Neutron Investigations of Rotational Motions in Monolayer and

Jan 5, 2009 - Tej Bhinde , Adam Y. Brewer , Stuart M. Clarke , and Tamsin K. Phillips , Thomas Arnold and Julia E. Parker. The Journal of Physical Che...
0 downloads 0 Views 871KB Size
Download PDF
4078

Langmuir 2009, 25, 4078-4083

Neutron Investigations of Rotational Motions in Monolayer and Multilayer Films at the Interface of MgO and Graphite Surfaces† J. Z. Larese,*,‡,§ T. Arnold,| A. Barbour,‡ and L. R. Frazier‡ Chemistry Department, UniVersity of Tennessee, KnoxVille, Tennessee 37996, USA, Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, Tennessee 37831, USA, Diamond Light Source, Harwell Science & InnoVation Campus, Chilton, OX11 0DE, U.K. ReceiVed September 5, 2008. ReVised Manuscript ReceiVed October 30, 2008 Recent experimental investigations of the rotational motion of methane and molecular hydrogen using inelastic neutron scattering (INS) measurements in combination with thermodynamic techniques have provided a unique view of the evolution of the interaction of these two molecules with the MgO (100) surface and graphite basal plane. Despite significant differences in the chemical and physical properties and surface symmetry of these two adsorbents, the dynamical behavior of the adsorbed films is remarkably similar. The interaction of a CH4 monolayer solid with MgO and graphite, as monitored by the behavior of the J ) 0 f J ) 1 free rotor transition, is so strong that there is no evidence for unhindered rotation of the molecule below 20 K. Using this same transition as a probe, H2 monolayer solids exhibit nearly free or significantly hindered motion on graphite and MgO (100) surfaces, respectively. Investigations of CH4 and H2 multilayer films on MgO find that once the film thickness exceeds ∼3 layers, the molecule-molecule interactions predominantly determine the dynamical properties of the molecular film furthest from the surface. INS signals indicate that the dynamical motion in thicker films is closely related to that observed in the bulk system. The results of these studies serve as a valuable pathway for developing a qualitatively accurate description of the potential energy surfaces that govern the microscopic properties of these systems.

Introduction Much attention has recently been focused on the physical properties of materials that are on the nanometer length scale (e.g., those systems where molecules are confined in a restricted geometry). Theory predicts that the physical properties of a single 2D molecular layer or ensembles restricted to spatial dimensions that are on the order of several nanometers in length are different from those of ordinary bulk (i.e.,3D) matter. (See McKenna1 and Rols et al.2,3 and references therein.) Using high-resolution thermodynamic methods coupled with neutron diffraction and inelastic neutron scattering (INS) and molecular modeling techniques, the structure, dynamics, and wetting properties of atomic and molecular films physisorbed on the surface of graphitic materials and the (100) surface of MgO nanocubes are investigated. A description of how this multifaceted approach can effectively provide microscopic insight into the evolution of the moleculemolecule (M-M) and molecule-surface (M-S) interactions by tracking the changes in the rotational dynamics of hydrogen and methane films is provided below. It will be shown that INS techniques are exceptionally well suited to monitor the changes in the barrier to molecular rotation as a function of layer thickness. The result is an accurate microscopic picture of how the substrate interaction behaves as a function of increasing film thickness. When a molecule within the adsorbed film moves further from the solid interface, the influence of the substrate diminishes, as a reflection of the decrease in the attractive interaction (e.g., the † Part of the Neutron Reflectivity special issue. * Corresponding author. E-mail: [email protected]. ‡ University of Tennessee. § Oak Ridge National Laboratory. | Diamond Light Source.

(1) McKenna, G. B. Eur. Phys. J. 2007, 141, 291–301 (special topics). . (2) Rols, S.; Cambedouzou, J.; Chorro, M.; Schober, H.; Agafonov, V.; Launois, P.; Davydov, V.; Rakhmanina, A. V.; Kataura, H.; Sauvajiol, J. L. Phys. ReV. Lett. 2008, 101, 065507. (3) Rols, S.; Jobic, H.; Schober, H. C. R. Phys. 2007, 8, 777–788.

