44249-AC5
Fundamental Aspects of Adsorption on Individual Carbon Nanotubes
The interaction of vapors with carbon nanotubes is a rich topic, the understanding of which is crucial for developing chemical sensors as well as for using nanotubes and graphene in electronics and other applications in which they are inevitably sensitive to their environment. A number of studies to date have revealed complex and often surprising sensitivity of the electrical properties to, for instance, noble gases and oxygen. The origin of the adsorbate-electron coupling remains murky, with many mechanisms having been proposed. In addition, fundamental questions arise concerning the behavior of adsorbed molecules on a cylindrical surface: for instance, are there phase transitions on the nanotube surface at reduced temperatures? Additionally, in the important case of oxygen, the magnetic nature and electronegativity of the molecules should have interesting consequences.
To address these questions, supported by this proposal, we developed nanotube “yoctobalance” devices, each consisting of an individual single-walled nanotube freely suspended across a 1 micron-wide trench between platinum electrical contacts. The vibrational resonant frequency of such a nanotube (usually in the range 50-500 MHz) can be determined by driving its motion electrostatically using an ac voltage applied relative to the conducting silicon substrate gate and determining the amplitude via the change in capacitance as the nanotube moves relative to the gate (a technique developed by the McEuen group at Cornell). The devices are held at a fixed temperature in a cell in a liquid nitrogen cryostat while being exposed to a varying vapor pressure. From the downward frequency shift the adsorbed density can be determined, to a precision presently of order 500 yg (a few atoms) at 77 K. This yields precise isotherms of adsorbate density vs pressure. Simultaneously, the electrical properties can be monitored and hence related directly to the adsorbate density. Refinement of the technique should allow detection of single atoms.
We are now preparing papers on our main results which are as follows. We identified vibrational resonances (amidst circuit resonances) in multiple devices by their parabolic dependence on dc gate voltage, which results from the tension in the nanotube increasing with electric field. Apparent quality factors are ~500 in vacuum at 77 K, while the resonances wash out at pressures of tens of mbar. Typically two or three modes are visible. The variation of resonance frequency with pressure for Ar and Kr at 77 K is consistent with formation of a monolayer, and indicates behavior similar in many ways to that seen in traditional graphite isotherms. The ratio of mass shifts due to monolayers of Ar and Kr on the same nanotube is consistent with that of their densities on graphite. In the case of Ar, which is above the critical point, the isotherms are smooth, and the isosteric heat obtained from the temperature dependence of the pressure at fixed coverage (frequency) agrees with expectations, varying between nanotubes depending on their diameter with a typical value of about 1100 K. The behavior can be stable and reproducible for months, allowing in-depth studies of different gases and conditions on an identical substrate.
The behavior of Kr, with its higher triple point, is remarkable. We see three dramatic first-order phase transitions within the first Kr monolayer in isotherms at 77 K, which imply vapor, commensurate solid, liquid and incommensurate solid phases according to the conventional phase diagram of Kr on graphite. We can now study such questions as the effects of cylindrical commensurability on two dimensional phase transitions and the approach to one dimension, with anticipated loss of long-range order and broadening of the transitions into crossovers.
Another interesting result is that in at least one device we see adsorption of 5-10% of a monolayer of Ar at a lower pressure, corresponding to about twice the isosteric heat for the full monolayer – just as would be expected for binding in the grooves between two nanotubes. This implies that in this regime we can study the behavior of one-dimensional chains of atoms.
We have also done preliminary measurements of the electronic properties. Surprisingly, a monolayer of Ar produces a conductance change of 10-50%, in spite of its nominal inertness. This confirms that the technique will allow us to carry out the needed thorough studies of doping and scattering effects of all kinds of adsorbates, in particular oxygen and other relatively active species.
This was the first external grant obtained by the PI as junior faculty and allowed him to begin this project in a new area, in collaboration with emeritus Professor Oscar Vilches who brings unique expertise in measuring low temperature adsorption isotherms and calorimetry. Graduate student Zenghui Wang was fully supported by the grant and his thesis will be based on the work described here. The project continues under NSF grant 0606078, “One-dimensional and two-dimensional adsorbate systems on carbon nanotubes”.
