
Titanium tetrachloride (also called titanium (IV) chloride or TiCl4) is a strong Lewis acid. It is highly oxophilic and reacts instantly with water to form TiO2 and HCl vapours. It is a clear, colourless liquid at RT. Less pure samples may have a colour or solid particulates but it can be re-purified by distillation. It is not compatible with ethereal solvents as it is known to form crystalline diether complexes. It is non-polar and is soluble in/compatible with benzene and chlorinated solvents. I recently used TiCl4 in an attempt to catalyze a reaction.
It is a bit tricky to use, but does not require strict anhydrous conditions to the point of using a syringe or cannula. I was able to remove it from the bottle using a Pasteur pipette. Further I was able to comfortably monitor the result by C-NMR by placing my compound in CDCl3 and then adding TiCl4 before transfer to an NMR tube. I was concerned about HCl escaping from the solution in the NMR tube so I sealed the top with para-film but this proved not to be a problem. I thought a TiO2 precipitate might form which could interfere with the NMR by disturbing the homogeneity but this too turned out to be not a problem and I was able to view the signals of my start material and the product.
It is a bit tricky to use, but does not require strict anhydrous conditions to the point of using a syringe or cannula. I was able to remove it from the bottle using a Pasteur pipette. Further I was able to comfortably monitor the result by C-NMR by placing my compound in CDCl3 and then adding TiCl4 before transfer to an NMR tube. I was concerned about HCl escaping from the solution in the NMR tube so I sealed the top with para-film but this proved not to be a problem. I thought a TiO2 precipitate might form which could interfere with the NMR by disturbing the homogeneity but this too turned out to be not a problem and I was able to view the signals of my start material and the product.

I felt that the reaction had not worked cleanly enough. Given the oxophilicity of TiCl4, this could be due to a methoxy or a nitro group (both of which were present on my compound but on the other half of the molecule to where I aimed for the reaction). TiCl4 is known to reduce nitro groups and in my case it looked like mine had been replaced with an amino and an alkene.

I decided to try with a homologue compound that contained a nitrile group in the place of the methoxy. I ran this reaction in a flask in 1,2-DCE (above) instead of an NMR tube in CDCl3. With the first compound, the mixture instantly looked like brown gunk. However, with this second compound, a beautiful bright yellow precipitate was formed (right). The start material was white so I think that this yellow precipitate was a Ti(IV)Cl4 complex where the nitrile group has acted as either one or two ligands. When I worked up this reaction by addition of water, the yellow vanished. Rotavapping the organic layer returned over 80% of my start material as a pure white solid and it was possible to make the yellow complex again from the recovered product. This yellow colour seems to be from delocalization of d-orbital electrons in the Ti(IV), rather than freely moving conjugated electrons in the organic compound. It therefore seems that TiCl4 prefers to coordinate with nitriles groups more than methoxy groups, which in a sense can 'protect' the rest of the molecule from the oxophilic attack of TiCl4.
The Ti(IV)Cl4 complex might be breaking up by the action of water, or it may be the HCl produced by the addition of water which removes the ligands from the Ti(IV). If I could suck off the 1,2-DCE/TiCl4 in a vacuum trap I could run a crystal structure on the yellow product and determine its configuration. While I have not so far had the resources to follow this up, it may be useful to other workers.
The Ti(IV)Cl4 complex might be breaking up by the action of water, or it may be the HCl produced by the addition of water which removes the ligands from the Ti(IV). If I could suck off the 1,2-DCE/TiCl4 in a vacuum trap I could run a crystal structure on the yellow product and determine its configuration. While I have not so far had the resources to follow this up, it may be useful to other workers.