Lennard Jones 6-12 potential) and the screening of the interaction with the surface by the intervening layers. Thus, there is a continuous rebalancing and evolution of the M-M and M-S interactions as the adsorbed film grows, into which our neutron and modeling studies provide effective insight not available using thermodynamic techniques alone. Our goal is to aid in the development of theoretical descriptions of these interactions that can ultimately be used in theoretical and numerical modeling studies that are both predictive and quantitatively accurate. A basic understanding of these systems (i.e., CH4 and H2 on MgO and graphite) is an important component for advancing an energy plan that includes the synthesis of advanced materials for the incorporation of gas storage and conversion to address our future energy needs.

Experimental Methods Material Synthesis and Preparation. High-quality, chemically pure, nominally single-facetted ((100) surface exposed) cubes of MgO produced using a unique vapor-phase process were employed in this study.4 Figure 1 is an electron micrograph that displays the typical size distribution (∼200 nm cube length with 10% variation) of our MgO particles. The powders are heat treated in vacuo (p e 10-7 Torr) at 950 °C for about 36 h. This procedure removes any physisorbed molecules (e.g., CO, CO2, and H2O) and homogenizes the (100) surface. After the heat treatment, all of the handling and loading of the experimental cells takes place in an UHP inert-gasfilled glovebox. These procedures are described in detail elsewhere.5-7 Adsorption Isotherms. To identify the thermodynamic behavior of the system, we perform a set of volumetric adsorption isotherms (4) Kunnmann, W.; Larese, J. Z. Novel Method for the Generation of High Density (Pure and Doped) Magnesium Vapors which Bypass the Liquidus Phase. U.S. Patent 6,179,897, 2001. (5) Arnold, T.; Cook, R. E.; Larese, J. Z. J. Phys. Chem. B 2005, 109, 8799– 8805. (6) Freitag, A.; Larese, J. Z. Phys. ReV. B 2000, 62, 8360–8365. (7) Yaron, P. N.; Telling, M. T. F.; Larese, J. Z. Langmuir 2006, 22, 7203– 7207.

10.1021/la802929b CCC: $40.75  2009 American Chemical Society Published on Web 01/05/2009

Rotational Motions in Monolayer and Multilayer Films

Figure 1. Typical scanning electron micrograph of MgO nanocubes produced using the Kunnmann-Larese vapor-phase method.4 The large cube in the inset at the upper right is a high-resolution TEM micrograph and is ∼200 nm on an edge with the (100) face exposed.

Figure 2. Typical adsorption isotherms for (a) methane (T ) 77.2 K) and (b) hydrogen [T ) 9.38 K) on MgO (100) surfaces. Note the stepwise layer growth in both cases that can be followed to ∼5 steps by inspection. The area per molecule for H2 on MgO is approximately one-half that of CH4 on MgO, resulting in the difference in the amount adsorbed step height observed in the isotherms.

over a wide range of temperatures to determine the surface area, heats of adsorption, isothermal compressibility, isosteric heat, and location of possible phase transitions. A detailed discussion of these thermodynamic procedures and methods can be found elsewhere.5,7-10 Figure 2a,b displays a typical isotherm for methane and hydrogen adsorption on the MgO powders, respectively. The pronounced five or six observable steps are indicative of a wetting sequence that proceeds layer by layer. In the case of methane, a nearly perfect match of the MgO (100) lattice constant to the (100) plane of the methane (II) solid phase is responsible for the wetting behavior. Much of the thermodynamic work on methane6 and hydrogen10,11 on MgO (100) can be found elsewhere. Neutron Scattering Measurements. CH4 on MgO. The adsorption isotherm in Figure 2a is an example taken from a comprehensive thermodynamic study addressing the (8) Arnold, T.; Chanaa, S.; Clarke, S. M.; Cook, R. E.; Larese, J. Z. Phys. ReV. B 2006, 74, 085421. (9) Arnold, T.; Cook, R. E.; Chanaa, S.; Clarke, S. M.; Farinelli, M.; Yaron, P.; Larese, J. Z. Physica B 2006, 385, 205–207. (10) Larese, J. Z.; Frazier, L.; Adams, M. A.; Arnold, T.; Hinde, R. J.; RamirezCuesta, A. Physica B 2006, 385, 144–146. (11) Frazier, L. R. Molecular Hydrogen Adsorbed on MgO (100) Surfaces: A Thermodynamic Study. Ph D. Thesis, University of Tennessee, Knoxville, TN, 2008.

