Breaking Carbon's Tetrahedral Mold

The tetrahedral geometry of four-coordinate carbon atoms is a steadfast property of organic compounds. Nevertheless, chemists are still attracted to the possibility of twisting or squashing a tetrahedron to create methane derivatives or molecules with quaternary carbons that are as flat as a pancake.

In one of the latest attempts, Stephen T. Liddle and coworkers at the University of Nottingham, in England, have prepared and determined the X-ray structure of a predicted type of planar dilithium methane derivative (Angew. Chem. Int. Ed. 2010, 49, 5570). As in previous attempts to make planar structures, the carbon atom in the compound doesn’t have perfectly square-planar geometry, Liddle says, and the bonding is considerably more ionic than covalent.

But the Nottingham team’s synthetic strategy could facilitate synthesis of other square-planar carbon compounds. The research also serves as a bridge between computational and experimental chemistry that could help improve understanding of chemical structure and bonding.

The story of square-planar, four-coordinate carbon compounds began in 1874, when Dutch chemist and future inaugural winner of the Nobel Prize in Chemistry Jacobus H. van’t Hoff first deduced the relationship between molecular structure and the optical activity of organic compounds. Van’t Hoff pointed out that the bonds between carbon and its four adjacent neighbors are directed toward the corners of a tetrahedron, with bond angles of 109.5° when all four substituents are the same, as in CH₄.

The flat world of four-coordinate carbon started to get serious attention in 1970 when future Nobel Laureate Roald Hoffmann of Cornell University and colleagues described a theoretical model for square-planar carbon (J. Am. Chem. Soc. 1970, 92, 4992). Their goal wasn’t to predict square-planar compounds but instead to consider how to stabilize a planar transition state for stereochemical isomerization experiments. Hoffmann and coworkers acknowledged the difficulty of their quest: “It would seem too much to hope for a simple carbon compound to prefer a planar to a tetrahedral structure,” they wrote.

The Hoffmann model includes an sp²-hybridized central carbon—like that found in alkenes and aromatic rings—rather than the sp³-hybridized carbon in tetrahedral geometry. The tetravalent sp² carbon forms two two-electron bonds with hydrogen atoms and a two-electron, three-center bond with two hydrogen atoms (H–C–H) that uses only the hydrogens’ electrons. The carbon atom’s remaining two valence electrons reside as a lone pair in the unused p orbital that sticks out perpendicular to the CH₄ plane. The four C–H bonds are considered equal because of resonance.

Hoffmann and his colleagues suggested that this square-planar form could be stabilized by delocalizing the lone pair and by using an electropositive substituent such as lithium in place of hydrogen to compensate for the electron deficiency of the σ–bonding system. They also suggested that a small ring such as cyclopropane conjoined with another ring could reduce the unfavorable energy of a square-planar carbon transition state, as could creating a molecule with a saturated carbon at the heart of a fused-ring system.

The next breakthrough came when Paul von Ragué Schleyer, now at the University of Georgia, led a team of chemists that built on the Hoffmann model to predict the first plausible square-planar methane derivative computationally (J. Am. Chem. Soc. 1976, 98, 5419). Schleyer and coworkers showed that sequentially replacing the hydrogen atoms on one carbon of cyclopropane with alkali-metal atoms such as lithium could do the trick.

The researchers pointed out that disubstituted planar methane derivatives could have cis and trans forms. And they predicted that the cis planar form, in which the substituent atoms are 90° apart, would be substantially more stable and more likely to be made than the trans planar form, in which the substituents are opposite each other (180° apart).

Making a square-planar carbon compound is easy to do computationally, Schleyer points out, but difficult experimentally. Hundreds of predicted examples are now reported in the chemical literature, he notes, and a few of them have been detected in gas-phase experiments. But most haven’t been and never will be made—these are not your ordinary organic compounds.

For example, in 1991, Schleyer and Alexander I. Boldyrev, who is now at Utah State University, designed a set of square-planar carbon compounds, including CSi₂Al₂. In 2000, Boldyrev and Lai-Sheng Wang, now at Brown University, led a team that detected CSiAl₃ ^– and other analogs in gas-phase experiments.

But when it comes to isolable compounds, only a few dozen examples of nearly planar cis four-coordinate carbon compounds have been prepared. For example, Schleyer led a research team that set out to make the proposed lithiated cyclopropanes and succeeded in preparing dimeric versions with carbon atoms that approach square-planar geometry (J. Am. Chem. Soc. 1996, 118, 6924).

Since then, several research groups have reported dilithium methane derivatives containing cis near-planar carbon that are dimers or contain higher order lithium clusters. In addition, Gerhard Erker of the University of Münster, in Germany, and coworkers have developed zirconocene-based methane derivatives, which are stabilized by forming a carbon-carbon double bond and aren’t perfectly planar (Chem. Soc. Rev. 1999, 28, 307).

