Charge-transfer measurements provide new angle on diamond surface conductivity

Undoped diamond is normally an excellent electrical insulator. Its bandgap is about 5.5 eV, too big for a substantial number of electrons to enter the conduction band from the valence band. But in 1989 Maurice Landstrass and Kramadhati Ravi noticed that their synthetic diamond films conducted electricity much better than they expected. ¹ Many researchers since then have tried to work out the origin of diamond films’ conductivity, but a convincing proof of a mechanism has been slow in coming, even as the effect has been exploited in designs for diamond-based electronics.

Now, Case Western Reserve University’s John Angus, his student Vidhya Chakrapani, and their colleagues have shown that diamond immersed in water can establish a charge-transfer equilibrium with the surrounding liquid. ² They interpret their results to support an earlier explanation of Landstrass and Ravi’s 1989 experiment: Electrons transfer from the diamond to a film of atmospheric moisture adsorbed on the diamond surface, leaving behind the positively charged holes known to be responsible for diamond’s conductivity.

Like many other diamond researchers, Landstrass and Ravi made their diamond films by chemical vapor deposition. They ran an electric current through a mixture of methane and hydrogen gases to create a plasma. Under the right conditions, the methane molecules’ carbon atoms were deposited on a substrate to form the diamond structure. That technique produces hydrogen-terminated diamond crystals: The carbon atoms at the surface, instead of having dangling chemical bonds, are bound to H atoms. Undoped diamond needs to be H-terminated to conduct electricity.

In 2000, Lothar Ley and colleagues at the University of Erlangen in Germany showed that diamond conductivity also requires contact with air. ³ They were the ones who suggested that charge transfer to an adsorbed water film could quantitatively explain diamond films’ unexpectedly high conductivity. In their model, the electrons entering the water film react with H₃O⁺ ions (naturally present in all water) to produce H₂ molecules. But their explanation was greeted with some skepticism, in part because it wasn’t clear that such water films form on hydrophobic H-terminated diamond.

“From the chemistry side”

The Case Western Reserve group’s experiments suggest a slightly different fate for the diamond surface’s lost electrons—that they react with H₃O⁺ ions and dissolved O₂ molecules to produce H₂O. Chakrapani took diamond powder, reacted it with a hydrogen plasma to create the necessary hydrogen layer, and let it equilibrate with ambient air. Atmospheric moisture is slightly acidic because dissolved carbon dioxide creates carbonic acid. When she mixed the powder into an even more acidic solution, Chakrapani saw an immediate increase in the pH, which means that H₃O⁺ ions were removed. She also found that the amount of dissolved O₂ decreased and the diamond particles acquired a net positive charge.

Starting with an alkaline solution, she observed the reverse—a decrease in pH and the production of O₂—which indicated that the diamond particles had already lost some electrons during their interaction with the air, so they had space in their valence band to acquire more.

Those observations were consistent with the Case Western Reserve group’s proposed reaction, which consumes electrons, O₂ molecules, and H₃O⁺ ions. Further confirmation came when Chakrapani repeated the experiment with a deoxygenated acidic solution. The pH barely changed—proof that oxygen is necessary for the electron-transfer reaction to take place.

The team also addressed the issue of hydrophobicity. Chakrapani partially immersed H-terminated diamond crystals in acidic solutions and measured the angle of the meniscus-like contact between the liquid and the crystal. That contact angle is a measure of the hydrophobicity or hydrophilicity of a surface: The more readily the water covers the surface, the smaller the angle. (For further explanation and a picture, see the Quick Study by Laurent Courbin and Howard A. Stone, Physics Today, February 2007, page 84 .) Chakrapani found that electron transfer from the diamond to the solution makes the diamond surface less hydrophobic—probably because the negatively charged liquid and the positively charged diamond attract each other.

Explains Ley, “The contribution of Angus’s experiments lies in the fact that they approach the problem from the chemistry side. They thereby demonstrate that the chemical consequences of the model—change in pH, wetting angle, O₂ concentration, and so forth—are correct. Before, those things had to be inferred indirectly from electrical measurements performed under different conditions.”

Energetics

For an electron to move from the diamond to the solution, the transfer and the subsequent reaction must be energetically and entropically favorable. More free energy must be released by the reaction in the solution than is expended to remove the electron from the diamond. Both the Case Western Reserve group’s O₂ reaction and the Erlangen group’s H₂ reaction consume H₃O⁺ ions, so both reactions are more favorable in acidic solutions, where H₃O⁺ ions are plentiful, than in basic solutions, where they are scarce. The electrochemical potential (free energy per electron) of the O₂ reaction varies from –5.66 eV at pH = 0 to –4.83 eV at pH = 14, as shown in the figure. For the H₂ reaction, the range is –4.62 eV to –3.79 eV.

Which of the reactions takes place, if either, depends on the energy of the most energetic electrons at the diamond surface, which in turn depends on the surface environment. The Case Western Reserve researchers took that energy to be –5.2 eV, the valence-band maximum of the H-terminated diamond surface submerged in water. That number is accurate for their experiments, in which the diamond was submerged in water. For electrons with energy –5.2 eV, only the O₂ reaction is favored, and only in acidic solutions—consistent with the group’s observations.

The Erlangen group considers –4.2 eV, the valence-band maximum for the H-terminated diamond surface under ultrahigh vacuum, to be more appropriate for a diamond surface in contact with just a thin film of water. Electrons of that energy could drive both the H₂ and O₂ reactions. Researchers from both groups admit that there is some uncertainty in the details of the energy scheme, but they agree that the difference is a minor one. “As I see it, we are really on the same page on the basic issue: that the effect is electrochemical in origin,” says Angus.

Both groups are now studying electron transfer in other systems: The Erlangen group uses carbon fullerenes instead of water as the electron acceptor, while the Case Western Reserve group is looking at transfer from materials other than diamond. Angus and colleagues have been able to use pH to tune both the conductivity of semiconducting carbon nanotubes—a phenomenon that’s been observed before by others but not attributed to the charge-transfer effect—and the distinctive yellow-band luminescence of gallium nitride. As Angus explains, “The biggest surprise is that we saw the effect in other semiconductors as soon as we looked for it.”