Langmuir, Vol. 25, No. 7, 2009 4079

Figure 3. Plot of the INS spectrum from the bulk CH4 phase (II) (upper trace) and from monolayer CH4 (bottom trace) on the MgO (100) surface at T ) 4.0 K collected on the IRIS spectrometer at ISIS. Each trace took approximately 12 h to record and represents a sum over all detectors. Note that in the upper trace the J ) 0 f J ) 1 free rotor transition (discussed in the text) is found at ∼1.1 meV whereas there are no equivalent excitations observed in the methane monolayer film. This figure is adapted from Figure 7 in ref 16.

wetting properties of CH4 on the MgO (100) surface.6 Details describing the growth and structural properties of monolayer12 and multilayer film behavior,13,14 the diffusion within the solid films,15 and the low-temperature dynamics16 can be found elsewhere. In the following section, emphasis is placed on understanding how the rotational mobility of methane films varies as the film thickens. This is accomplished by examining the evolution of the INS response in the neighborhood of the J ) 0 f J ) 1 free rotor line of bulk methane (which experimentally appears around 1.1 meV) as shown in the upper part of Figure 3. Press and co-workers17-21 have examined the low-temperature structure and tunneling properties in detail. A simple model can be used to calculate the energy difference associated with the J ) 0 f J ) 1 transition for an isolated methane molecule; a value of ∼1.3 meV is derived (energy of the Jth level given by J(J + 1)B, where B ) 0.65 meV for methane). Clearly, the rotational motion of an isolated methane molecule is modified once some neighboring methane molecules are introduced through the formation of a cluster, a 3D solid, or the deposition of a condensed film on a solid substrate. Our previous studies of the methane on the MgO system focused on the rotational tunneling behavior of the CH4 monolayer. (See the lower spectrum in Figure 3.) Using symmetry arguments and a simple model, we proposed that below 20 K molecules in the monolayer 2 × 2 R45° solid phase were orientationally ordered with the C2V axis of the molecules perpendicular to the MgO (100) surface plane (sometimes referred to as the dipod orientation).16,22 Recent DFT studies demonstrated that this C2V orientation is indeed the lowest-energy configuration and that the molecules are arranged in an interdigitated network (i.e., similar to a meshed gear; see Figure 4b). For comparison, Figure (12) Coulomb, J. P.; Madih, K.; Croset, B.; Lauter, H. J. Phys. ReV. Lett. 1985, 54, 1536–1538. (13) Gay, J. M.; Suzanne, J.; Coulomb, J. P. Phys. ReV. B 1990, 41, 11346– 11351. (14) Madih, K.; Croset, B.; Coulomb, J. P.; Lauter, H. J. Europhys. Lett. 1989, 8, 459–464. (15) Gay, J. M.; Stocker, P.; Degenhat, D.; Lauter, H. J. Phys. ReV. B 1992, 46, 1195. (16) Larese, J. Z. Physica B 1998, 248, 297–303. (17) Asmussen, B.; Gerlach, P.; Press, W.; Prager, M.; Blank, H. J. Chem. Phys. 1989, 90, 400–405. (18) Heidemann, A.; Lushington, K. J.; Morrison, J. A.; Neumaier, K.; Press, W. J. Chem. Phys. 1984, 81, 5799–5804. (19) Heidemann, A.; Press, W.; Lushington, K. J.; Morrison, J. A. J. Chem. Phys. 1981, 75, 4003–4009. (20) Huller, A.; Press, W. Phys. ReV. B 1981, 24, 17–28. (21) Press, W.; Kollmar, A. Solid State Commun. 1975, 17, 405–408. (22) Larese, J. Z.; Hastings, J. M.; Passell, L.; Smith, D.; Richter, D. J. Chem. Phys. 1991, 95, 6997–7000.

4080 Langmuir, Vol. 25, No. 7, 2009

Larese et al.

Figure 4. (a) Structure of the bulk methane solid in phase (II) from Grieger et al.40 There are four molecules in the unit cell. One of these four molecules is orientationally disordered as shown by the blue circle. (b) Top view of the orientationally ordered, minimum-energy configuration of the monolayer methane solid at low temperatures discussed in the text. Note how the interdigitated network has a pair of hydrogen atoms oriented at right angles to its nearest neighbors. This is similar to the central plane of C2V-oriented molecules in the center of the cubic box identified by the solid line on the left.

