735W and solute transport in soils are controlled by the water dynamics in the porous system. Classical techniques for the investigation of pore space are the determination of pore size distributions from analysis of water retention curves or gas and Hg adsorption isotherms. As an extension of these methods, NMR relaxometry is particularly convenient since it probes pore spaces by direct investigation of the dynamics of included water. Th e physical basis is the nuclear magnetic resonance of 1 H 2 O. In a magnetic fi eld with a fl ux density B, proton spins (here, 1 H 2 O) possess two states: precession parallel and antiparallel around the direction of B. For an ensemble of spins in thermal equilibrium, the parallel state is more highly populated than the antiparallel one. Th is ensemble can be excited by absorption of electromagnetic radiation of the Larmor frequency ν = Bγ H /2π, where γ H is the gyromagnetic ratio of 1 H (γ H = 267.5 × 10 6 rad s −1 ), and B is the absolute value of the magnetic fl ux density B, i.e., B = |B| (units of T). After excitation, two kinds of relaxation processes take place: spin-spin or T 2 relaxation and spin-lattice or T 1 relaxation, which restores thermal equilibrium. In water molecules, both are mainly controlled by dipole-dipole interactions between the water molecules and the matrix, i.e., by translational and rotational motion. Th e exact physics of these processes are well derived in, e.g., Callaghan (1991) and Blümich (2000).Since the relaxation properties of water in porous media are also strongly infl uenced by pore sizes and geometries, NMR relaxometry methods are very convenient for probing the pore space of natural porous media like rocks or soils (Kleinberg et al., 1994;Hall et al., 1997;Votrubová et al., 2000). Th us, numerous studies have used the T 2 relaxation time distribution for such purposes. Th is approach has been investigated intensively and is routinely used for purposes like oil well logging (Dunn et al., 2002), but it has also been applied successfully on natural and model soils (e.g., Hall et al., 1997;Todoruk et al., 2003;Mikutta et al., 2004;Jaeger et al., 2006;Keating and Knight, 2007). Measured T 2 data are apparent quantities, however, since they are aff ected additionally by diff usion of the water molecules in magnetic fi eld gradients (Barrie, 2000). Th is can be demonstrated by variation of the echo time T E , which is an experimental parameter. With shorter T E , the measured T 2 relaxation times increase due to the decreasing infl uence of water motion in inhomogeneous magnetic fi elds and local magnetic fi eld gradients, which are present at phase boundaries in porous systems (Dunn et al., 2002). Th erefore, if optimization of measurement time is not an issue, as in well logging, the determination of the spin-lattice relaxation time T 1 is A : BET, Brunauer-Emmett-Teller; CT, computed tomography; NMR, nuclear magnetic resonance.
S S : A IThis study used nuclear magne c resonance (NMR) relaxometry at diff erent Larmor frequencies to inves g...