Lignin-Based Porous Supraparticles for Carbon Capture

Multiscale carbon supraparticles (SPs) are synthesized by soft-templating lignin nano- and microbeads bound with cellulose nanofibrils (CNFs). The interparticle connectivity and nanoscale network in the SPs are studied after oxidative thermostabilization of the lignin/CNF constructs. The carbon SPs are formed by controlled sintering during carbonization and develop high mechanical strength (58 N·mm–3) and surface area (1152 m2·g–1). Given their features, the carbon SPs offer hierarchical access to adsorption sites that are well suited for CO2 capture (77 mg CO2·g–1), while presenting a relatively low pressure drop (∼33 kPa·m–1 calculated for a packed fixed-bed column). The introduced lignin-derived SPs address the limitations associated with mass transport (diffusion of adsorbates within channels) and kinetics of systems that are otherwise based on nanoparticles. Moreover, the carbon SPs do not require doping with heteroatoms (as tested for N) for effective CO2 uptake (at 1 bar CO2 and 40 °C) and are suitable for regeneration, following multiple adsorption/desorption cycles. Overall, we demonstrate porous SP carbon systems of low cost (precursor, fabrication, and processing) and superior activity (gas sorption and capture).


S2
Lignins. Softwood Kraft lignin (BioPiva 100) was obtained from UPM (Finland), which was purified from black liquor using LignoBoost technology. The molecular weight of this lignin material (Mw=5250 g·mol -1 , Mn=1190 g·mol -1 , Mw/Mn=4.4) has been characterized in our previous work. 1 Figure S1. SEM images of (a) prepared LP420, (b) carbonized LP420 with no pre-oxidization step, and (c-f) carbonized LP420 prepared after pre-oxidization at a heating rate of 5 ºC·min -1 (c),   The chemical composition of pristine, oxidized and carbonized LP420 and CNF are shown in  In order to illustrate the dependence of theoretical pressure drop on particle size, it was assumed that carbon particles, ranging from 420 nm to 1.2 mm, were installed in a packed column with a diameter of 0.02 m (dt). The normalized pressure loss (ΔP·m -1 ) was calculated by using Ergun equation (Eq. 1). 3 The average diameter (dp) of the particles was adopted in the calculation regardless of the polydispersity.
The bulk gas flow (dashed arrow) and the gas diffusion (solid arrow) associated with the operation of the adsorption column packed with carbon SPs are illustrated in Figure S6. The meso-S7 scaled interstitial spaces emerges from the assembly of LPs (light green spaces), which are defined by the LP size. The interstitial spaces among the packed carbon SPs (light indigo spaces) are the mm-scale, which allow the gas flow through (dashed arrow) and is factored in the pressure drop.
The uniformity of the LPs influence the homogeneity of mesopore size while the gas diffusion is independent of mesopore size. We speculate that when the  The particles were considered as spherical, namely Φ=1, since carbon SPs are round-shaped.
The void fraction of the packed bed is usually between the random loose packing and random close S8 packing. The packing density fall between the maxmim packing density of 63.4 % and the minmum packing density of 55 %. 4 The void fraction was chosen from the ratio of the packing bed diamater to the sphere diameter (dt/dp ratio). An average void fraction () of 38% was assumed given that the dt/dp ratio vary from 16.7 to several thousands in this work. 5 The condition of the flue gas is summarized in Table S2. The superficial velocity of the flue gas in the column is 1 m·s -1 . The normalized pressure loss (ΔP·m -1 ) as a function of the particle diameter are shown in Figure   S7.

S9
ΔP= The pressure drop, dp=the particle diameter, L=the height of the bed, ρ=the density of the fluid, µ=the fluid viscosity, Ɛ=the void space of the bed, µ0=the fluid superficial velocity, Φ=sphericity of the particle, dt=the diameter of a packed bed, Figure S7. The theoretical pressure loss (ΔP·L -1 ) in a packing column as a function of the particle diameter (dp). S10 Figure S8.  XPS spectra of carbon SPs-800 and the SPs activated by steam and ammonia-steam after 3h are shown in Figure S11a. The XPS spectra of all carbon samples contained only three apparent peaks assigned to C 1s (285 eV), N 1s (400 eV), and O 1s (532 eV). 6 The N 1s peaks were deconvoluted (inset of Figure S11a). The peaks at 398, 400 and 401.5 eV were attributed to the pyridinic-N (N-6), pyrrolic-N (N-5) and quaternary-N (N-Q), respectively. 7 As listed in Table S3, 4.7 wt. % of elemental N was successfully incorporated to the carbon surface, which was almost the same as the bulk nitrogen content (4.78 wt. %) obtained from elemental analysis. Figure S11b    Figure S12. Distribution of the overall yield of steam-activated carbon SPs.
CO2 adsorption-desorption measurements following three cycles are shown in Figure S13. Carbon SPs were degassed in N2 at 200 °C for 120 mins. The adsorption was carried out at 40 °C under CO2 atmosphere and then the adsorbed CO2 was released by purging N2 at 120 °C. In the third cycle, carbon SPs showed a CO2 uptake capacity of 75 mg CO2·g -1 , which is very close to the initial value, 77 mg CO2·g -1 . In this cyclic measurement, carbon SPs maintained 97-99% of the adsorption capacity after three adsorption-desorption cycles. S15 Figure S13. Protocol used in gravimetric CO2 adsorption-desorption measurements of three cycles.