Figure 5. INS response of the multilayer CH4 film deposited on MgO (100) nanocubes. All of the traces were recorded at 4.0 K on the IRIS spectrometer at ISIS (each trace took approximately 12 h to record). The numbers on the left-hand side of the figure represent the nominal methane surface coverage. All of the traces have the INS signal from the bare MgO substrate subtracted away. The green arrow indicates where the signal in the bulk phase (II) solid appears.

4a shows the structure and orientation of methane molecules in the unit cell of the phase (II) bulk methane solid. Note the similarity of the orientation of the molecules in the central layer of the phase (II) solid with those of the monolayer film. The strength of the interaction with its neighbors and/or the substrate dictates the magnitude of the change in the hindering potential (compare the two INS traces in Figure 3). Figure 5 displays the coverage dependence of the INS signal in the neighborhood of the energy transfers between 0.6 and 1.2 meV for methane on MgO. These measurements were performed on the inverted geometry, time-of-flight (TOF) crystal analyzer backscattering spectrometer, IRIS, at ISIS, a spallation neutron source at the Rutherford Appleton Laboratory using PG (002) reflection wherein an energy resolution of ∼17 µeV is obtained over an adjustable energy-transfer window of about 1.5 meV. The INS signals shown represent the sum of the TOF intensity from all of the detectors. We note that all of the traces in Figure 5 represent the difference between the inelastic response of the bare MgO substrate and the response of the substrate plus the adsorbed film. There are several features in Figure 5 worth noting. First, for surface coverage at and below monolayer completion (where a commensurate 2 × 2 monolayer

solid grows islandlike) there is essentially no INS intensity observed in the neutron energy-transfer range for 0.6 < E (mev) < 1.2. (See Figure 3 as well.) This absence of signal in the ∼1.1 meV range is what is expected for a rotationally immobile solid. The second point of interest is the appearance of two distinct features at ∼0.9 and ∼1.0 meV in the INS spectrum when the film thickness exceeds about 1.2 solid layers. These well-defined excitations remain reasonably fixed in energy transfer and increase in intensity proportionally with increasing film thickness from about 1.2 to about 2.5 equivalent layers. Third, at coverages above ∼3.0 equivalent layers, a third excitation centered at ∼1.1 meV appears in the spectrum. An additional increase in molecules in the film beyond this point serves only to increase the intensity of this excitation at ∼1.1 meV. This feature that dominates the spectrum for the thickest films studied (∼5 or 6 equivalent layers) is located very close to the INS signature of the J ) 0 f J ) 1 transition of the bulk solid. It is important to note that, under the thermodynamic conditions where these INS data have been collected, diffraction measurements produced no evidence that any bulk solid is present. (Although the vapor pressure in the cell is too low to measure at these temperatures, we performed diffraction experiments with CD4 that showed the layered films form without bulk signals.23) H2 on MgO. To follow the rotational motions for molecular hydrogen adsorbed on MgO (100) surfaces and make an equivalent comparison with the methane layering behavior, we used TOSCA, a TOF crystal analyzer spectrometer also at ISIS, to record the INS features. The energy resolution of TOSCA is ∼2% (∆E/E) over an energy-transfer range from ∼2 to 500 meV. In the H2 case, it is the ortho-to-para (O f P) rotational transition that falls in an energy range near 15 meV that was used to track the changes with coverage of the molecular hydrogen rotational motion. This transition (analogous to the methane transition) is also from the J ) 0 to the J ) 1 state where again the energy of the Jth level of a spherical rotor is given by J(J + 1)B and B ) 7.35 meV and is the rotational constant for molecular hydrogen.24 It is found that for bulk molecular hydrogen or H2 adsorbed on the graphite basal plane (where it also behaves essentially as a free rotor, see Figure 6) the O f P transition is observed at ∼14.7 meV. On the MgO (100) surface, however, the hydrogen experiences a greater barrier to rotation, hence the transition appears at a lower energy, ∼11.3 meV.10 Recent theoretical investigations confirm that the interaction of hydrogen with the MgO surface steers the molecule to an adsorption site directly above a Mg cation and that the motion has a more planar (2D) motion.10,25The INS response of monolayer films of H2 and the HD film on MgO and H2 on the graphite basal plane in the neutron energy-transfer (23) Larese, J. Z.; Harada, M.; Passell, L.; Krim, J.; Satija, S. Phys. ReV. B 1988, 37, 4735–4742. (24) Silvera, I. F. ReV. Mod. Phys. 1980, 52, 393–452. (25) Larese, J. Z.; Arnold, T.; Frazier, L.; Hinde, R. J.; Ramirez-Cuesta, A. J. Phys. ReV. Lett. 2008, 101, 165302.

