A monomeric Complex

To identify the source of the hydrogen atom, several experiments were carried out.

They show that toluene cannot be the source of the hydrogen atom. After stirring the reaction mixture for 20 h, there was evidence of a compound which gave a very broad signal at -39 ppm in the ³¹P{¹H} NMR spectrum. From this intermediate compound the reaction to the final product (δ31P = -28.8 ppm) proceeded very slowly

at -25 °C. From the NMR spectrum it was concluded that the intermediate compound probably is some kind of lithium complex. When the same reaction (according to Equation 2-5) was carried out in pentane instead of toluene it was possible to crystallise the intermediate lithium complex [(tmeda)Li{Ph2PCH2S(NSiMe3)2}] (11) (Equation 2-6). The ³¹P{¹H} NMR spectrum of the crystals shows a very broad signal at -39 ppm. This indicates 11 to be the intermediate in the synthesis of 10.

These results were unexpected regarding the strategy and are probably due to the different electronic situation in the S(NSiMe3)2 moiety. Compound 11 crystallises from pentane as a monomer in the monoclinic space group C2/c with half a pentane molecule in the asymmetric unit (Figure 2-15). A comparison of the bond lengths and angles with Ph2PCH2S(NSiMe3)(HNSiMe3) (10) can be found in Table 2-6.

pentane

-78 °C P S

N N SiMe₃

SiMe₃ Li N [(tmeda)Li(CH₂PPh₂)] + S(NSiMe₃)₂ N

Equation 2-6: Preparation of [(tmeda)Li{Ph2PCH2S(NSiMe3)2}] (11).

The coordination mode in this compound is new compared to the complexes 1-8. The lithium cation is fourfold N-coordinated and the phosphorus atom is not taking part in the coordination as the Li–P distance of 3.23 Å is too long to be regarded a bond. Nevertheless, an orientation towards the lithium ion can be observed which is due to electrostatic attraction. Furthermore, the S1–C13–P1 angle of 108.79(9)° indicates the inclination of the phosphorus atom towards the lithium cation. This finding is confirmed by the ³¹P{¹H} NMR spectrum which shows a very broad signal at -39 ppm. The line broadening can only be explained with a (weak) Li–

P contact in solution. The significant quadropolar moment of the lithium nucleus broadens the phosphorus signal. This long-range interaction might also be the reason for the protonation of the ligand when the reaction is conducted in toluene.

Figure 2-15: Molecular structure of [(tmeda)Li{Ph2PCH2S(NSiMe3)2}] (11). Hydrogen atoms are omitted for clarity.

In the solid state both phenyl rings are twisted by 90° with respect to each other, facilitating close packing of the molecules in the crystal. The central (SN2Li) ring is almost perfectly planar with the phosphorus atom residing above this plane. It is aligned with C13, S1, Li1 and both nitrogen atoms of the TMEDA molecule with respect to the N1–S1–N2 bisection.

Table 2-6: Selected bond lengths [Å] and angles [°] in 10 and 11

10 11 10 11

S1–N1 1.6520(9) 1.6070(15) N1–S1–N2 109.26(5) 103.72(8) S1–N2 1.7184(10) 1.6032(16) S1–C13–P1 113.54(6) 108.79(9) S1–C13 1.8092(12) 1.8338(18) S1–N1–Li1 --- 89.41(12) P1–C13 1.8526(11) 1.8557(18) S1–N2–Li --- 90.63(12) N1–Si1 1.7421(10) 1.7150(16) S1–N1–Si1 123.18(6) 117.55(9)

N1–Li1 --- 2.071(4) N1–Li1–N2 --- 75.80(12)

N2–Li1 --- 2.039(4) N3–Li1–N4 --- 83.19(13)

N3–Li1 --- 2.150(4) N1–Li1–N3 --- 132.61(18)

N4–Li1 --- 2.129(4)

