Structural coarsening of aspen wood by hydrothermal pretreatment monitored by small- and wide-angle scattering of X-rays and neutrons on oriented specimens

Abstract

Structural changes across multiple length scales associated with hydrothermal pretreatments of biomass were investigated by using small- and wide-angle X-ray and neutron scattering on oriented specimens. Isotropic and anisotropic scattering components were numerically separated and then interpreted as contributions from matrix and cellulose components, respectively. Equatorial diffraction peaks present in the isotropic component became sharper after hydrothermal treatments or ammonia treatment. Before pretreatment the wet cell wall was found to be homogeneous in the 10–100 nm range and scattering below Q = 0.5 (nm⁻¹) was dominated by surface scattering from the lumen. After pretreatment with acid or steam at 160 or 180 °C, density fluctuation developed in the cell wall at length scales above 10 nm, most likely due to lateral coalescence of microfibrils that partially co-crystallize to give larger apparent crystal sizes. A density fluctuation up to about 100 nm appeared in the isotropic component after acid and steam pretreatments due to morphological changes in the hemicellulose and lignin matrix.

Appendix

A rough estimation of the specific surface area S (area/volume) from the SANS data at low Q end will be shown. The application of Porod’s law of X-ray scattering by the internal surface at high Q (Porod 1951; Nishiyama and de Souza Lima 2013) to neutron gives

$$I = 2\pi \left( {\Updelta \rho } \right)^{2} S\frac{1}{{Q^{4} }}$$

where I is the scattered intensity in absolute scale (cm⁻¹) and Δρ the scattering length density contrast of the surface. From the data, the experimental prefactor A = 2π(Δρ)²S was 2.5, 5.0, 1.5, 1.5 (×10²⁴ cm⁵) for native, acid, steam and AFEX treated samples respectively. Since the azimuthal spread was typically 20°, these values are equivalent to their one-third in a random orientation case. If we use a (Δρ)² between cell wall and lumen filled with heavy water as 2 × 10²⁰ (cm⁻⁴), the specific surface

$$S = \frac{A}{3}\frac{{\left( {\Updelta \rho } \right)^{2} }}{2\pi } \approx 0.2 \times 10^{20} A$$

The value S varies between 3 and 10 (×10³cm⁻¹).

If we consider a typical wood fiber with square section 10 μm wide and a lumen of 6 μm wide, the specific surface will be

$$\sum = \frac{S}{V} = \frac{4 \times 6 \times length}{10 \times 10 \times length}\left[ {\mu m^{ - 1} } \right] \approx 2400\left[ {cm^{ - 1} } \right]$$

Thus the lumen surface area and the surface area calculated from neutron scattering prefactor agree in magnitude.