Rotational Motions in Monolayer and Multilayer Films

Figure 6. INS spectrum recorded at ∼10 K illustrating the behavior of the J ) 0 to J ) 1 O f P transition for an adsorbed H2 monolayer film on graphite (blue trace), MgO (100) (red trace), and a monolayer HD film adsorbed on MgO (100) (green trace). The excitation located at ∼14.7 meV is located at nominally the same place as this transition in bulk H2, as discussed in the text. These spectra were recorded using the TOSCA spectrometer at ISIS and represent a summation of the forward and backscattering detectors; the H2 scans took ∼12 h, and the HD scan required ∼36 h.

Figure 7. (Left) INS response of the multilayer H2 film deposited on MgO (100) nanocubes as recorded on TOSCA at ISIS. All of the traces were recorded at ∼10.0 K. The numbers on the right-hand side of the figure represent the nominal hydrogen surface coverage. All of the traces have the INS signal from the bare MgO substrate subtracted away. (Right) INS response of the multilayer H2 film deposited on MgO (100) nanocubes using the IRIS high-resolution instrument where the energy window is offset to center the excitation in the window. The spectrum has a hydrogen surface coverage of 0.8 monolayer.

(loss) range of 2 < E (meV) < 35 is displayed in Figure 6. We note that the substitution of deuterium for a hydrogen atom in the molecule downshifts the energy of O f P to ∼8 meV, which is due to the change in reduced mass. At a surface coverage of ∼0.8, where the c(2 × 2) solid phase of H2 completely covers the MgO (100) surface, a companion experiment (shown in right panel in Figure 7) was performed using the IRIS spectrometer mentioned above. By exploiting the much higher energy resolution and by shifting the center of the energy-transfer window, we were able to record this O f P excitation near 11 meV. This peak is found to be instrumentally limited (i.e., energy width ∼50 µeV). To our knowledge, this is the first such application of the IRIS instrument in this manner and

