Diffusion and Transport of CO2

Application of quasielastic neutron scattering (QENS) in tandem with molecular dynamics (MD) simulations brings insights into the dynamics and transport of CO2 in porous media [45]. Whilst the self-diffusion of molecules containing atoms that have appreciable incoherent neutron-scattering cross sections can be measured directly using QENS, the neutron-scattering cross section of CO2 is coherent and so the coherent QENS approach must be used to monitor the CO2 and the transport diffusivity extracted from this. Comparisons between MD simulation and QENS experiments involve the diffusing molecule’s self-diffusivity, however, in practical separations and catalytic applications, it is the transport diffusivity that is of greater importance. The transport diffusivity involves the response to a chemical potential gradient and its direct determination calls for non-equilibrium experiments. At the molecular level the dependence of the transport diffusivity and the so-called cor­rected diffusivity needs to be resolved. Linear response theory allows the transport

image029

Fig. 3.4 Left The orthorhombic (Pnma) V(O)(bdc) structure displayed along the z axis, highlighting the one-dimensional pore system (top). Labels of the different atoms of the V(O)(bdc) structure (bottom). Right Evolution of the experimental diffusion coefficients (corrected = A and transport = □) and simulated diffusion coefficients (self = •, corrected = ▲, and transport = A) as a function of the CO2 concentration in V(O)(bdc). The error bars for the simulations are 16, 7, and 12 % for low, intermediate, and high loading, respectively, while the experimental data are defined within an average error bar of 20 %. Reprinted with permission from (F. Salles, H. Jobic, T. Devic, P. L. Llewellyn, C. Serre, G. Ferey, G. Maurin, ACS Nano 4, 143 (2010)) [48]. Copyright (2010) American Chemical Society

and corrected diffusivity to be determined directly under equilibrium conditions, where chemical potential gradients are absent. Chapter 2 contains details of the relationship between the self-diffusivity, as measured using incoherent QENS, and the transport diffusivity, Dt, as derived from coherent neutron-scattering. Experi­mentally, this objective can be accomplished by using coherent QENS, which probes the collective motion of guest molecules at equilibrium [46]. The same objective can be accomplished using equilibrium MD simulations.

The joint MD-experimental QENS approach in which simulated CO2 dynamic properties are validated allows further details of the CO2 dynamics and transport in the host to be obtained, such as first demonstrated for CO2 in the NaX and NaY faujasite zeolites [47]. The joint coherent QENS-MD approach was first applied to MOFs in the study of the isostructural Cr(OH)(bdc) and V(O)(bdc) materials and also known as MIL-53(Cr) and MIL-47(V), respectively [48, 49]. This work built on the study of H2 self-diffusion in these materials as detailed in Chap. 2. V(O)(bdc) contains corner-sharing V4+O4O2 octahedra connected by bdc linker groups yielding one-dimensional channels (Fig. 3.4). Consequently, V(O)(bdc) has a rel- atively-high working capacity for CO2 uptake with moderate selectivity for CO2 in the absence of functional groups within the channel wall. The concentration- dependent self corrected (calculated from MD) and transport diffusivities (measured using QENS) of CO2 in the rigid V(O)(bdc) material were determined [48], revealing the three-dimensional diffusion of CO2 through the channels.

Cr(OH)(bdc) differs from V(O)(bdc) by the substitution of p2-O groups located at the M-O-M links in V(O)(bdc) by ^-OH groups in Cr(OH)(bdc). This

image030

CCVunit cell C02/unit cell

Fig. 3.5 Structural switching of the Cr(OH)(bdc) system induced by CO2 adsorption between large-pore (Imcm) (a) and narrow-pore (C2/c) (b) forms. Experimental (c) and simulated (d) transport diffusivity (Dt) as a function of CO2 concentration. Corrected diffusivities (D0) simulated for the rigid narrow (•) and large pore (▼) forms, along with D0 simulated using a composite approach (a) and experimentally determined (■). Top: Reprinted with permission from (A. V. Neimark, F.-X. Coudert, C. Triguero, A. Boutin, A. H. Fuchs, I. Beurroies, R. Denoyel, Langmuir 27, 4734 (2011)) [50]. Copyright (2011) American Chemical Society. Bottom: Reprinted from (F. Salles, H. Jobic, A. Ghoufi, P. L. Llewellyn, C. Serre, S. Bourrelly, G. Ferey, G. Maurin. Angew. Chem. Int. Edition 48, 8335 (2009)) [49]

difference makes Cr(OH)(bdc) highly selective for CO2, in a specific pressure range, which is a consequence of the large-scale breathing modes exhibited by the material. This allows it to switch from a large pore (LP) form to a narrow pore (NP) structure upon CO2/CH4 adsorption, with the NP structure able to trap CO2 and not CH4. As with the V(O)(bdc) study, the combined coherent QENS-MD approach was used to study the concentration-dependent self- and transport diffusion of CO2 in Cr(OH)(bdc). This work found a single-file diffusion regime in the material at high CO2 loading, a phenomenon not previously shown for any MOF (Fig. 3.5).