The N1–S1–N2 angle of 103.72(8)° is slightly more acute than in the lithium complexes 1-8. The Li–N distances range from 2.039(4) Å (Li1–N2) to 2.150(4) Å (Li1–N3) with the bonds from the TMEDA molecule being marginally longer than the bonds from the diimido moiety. As the phosphorus side-arm is not donating to the Li cation, it is free for binding to any other soft metal, thus providing the opportunity to generate heterobimetallic complexes just like the earlier reported [(thf)2Li{(NtBu)3SMe}ZnMe2].^[82]

To check whether or not the solvent toluene was the source of the hydrogen atom in 10, a sample of crystalline [(tmeda)Li{Ph2PCH2S(NSiMe3)2}] (11) was dissolved in toluene-d⁸ and the NMR tube was melted off. The idea behind this was that the incorporation of deuterium into Ph2PCH2S(NSiMe3)(HNSiMe3) (10) would be detectable in the NMR spectra. Curiously, the spectra did not change at all, i. e. pure 11 could not be converted into 10 by this method. Consequently, it could not be resolved why the deprotonation of TMEDA only occurs in toluene and not in pentane.

The different polarity of the two solvents and therefore a distinct activation of the lithium complex 11 could be a reason. When the synthesis of [Li{Ph2PCH2S(NSiMe3)2}]2 (5) is carried out in toluene, the outcome is the same as with pentane. However, if 5 is dissolved in pentane together with two equivalents of TMEDA and stored at room temperature for several weeks, the protonated species Ph2PCH2S(NSiMe3)(HNSiMe3) (10) is isolated.

Conclusion

With the free ligand at hand, it should now be easily possible to obtain a great variety of mono- and bimetallic complexes directly rather than following the metathesis or salt elimination route. This can be problematic as some sulphur diimido compounds undergo ligand scrambling with metal halogenides as was already discussed at the beginning of this chapter. Now the reaction of 10 with metal amides or metal hydrides will hopefully give new metal complexes. 10 would be an excellent starting material for such reactions as it can be prepared in nearly quantitative yield and great purity.

However, the reaction has to be further investigated to prove the source of the hydrogen.

However, 11 could also be a very good starting material for metal exchange reactions. As the phosphorus atom does not take part in the complexation of lithium, it is free to coordinate to other metals. Thereby, a precoordination is possible that

brings the metal in close proximity to the nitrogen atoms. Lithium could then leave the complex as a TMEDA/ligand adduct and the metal exchange would be complete.

In conclusion, it can be stated that the phosphorus side-arm on the sulphur diimide can be modified in a straightforward way. The NSCP ligands are indeed tridentate, containing hard nitrogen and soft phosphorus donor sites. Although the bite of the ligand system is not optimized for lithium cations, the complexes formed are quite stable. Even the softer phosphorus site coordinates the hard lithium cation, yet in solution. It seems that the {R2PCH2S(NR’)2}^- anions are indeed the ligands which are complexing as good as envisaged. Further, the formation of dimers seems to be favoured as it helps to balance the electron deficiency of the metal cations.

There is a wide range of methyl phosphanes which can be deprotonated and reacted with a sulphur diimide. Thus, the steric properties of the obtained lithium complexes can be tuned. It is even possible to introduce a stereocentre at the phosphorus atom (3, 6) or at the carbon atom (7) of the connecting CH2 bridge. In addition, different sulphur diimides can be used which are responsible for the distinct electronic properties of the compounds. In addition, the side-arm can be extended by oxidation of the phosphorus atom without loosing the soft donor site.

It has been shown, that it is possible to generate a complex of choice by choosing the appropriate phosphane, diimide and solvent. Thereby, a metal-free ligand (10) and a monomeric complex (11) were synthesised. There is now a library of different building blocks that can be combined in order to synthesise the appropriate compound for the desired application. The yields are good and 10 and 11 are excellent starting materials for subsequent metalation or metal exchange reactions. The new ligand system can thereby be fitted to the needs of the synthetic chemist, giving the opportunity to choose from a variety of possible derivatisations.

Besides, the phosphorus side-arm is in general flexible in solution and can bind to both lithium atoms. This process is fast on the NMR time scale and cannot be frozen out even at very low temperatures (120 K).