NMR relaxation in modern wood and treated and untreated wood from a XVI century wooden statue

Longitudinal (T1) relaxation data are presented for European lime (Tilia cordata Mill) wood samples, a modern sample and two from a XVI century wooden statue, one of which samples is treated with Paraloid B72 for preservation. The 1H signals from liquid and solid are separated in the free induction decays (FID). T1 distributions are obtained by UPEN and show two resolved peaks for the liquid component of either untreated wood but a wide distribution for the liquid of the treated sample. In each case, the solid component showed a single peak, which for either untreated sample was at about the position of the longer T1 liquid peak. We compute geometric-mean times, T1-gm, 2.1 and 36 ms for the liquid peaks, with 62% of the liquid signal in the longer peak, and 38 ms for the one solid peak. The near-equality of the last two times suggests substantial magnetization exchange between liquid and solid. Using the solid–liquid 1H ratio a =3.8 from the FID analysis, we can test whether the data fit a model for simple two-site exchange, which they do. The value computed from this model for the liquid’s intrinsic T1-gm is 4.8 ms, which is very close the total-liquid (both peaks) rate-average T1-ra, 4.4 ms, which in the model should be equal to the intrinsic time. For the modern wood sample, the liquid times are 76 and 3.9 ms, and the solid is 77 ms, substantially identical to the longer liquid time. The two-site exchange model gives the intrinsic liquid T1=8.1 ms, and T1-ra=7.1 ms, again in close agreement. It is clear that the untreated old sample and the modern sample both have two resolved peaks separated by about the same factor in T1, but the modern wood has longer times and a larger fraction of the liquid signal in the peak at longer times, even though they originally come from the same kind of wood. On the other hand, the T1 distribution for the liquid of the treated old sample is quite different, without resolved peaks and going to both longer and shorter times. The solid curve for the treated sample is in the same general time range of the other two but is significantly wider. The data for the treated sample do not fit a simple magnetization exchange model.