table 1.Under the conditions employed, reaction afforded the 3-substituted derivatives 15a-c in 46-71% yields after subsequent trapping with 4-methoxybenzoyl chloride, methyl iodide or phenyl disulfide, respectively (Table 1, entries 1-3 and Fig. S13 †).In the light of cyanocuprate 2 having been shown to promote the quantitative homocoupling of N,N-diisopropylbenzamide in the presence of the oxidant PhNO 2 , 12 so obviating the need for the inclusion of an additional transition metal-based catalyst, it was decided to test the efficacy of 10 in this respect as a safer alternative to the cyanocuprate.Accordingly, the formation of 15d was observed in 69% yield (entry 4).
The response of 2,3-dichloropyridine 14 was tested next (Table 2).This is known to undergo lithiation at the 4-position This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

Conclusions
The preparation of (TMP) 2 Cu(CN)Li 2 (THP) 8 served to extend the family of TMP-incorporating lithiocuprates based on the formulation (TMP) 2 Cu(CN)Li 2 (L) (L = OEt 2 1, THF 2) and to establish the prevalence of approximately flat dimers for these complexes in the solid state. 12,21The use of thiocyanate instead of cyanide was then probed to investigate the former as a convenient, cheap and safe alternative that avoids redirecting reactivity in ways recently described when cyanide has been replaced by halide. 13The resulting complexes showed interesting structural variability; the use of Et 2 O incurring an essentially planar dimer akin to those of 1, 2 and 8 but based on a (LiNCS) 2 core, while THF and TMP gave boat and chair conformers, respectively.These aggregates were best viewed as incorporating Li[S-CuN] moieties.Consistent with recently developed theoretical views on cuprate reactivity, 23 solid state structures were not necessarily retained in solution, with at least some Gilman cuprate formation noted for each thiocyanate system in benzene.Lastly, preliminary synthetic investigations successfully applied 10 to the selective deprotometalation and homocoupling of halopyridines, with ongoing work seeking to establish the extent of the synthetic portfolio of these new reagents.

Experimental section
General synthetic and analytical details

Synthesis and characterization of chloropyridine derivatives
A stirred solution of LiTMP was prepared at 0 °C by sequentially treating THF (5 mL) with 2,2,6,6-tetramethylpiperidine (0.68 mL, 4 mmol) and n BuLi (2.5 mL, 1.6 M hexanes solution, 4 mmol).This reagent was then treated with copper(I) thiocyanate (0.24 g, 2 mmol).The mixture was stirred for 15 min at 0 °C before the introduction of 2-chloropyridine (13, 0.19 mL, 2 mmol) or 2,3-dichloropyridine (14, 0.30 g, 2 mmol).After 2 h. at RT, the electrophile (4 mmol) was added either as a neat liquid or as a solution in THF (5 mL).The mixture was stirred overnight at RT before addition of a 1 M aqueous solution of NaOH (20 mL) and extraction with Et 2 O (2 × 20 mL).After washing the organic phase with an aqueous saturated solution of NH 4 Cl (10 mL) and drying over anhydrous Na 2 SO 4 , the solvent was evaporated under reduced pressure and the product isolated after purification by flash chromatography on silica gel (the eluent is given in the product description).
Reactions and manipulations were carried out under an inert atmosphere of dry nitrogen, using standard double manifold and glove-box techniques.Solvents were distilled off sodium (toluene) or sodium-potassium amalgam (Et 2 O, THF, hexane) immediately prior to use.Copper(I) thiocyanate and tetrahydropyran (THP, Sureseal) were purchased from Aldrich and 2,2,6,6-tetramethylpiperidine (TMPH) was purchased from Alfa Aesar.The amine was stored over molecular sieve (4 Å) while the other chemicals were used as received.n BuLi (2.5 mL, 1.6 M in hexanes) was purchased from Acros and used as received.IR spectra were collected on a Bruker Alpha FT IR spectrometer.NMR data were collected on a Bruker Avance III HD 500 MHz Smart Probe FT NMR spectrometer (500.200MHz for 1 H, 125.775MHz for 13 C, 194.397 for 7 Li).Spectra were obtained at 25 °C and chemical shifts are internally referenced to d 6 -benzene and calculated relative to TMS except for7 Li, for which an external reference was used (1 M LiCl in D 2 O).Chemical shifts are expressed in δ ppm.The following abbreviations are used: br = broad, m = multiplet, s = singlet, G = Gilman cuprate, L = Lipshutz-type cuprate.

Table 1
Results of the reaction of 10 with 2-chloropyridine 13 in THF and subsequent reaction with electrophiles or in the presence of PhNO 2 a Isolated yield.b 55% yield by using CuCl instead of CuSCN.c Yield estimated by 1 H NMR spectroscopy due to volatile nature of the product.

Table 2
Results of the reaction of 10 with 2,3-dichloropyridine 14 in THF and subsequent reaction with electrophiles or in the presence of PhNO 2.
a Yield estimated by 1 H NMR spectroscopy due to volatile nature of the product.b Isolated yield.