Langmuir, Vol. 25, No. 7, 2009 4081 serves to underline the capability and flexibility of such an inverted geometry crystal analyzer instrument on a spallation source.26 The left-hand panel of Figure 7 displays the coverage dependence of the INS signals for molecular hydrogen films on the MgO surface at 10 K around 15 meV (i.e., where the O f P transition energy appears for bulk H2). Although there is a change in shape of 11.3 meV excitation in the submonolayer and monolayer completion regime that appears to be associated with an attendant uniaxial compression of the c(2 × 2) solid, where some of the molecules are squeezed out of registry from their position atop Mg sites (a devil’s staircase sequence forming p(2 × 4) and p(2 × 6) solids10,27), we will not discuss this behavior in detail here. The displacement of the molecules off of the Mg cation sites results in a reduction in the rotational hindering potential for some of the molecules within the film and produces a shoulder on the high-energy (less-hindered) side of the 11.3 meV excitation. However, a new excitation emerges in the spectrum near 15 meV when the coverage exceeds ∼2 nominal layers. This new feature coexists with the ∼11.3 meV excitation and grows steadily in intensity with increasing hydrogen coverage. This behavior is reminiscent of the thicker methane films on MgO described above. These emergent excitations correspond nicely to the O f P transition recorded for H2 in the bulk or adsorbed on the graphite basal plane (the latter shown in Figure 6) and therefore must correspond to the response of molecules that experience a (weaker) local potential than those closest to the substrate. As in the case of methane, the H2 chemical potential is lower than that necessary to produce small crystallites of bulk hydrogen, and recent diffraction experiments performed at these temperatures also show no signs of bulk diffraction peaks.10 This assertion is also confirmed by extensive thermodynamic measurements recently performed in a companion study11 at the temperatures of interest. Hence, the rotational behavior of both methane and hydrogen films on the MgO (100) surface is similar when the film thickness exceeds ∼3 layers. We discuss this behavior in more detail in the Discussion and Summary section below. Graphite Studies. CH4 and H2 on Graphite. Some brief comments should be made about the dynamic behavior of the equivalent CH4 and H2 films adsorbed on the graphite basal plane in the same context as discussed above. Numerous studies of the tunneling properties of monolayer methane on graphite and carbon-based materials28-30 and of hydrogen adsorption on graphite31 surfaces have been performed in the past. It is also known that these films grow layer by layer on graphite, but uniform film growth beyond ∼3 to 4 layers is complicated by capillary condensation (i.e., for exfoliated graphite materials such as Grafoil/Papyex and crystalline graphite foam). The current and earlier measurements of methane on graphite also found no free rotor behavior in the monolayer region as evidenced by the absence of the 0 f 1 free rotor transition in the INS signal. (See lower traces in Figure 8.) Above ∼1.5 to 2 layers, a broad INS signal between ∼0.7 to 1.0 meV appears. This ill-defined signal suggests that there is a significant spread in the rotational barrier experienced by individual molecules within the film. Thus, the adsorption sites appear to be less well defined when compared to those for CH4 on MgO (100). An earlier structural study of multilayer methane on graphite confirms that in the 2 to 3 layer regime the layers do not grow in registry with one another and that the films are most likely composed of several equivalent intralayer domains.32 Variations in the local structure within and between layers produce a variation in the magnitude of the rotational barrier as one moves (26) Carlile, C. J.; Adams, M. A. Physica B 1992, 182, 431–440. (27) Degenhardt, D.; Lauter, H. J.; Haensel, R. Jpn. J. Appl. Phys. 1987, 26, 341–342. (28) Rayment, T.; Thomas, R. K.; Bomchil, G.; White, J. W. Mol. Phys. 1981, 43, 601–620. (29) Newbery, M. W.; Rayment, T.; Smalley, M. V.; Thomas, R. K.; White, J. W. Chem. Phys. Lett. 1978, 59, 461–466. (30) Smalley, M. V.; Huller, A.; Thomas, R. K.; White, J. W. Mol. Phys. 1981, 44, 533–555. (31) Silvera, I. F.; Nielsen, M. Phys. ReV. Lett. 1976, 37, 1275–1278. (32) Larese, J. Z.; Harada, M.; Passell, L.; Krim, J.; Satija, S. Phys. ReV. B 1988, 37, 4735–4742.

4082 Langmuir, Vol. 25, No. 7, 2009

Larese et al.

Figure 9. Cartoon that schematically illustrates the possible dynamic behavior of the rotationally disordered molecules within the near surface layers of methane on the MgO (100) surface. The blurred methane molecules are meant to represent the molecules executing hindered rotation.

Figure 8. INS response of the multilayer CH4 film deposited on the graphite basal plane as measured using the IRIS spectrometer at ISIS using the PG (002) reflection, representing the summation of all of the detectors. All of the traces were recorded at 2.0 K; they took ∼6 h to record and are vertically displaced for ease of view. The numbers on the right-hand side of the figure (where the INS signals fall essentially to zero) represent the nominal methane surface coverage. All of the traces have the INS signal from the graphite substrate subtracted away. Note how there are no discrete, well-defined excitations in the neighborhood of 1.0 meV as there were in the MgO case.

from site to site within the multilayer film, thereby producing the broad INS feature. The INS spectrum of monolayer H2 on graphite shown in Figure 6 is for a 3 × 3 R 60° commensurate solid where molecules occupy every other hexagon.33 The depth and corrugation of the H2 on graphite substrate potential is apparently smaller than the contribution by the M-M interaction. Thus, the excitation shown in Figure 6 is similar to the INS response recorded for the bulk H2 solid.31,34 Increases in H2 coverage on graphite produce a concomitant increase in the intensity of the O f P signal at ∼14.7 meV; no unrelated rotational excitations appear, and hence no further discussion is necessary regarding a possible link between the evolution of the rotational dynamics and the layering behavior. The INS investigations described here lead one to the conclusion that the interaction strength of methane is stronger on graphite than on MgO (100). In contrast, the interaction of H2 is weaker with graphite than on MgO (100). One caveat to this generalization is derived from the nonideal nature of the graphite material (i.e., the tendency to exhibit capillary condensation complicates the issue somewhat, unlike the MgO case).