Other approaches to square-planar carbon include making fenestranes—one of Hoffmann and coworkers’ suggestions. Fenestranes are planar, fused aromatic-ring systems that have a central carbon atom and look like a window with four or more panes. A handful of stable fenestranes have been made, but the geometry of the central carbon in those compounds is stuck between tetrahedral and square-planar.

In the 1990s, Leo Radom, now at the University of Sydney, in Australia, and coworkers predicted that neutral saturated hydrocarbons called alkaplanes host a square-planar carbon. But alkaplanes, which can be thought of as three-dimensional analogs of fenestranes, are so complex that no one has been able to figure out how to make one.

The square-planar story now has a new chapter with Liddle and coworkers’ synthesis of the first monomeric dilithium methane compound and the first with trans geometry. Liddle’s team started with a methane derivative, H₂CR₂, where R is a bulky phosphinimine substituent, Ph₂P=NRʹ (Ph = phenyl, Rʹ = diisopropyl­phenyl). Phosphinimine is a bidentate ligand that other researchers have used in their attempts to make square-planar methanes. The Nottingham chemists deprotonated H₂CR₂ by treating it with tert-butyllithium, forming HCR₂Li, in which carbon and lithium are bridged by the two phosphinimine groups.

The researchers then carried out a second deprotonation by treating HCR₂Li with another equivalent of tert-butyllithium in the presence of the chelating ligand tetramethylethylenediamine. The result was successful formation of a dilithium methandiide in which one lithium atom is chelated by the diamine and the other is coordinated to the nitrogen atoms of the phosphinimine groups.

“The beauty of this discovery lies in its simplicity,” observes Robert E. Mulvey of the University of Strathclyde, in Scotland, who specializes in organolithium chemistry. “Liddle and coworkers lured carbon into a novel planar geometry—and trans planar at that—by using two different lithiation reactions.”

The first lithiation doesn’t require donor molecule support due to the steric and electronic stabilization provided by the front-facing bulky phosphinimino arms, Mulvey notes. The remaining hydrogen is left pointing backward and is accessible for the second lithiation; donor support from the diamine ligand prevents the usual dimer formation.

“That’s a clever approach,” Mulvey says. “This breakthrough will appeal to experimentalists and theoreticians alike and will no doubt stimulate more activity in the pursuit of novel organometallic structures.”

The trans Li–C–Li angle in the molecule is 161°, distorted from 180° expected for square-planar geometry, Liddle explains. And the carbon atom sits 0.007 Å out of the mean plane formed by the lithium and phosphorus atoms. But it is perhaps the flatest isolated four-coordinate carbon compound made so far, Liddle says.

The new compound doesn’t exactly follow the Hoffmann bonding model for covalent planar methane, Liddle adds. The Li–C–Li system is not a two-electron, three-center bond. It’s better described as two asymmetric ionic bonds, with one Li–C bond at 2.124 Å and the other at 2.531 Å as a consequence of the steric demands of the ligands. Despite the difference in bond lengths, the level of interaction between each lithium and carbon are surprisingly the same, Liddle notes.

Utah State’s Boldyrev ran some calculations of his own to check out Liddle’s compound. “The monomeric dilithium methandiide is a remarkable compound with highly unusual bonding surrounding a carbon atom,” Boldyrev says. “It’s not a clear-cut example of the four-coordinate planar carbon species, however.”

Only one of the lithium atoms is close enough to carbon to be counted as a ligand, giving a coordination number of three, instead of four, for carbon, Boldyrev notes. He points out that the bond distance alone can’t be a criterion for assigning coordination number of a central atom, however. “I think this work by Liddle and coworkers might spark a discussion on how to define coordination number in complicated cases.”

Beyond the quest for square-planar carbon, Liddle notes that these dilithium methandiides have become an important class of organometallic intermediates during the past decade. For example, Liddle’s group has used them as carbene ligands that form double bonds to the metal center in yttrium, erbium, and uranium complexes.

Carbon is normally tetravalent, but it forms stable divalent carbenes, which have a lone electron-pair on carbon and either have two substituents or are part of a ring, such as an N-heterocyclic carbene. Carbenes are now well-known fixtures in organometallic chemistry, but just two decades ago they were still curiosities.

Some chemists are now looking at zero-valent carbon as a curiosity. In this case, carbon’s four valence electrons remain localized as two lone pairs and two substituents supply electrons for bonding to carbon. Liddle notes that some of his group’s carbon-based ligands fall in this category. These carbon(0) species can accept electron donation from ligands and back-donate the lone pairs to the ligands, behaving just like a metal, a concept that’s challenging chemists’ beliefs about carbon.

Attempts to break the tetrahedral rule show how a thread of knowledge can lead from one new finding to another, Schleyer observes. “The work on planar four-coordinate carbon, which we first considered 35 years ago, justifies pure research that has no immediate uses,” he says. “Chemists should take away from Liddle’s study the message that four-coordinate carbon can be planar and not just tetrahedral or three-dimensional. I hope that experimental examples will now be realized more quickly.”