ARTICLES Probing Pore Size Distribution by Cryogenic- and Relaxation 2 H-NMR E. W. Hansen,* ,² C. Simon, ‡ R. Haugsrud, ‡ H. Raeder, ‡ and R. Bredesen ‡ Department of Chemistry, UiO, P.O. Box 1033 Blindern, 0315 Oslo, Norway, and SINTEF Materials Technology, P.O. Box 124, Blindern, 0314 Oslo, Norway ReceiVed: December 31, 2001; In Final Form: June 21, 2002 Cryogenic NMR and NMR spin-lattice relaxation time (NMRT) measurements of pore-confined water (D 2 O) have been performed using deuterium NMR to probe the pore size distribution (PSD) of silica materials and porous membranes. NMRT measurements were performed at a temperature slightly below the normal freezing point (277 K) of bulk water (D 2 O) to ensure that all interparticle water was frozen out. PSD derived from cryogenic NMR was in excellent agreement with PSD obtained from N 2 -adsorption measurements. Also, PSD obtained by NMRT revealed approximately the same average pore dimension as obtained by N 2 adsorption. However, the former experimental technique resulted in somewhat narrower PSD than obtained by cryogenic NMR and N 2 adsorption and is discussed in the text. An attempt to determine the PSD of a TiO 2 membrane on a silica support by NMRT will also be discussed. The main results obtained in this work suggest that a combined use of cryogenic NMR and NMRT may give information on both PSD and pore-connectivity. 1. Introduction Room temperature 1 H NMR relaxation time (NMRT) mea- surements of pore-confined fluids have been demonstrated as an important method for determining the pore-size distribution (PSD) of consolidated materials, like sandstone and concrete. 1-8 The technique is a valuable supplement to other principal tech- niques as mercury intrusion porosimetry (MIP), gas adsorption/ desorption, small-angle scattering (light, X-ray, and neutron), and optical/electron microscopy. Of particular importance, NMR is nondestructive and does not require any kind of drying or vacuum treatment before testing. Actually, drying or evacuation may cause artificial changes to the structure of some materials. An alternative technique to probe PSD of both consolidated and nonconsolidated materials is cryogenic NMR, as discussed by Hansen and co-workers. 9-12 Concerning microporous materials, Fraissard et al. 13 explored several gases to probe PSD and discovered that the 129 Xe NMR chemical shift is exceptionally sensitive to pore size. Also, pulse field gradient (PFG) NMR 14,15 has been reported to probe pore characteristics of porous materials. PSD derived from NMRT is based on a correlation between relaxation times (T 1 and T 2 ) of pore-confined fluid and the surface-to-volume ratio of the pore system. 5,16 For a certain pore geometry, like spherical and cylindrical pores, these relaxation rates are, to a first-order approximation, proportional to the inverse pore radius. A majority of reported measurements found in the literature have been performed on porous materials with simplified pore systems and narrow pore size distributions such as zeolites, different leached glasses, and packed spheres. 5,16-20 Again, it is of importance to emphasize that when attempting to derive PSD from NMRT a second and independent experi- mental technique is required for calibration purposes; that is, a correlation between the actual NMR parameter to be monitored and pore size must be established. Bhattacharja et al. 19 used a surface relaxation parameter based on surfaces determined by nitrogen adsorption in their NMR pore size measurements. Gallegos et al. 8 concluded that an assumption or correction concerning pore geometry or surface layer volume was required. Another difficulty relates to the existence of paramagnetic impurities, which may well affect the proton relaxation time of pore-confined fluids. For instance, it is known that the presence of paramagnetic ions such as iron or manganese on the grain surface is responsible for the enhancement of proton relaxation rates. In particular, Kleinberg et al. 21 emphasized that the proton relaxation rate of pore-confined fluids is not dominated by geometric restrictions (pore dimension) but by nucleus-electron interactions. Deuteron has a nuclear spin of 1 and, hence, possesses a nuclear quadrupole moment, implying that its relaxation behavior is affected by electric field gradients. Its magnetogyric ratio is about 6.5 times smaller than that of a proton. Theoretically, the paramagnetic contribution to 1/T 1 is proportional to the square of the nuclear magnetogyric ratio, suggesting the relaxation time of deuteron to be much less affected by the existence of paramagnetic impurities than protons. Glasel at al. 22 have shown that the deuteron relaxation rate of D 2 O mixed with glass beads (of dimension 20-660 µm) depends linearly on the inverse dimension of the glass beads. With this in mind, we decided to characterize some commercial silica samples (reference samples) by cryogenic NMR, NMRT, and N 2 adsorption. To the best of our knowledge, a critical comparison of PSD, as derived from these experimental techniques, have not been well documented. An important * To whom correspondence should be addressed. ² Department of Chemistry, UiO. ‡ SINTEF Materials Technology. 12396 J. Phys. Chem. B 2002, 106, 12396-12406 10.1021/jp0146420 CCC: $22.00 © 2002 American Chemical Society Published on Web 11/07/2002