Discussion and Summary Films of methane and molecular hydrogen adsorbed on highly uniform MgO (100) surfaces exhibit strikingly similar rotational dynamics. An increased barrier to rotational motion is experienced by molecules residing near the substrate (when compared to the dynamics within the bulk solids), and INS investigations enable (33) Freimuth, H.; Wiechert, H.; Lauter, H. J. Surf. Sci. 1987, 189-190, 548– 56. (34) Silvera, I. F.; Hardy, W. N.; McTague, J. P. Discuss. Faraday Soc. 1969, 54–60.

a direct examination of the decay in the attractive hold of the surface as the films thicken. Why do these systems behave this way? Thermodynamic studies confirm that it is possible to observe the stepwise growth of solid films ∼5 or 6 layers thick of methane and hydrogen on MgO (100) surfaces. In the temperature range explored here, bulk methane forms a face-centered cubic structure in the phase (II) solid whereas bulk hydrogen forms an hcp solid. Multilayer films need to be at least 3 to 4 layers thick for methane (abc stacking) and 2 to 3 layers thick for hydrogen (aba stacking) before any molecules within the layered solid will experience near-neighbor M-M interactions similar to those experienced in the bulk. Because the interaction of the layer closest to the surface is modified by M-S interactions, it should not be surprising that the film thickness has to increase beyond the minimum film thickness (for either abc or aba stacking) before the rotational dynamics within the film exhibits any similarities of the M-M interactions present in the bulk solids. The multilayer INS rotational features discussed above represent the response of molecules at positions within the solid films where the local interaction (and structure) is closely related to those in the bulk solid. Furthermore, we noted the striking similarity of the monolayer and central plane of the phase (II) methane bulk structure in Figure 4a,b. Although thermodynamic studies exhibit no evidence of bulk formation, diffraction measurements of multilayer films using CD4 and D2 on MgO exhibit line shapes consistent with layered films and not the sharp, symmetric Bragg peaks associated with the formation of bulk crystallites.13,16 Earlier studies of multilayer ethylene (C2D4) adsorbed on graphite are excellent examples of how diffraction can be used to identify when film growth takes place simultaneously with bulk crystallites.35,36 Why do the two well-defined excitations appear in the INS traces in Figure 5 in the methane films on MgO when the film thickness increases beyond ∼1.5 nominal layers? These new excitations are related to a reduction in the local rotational barrier and an attendant increase in the dynamical motion of molecules either in the first layer (∼0.9 meV) or in the second layer (∼1.0 meV). There are two ways to produce this reduction in barrier height. Molecules can be located further from the substrate than they would be in the second adsorbed layer, or molecules in the first layer can move further from the MgO surface because they are attracted to a sufficient number of neighboring molecules in the second layer.37 (See the cartoon in Figure 9.) Remember, the greater the rotational barrier, the lower in energy the INS excitation (35) Mochrie, S. G. J.; Sutton, M.; Birgeneau, R. J.; Moncton, D. E.; Horn, P. M. Phys. ReV. B 1984, 30, 263–273. (36) Sutton, M.; Mochrie, S. G. J.; Birgeneau, R. J. Phys. ReV. Lett. 1983, 51, 407–410. (37) Drummond, M. L.; Sumpter, B. G.; Shelton, W. A.; Larese, J. Z. J. Phys. Chem. C 2007, 111, 966–976.

