More progress has been made in the second project, that for the synthesis of the Group 5 four legged piano stools. Two routes have been used and are shown below in Scheme 1. Preliminary results show that the tin reagent at the right of the scheme is much more selective than the corresponding silicon reagent that we previously prepared and characterized. However, the use of the lithium reagent at left gives much cleaner reaction than the potassium salt previously used and respectable unoptimized yields are obtained.

Scheme 1. Synthetic routes to four legged piano stools.       

Both complexes are effective catalysts for the silylcyanation of butyraldehyde.

R= C3H7

NMR studies show nearly complete conversion at ambient temperatures in 18 hours, at which time no reaction is observed by NMR in the control.

Silica gel bound zirconium complexes.

Drying silica gel at 3 mtorr at different temperatures was studied by 29Si NMR and IR spectroscopy.

Figure 1. MAS 29Si NMR spectra of silica gel dried at different temperatures

As drying is performed at higher temperature, the MAS 29Si NMR in Figure 1shows better resolution of the Q2, Q3 and Q4 signals. The ATR FT-IR spectra of these samples in Figure 2 shows a diminishment of the H-bonded O-H groups (broad band ca. 3500 cm-1) and the presence of non-H-banded OH groups (sharp peak ca. 3700 cm-1)

Figure 2. IR absorption spectra of silica gel dried at different temperatures.

In our studies we discovered a novel application of trimethyl silyl chloride in treating silica gel.  We first contemplated the use of trimethyl silyl chloride as a passivating agent, to remove sites of Brønsted acidity.

surface bound-Si- OH + TMSCl ---> HCl + surface TMS

The lower two traces in Figure 3 show the reaction of TMSCl after the surface has reacted with Zr(NMe2)4. The signal for the dimethyl amido groups has split in to a two peaks, one nearly 10 PPM downfield and a signal for the trimethyl silyl group is present at about d 0. The middle trace (labeled 500 SiO2-TMSCl) shows a single peak at d 0 when TMSCl is reacted with the dried silica gel. The upper two traces show the results of reaction of the silylated silica gel with Zr(NMe2)4: in both cases a single signal at d 40 results with only a small signal at 0.  The CP-MAS 29Si NMR spectra of these samples are consistent with the 13C NMR:

Figure 4 (uppermost trace) shows the silicon from TMS (near 0 PPM) only when TMSCl is the last reagent to be added. Infrared spectra show that TMSCl is very efficient at reacting with surface hydroxyl groupsFigure 2, trace b shows that OH groups remain after reaction with Zr(NMe2)4, but subsequent reaction with TMSCl (trace c) removes those OH groups. However, the reaction of  TMSCl after the  dimethylamidozirconium reaction changes the coordination sphere of the Zr as noted by changes in the 13C NMR (Figure 3 (500 Si-Zr(NMe2)-TMSCl) and the IR spectra  (Figure 2., traces b and c).

Figure 3. CP-MAS 13C NMR spectra of silica gel reacted with Zr(NMe2)4
and TMSCl with different orders of addition.

Since the top three traces in Figure 3 demonstrate that Zr(NMe2)4 react with the surface bound TMS groups (note the formation of one signal at d 40), we believe that subsequent reaction of TMSCl with surface bound Zr(NMe2)x  groups results in substitution of Cl for one or more dimethylamido groups. 

Figure 4. 29Si NMNR of functionalized silica gel


surface –Si-O-Zr(NMe2)x + TMSCl ---> surface –Si-O-Zr(NMe2)x-1(Cl) + TMS-NMe2

This is supported by Figure 3, trace SiO2-Zr(NMe2)-TMSCl (where two peaks are observed after TMS-Cl reaction).

The materials are being tested as catalysts in silylcyanation in standard bulk scale reactions as well as catalysts in microfluidic systems as well as potential chromatographic media in microfluidic systems.