Rotational Motions in Monolayer and Multilayer Films

will appear. Previous diffraction studies confirm that thicker films grow commensurately with one another on the MgO surface.13,14,16 There is little doubt that much of the behavior described here takes place because the films predominantly grow in a close-packed, nearly ideal layering sequence. Naturally, the microscopic details of the process are much more complicated. Recent DFT calculations37 for methane on MgO in the 2 to 3 layer regime have provided additional insight into the multilayer behavior; however, on the basis of the DFT results, it is clear that additional work is necessary to refine the M-S contribution to the interaction energy. Recently, Stimac and Hinde noted that the C2V orientation in the monolayer solid requires the hexadecapole moment to be included in the potential.38 Once film growth exceeds 3 to 3.5 layers, the layered film exhibits only minor structural differences with the bulk methane phase (II) (fcc) solid. Changes in the low-energy tunneling portion of the INS spectrum also confirm that this is the case.39 Molecular hydrogen on MgO at coverages below ∼0.8 layers forms a c(2 × 2) solid film and a single, well-defined excitation corresponding to the O f P transition. The instrumentally limited excitation in the right panel of Figure 7 demonstrates that all of the H2 molecules experience the same rotational barrier. This finding agrees with recent investigations25 that suggest the rotational dynamics in the c(2 × 2) solid phase is more planar than spherical and is dictated by the interaction with the Mg cation. The energy shift in the INS spectrum upon substitution of deuterium for one of the hydrogen atoms in the molecule confirms the rotational character of the motion. In a recent paper, a more complete discussion of the behavior reported in the monolayer regime is provided.25 The behavior of H2 films greater than ∼2 or 3 layers thick can be understood using the same logic that we used to describe the methane films. Namely, some molecules within the film experience a local potential barrier that is lowered either because of an increase in the relative strength of the M-M interactions or because the forces felt by a molecule are similar to the local environment in the bulk crystal. This results in a new INS excitation at approximately the O f P transition (i.e., ∼14.7 meV) of bulk hydrogen. (38) Stimac, P. J.; Hinde, R. J. Eur. Phys. J. D 2008, 46, 69–76. (39) Larese, J. Z.; Marero, D. M. Y.; Sivia, D. S.; Carlile, C. J. Phys. ReV. Lett. 2001, 87, 206102. (40) Grieger, S.; Friedrich, H.; Guckelsberger, K.; Scherm, R.; Press, W. J. Chem. Phys. 1998, 109, 3161–3175.

Langmuir, Vol. 25, No. 7, 2009 4083

Conclusions In summary, we have used INS to investigate and compare the rotational dynamics of CH4 and H2 films on the (100) surface of MgO and the basal plane of graphite. We observed that in the case of MgO the significant interaction with the surface produces a sizable barrier to thte free rotation of molecules in monolayer films. In the methane studies on both substrates, the barrier height is large enough that no free rotor excitations are observed. The behavior of hydrogen on graphite is such that the rotational barrier differs only slightly from that of the bulk or gas-phase value. For hydrogen on MgO (100), the decrease in the energy difference between J ) 0 and J ) 1 is directly related to the increased height of the surface barrier created by the interaction with the Mg cation. Evidence is gathered from the evolution of the INS signals in films ∼3 to 4 layers thick suggesting that the local M-M interactions within the layered film converge toward the M-M forces experienced in the bulk solid even though no evidence for bulk crystallite formation is found. This behavior is consistent with film growth that closely resembles complete wetting. Future molecular modeling studies will be aimed at understanding the evolution of the dynamic behavior of these multilayer films more completely. Acknowledgment. There is little doubt that much of this work was stimulated by the seminal neutron investigations of the dynamics of methane adsorption on graphite by R. K. Thomas and his numerous co-workers many years ago. Since that time, improvements in materials synthesis, neutron instrumentation, and numerical modeling have made it possible to investigate and understand many of the issues that undoubtedly challenged these pioneers. J.Z.L. is especially indebted to R. K. Thomas for his numerous comments, suggestions, and help over the past 20 years as well as the encouragement of T.A. to spend a postdoctoral term in the J.Z.L. research group. We thank M. Adams, L. Bruch, C. J. Carlile, R. Cook, M. Drummond, M. Felty, J. M. Hastings, R. J. Hinde, D. Martin y Marrero, A. Novaco, S. Parker, L. Passell, A. Ramirez-Cuesta, D. Richter, P. Stimac, B. Sumpter, M. Telling, and P. Yaron for numerous helpful discussions and the Rutherford Appleton Laboratory and the Institute Laue Langevin for their hospitality and for access to their neutron scattering facilities. This work was made possible by the generous support of the U.S. DOE, Basic Energy Science, under contract no. DE-AC05-00OR22725 with ORNL (managed and operated by UT-Battelle, LLC) and by the U.S. NSF under award DMR0412231 and through J.Z.L. startup funds at the University of Tennessee, Knoxville. LA802929B